OpsCenter 4.1 User Guide Documentation - DataStax
OpsCenter 4.1 User Guide
Documentation
December 18, 2014
В©
2014 DataStax. All rights reserved.
Contents
Contents
About OpsCenter............................................................................................................. 5
Key features........................................................................................................................................ 5
Installation........................................................................................................................ 7
Installing OpsCenter............................................................................................................................ 7
Installing the OpsCenter RPM package................................................................................... 7
Installing the OpsCenter deb package..................................................................................... 7
Installing the OpsCenter on Mac OS X or any Linux distribution.............................................8
Installing DataStax agents...................................................................................................................9
Manually deploying agents - tarball....................................................................................... 11
Manually deploying agents - rpm........................................................................................... 11
Manually deploying agents - deb........................................................................................... 12
Automatic installation of DataStax agents..............................................................................13
Configuring JAVA_HOME for DataStax agents..................................................................... 13
OpsCenter and DataStax agent ports...............................................................................................14
Installation and configuration locations............................................................................................. 15
Debian and Ubuntu Package install locations........................................................................15
CentOS, OEL, and RHEL Package install locations.............................................................. 15
Binary Tarball distribution install locations............................................................................. 15
Starting, restarting, and stopping OpsCenter....................................................................................16
Starting and restarting DataStax agents........................................................................................... 16
Upgrading OpsCenter....................................................................................................18
Configuration.................................................................................................................. 19
Configuring user access....................................................................................................................19
About user access roles.........................................................................................................19
Assigning or modifying user access roles..............................................................................20
Removing a user.................................................................................................................... 20
Configuring SSL................................................................................................................................ 21
Enabling SSL - package installations.....................................................................................21
Enabling SSL - tarball installations........................................................................................ 23
Disabling SSL - package installations.................................................................................... 24
Disabling SSL - tarball installations........................................................................................25
Enabling HTTPS................................................................................................................................26
Configuring events and alerts........................................................................................................... 27
Enabling email alerts.............................................................................................................. 28
Enabling alerts posted to a URL............................................................................................29
Verifying that events are posting correctly............................................................................. 30
Configuring data collection and expiration........................................................................................ 31
Estimating the amount of data generated.............................................................................. 31
Controlling data collection...................................................................................................... 31
Storing collection data on a different cluster..........................................................................32
Configuring OpsCenter definition file updates.................................................................................. 33
Advanced configuration..................................................................................................................... 33
OpsCenter configuration properties........................................................................................34
Cluster configuration properties..............................................................................................37
2
Contents
DataStax Agent configuration.................................................................................................38
OpsCenter updater properties................................................................................................ 39
Backing up opscenterd's configuration files for failover.................................................................... 39
Using custom start/stop scripts for DataStax Enterprise and Cassandra......................................... 40
Example scenarios............................................................................................................................ 40
Configuring for multiple regions............................................................................................. 40
Using OpsCenter............................................................................................................43
Overview............................................................................................................................................ 43
Managing clusters............................................................................................................................. 43
Creating a cluster................................................................................................................... 43
Adding an existing cluster...................................................................................................... 47
Node monitoring and administration................................................................................................. 47
Ring View................................................................................................................................47
List View................................................................................................................................. 52
Node management operations............................................................................................... 53
Cluster administration........................................................................................................................ 57
Generating a report................................................................................................................ 57
Collecting diagnostic data.......................................................................................................57
Adding a node to a cluster.....................................................................................................57
Configuring a cluster.............................................................................................................. 58
Removing a cluster.................................................................................................................59
Rebalancing a cluster............................................................................................................. 59
Restarting a cluster................................................................................................................ 59
Modifying a cluster setting......................................................................................................60
Performance metrics......................................................................................................................... 60
Using performance metrics.....................................................................................................60
Cluster performance metrics.................................................................................................. 63
Pending task metrics.............................................................................................................. 64
Alert metrics.......................................................................................................................................71
Advanced system alert metrics.............................................................................................. 75
Managing DataStax Enterprise Management Services.................................................................... 78
Repair Service........................................................................................................................ 78
Capacity Service..................................................................................................................... 82
Managing backups and restoring from backups............................................................................... 84
Scheduling a backup.............................................................................................................. 84
Restoring from a backup........................................................................................................ 84
Using custom scripts before and after backups.....................................................................85
Data modeling................................................................................................................................... 85
Keyspaces...............................................................................................................................85
Managing column families...................................................................................................... 86
Browsing data......................................................................................................................... 87
Troubleshooting............................................................................................................. 90
Internet Explorer web browser not supported...................................................................................90
The SSTables in this snapshot '<tag>' are not compatible.............................................................. 90
OpsCenter data growing too large....................................................................................................90
Cannot create a keyspace................................................................................................................ 90
Error exceptions.ImportError:libssl.so.0.9.8.......................................................................................90
Python used to run OpsCenter not built with SSL............................................................................91
DataStax agent port setting conflict.................................................................................................. 91
Limiting the metrics collected by OpsCenter.................................................................................... 91
Java not installed or JAVA_HOME environment variable not set.....................................................92
Insufficient user resource limits errors.............................................................................................. 92
3
Contents
Installing EPEL on CentOS 5.x or RHEL 5.x....................................................................................92
Problems with provisioning................................................................................................................92
General troubleshooting steps................................................................................................92
Invalid repository credentials.................................................................................................. 92
Timed out waiting for Cassandra to start...............................................................................93
The following packages are already installed........................................................................ 93
Agents cannot connect to opscenterd....................................................................................93
Removing all Cassandra or DSE files after failed provisioning..............................................93
Running sstableloader results in broken data distribution................................................................ 93
Timeout connecting to Cassandra 2.0 clusters.................................................................................93
Sophos Web Protection breaks browser access to OpsCenter on Windows....................................94
OpsCenter API reference.............................................................................................. 95
Release Notes................................................................................................................ 96
4.1.4................................................................................................................................................... 96
4.1.3................................................................................................................................................... 96
4.1.2................................................................................................................................................... 97
4.1.1................................................................................................................................................... 97
4.1.0................................................................................................................................................... 97
4.0.3................................................................................................................................................... 98
4.0.2................................................................................................................................................... 99
4.0.3................................................................................................................................................... 99
4.0.1................................................................................................................................................. 100
4.0.................................................................................................................................................... 100
Using the docs.............................................................................................................101
Feedback....................................................................................................................... 102
4
About OpsCenter
About OpsCenter
DataStax OpsCenter is a visual management and monitoring solution for Apache Cassandra and DataStax
Enterprise. OpsCenter provides architects, DBA’s, and operations staff with the capabilities to intelligently
and proactively ensure their database clusters are running well and that administration tasks are simplified.
The DataStax agents are installed on the Real-time (Cassandra), Analytics (Hadoop), and Search (Solr)
nodes. They use Java Management Extensions (JMX) to monitor and manage each node. Cassandra
exposes a number of statistics and management operations through JMX. Using JMX, OpsCenter obtains
metrics from a cluster and issues various node administration commands, such as flushing SSTables or
doing a repair.
Agent
THRIFT
OpsCenter
server
Node0
HTTP
STOMP
HTTP
STOMP
Cluster1
Agent
Agent
Node2
Node1
Browser
THRIFT
JMX
Key features
OpsCenter offers a number of features to help manage both DataStax Enterprise and Apache Cassandra
clusters and make your life easier.
The key features of OpsCenter include:
Dashboard
•
•
•
A Dashboard that displays an overview of commonly watched performance metrics
Adding your favorite graphs to the dashboard
An Overview that condenses the dashboards of multiple clusters (not visible when monitoring a single
cluster)
Configuration and administration
•
•
•
•
•
•
Basic cluster configuration
Administration tasks, such as adding a cluster, using simple point-and-click actions
Visual creation of clusters
Multiple cluster management from a single OpsCenter instance using agents
Rebalancing data across a cluster when new nodes are added
Downloadable PDF cluster report
5
About OpsCenter
Alerts and performance metrics
•
•
•
Built-in external notification capabilities
Alert warnings of impending issues
Metrics are collected every minute from Cassandra, Analytics, and Search nodes and stored in a
keyspace created by OpsCenter
Backup operations and restoring from backups
•
•
Automatic backup operations, including scheduling and removing of old backups
Restoring from backups
Enterprise-only functionality
Enterprise functionality in OpsCenter is only enabled on DataStax Enterprise clusters. The following is a list
of features included: (each is linked to another docs page)
•
•
•
•
•
•
•
•
6
DataStax Enterprise Management Services
Alerts
Data Backup and Restore
Management en Masse
Viewing historical metrics more than one week in the past
Rebalance
Diagnostics tarball
Hadoop jobtracker integration
Installation
Installation
Installing OpsCenter
Installing the OpsCenter RPM package
Install the DataStax OpsCenter using Yum repositories on RedHat Enterprise Linux (RHEL), CentOS, and
Oracle Linux (OL) distributions.
For a complete list of supported platforms, see DataStax OpsCenter – Supported Platforms.
Before you begin
•
•
•
Yum package management utility.
For CentOS or RHEL 5.x, EPEL.
Python 2.6+
About this task
The CentOS, RHEL, and OL OpsCenter packaged releases create an opscenter user. OpsCenter runs
as a service and runs as this user. The service initialization script is located in /etc/init.d. If the
OpsCenter machine reboots, OpsCenter restarts automatically.
Procedure
1. Open the Yum repository specification /etc/yum.repos.d for editing. For example:
$ sudo vi /etc/yum.repos.d/datastax.repo
2. In this file, add the repository for OpsCenter.
[opscenter]
name = DataStax Repository
baseurl = http://rpm.datastax.com/community
enabled = 1
gpgcheck = 0
3. Install the OpsCenter package.
$ sudo yum install opscenter
For most users, the out-of-box configuration should work just fine, but if you need to you can configure
OpsCenter differently.
4. Start OpsCenter:
$ sudo service opscenterd start
5. Connect to OpsCenter in a web browser using the following URL:
http://opscenter-host:8888/
6. Next you can add an existing cluster or provision a new one.
Installing the OpsCenter deb package
Install the DataStax OpsCenter using APT repositories on Debian or Ubuntu distributions.
For a complete list of supported platforms, see DataStax OpsCenter – Supported Platforms.
7
Installation
Before you begin
•
•
APT Package Manager is installed.
Python 2.6+
About this task
The OpsCenter Debian and Ubuntu packaged releases runs as a service from root. The service
initialization script is located in /etc/init.d. If the machine reboots, OpsCenter restarts automatically.
Procedure
1. Modify the aptitude repository source list file (/etc/apt/sources.list.d/
datastax.community.list).
$ echo "deb http://debian.datastax.com/community stable main" | sudo tee a /etc/apt/sources.list.d/datastax.community.list
2. Add the DataStax repository key to your aptitude trusted keys:
$ curl -L http://debian.datastax.com/debian/repo_key | sudo apt-key add 3. Install the OpsCenter package using the APT Package Manager:
$ apt-get update
$ apt-get install opscenter
For most users, the out-of-box configuration should work just fine, but if you need to you can configure
OpsCenter differently.
4. Start OpsCenter:
$ sudo service opscenterd start
5. Connect to OpsCenter in a web browser using the following URL:
http://opscenter-host:8888/
6. Next you can add an existing cluster or provision a new one.
Installing the OpsCenter on Mac OS X or any Linux distribution
Install the DataStax OpsCenter on Mac OS X or any Linux Distribution using the OpsCenter binary tarball.
About this task
For a complete list of supported platforms, see DataStax OpsCenter – Supported Platforms.
Before you begin
•
Python 2.6+
Procedure
1. Download the tarball distribution of OpsCenter.
$ curl -L http://downloads.datastax.com/community/opscenter.tar.gz | tar xz
Files for OpsCenter and a single DataStax agent are now in place.
2. Change to the opscenter<version-number> directory.
$ cd opscenter-<version-number>
3. Start OpsCenter from the install location:
8
Installation
$ bin/opscenter
Note: Use bin/opscenter -f to start OpsCenter in the foreground.
4. Connect to OpsCenter in a web browser using the following URL:
http://opscenter-host:8888/
5. Next you can add an existing cluster or provision a new one.
Installing DataStax agents
DataStax agents must be installed on every managed node in a cluster and are necessary to perform most
of the functionality within OpsCenter.
About this task
Once you've added the cluster to OpsCenter, you'll see the status of the agents in the Dashboard.
If your cluster is running DSE 3.2 or higher, the agents are already installed and should automatically
connect. If you don't see that the agents have successfully connected to OpsCenter on the Dashboard
after about five minutes, or if you are running DSE 3.1.x or lower, select Fix on the Dashboard.
Before you begin
•
•
•
Root or sudo access to the machines where the agents will be installed.
JMX connectivity is enabled on each node in the cluster.
You either configured the SSH port, or you accepted the default SSH port (22) for node-agent
communication.
Procedure
1. Open a browser window and go to the OpsCenter URL at http://<opscenter_host>:8888 where
<opscenter_host> is the IP or hostname of the OpsCenter machine.
When you start OpsCenter for the first time, you will be prompted to connect to your cluster:
2. Open a browser window and go to the OpsCenter URL at http://<opscenter_host>:8888/
where <opscenter_host> is the IP or hostname of the OpsCenter machine.
http://110.123.4.5:8888/
9
Installation
3. In Add Cluster, enter the Hostnames or IP addresses of two or three nodes in the cluster and set the
JMX set JMX and Thrift ports credentials, and then click Save Cluster.
After OpsCenter connects to the cluster, a Fix link appears near the top of the Dashboard.
4. Start installing the agents by clicking the Fix.
5. In Install Node Agents, click Enter Credentials.
6. In Node SSH Credentials, enter a username that has root privileges or sudo access to all of the nodes
in your cluster, plus any other required credentials, and then click Done.
7. In the Install Nodes Agent dialog, click Install on all nodes.
8. If prompted, click Accept Fingerprint to add a node to the known hosts for OpsCenter.
Results
DataStax agents have been deployed and configured for each managed node in the cluster.
If you are unable to install the agents through the OpsCenter UI, you can manually install the agents by
following these instructions.
10
Installation
Manually deploying agents - tarball
Install agents on nodes running Cassandra or DataStax Enterprise clusters.
Before you begin
•
•
•
•
Your Cassandra or DataStax Enterprise cluster is up and running.
OpsCenter is installed and configured.
JMX connectivity is enabled on each node in the cluster.
SYSSTAT Utilities (needed for the collection of I/O metrics).
Procedure
1. Download the DataStax agent tarball, expand and unarchive it.
$ curl -L http://downloads.datastax.com/community/datastax-agent-<versionnumber>.tar.gz | tar xz
2. Change into the agent directory.
$ cd datastax-agent-<version-number>
3. In address.yaml set stomp_interface to the IP address that OpsCenter is using. (You may have to
create the file.)
$ echo "stomp_interface: <reachable_opscenterd_ip>" >> ./conf/address.yaml
4. If SSL communication is enabled in <install_location>/conf/opscenterd.conf, use SSL in
address.yaml.
$ echo "use_ssl: 1" >> ./conf/address.yaml
5. Start the agent.
$ bin/datastax-agent
Use the -f flag to run in the foreground.
Manually deploying agents - rpm
You can install agents on Linux nodes using Yum packages.
Before you begin
•
•
•
•
Root or sudo access to the machines where the agents will be installed.
Your Cassandra or DataStax Enterprise cluster is up and running.
OpsCenter is installed and configured.
JMX connectivity is enabled on each node in the cluster.
Procedure
In a terminal for both 32- and 64-bit systems:
1. Add the DataStax Yum repository in the /etc/yum.repos.d/datastax.repo file.
[datastax]
name = DataStax Repo for Apache Cassandra
baseurl = http://rpm.datastax.com/community
enabled = 1
gpgcheck = 0
2. Install the DataStax agent.
11
Installation
# yum install datastax-agent
3. In address.yaml set stomp_interface to the IP address that OpsCenter is using. (You may have to
create the file.)
$ echo "stomp_interface: <reachable_opscenterd_ip>" | sudo tee -a /var/lib/
datastax-agent/conf/address.yaml
4. If SSL communication is enabled in /etc/opscenter/opscenterd.conf, use SSL in
address.yaml.
$ echo "use_ssl: 1" | sudo tee -a /var/lib/datastax-agent/conf/address.yaml
5. Start the DataStax agent.
$ sudo service datastax-agent start
Manually deploying agents - deb
You can install agents on Linux nodes using APT packages.
Before you begin
•
•
•
•
Root or sudo access to the machines where the agents will be installed.
Your Cassandra or DataStax Enterprise cluster is up and running.
OpsCenter is installed and configured.
JMX connectivity is enabled on each node in the cluster.
Procedure
1. Add the DataStax repository to the /etc/apt/sources.list.d/datastax.community.list file
(if you have already not done so).
echo "deb http://debian.datastax.com/community stable main" | \
sudo tee -a /etc/apt/sources.list.d/datastax.community.list
2. Add the DataStax repository key to your Aptitude trusted keys.
$ curl -L https://debian.datastax.com/debian/repo_key | sudo apt-key add Note: If you have trouble adding the key, use http instead of https.
3. Install the DataStax agent.
$ sudo apt-get update
$ sudo apt-get install datastax-agent
4. In address.yaml set stomp_interface to the IP address that OpsCenter is using. (You may have to
create the file.)
$ echo "stomp_interface: <reachable_opscenterd_ip>" | sudo tee -a /var/lib/
datastax-agent/conf/address.yaml
5. If SSL communication is enabled in /etc/opscenter/opscenterd.conf, use SSL in
address.yaml.
$ echo "use_ssl: 1" | sudo tee -a /var/lib/datastax-agent/conf/address.yaml
6. Start the DataStax agent.
$ sudo service datastax-agent start
12
Installation
Automatic installation of DataStax agents
When you install DataStax Enterprise 4.0 or later, the DataStax agent is installed automatically for you on
the nodes of your cluster.
About this task
After installing, configuring, and running a DSE cluster and OpsCenter, you connect to OpsCenter in a
web browser and are automatically asked wether to provision a new cluster or connect to an existing one.
In either case. OpsCenter will connect to nodes with agents already deployed (DSE 4.0 and greater) or
deploy agents to nodes and then connect.
Procedure
1.
2.
3.
4.
Install DSE 4.0 or greater.
Install OpsCenter 4.0 or greater.
Start your cluster and the OpsCenter daemon.
Open the URL for OpsCenter in your a web browser.
http://localhost:8888/
A dialog displays giving you two options.
5. Select Using Existing Cluster.
A dialog displays.
6. Add the hostnames or IP addresses of the nodes in your cluster.
localhost
7. Select Save Cluster.
Configuring JAVA_HOME for DataStax agents
DataStax agents do not pick up the environment variables of the currently logged-in user by default.
About this task
For example, if Java is not in the machine's PATH, you may notice errors in the agent log on start-up:
13
Installation
nohup: cannot run command 'java': No such file or directory
Procedure
•
On the Cassandra nodes where the agents are installed, create the file /etc/default/datastaxagent and set the environment variables for JAVA_HOME and any other custom environment
variables that the agent may need. For example:
export JAVA_HOME = /usr/bin/java
OpsCenter and DataStax agent ports
A list of the default port numbers used by OpsCenter and the DataStax Agents:
Port
Description
OpsCenter ports
8888
OpsCenter website. The opscenterd daemon listens on this port for HTTP requests
coming directly from the browser. Configurable in opscenterd.conf.
50031
OpsCenter HTTP proxy for Job Tracker. The opscenterd daemon listens on this port for
incoming HTTP requests from the browser when viewing the Hadoop Job Tracker page
directly. (DataStax Enterprise only)
61620
OpsCenter monitoring port. The opscenterd daemon listens on this port for TCP traffic
coming from the agent.
DataStax agents ports (on the monitored nodes)
7199
JMX monitoring port. Each agent opens a JMX connection to its local node (the
Cassandra or DataStax Enterprise process listening on this port). The JMX protocol
requires that the client then reconnect on a randomly chosen port (1024+) after the initial
handshake.
8012
Hadoop Job Tracker client port. The Job Tracker listens on this port for job submissions
and communications from task trackers; allows traffic from each Analytics node in a
DataStax Enterprise cluster.
8012
Hadoop Job Tracker Thrift port. The Job Tracker listens on this port for Thrift requests
coming from the opscenterd daemon. (DataStax Enterprise only)
8012
Hadoop Job Tracker website port. The Job Tracker listens on this port for HTTP requests.
If initiated from the OpsCenter UI, these requests are proxied through the opscenterd
daemon; otherwise, they come directly from the browser. (DataStax Enterprise only)
8012
Hadoop Task Tracker website port. Each Task Tracker listens on this port for HTTP
requests coming directly from the browser and not proxied by the opscenterd daemon.
(DataStax Enterprise only)
61621
DataStax agent port. The agents listen on this port for SSL traffic initiated by OpsCenter.
22
SSH port. Configurable in opscenterd.conf.
Solr Port and Demo applications port
8983
Solr Port and Demo applications port.
Cassandra client port
9160
14
Each agent makes Thrift requests to its local node on this port. Additionally, the port can
be used by the opscenterd daemon to make Thrift requests to each node in the cluster.
Installation
Installation and configuration locations
Debian and Ubuntu Package install locations
File locations for Debian and Ubuntu package installs.
Directory
Description
/var/lib/opscenter
SSL certificates for encrypted agent/dashboard
communications
/var/log/opscenter
Log directory
/var/run/opscenter
Runtime files
/usr/share/opscenter
JAR, agent, web application, and binary files
/etc/opscenter
Configuration files
/etc/init.d
Service start-up script
CentOS, OEL, and RHEL Package install locations
File locations for RHEL-based package installs.
Directory
Location
/var/lib/opscenter
SSL certificates for encrypted agent/dashboard
communications
/var/log/opscenter
Log directory
/var/run/opscenter
Runtime files
/usr/share/opscenter
JAR, agent, web application, and binary files
/etc/opscenter
Configuration files
/etc/init.d
Service startup script
Binary Tarball distribution install locations
File locations for binary-based installs.
Directory
Location
/agent
Agent installation files
/bin
Startup and configuration binaries
/content
Web application files
/conf
Configuration files
/doc
License files
/lib and /src
Library files
/log
OpsCenter log files
/ssl
SSL files for OpsCenter to agent communications
15
Installation
Starting, restarting, and stopping OpsCenter
Commands for each type of installation.
About this task
Packaged installations include startup scripts for running OpsCenter as a service. The available service
opscenterd options are:
service opscenterd start|stop|status|restart|force-reload
Procedure
The following list shows start, stop, and restart instructions for the supported platforms:
• To start DataStax OpsCenter:
•
•
•
Packaged installs: sudo service opscenterd start
Tarball installs: bin/opscenter (Use -f to start in the foreground.)
Windows installs: Start the OpsCenter Service from the Control Panel.
Note: By default, DataStax Enterprise services on Windows start automatically.
•
To stop DataStax OpsCenter:
•
•
•
Packaged installs: sudo service opscenterd stop
Tarball installs: Find the OpsCenter Java process ID (PID) and kill the process using its PID
number:
ps -ef | grep opscenter
sudo kill <pid>
• Windows installs: Stop the OpsCenter Service from the Control Panel.
To restart DataStax OpsCenter:
•
Packaged installs:
•
sudo service opscenterd restart
Tarball installs:
Find the OpsCenter process ID (<pid>), kill the process using its PID number, and then start the
OpsCenter:
ps -ef | grep opscenter
sudo kill <pid>
•
bin/opscenter (Use -f to start in the foreground.)
Windows installs:
Restart the OpsCenter Service from the Control Panel.
Starting and restarting DataStax agents
Commands for each type of installation.
Procedure
•
To start the DataStax agent:
•
•
•
16
Packaged installs: The DataStax agent starts automatically.
Tarball installs: $ install_location bin/datastax-agent (Use -f to start in the
foreground.)
Windows installs: Start the DataStax Agent Service from the Control Panel.
Installation
•
To restart the DataStax agent:
•
•
Packaged installs: $ sudo service datastax-agent restart
Tarball installs: Find the DataStax agent Java process ID (PID), kill the process using its PID
number, and then start the DataStax agent:
ps -ef | grep datastax-agent
sudo kill <pid>
•
$ bin/datastax-agent (Use -f to start in the foreground.)
Windows installs: Restart the DataStax Agent Service from the Control Panel.
17
Upgrading OpsCenter
Upgrading OpsCenter
See the Upgrade Guide for detailed instructions on upgrading OpsCenter.
18
Configuration
Configuration
Configuring user access
By default, access control is disabled. Any user that knows the OpsCenter URL can view all objects and
perform all tasks. To control access, you configure authentication for OpsCenter users by performing these
tasks:
•
•
•
Assign passwords.
Add users.
Set access roles using the set_passwd.py utility.
About user access roles
OpsCenter provides two access roles: admin and user.
Admin role privileges
•
Alerts
•
• add
• delete
• modify
Cluster operations
•
• add nodes to a cluster
• configure the cluster (all at once rather than a single node at a time)
• rebalance
• restart the cluster
Column families
•
• add column metadata
• create
• delete column metadata
• delete index
• drop
• truncate
• modify
Keyspaces
•
• create
• drop
• modify
Node
•
•
•
•
•
•
•
•
cleanup
compact
configure
decommission
drain
flush
move
perform garbage collection
19
Configuration
•
•
• repair
• restart
• start
• stop
Install the DataStax agent on Cassandra nodes
OpsCenter configuration
•
•
• add an existing cluster to OpsCenter
• delete a cluster from OpsCenter
• edit the config for a cluster
OpsCenter is monitoring
Provisioning
•
•
•
• add nodes to an existing cluster
• provision a new cluster (local or EC2)
Run a one-off backup
Run a restore of a backup
Scheduled backups
•
•
•
add
delete
modify
User role privileges
Users assigned the user role can perform all other OpsCenter tasks.
Assigning or modifying user access roles
The first time you assign an access role to an administrator or user, OpsCenter generates a password
file and enables access control. Authentication is required to access OpsCenter for viewing objects and
performing tasks.
About this task
To create or modify access roles:
Procedure
1. Run the set_passwd.py utility.
For example, to create user johndoe with admin role privileges:
$ python /usr/share/opscenter/bin/set_passwd.py johndoe admin
Please enter a password
Password:
for 'johndoe'.
2. After configuring authentication, restart OpsCenter.
$ service opscenterd restart
Restarting is required only when you create the first user (because it enables access control). No restart
is required for adding, modifying, or removing users.
Removing a user
About this task
To remove a user:
20
Configuration
Procedure
1. Edit the OpsCenter password file:
• Packaged installs: /etc/opscenter/.passwd
• Binary installs: <install_dir>/passwds
2. Delete the line of the user that you want to remove (<username>:<password_hash>:<role>).
johndoe:5e8848...42d8:admin
Restarting is not required to remove a user. Restarting is required to delete the password file. Deleting
the password file disables access control.
Note: If you delete all the users, you cannot access OpsCenter.
Configuring SSL
OpsCenter uses Secure Socket Layer (SSL) to encrypt the communication protocol and authenticate traffic
between DataStax agents and the main OpsCenter daemon.
By default SSL is disabled, which should only be done if you are running OpsCenter and DataStax
Enterprise under the following conditions:
•
•
•
On a secure internal network.
In a development environment where agents and OpsCenter run on the same computer free from
network threats.
In a situation where you are not concerned about someone listening to OpsCenter traffic.
Otherwise, you should enable SSL.
Enabling SSL - package installations
To enable SSL for package installations, you edit the confiugration file and run a script to generate the
keys used by OpsCenter and the agents.
Before you begin
•
The Python interface for the OpenSSL library (pyOpenSSL). With package installs (rpm or deb) of
OpsCenter, the python-openssl package is installed as a dependency. However, this is not the case
with CentOS 5.x installs.
Procedure
1. Ensure that a version of pyOpenSSL compatible with the version of libssl installed is a requirement
for any secure communications in OpsCenter.
•
•
Opscenter ships with pyOpenSSL 0.10, compiled for libssl 0.9.8, so if you are using libssl
0.9.8 on the machine running opscenterd, no further action should be required.
If you are using libssl 1.x, you need to ensure that pyOpenSSL 0.10+ is installed and compiled
properly.
a) (Optional) Determine the version of pyOpenSSL installed.
$ python -c "import OpenSSL; print OpenSSL.__version__"
b) (Optional) Manually install pyOpenSSL.
$ sudo easy_install pyOpenSSL
2. Run the OpsCenter setup.py script:
$ sudo /usr/share/opscenter/bin/setup.py
21
Configuration
The script generates the SSL keys and certifcates used by the OpsCenter daemon and the agents to
communicate with one another in the following directory.
/var/lib/opscenter
3. Open opscenterd.conf in an editor and add two lines to enable SSL.
$ sudo vi /etc/opscenter/opscenterd.conf
[agents]
use_ssl = true
4. Restart the OpsCenter daemon.
If you want to connect to a cluster in which agents have already been deployed, you can log in to each of
the nodes and reconfigure the address.yaml file (see steps below).
5. Reconfigure the agents on all nodes.
a) Copy /var/lib/opscenter/ssl/agentKeyStore from the OpsCenter machine to /var/lib/
datastax-agent/ssl/agentKeyStore on each node in the cluster.
$ scp /var/lib/opscenter/ssl/agentKeyStore [email protected]<node>:/var/lib/datastaxagent/ssl/
Where <node> is either the host name of the node or its IP address and <user> is the userid on the
node.
b) Log into each node in the cluster using ssh.
$ ssh <user>@<node>
c) Edit the address.yaml file, changing the value of use_ssl to 1.
$ sudo vi /var/lib/datastax-agent/conf/address.yaml
use_ssl: 1
d) Restart the agent.
$ sudo service datastax-agent restart
If you do not want to edit all the node configuration files by hand, you can follow the agent installation
procedure.
6. Once opscenterd and all agents have been configured and restarted, verify proper connection via the
dashboard.
22
Configuration
Enabling SSL - tarball installations
To enable SSL for tarball installations, you edit the confiugration file and run a script to generate the keys
used by OpsCenter and the agents.
Procedure
1. Ensure that a version of pyOpenSSL compatible with the version of libssl installed is a requirement
for any secure communications in OpsCenter.
•
•
Opscenter ships with pyOpenSSL 0.10, compiled for libssl 0.9.8, so if you are using libssl
0.9.8 on the machine running opscenterd, no further action should be required.
If you are using libssl 1.x, you need to ensure that pyOpenSSL 0.10+ is installed and compiled
properly.
a) (Optional) Determine the version of pyOpenSSL installed.
$ python -c "import OpenSSL; print OpenSSL.__version__"
b) (Optional) Manually install pyOpenSSL.
$ sudo easy_install pyOpenSSL
2. Run the OpsCenter setup.py script:
$ sudo <install_location>/bin/setup.py
The script generates the SSL keys and certifcates used by the OpsCenter daemon and the agents to
communicate with one another in the following directory.
<install_location>/ssl
3. Open opscenterd.conf in an editor and add two lines to enable SSL.
$ sudo vi <install_location>/opscenterd.conf
[agents]
use_ssl = true
4. Restart the OpsCenter daemon.
23
Configuration
If you want to connect to a cluster in which agents have already been deployed, you can log in to each of
the nodes and reconfigure the address.yaml file (see steps below).
5. Reconfigure the agents on all nodes.
a) Log into each node in the cluster using ssh.
$ ssh <user>@<node>
Where <node> is either the host name of the node or its IP address and <user> is the userid on the
node.
b) Edit the address.yaml file, changing the value of use_ssl to 1.
$ sudo vi <install_location>/conf/address.yaml
use_ssl: 1
c) Restart the agent.
$ sudo <install_location>/bin/datastax-agent
If you do not want to edit all the node configuration files by hand, you can follow the agent installation
procedure.
6. Once opscenterd and all agents have been configured and restarted, verify proper connection via the
dashboard.
Disabling SSL - package installations
To disable SSL for package installations, you modify the OpsCenter configuration file and restart
OpsCenter.
About this task
By default SSL is turned off in OpsCenter. You would only need to perform this task if you have configured
the agents on a cluster to use SSL earlier and now wished to turn SSL off.
Procedure
1. Open opscenterd.conf in an editor and add two lines to enable SSL.
24
Configuration
$ sudo vi /etc/opscenter/opscenterd.conf
[agents]
use_ssl = false
2. Restart the OpsCenter daemon.
3. Reconfigure the agents.
a) Log into each node in the cluster using ssh.
$ ssh <user>@<node>
b) Edit the address.yaml file, changing the value of use_ssl to 0.
$ sudo vi /var/lib/opscenter/address.yaml
use_ssl: 0
c) Restart the agent.
$ sudo service datastax-agent restart
If you do not want to edit all the node configuration files by hand, you can follow the agent installation
procedure.
4. Once opscenterd and all agents have been configured and restarted, verify proper connection via the
dashboard.
Disabling SSL - tarball installations
To disable SSL for tarball installations, you modify the OpsCenter configuration file and restart OpsCenter.
About this task
By default SSL is turned off in OpsCenter. You would only need to perform this task if you have configured
the agents on a cluster to use SSL earlier and now wished to turn SSL off.
Procedure
1. Open opscenterd.conf in an editor and add two lines to enable SSL.
25
Configuration
$ vi <install_location>/conf/opscenterd.conf
[agents]
use_ssl = false
2. Restart the OpsCenter daemon.
3. Reconfigure the agents.
a) Log into each node in the cluster using ssh.
$ ssh <user>@<node>
Where <node> is either the host name of the node or its IP address and <user> is the userid on the
node.
b) Edit the address.yaml file, changing the value of use_ssl to 0.
$ sudo vi <install_location>/conf/address.yaml
use_ssl: 0
c) Restart the agent.
$ sudo <install_location>/bin/datastax-agent
If you do not want to edit all the node configuration files by hand, you can follow the agent installation
procedure.
4. Once opscenterd and all agents have been configured and restarted, verify proper connection via the
dashboard.
Enabling HTTPS
You can enable or disable Hypertext Transfer Protocol Secure (HTTPS) support in OpsCenter.
Procedure
1. Open the OpsCenter configuration file, opscenterd.conf, located in one of these directories:
26
Configuration
•
•
•
Package installations: /etc/opscenter/opscenterd.conf
Binary tarball installations (Linux and Mac OSX): install_location/conf/opscenterd.conf
Windows installations: Program Files (x86)\DataStax Community\opscenter\conf
\opscenterd.conf
2. Scroll to the [webserver] section.
This snippet from opscenterd.conf shows the [webserver] section that you change:
[webserver]
port = 8888
interface = 127.0.0.1
# The following settings can be used to enable ssl support for the
opscenter
# web application. Change these values to point to the ssl certificate and
key
# that you wish to use for your OpsCenter install, as well as the port you
would like
# to serve ssl traffic from.
#ssl_keyfile = /var/lib/opscenter/ssl/opscenter.key
#ssl_certfile = /var/lib/opscenter/ssl/opscenter.pem
#ssl_port = 8443
3. Remove the comment markers (#) in front of ssl_keyfile, ssl_certfile, and ssl_port.
You can use the default values for the ssl_keyfile and ssl_certfile or replace them with the path to your
own private and public certificates.
4. Save opscenterd.conf and restart OpsCenter.
Configuring events and alerts
The OpsCenter Event Log page displays a continuously updated list of events and alerts.
The following list reflects the most detailed logging level available for Cassandra, DataStax Enterprise, and
OpsCenter events:
•
•
•
•
•
•
DEBUG(0)
INFO (1)
WARN (2)
ERROR (3)
CRITICAL (4)
ALERT (5)
Events
Data for these events is stored in the events and events_timeline column families in the OpsCenter
keyspace:
Event
Code
Description
COMPACTION
0
Major compaction has occurred.
CLEANUP
1
Unused keys have been removed or cleaned up.
REPAIR
2
A repair operation has been initiated.
FLUSH
3
Memtables have been flushed to disk.
DRAIN
4
The commit log has been emptied, or drained.
27
Configuration
Event
Code
Description
DECOMMISSION
5
A leaving node has streamed its data to another node.
MOVE
6
Like NODE_MOVE; a new token range has been assigned.
NODE_DOWN
13
A node has stopped responding.
NODE_UP
14
An unresponsive node has recovered.
NODE_LEFT
15
A node has left, or been removed from, the ring.
NODE_JOIN
16
A node has joined the ring.
NODE_MOVE
17
A node has been assigned a new token range (the token
has moved).
OPSC_UP
18
OpsCenter has been started and is operating.
OPSC_DOWN
19
OpsCenter was stopped or stopped running.
GC
20
Java garbage collection has been initiated.
Alerts
Optionally, you can configure OpsCenter to send alerts for selected levels of events. These alerts can be
provided remotely by email, or through HTTP to a selected URL. Alerts are disabled by default.
Alerts are triggered only by events from the OpsCenter API/UI. For example, a nodetool move operation
submitted from the command line does not trigger an alert. However, a move operation launched using
Dashboard > List View > Actions > Movecontrols in the OpsCenter does trigger an alert.
All alerts contain the following information about each event captured:
Field
Description
Example
api_source_ip
IP that originally sent the request.
67.169.50.240
target_node
Destination of a STREAMING action.
10.1.1.11
event_source
Component that caused the event.
OpsCenter (i.e., restart, start)
user
OpsCenter user that caused the event.
opscenter_user
time
Normal timestamp for the event.
1311025650414527
action
Type of event (see above table)
20
message
Description of the event.
Garbage Collecting node 10.1.1.13
level
Numerical code for the log level.
1
source_node
Node where the event originated.
10.1.1.13
level_str
Logging level of the event.
INFO
Enabling email alerts
OpsCenter can post alerts to selected email addresses.
About this task
To enable email alerts, you must edit the <config_location>/event-plugins/email.conf file and
provide valid SMTP server host and port information. This file is located in the following directories:
28
Configuration
Procedure
1.
2.
3.
4.
5.
Make sure that you have valid SMTP mail accounts to send and receive alerts.
On the OpsCenter daemon host, open the email.conf file for editing.
Set enabled to 1.
Provide valid values for your SMTP host, port, user, and password.
Enable Secure Sockets Layer (SSL) or Transport Layer Security (TLS) protocol on your system if you
want secure communications. Typically, SSL is required.
6. Provide valid values for the to_addr and from_addr email addresses. The to_addr value is the account
that will receive alerts.
7. Optionally, set the level of alerts to send and the desired subject line.
8. Save <config_location>/event-plugins/email.conf and restart the OpsCenter daemon.
To send alerts to multpile email addresses, create a different email conf file with settings for each email
address. All conf files are loaded so you can name them email1.conf, email2.conf, and so on.
Example
In a system with email alerts enabled for critical and alert-level events, <config_location>/eventplugins/email.conf looks like:
[email]
enabled = 1
# levels can be comma delimited list of any of the following: #
DEBUG,INFO,WARN,ERROR,CRITICAL,ALERT # If left empty, will listen for all
levels levels = CRITICAL,ALERT
smtp_host = smtp.gmail.com
smtp_port = 465
smtp_user = [email protected]
smtp_pass = *********
smtp_use_ssl = 1
smtp_use_tls = 0
to_addr = [email protected]
from_addr = [email protected]
subject = OpsCenter Event
Enabling alerts posted to a URL
OpsCenter can post alerts to a URL if you provide a correctly formatted POST script.
About this task
OpsCenter can be configured to send alerts within an HTTP POST request to a specified URL. For
example, a simple PHP script containing print_r($_POST); will echo the received POST request. An
example POST request is:
POST / HTTP/1.0
Host: localhost
User-Agent: Twisted PageGetter
Content-Length: 184
Content-type: application/x-www-form-urlencoded
connection: close
target_node=None&event_source=OpsCenter&success=None&level=1&level_str=INFO&api_source_i
+starting+up.&source_node=None
The request body contains fields described in Alerts.
29
Configuration
To enable URL posting on the OpsCenter side:
Procedure
1. Edit the posturl.conf file and provide a path to your script.
•
•
Package installations: /etc/opscenter/event-plugins
Binary tarball installations (Linux and Mac OSX): install_location/opscenter/conf/
event-plugins
• Windows installations: Program Files (x86)\DataStax Community\opscenter\conf
\event-plugins
2. Make sure your web server and posting script are configured to receive alerts.
3. On the OpsCenter daemon host, open posturl.conf for editing.
4. Set enabled to 1. For url, provide a valid path to your posting script.
url = http://50.1.1.11/postOPSCevents.php
5. Optionally, select the desired logging level. The default is to listen for all levels of events.
6. Save posturl.conf and restart the OpsCenter daemon.
Example
In a system with posting enabled for critical and alert-level events, posturl.conf looks like:
[posturl]
enabled = 1
url = http://10.1.1.11/postOPSCevents.php
# levels can be comma delimited list of any of the following:
# DEBUG,INFO,WARN,ERROR,CRITICAL,ALERT
# If left empty, will listen for all levels
levels = CRITICAL,ALERT
Verifying that events are posting correctly
You can set preferences to specify how the posting is handled on the receiving side.
About this task
Procedure
1. Post events to a file such as /tmp/events on the web server host.
2. Create a script.
URL: http://10.1.1.11/postOPSCevents.php
<?php
file_put_contents( '/tmp/events', print_r ( $_POST,true ),
FILE_APPEND );
?>
3. Deploy the script. You might need to restart the web server.
4. Launch a logged event, such as an OpsCenter restart or garbage compaction from Dashboard >
Cluster > List View.
Output to /tmp looks something like this:
30
Configuration
Array
( [api_source_ip ] => 67.169.50.240
[target_node ] => None
[event_source ] => OpsCenter
[user ] => None
[ time ] => 1311025598851602
[action ] => 20
[message ] => Garbage Collecting node 50.1.1.24
[level ] => 1
[source_node ] => 50.1.1.24
[level_str ] => INFO
)
Configuring data collection and expiration
OpsCenter collects system and column family metrics data for each node in your cluster.
OpsCenter creates its own keyspace within a cluster for storing collected metrics. This data can also be
stored on a cluster other than the one currently being managed by OpsCenter. Metrics data is collected
at regular intervals and stored within your cluster in a keyspace called OpsCenter. The column families
containing metric data continue to grow. You can configure how long you want to keep historical metrics.
Data expires after configurable time periods.
Estimating the amount of data generated
The following table provides guidance for estimating the amount of metrics data generated:
Number of days
Number of column families
monitored
MB per node
31
5
200
31
10
300
31
20
500
365
5
250
365
10
380
365
20
630
The default upper limit of data collected is 365 days.
Controlling data collection
To help control consumption of disk space, OpsCenter provides two ways to limit the growth of OpsCenter
performance data by:
•
•
Excluding specified keyspaces and column families from performance data collection
Shortening the time period after which performance data automatically expires
Excluding keyspaces and column families
By default, OpsCenter does not collect performance data for its own keyspace or the Cassandra system
keyspace. You can manually add any other keyspaces or column families that you do not want to monitor
in the [cassandra_metrics] section of the configuration file.
For example, to prevent data collection for the keyspace test as well as the column family
Keyspace1.Standard1, uncomment and edit the following values in the OpsCenter cluster configuration file
(<cluster_specific>.conf):
31
Configuration
[cassandra_metrics ]
ignored_keyspaces = system, OpsCenter, test
ignored_column_families = Keyspace1.Standard1
Column families are specified in the format:
<keyspace_name>.<column_family_name>.
Changing performance data expiration times
Performance data stored in OpsCenter expires after configurable time periods.
About this task
The default values are designed to provide efficient compaction and eventual deletion of the data, with
faster expiration times for the more granular, larger-volume data rollups.
•
•
•
One-minute rollups (1min_ttl) expire after after one week, or 604800 seconds.
Five-minute rollups (5min_ttl) expire after four weeks, or 2419200 seconds.
Two-hour rollups (2hr_ttl) expire after one year, or 31536000 seconds.
To change expiration time period:
In this example, the one-minute and five-minute rollups are set to expire twice as fast as the defaults, and
two-hour rollups are set to be kept indefinitely (expiration is disabled).
Procedure
1. Edit the conf/clusters/<cluster>.conf file.
2. Add the following time-to-live (ttl) values in seconds under a [cassandra_metrics] section:
[cassandra_metrics]
1min_ttl = 302400
5min_ttl = 1209600
2hr_ttl = -1
3. Restart OpsCenter.
Data collected after restarting OpsCenter expires according to the new setting. The data collected
before restarting OpsCenter expires according to the setting in effect when it was collected.
Storing collection data on a different cluster
If you do not want OpsCenter to store data in an OpsCenter keyspace on the cluster being managed, you
can store the data on a separate cluster. The seed nodes must be accessible without SSL or Kerberos
security.
Procedure
Add a section to the cluster configuration file.
•
•
/etc/opscenter/clusters/<MyCluster>.conf
install_location/conf/clusters/<MyCluster>.conf
[storage_cassandra]
seed_hosts = host1, host2
api_port = 9160
32
Configuration
Option
Description
seed_hosts
A comma-delimited list of at least one of the
nodes in the cluster where you want to store
collection data.
api_port
The Thrift port used by that cluster.
username
The appropriate username (if using Thrift
security).
password
The appropriate password (if using Thrift
security).
keyspace
The name of the keyspace to store the collection
data in. (OpsCenter by default.)
Configuring OpsCenter definition file updates
Definition files are used by OpsCenter to enable support for different versions of DataStax Enterprise,
DataStax Community, and Cassandra. They are updated independently of OpsCenter by downloading new
definitions from a central server at regular intervals.
OpsCenter ships with a set of files called definition files which can be updated independently of OpsCenter
itself, allowing additional support for newer versions of DataStax Enterprise, DataStax Community, and
Cassandra without the need to upgrade your version of OpsCenter.
The definition files are located in /etc/opscenter/conf/definitions for package installs, and
install_location/conf/definitions for tarball installs.
These definition files are updated every hour by default. The opscenterd process checks a central server
located at opscenter.datastax.com, and pulls down updates to the set of definition files specific to
that version of OpsCenter as needed.
This auto-update process can be disabled by setting auto_update to False in the [definitions]
section of opscenterd.conf. The interval can also be modified by setting the sleep option in the
[definitions] section of opscenterd.conf. The sleep option interval should be specified in
seconds.
Setting the update interval to 7200 seconds, or every 2 hours.
[definitions]
sleep = 7200
Advanced configuration
To configure advanced capabilities, you can manually modify the configuration files.
Note: The OpsCenter console is the most convenient way to configure basic OpsCenter settings.
•
•
opscenterd.conf: configures the properties for the OpsCenter daemon.
<cluster_specific>.conf: configures properties for each cluster monitored by OpsCenter. This
file is created when you add a cluster to the Opscenter.
33
Configuration
OpsCenter configuration properties
These properties are configured in the opscenterd.conf file. The location of the opscenterd.conf file
depends on the type of installation:
•
•
•
Packaged installs: /etc/opscenter/opscenterd.conf
Tarball installs: install_location/conf/opscenterd.conf
Windows installs: Program Files (x86)\DataStax Community\opscenter\conf
\opscenterd.conf
Note: After changing properties in this file, you must restart OpsCenter for the changes to take
effect.
[agents] ssh_port
The Secure Shell (SSH) port that listens for agent-OpsCenter communications. Add an [agents] section, if
one doesn't already exist, to the opscenterd.conf. In this section, add the ssh_port option and a value
for the port number:
[agents]
ssh_port = 2222
[agents] runs_sudo
Sets whether the DataStax Agent will be run using sudo or not. Setting this option to false means the agent
will not use sudo, and the agent user will not run using elevated privileges. Setting this option to true means
the agent will run using sudo, and elevated privileges.
The default setting is true.
[webserver] port
The HTTP port used for client connections to the OpsCenter web server. Default is 8888.
Optional HTTPS support. To enable, remove the comment markers (#) in front of properties prefixed with
ssl in the opscenterd.conf file, as described in Configuring HTTPS.
[webserver] interface
The interface that the web server uses to listen for client connections. The interface must be an externally
accessible IP address or host name.
[logging] level
The logging level for OpsCenter. Available levels are (from most to least verbose): TRACE, DEBUG, INFO,
WARN, or ERROR.
The OpsCenter log file is located in /var/log/opscenter/opscenterd.log.
[stat_reporter] interval
Reporting to DataStax Support. By default, OpsCenter periodically sends usage metrics about the cluster
to DataStax Support.
To disable the phone-home functionality, add the following lines to your opscenterd.conf file:
[stat_reporter]
interval = 0
Additional configuration metric collection properties are available in Metrics Collection Properties.
[authentication] passwd_file
Full path to the file for configuring password authentication for OpsCenter. If this file does not exist,
OpsCenter does not verify passwords.
To enable password authentication, use the set_passwd.py utility to create users and set their password
and role. OpsCenter currently has two available roles: admin or user.
[cassandra] auto_node_discovery
Enables or disables auto-discovery of nodes. When disabled, OpsCenter only attempts to contact nodes in
the seed list, and will not auto-discover nodes. By default this is True.
34
Configuration
Statistics reporter properties
A complete breakdown of the data OpsCenter communicates back to DataStax. The data is sent in a keyvalue JSON format.
The following information is recorded about the OpsCenter install:
install_id
This is a random uuid generated when OpsCenter starts for the first time. This is used for associating reports
from the same install.
is_paid
This is a flag indicating whether or not this is the free or enterprise version of OpsCenter.
opscenter_version
The version of OpsCenter in use.
opscenter_ram
The amount of RAM, in megabytes, on the OpsCenter machine.
opscenter_cores
The number of cores on the OpsCenter machine.
opscenter_os
The generic name of the operating system of the OpsCenter machine. For example, linux, windows, or
mac.
opscenter_os_sub
The specific name of the operating system of the OpsCenter machine. For example CentOS, Ubuntu, or
Debian.
opscenter_os_version
The operating system version of the OpsCenter machine.
opscenter_arch
The architecture of the OpsCenter machine.
opscenter_install_type
The type of install (package or tarball).
python_version
The version of python running on the OpsCenter machine.
opscenter_instance_type
The instance type the OpsCenter machine, if OpsCenter is running in EC2.
separate_storage
A flag indicating if OpsCenter is storing metrics in the cluster it is monitoring.
config_diff
A list of the OpsCenter config options that were modified to be different than the defaults. This includes the
names of the options that were changed but not the values of those options.
These statistics are collected about each cluster OpsCenter is monitoring:
cluster_id
An MD5 hash of the cluster name. Used for identifying unique clusters while maintaining anonymity.
conf_id
An MD5 hash of the file name the config for the cluster is stored in. Used for the same purposes as cluster_id.
partitioner
The partitioner the cluster is using.
snitch
The snitch the cluster is using.
keyspace_count
35
Configuration
The number of keyspaces in the cluster.
columnfamily_count
The number of column families in the cluster.
strategy_options
A list of the replication options used for each keyspace in the cluster.
cql3_cf_count
The number of column families created with CQL3 in the cluster.
node_count
The number of nodes in the cluster.
avg_token_count
The average number of tokens per node.
cassandra_versions
A list of the different Cassandra versions in the cluster.
bdp_version
A list of the different DataStax Enterprise versions in the cluster.
rack_map
A map of each rack in the cluster and how many nodes are in that rack.
dc_count
The number of datacenters in the cluster.
free_space
The amount of free disk space across the cluster.
used_space
The amount of used disk space across the cluster.
cluster_os
A list of the different operating systems used across the cluster.
cluster_ram
The average amount of ram per node in the cluster.
cluster_cores
The average number of cores per node in the cluster.
cluster_instance_types
A list of the EC2 instance types in the cluster, if EC2 is being used.
OpsCenter logging properties
Properties that control the location of OpsCenter's log files.
The following properties allow you to set the location of OpsCenter's log files.
[webserver] log_path
A log for all HTTP requests sent to opscenterd. By default /var/log/opscenter/http.log on
package installs, and install_location/logs/http.log on tarball installs.
[logging] log_path
The log file for opscenterd. By default /var/log/opscenter/opscenterd.log on package installs,
and install_location/logs/opscenterd.log on tarball installs.
[logging] log_length
Rotate the logs when the log file grows to this size, in bytes. By default, this is 10,000,000, or 10 MB.
[logging] max_rotate
The maximum number of logs to keep after rotation. By default, this is 10.
36
Configuration
[logging] level
The level of messages to log. Valid settings are TRACE, DEBUG, INFO, WARN, and ERROR, in descending
level of detail. By default, this is set to INFO.
[repair_service] log_directory
The directory that contains the log files for the repair service. By default /var/log/opscenter/
repair_service/ on package installs, and install_location/logs/repair_service/ on tarball
installs.
[repair_service] log_length
Rotate the logs when the log file grows to this size, in bytes. By default, this is 10,485,760, or 10 MB.
[repair_service] max_rotate
The maximum number of logs to keep after rotation. By default, this is 10.
Cluster configuration properties
These properties inform OpsCenter about the Real-time (Cassandra), Analytics (Hadoop), and Search
(Solr) nodes that it is monitoring.
Cassandra connection properties
These properties are configured in the cluster-specific opscenterd.conf file.
The location of the opscenterd.conf file depends on the type of installation:
•
•
•
Packaged installs: /etc/opscenter/clusters/cluster_specific.conf
Tarball installs: install_location/conf/clusters/cluster_specific.conf
Windows installs: Program Files (x86)\DataStax Community\opscenter\conf\clusters
\cluster_specific.conf
Note: After changing properties in this file, restart OpsCenter for the changes to take effect.
[jmx] port
The JMX (Java Management Extensions) port of your cluster. In Cassandra versions 0.8 and higher, the
JMX port is 7199.
[cassandra] seed_hosts
A Cassandra seed node is used to determine the ring topology and obtain gossip information about the
nodes in the cluster. This should be the same comma-delimited list of seed nodes as the one configured for
your Cassandra or DataStax Enterprise cluster by the seeds property in the cassandra.yaml configuration
file.
[cassandra] api_port
The Thrift remote procedure call port configured for your cluster. Same as the rpc_port property in the
cassandra.yaml configuration file. Default is 9160.
[cassandra] install_location
The directory in which Cassandra is installed. If install_location is not specified, OpsCenter looks in the
package-specific installation locations. For a tarball installation of DataStax Enterprise, the install_location
is <dse_install_location>/resources/cassandra.
[cassandra] conf_location
The location of the cassandra.yaml configuration file. If install_location is specified, but conf_location is
not, then conf_location is assumed to be install_location/conf/cassandra.yaml. If conf_location
is specified, it must be the absolute path to the Cassandra configuration file on all nodes. These settings
are cluster-wide and require that the specified locations be correct for every node.
Metrics Collection Properties
These properties are used to limit the keyspaces and column families for which you collect metrics.
[cassandra_metrics] ignored_keyspaces
37
Configuration
A comma-delimited list of Cassandra keyspaces for which you do not want to collect performance metrics.
By default, the system and OpsCenter keyspaces are excluded.
[cassandra_metrics] ignored_column_families
A comma-delimited list of Cassandra column families for which you do not want to collect performance
metrics. Entries should be in the form of keyspace_name.columnfamily_name.
Performance Data Expiration Properties
These properties set the expiration time for data stored in the OpsCenter keyspace. Each time period for
rolling up data points into summary views has a separate expiration threshold, or time-to-live (ttl) value
expressed in seconds. By default, shorter time periods have lower values that result in more efficient
expiration and compaction of the relatively larger volumes of data.Uncomment these properties to change
the default expiration periods for performance data. Properties and default values are:
1min_ttl = 604800
One-minute rollups expire after after one week, or 604800 seconds.
5min_ttl = 2419200
Five-minute rollups expire after four weeks, or 2419200 seconds.
2hr_ttl = 31536000
Two-hour rollups expire after one year, or 31536000 seconds.
DataStax Agent configuration
The address.yaml configuration file
The address.yaml file is located in /var/lib/datastax-agent/conf on package installations, and
in <install location>/conf on tarball installations. It contains configuration options for the DataStax
Agent.
Options
stomp_interface
Reachable IP address of the opscenterd machine. The connection made will be on stomp_port
stomp_port
The stomp_port used by opscenterd. The default setting is 61620.
local_interface
The IP used to identify the node. If broadcast_address is set in cassandra.yaml, this should be the same
as that; otherwise, it is typically the same as listen_address in cassandra.yaml. A good check is to confirm
that this address is the same as the address that nodetool ring outputs.
agent_rpc_interface
The IP that the agent HTTP server listens on. In a multiple region deployment, this is typically a private IP.
agent_rpc_broadcast_address
The IP that the central OpsCenter process uses to connect to the DataStax agent.
use_ssl
Whether or not to use SSL communication between the agent and opscenterd. Affects both the STOMP
connection and agent HTTP server. Corresponds to [agents].use_ssl in opscenterd.conf. Setting
this option to true turns turns on SSL connections. The default setting is 0.
cassandra_conf
The agent will attempt to auto-detect the location of the cassandra.yaml file via JMX, but if it cannot this
needs to be set to the full path of cassandra.yaml. By default /etc/cassandra/cassandra.yaml on
package installs or <install_location>/conf/cassandra.yaml on tarball installs.
cassandra_install_location
38
Configuration
The location where Cassandra is installed for tarball installs if OpsCenter is unable to auto-detect the install
location.
cassandra_log_location
The location of Cassandra's system.log file. This is only used for the diagnostics tarball, and should only
be set if system.log is in a non-standard location.
Advanced options
metrics_enabled
Whether or not to collect and store metrics for the local node. Setting this option to false turns off metrics
collection. The default setting is true.
thrift_port
Port used to connect to local thrift server. The default setting is 9160. This information will be sent by
opscenterd for convenience, but can be configured locally as needed.
jmx_host
Host used to connect to local JMX server. The default setting is localhost. This information will be sent
by opscenterd for convenience, but can be configured locally as needed.
jmx_port
Port used to connect to local JMX server. The default setting is 7199. This information will be sent by
opscenterd for convenience, but can be configured locally as needed.
api_port
Port the local HTTP server will bind to. The default setting is 61621. This option needs to be identical across
all agents, and set explicitly in opscenterd.conf if changed.
runs_sudo
Sets whether the DataStax Agent will be run using sudo or not. Setting this option to false means the
agent will not use sudo, and the agent user will not run using elevated privileges. Setting this option to true
means the agent will run using sudo, and elevated privileges.
The default setting is true.
OpsCenter updater properties
These properties are for configuring the OpsCenter updater, which updates the definition files that enable
support for different releases of DataStax Enterprise, DataStax Community, and Cassandra.
[definitions] auto_update
Enables the updater process when set to true, and disables the updater process when false. The default
value is true.
[definitions] sleep
The interval in seconds between updates. The default value is 3600, or 1 hour.
Backing up opscenterd's configuration files for failover
This task describes how to manually backup the configuration files of opscenterd in case of a failure of
the primary opscenterd instance.
About this task
There is no automatic failover for OpsCenter. However, both opscenterd and the agents it manages can
be restarted independently of each other. Most of the data used by OpsCenter is stored in the DataStax
Enterprise cluster, allowing you to start a secondary opscenterd instance in case the primary instance
fails. In this case, the secondary instance should match the configuration of the primary instance. The
39
Configuration
IP address in the agent's stomp_interface setting in address.yaml needs to be updated to the
secondary opscenterd instance.
Procedure
1. Copy the contents of the SSL configuration directory to the secondary machine.
•
•
/var/lib/opscenter/ssl (package installs)
install_location/ssl (tarball installs)
$ scp /var/lib/opscenter/ssl/* secondary:/var/lib/opscenter/ssl
2. Copy the contents of the main configuration directory to the secondary machine.
•
•
/etc/opscenter (package installs)
install_location/conf (tarball installs)
$ scp /etc/opscenter/* secondary:/etc/opscenter
3. Update each agent's stomp_interface in /var/lib/datastax-agent/conf/address.yaml
(package installs) or install_location/conf (tarball installs) to use the secondary machine's IP
address.
stomp_interface: secondary_opscenterd_ip
Using custom start/stop scripts for DataStax Enterprise and Cassandra
OpsCenter allows users to start and stop the DataStax Enterprise/Cassandra process on each node in
a visual way. The agent will attempt to automatically determine the best way to do this, but cannot not
do so in all cases. You can customize the startup or shutdown of a node using the start-cassandra
and stop-cassandra scripts located in /usr/share/datastax-agent/bin (package installs) or
<install_location>/bin (tarball installs).
Procedure
1. Rename the example script in /usr/share/datastax-agent/bin (package installs) or
<install_location>/bin (tarball installs) to remove the .example extension.
•
•
start-cassandra.example: example startup script
stop-cassandra.example: example shutdown script
$ cd /usr/share/datastax-agent/bin
$ mv start-cassandra.example start-cassandra
2. Edit the script to customize the behavior. The script should return an exit code of 0 when successful,
and a non-zero value if it fails.
3. Make the script executable.
$ chmod 755 start-cassandra
Example scenarios
Examples of OpsCenter deployments.
Configuring for multiple regions
OpsCenter can operate in multiple regions or IP forwarding deployments. Use the following approach for
deployments where a public IP forwards to a private IP on the agent, but that machine is not aware of (that
is, can't bind to) the public IP.
40
Configuration
About this task
To configure DataStax agents for multiple regions or IP forwarding:
Procedure
1. Open the address.yaml file for editing.
•
Packaged installs:
•
/var/lib/datastax-agent/conf directory
Tarball installs:
install_location/conf directory
2. Add the following option to the address.yaml file:
•
local_interface: (Optional) The IP used to identify the node. If broadcast_address is set in
cassandra.yaml, this should be the same as that; otherwise, it is typically the same as
listen_address in cassandra.yaml. A good check is to confirm that this address is the same as the
address that nodetool ring outputs.
• agent_rpc_interface: The IP that the agent HTTP server listens on. In a multiple region
deployment, this is typically a private IP.
• agent_rpc_broadcast_address: The IP that the central OpsCenter process uses to connect to the
DataStax agent.
3. Repeat the above steps for each node.
Example
here is the configuration for a three node cluster that spans two regions:
Region: us-west
Availability Zone: us-west-2
Node1
public IP: 198.51.100.1
private IP: 10.11.12.1
Cassandra (cassandra.yaml)
broadcast_address: 198.51.100.1
listen_address:
10.11.12.1
Agent (address.yaml)
local_address:
198.51.100.1
agent_rpc_interface:
10.11.12.1
agent_rpc_broadcast_address: 198.51.100.1
OpsCenter (opscenterd.conf )
interface: 198.51.100.1
Node2
public IP: 198.51.100.23
private IP: 10.11.12.15
Cassandra (cassandra.yaml)
broadcast_address: 198.51.100.23
listen_address:
10.11.12.15
Agent (address.yaml)
local_address:
198.51.100.23
agent_rpc_interface:
10.11.12.15
agent_rpc_broadcast_address: 198.51.100.23
Region: us-east
Availability Zone: us-east-1
Node1
41
Configuration
public IP: 203.0.113.20
private IP: 10.11.13.28
Cassandra (cassandra.yaml)
broadcast_address: 203.0.113.20
listen_address:
10.11.13.28
Agent (address.yaml)
local_address:
203.0.113.20
agent_rpc_interface:
10.11.13.28
agent_rpc_broadcast_address: 203.0.113.20
42
Using OpsCenter
Using OpsCenter
OpsCenter is a Web application for monitoring and administering all nodes in a Cassandra cluster from one
centralized console. It runs on the client-side in a web browser.
Overview
The major areas of functionality.
At the top of every functional area of OpsCenter, you can access these functions:
•
•
•
New Cluster: create a new cluster or add an existing cluster.
Feedback: an online form that sends your evaluation of OpsCenter or any comments to us.
Report: information about clusters that OpsCenter manages in PDF format.
OpsCenter is divided into these main functional areas:
•
•
•
•
•
•
•
•
•
•
•
Overview - Survey each cluster's Dashboard in this condensed view. Displayed when multiple clusters
are present.
Dashboard - View graphs of the most commonly watched Cassandra performance metrics.
Software update notification - Be notified when an upgrade to OpsCenter or Cassandra or DSE on
any of your clusters is available. The actual upgrade processes will be manual. When a new version is
available a link at the top of the Dashboard appears. Clicking on it displays a dialog with a link to the
new version of the software.
Cluster - See your cluster from different perspectives and perform certain maintenance operations on
cluster nodes.
Cluster administration - Add, modify, or remove a cluster from OpsCenter. Available in OpsCenter
Enterprise Edition only.
Performance - Monitor a number of Cassandra cluster performance metrics. Real-time and historical
performance metrics are available at different granularities: cluster-wide, per node, or per column
family.
Alerts - Configure alert thresholds for a number of Cassandra cluster-wide, column family, and
operating system metrics. Available in OpsCenter Enterprise Edition only.
Scehma - Create and manage keyspaces and the column families within them.
Data backups - Visually take, schedule, and manage backups across all registered clusters. Restore to
clusters from backups. Available in OpsCenter Enterprise Edition only.
Data Explorer - Browse through column family data.
Event Log - View the most recent OpsCenter log events, such as OpsCenter startup and shutdown.
Managing clusters
The New Cluster command allows you to create new clusters or add existing clusters to OpsCenter.
Creating a cluster
Instructions on how to provision a new cluster using OpsCenter.
About this task
OpsCenter can provision new Cassandra and DataStax Enterprise clusters. To do this, each node in the
cluster must meet the following requirements:
•
•
Have Oracle Java 7 installed.
Cannot have Cassandra or DSE installed.
43
Using OpsCenter
•
•
Have a user capable of using sudo for local clusters, unless you've configured OpsCenter to not use
sudo (see the runs_sudo option described in OpsCenter configuration properties).
When provisioning on EC2 nodes, the OpsCenter machine needs access to port 61621 on the
managed nodes, and the nodes need access to port 61620 on the OpsCenter machine.
Procedure
1. Click New Cluster.
2. Click Create Brand New Cluster.
The Create Cluster dialog displays.
3. (Optional) Choose whether to provision a cluster in the cloud or to a local machine by selecting the
Cloud or Local. The Cloud option only appears if you are running OpsCenter on an EC2 instance.
The Cloud and Local buttons only appear if OpsCenter is running in the cloud.
4. Fill out the form as appropriate.
44
•
If you select Local the Local pane displays. The DataStax credentials are the ones which were
emailed to you, when you signed up, while the node credentials are the user id on each node with
sudo privileges.
•
If you select Cloud the Cloud pane displays. (OpsCenter currently only supports the AWS EC2
cloud.)
Using OpsCenter
Fill out the form as appropriate.
Table 1: New cluster fields
Field
Description
Name
My Cluster
Package
The version of DSE or Cassandra to install on the
nodes.
DataStax Credentials
<userid> and <password> that were in the email
you received from DataStax when registering to
download DataStax Enterprise.
Nodes (Local only)
A list of existing machines on which to install the
cluster. (Only when provisioning locally.)
Total Nodes (Cloud only)
Total number of DSE or Cassandra nodes for the
cluster. (Only when provisioning to the cloud.)
# Solr Nodes
Total number of Solr nodes for the cluster.
# Hadoop Nodes
Total number of Hadoop nodes for the cluster.
Node Credentials (Local only)
The <userid> and <password> of the user with
sudo permission on the nodes. (Only when
provisioning locally.)
Private SSH Key (Local only)
The private SSH key to use instead of Node
Credentials. (Only when provisioning locally.)
Amazon EC2 Credentials (Cloud only)
The <access-jey-id> and <secret-access-key> to
use to authenticate on AWS EC2.
Availability Zone (Cloud only)
Which availability zone to use to create the
cluster. (The dropdown list is only populated after
entering your EC2 credentials.)
45
Using OpsCenter
Field
Description
Size (Cloud only)
Which size image to use.
AMI (Cloud only)
Which image to use.
Use OpsCenter specific security group (Cloud
only)
Determines whether OpsCenter creates its own
specific security group or allows you to select one
which is available using your EC2 credentials.
Use OpsCenter specific keypair (Cloud only)
Determines whether OpsCenter creates its own
specific keypair or allows you to select one which
is available using your EC2 credentials.
5. Click Build Cluster.
Results
The new cluster is now available.
If the agent fails to install, check opscenterd.log and /var/log/datastax-agent/installer.log
on the affected nodes. Also verify that the correct ports are open between machines as described in
OpsCenter and DataStax agent ports.
If the agent installs successfully, but there is an issue with the Cassandra or DSE setup process, check
opscenterd.log, /var/log/datastax-agent/agent.log, /var/log/cassandra/output.log,
and /var/log/cassandra/system.log for any errors.
Retrying a failed install
If an error occurs while creating a cluster or adding new nodes, you can use the Retry link to easily bring
back the form once the cause of the problem has been determined and resolved.
If you are installing DSE on existing machines (that is, the Local tab), you need to ensure that all all
relevant software has been uninstalled beforehand [link to resetting instances].
If you are installing on new EC2 instances (that is, the Cloud tab), brand new instances are created when
retrying; existing instances that were part of the failure need to be terminated manually.
Accessing Amazon EC2 instances created by OpsCenter
Instructions on how to log in to EC2 instances created by OpsCenter.
About this task
If you are running OpsCenter on Amazon EC2, you can use ssh to log in to the instances created
by OpsCenter. Using the default AMI, the username is ubuntu. The private key is located in /var/
lib/opscenter/ssl/ and is named after the region in which the opscenterd instance is running
appended with -OpsCenterProvisioningKeyPair.pem. For example, the private key might be
US_West_(Northern_California)-OpsCenterProvisioningKeyPair.pem.
Due to the way SSH handles permissions on the private key file, you must use sudo to call ssh unless you
make a copy of the private key and move it to location owned by a non-root user.
Procedure
Use SSH to log in to the EC2 instance.
$ sudo ssh -i "/var/lib/opscenter/ssl/US_West_(Northern_California)OpsCenterProvisioningKeyPair.pem" [email protected]
46
Using OpsCenter
Adding an existing cluster
Instructions on how to add an existing cluster to OpsCenter.
Procedure
1. Click New Cluster.
2. Click Manage Existing Cluster.
The Add Cluster dialog appears.
3. Enter at least one hostname or IP address for the a node in the cluster.
ec2-123-45-6-789.us-west-1.compute.amazonaws.com
ec2-234-56-7-890.us-west-1.compute.amazonaws.com
4. If you are not using the default JMX or Thrift ports, enter the appropriate port numbers.
5. If required, click Add Credentials and enter the username and password for JMX or Thrift ports.
6. (Optional) You can check the DSE security (kerberos) is enabled on my cluster and enter the service
name.
7. (Optional) You can check the Client node encryption is enabled on the cluster and enter your PEM
encoded certificate in CA Certificate File Path.
If you have a CER encoded certificate, use the following command to convert it.
$ openssl x509 -inform der -in certificate.cer -out certificate.pem
8. (Optional) You can check the Validate SSL Certifcates and enter the Truststore File Path and Trustore
Password.
For more information about enabling Kerberos see Security in the DSE Documentation.
9. Click Add Cluster.
Node monitoring and administration
In the Cluster section of OpsCenter, you select different views of the nodes comprising your Cassandra
cluster and then perform node management.
Ring View
The Ring View displays the cluster as a ring of nodes from which node health, data distribution, and
datacenter balance is determined at a glance within a single visualization.
•
•
•
Nodes are positioned around the ring according to their assigned token. In the case of
ByteOrderedPartioner or vnodes, nodes are positioned based on what percentage of data they
own.
The color of each node represents its health, which is determined by system load average (the number
shown by the uptime command). Per core: 0–0.999 is Normal (green); 1–5 is Medium (yellow); 5+ is
High (red).
The size of each node represents its data size, relative to all other nodes in the cluster.
47
Using OpsCenter
48
Using OpsCenter
Health summary
The health summary pane, located above the rings, contains a cluster-wide summary of the data
represented within the rings. You can quickly get an idea of the health of your cluster without having to
manually scan each ring. This is especially useful with larger clusters.
Hovering over a number in the health summary will highlight the nodes included in that total. You can
easily identify potential problem nodes, as well as whether or not multiple nodes within a single replica set
are having issues.
Clicking on the totals in the health summary presents a list of nodes included in the total.
Node details
Hovering over a node displays some basic details about that node. These details are updated in real time.
49
Using OpsCenter
Clicking a node reveals a dialog displaying more information so you can run various operations on the
node.
Streams
An indicator will be displayed inside of the ring whenever any nodes in that data center are streaming data
to or from another node in the cluster. The indicator distinguishes between streams contained within that
data center ("intra-dc streams") and streams between data centers ("inter-dc streams").
50
Using OpsCenter
Clicking on the indicator in any data center opens the Streams dialog, which gives details on all the
streams in the cluster.
Node positioning
The goal of positioning nodes in a ring is to visually represent whether a datacenter is balanced or not (that
is, data is more likely to be evenly distributed across nodes). In a healthy ring, nodes will be spread out
evenly around the ring.
When a cluster uses RandomPartitioner or Murmur3Partitioner for its snitch, its nodes are
positioned around the ring according to their assigned token, but there are some cases where positioning
by token does not make sense:
•
•
If vnodes are enabled, each node is made up of multiple virtual nodes (256 by default), so positioning
by token would mean having hundreds of times as many nodes around the ring.
If a partitioner that doesn't use consistent hashing is used, such as ByteOrderedPartitioner, data
is not guaranteed to be distributed evenly, so positioning by token also has no guaranteed value
In those cases, nodes are positioned based on the percentage of data they own in the ring, so a healthy
ring is still represented by nodes being evenly spaced out.
51
Using OpsCenter
Unknown datacenter list
Rings are displayed by datacenter. This information is gathered from the agent that runs on each node. If
the agent is not installed or has not connected properly, or there are any other problems getting datacenter
information for a node, it is displayed in a special list above all rings.
List View
List View is an alternative to Ring View, which allows for quicker access to data and more flexibility in how
that data is viewed. All data is updated in realtime.
By selecting any of the checkboxes next to each node, you can run any operation on any number of nodes
in your cluster. For more information, see the Node management operations section.
Columns can be sorted in ascending or descending order (by clicking on the column label) to view which
nodes have the most data, the highest CPU load, and so forth.
The status column displays whether a node is:
•
•
•
up or down
in a special mode (for example, joining, draining, or moving)
running any tasks, such as compactions
Viewing node details
To see more details about a single node, click the row in the list.
52
Using OpsCenter
A Node dialog displays.
Node management operations
You can run operations (or actions) on nodes in an easy to use, visual way, that takes the guess work out
of the proper way to do things.
Managing nodes en masse
Most node management operations can be run on multiple nodes of your choosing (for example, all the
nodes in a cluster, all the nodes in a single datacenter, or a handful of problem nodes. The operations
run in a rolling fashion and do not continue on to the next node until the previous one has completed
successfully. If the operation fails on a node, the entire process stops.
To run an operation on multiple nodes, select those nodes in List View and choose the appropriate action.
53
Using OpsCenter
Notifications appear when the operation starts and completes.
Managing single nodes
To run an operation on a single node, click that node from Ring View or List View and choose the action
you want to run from the Actions dropdown:.
54
Using OpsCenter
Operations details
View Metrics (single node only)
Redirects you to the Performance area of OpsCenter where you can select metrics graphs and configure
performance views for the selected node.
View Replication (ring view, single datacenter only)
Shows the replication relationships between the selected node and other nodes in the cluster, based on
the selected keyspace.
Configure (single node only)
Modify settings specified in cassandra.yaml for the selected node.
55
Using OpsCenter
When the configuration is saved, an option will be presented to restart the DSE or Cassandra process in
order for the changes to take effect.
Start/Stop
Starts or stops the DSE or Cassandra process on the node.
Restart
Restarts the DSE or Cassandra process on the node. If running on multiple nodes, each node is started as
soon as the start command for the previous node returns. If you want the operation to wait for thrift to be
active on each node before continuing, use the Rolling Restart action.
Cleanup
Removes rows which the node is no longer responsible for. This is usually done after changing the partitioner
tokens or the replication options for a cluster.
Compact
Performs a major compaction, which is not a recommended procedure in most Cassandra clusters.
Flush
Causes the recent writes currently stored in memory (memtables) to be flushed to disk as persistent
SSTables.
Repair
Makes a node consistent with its replicas by doing an in-memory comparison of all the rows of a column
family and resolving any discrepancies between replicas by updating outdated rows with the current data.
Perform GC
Forces the Java Virtual Machine (JVM) on the selected node to perform a garbage collection (GC).
Decommission (single node only)
56
Using OpsCenter
Removes a node from the cluster and streams its data to neighboring replicas.
Drain (single node only)
Causes the recent writes currently stored in memory (memtables) to be flushed to disk as persistent
SSTables and then makes the node read-only. The node stops accepting new writes. This is usually done
when upgrading a node.
Move (single node only)
Changes the partitioner token assignment for the node, thus changing the range of data that the node is
responsible for. Not enabled if vnodes are enabled.
Cluster administration
OpsCenter manages multiple DataStax Enterprise or Apache Cassandra clusters with a single install of the
central opscenterd server.
OpsCenter Cassandra / DSE Support
See the Upgrade Guide for information on the versions that we officially support and test.
Generating a report
To generate a PDF report about the cluster being monitored, click Report at the top of the OpsCenter
interface. The report shows the version of OpsCenter, number of clusters and nodes being monitoring,
gigabytes of storage used, name of the cluster, and information about nodes in the cluster. The node
information includes:
•
•
•
•
•
Node name and IP address
Cassandra software version
DataStax software version
Memory usage
Operating system running on the node
You can save or print the PDF report.
Collecting diagnostic data
Download a compressed tarball that contains information about the OpsCenter daemon and all the nodesin
a specified cluster.
About this task
You can attach this diagnostic data to support tickets.
Procedure
1. Click Diagnostics to download the tarball.
You see the message: "Collecting cluster data; please wait, this may take a few minutes..."
2. Save the tarball to your local machine.
diagnostics.tar.gz
Depending on your browser settings, you may or may not be prompted for a file directory to save the
tarball in.
Adding a node to a cluster
To add a node to a cluster.
57
Using OpsCenter
Procedure
1. Click Add Node in a cluster view.
2. If adding a node to a DSE cluster, select the type of node: Hadoop, Cassandra, or Solr.
3. (Optional) Choose whether to add a node in the cloud or on a local machine by selecting the Cloud or
Local. The Cloud option only appears if you are running OpsCenter on an EC2 instance.
4. Fill out the form as appropriate, with the following fields:
Option
Value
Package
The version of DSE to install on the node.
DataStax credentials
The username and password you received when
registering to Download DSE.
Nodes
The hostname or IP address, token, and software
to install on the node (from Cassandra, Solr, and
Hadoop). You can add more than one node by
clicking Add.
Node credentials (sudo) (Local only)
The username and password to authenticate on
the host. (Optional) the private SSH key to use to
use for authentication.
Amazon EC2 Credentials (Cloud only)
The <access-jey-id> and <secret-access-key> to
use to authenticate on AWS EC2.
Availability Zone (Cloud only)
Which availability zone to use to create the
cluster. (The dropdown list is only populated after
entering your EC2 credentials.)
Size (Cloud only)
Which size image to use.
AMI (Cloud only)
Which image to use.
Use OpsCenter specific security group (Cloud
only)
Determines whether OpsCenter creates its own
specific security group or allows you to select one
which is available using your EC2 credentials.
Use OpsCenter specific keypair (Cloud only)
Determines whether OpsCenter creates its own
specific keypair or allows you to select one which
is available using your EC2 credentials.
5. Click Add Nodes.
Configuring a cluster
You can manage cassandra.yaml from OpsCenter.
About this task
If the cluster exists in multiple datacenters, you can configure cassandra.yaml for a single datacenter
or for all nodes in a cluster. To manage cassandra.yaml for a single node by clicking on the Actions
dropdown for a node.
Procedure
1. Click Configure Cluster in any of the Cluster views.
2. Edit the value for any of the options.
For a description of the options in the file, see the documentation for the version of Cassandra or DSE
which the cluster or node is running. For example, cassandra.yaml.
58
Using OpsCenter
3. When finished editing, click Save.
4. You will be prompted to perform a rolling restart on the nodes in your cluster, or you may click Cancel if
you do not wish to do a restart.
Removing a cluster
To remove a cluster.
About this task
This removes the cluster from OpsCenter; it does not delete the cluster.
Procedure
1. Click Edit Cluster.
2. Click Delete Cluster.
When you delete a cluster, any EC2 nodes are not deleted.
Rebalancing a cluster
Cluster rebalancing is a process that makes sure each node in a Cassandra cluster is managing an equal
amount of data.
About this task
Currently, OpsCenter only supports rebalancing on clusters using the random partitioner or murmur
3 partitioner. Ordered partitioners are not supported. When using the random partitioner or murmur 3
partitioner, a rebalance is usually required only when you have changed the cluster topology in some way,
such as adding or removing nodes or changing the replica placement strategy.
A cluster is considered balanced when each node is responsible for an equal range of data. This is done
by evaluating the partitioner tokens assigned to each node to make sure that the data ranges each node is
responsible for are even. Even though a cluster is considered balanced, it is still possible that one or more
nodes have more data than the others. This is because the size of the rows is not taken into account, only
the number of rows managed by each node.
Procedure
1. In the Cluster section of OpsCenter, select Ring View, Physical View or List View.
The dashboard displays the specified view of your cluster.
2. Click the Rebalance Cluster button.
OpsCenter checks if the token ranges are evenly distributed across the nodes in your cluster.
3. If your cluster is already balanced, then there is nothing for OpsCenter to do. If the cluster does require
rebalancing, OpsCenter performs the following steps:
•
•
•
Calculates appropriate token ranges for each node and identify the nodes that need to move.
Makes sure that there is the appropriate free space to perform the rebalancing.
Moves nodes, one node at a time so as to lessen the impact on the cluster workloads. A move
operation involves changing the partitioner token assignment for the node, thus changing the range
of data that the node is responsible for. A move will stream data from other nodes.
• After a move is complete on a node, runs cleanup. A cleanup operation removes rows that the node
is no longer responsible for.
4. If you cancel a rebalance operation before all nodes are moved, you can resume it at a later time by
clicking the Rebalance Cluster button again.
Restarting a cluster
To restart a cluster.
59
Using OpsCenter
About this task
You can start, stop, or restart the Cassandra or DSE service on any node. This can be done via the
Actions dropdown on a node.
There is also rolling restart functionality for the entire cluster. This can be done via the Restart Cluster in
any of the Cluster views.
Procedure
1. In any cluster view, click on a node.
2. In the contextual menu select Restart from the Actions dropdown.
Modifying a cluster setting
To modify a cluster setting.
Procedure
1.
2.
3.
4.
5.
Click Edit Cluster.
Change the IP addresses of cluster nodes.
Change JMX and Thrift listen port numbers.
Click Add credentials if the ports require authentication.
(Optional) You can check the DSE security (kerberos) is enabled on my cluster and enter the service
name.
6. (Optional) You can check the Client node encryption is enabled on the cluster and enter your PEM
encoded certificate in CA Certificate File Path.
If you have a CER encoded certificate, use the following command to convert it.
$ openssl x509 -inform der -in certificate.cer -out certificate.pem
7. (Optional) You can check the Validate SSL Certifcates and enter the Truststore File Path and Trustore
Password.
For more information about enabling Kerberos see Security in the DSE Documentation.
8. Click Save Cluster.
Performance metrics
In the Performance area of OpsCenter, you monitor a number of performance metrics about a Cassandra
cluster. Real-time and historical performance metrics are available at different granularities: cluster-wide,
per node, or per column family.
Using performance metrics
Select Performance in the OpsCenter Console to view these types of metrics:
•
•
•
Cluster Performance Metrics
Pending Task Metrics
Column Family Metrics
When you add a graph, you choose the Metric and the source that OpsCenter uses to collect the data for
the graph:
•
•
•
60
Cluster wide
All nodes
The node running Opscenter
Using OpsCenter
Several commonly-used performance metrics graphs are displayed initially. Data appears in the graphs
after you set alerts.
You can save, delete, and choose the default view of graphs. Click the link to save presets at the top of
the Performance area. The Save, Delete, and Make Default menu options are available after saving more
than one view.
Creating and editing performance graphs
How to add and edit performance graphs.
About this task
Graphs can be added containing multiple metrics provided the metrics use the same unit. For example, a
graph can contain multiple metrics showing utilization as a percentage, like CPU and disk utilization. Other
metrics like write or read requests for a cluster or the operating system load for a node cannot be added to
the utilization graph. Metrics can be added to a graph for a cluster or one or more nodes.
Procedure
1. Click Performance in the left pane.
2. Click Add Graph.
3. In the Add Metric dialog, select the metric you want to add from the Metric drop down, then the
nodes you want to monitor in the Nodes drop down. To select multiple individual nodes, first select a
node, then click in the Node drop down again to select an additional node. You can optionally specify
particular column families by clicking the Col. Family selector.
4. Click Add Metric.
61
Using OpsCenter
5. To add additional metrics that are measured using the same unit, click Add Metric in the Metrics On
This Graph dialog. You may edit the details of a metric by clicking the Edit icon, or delete a metric by
clicking the Trash icon.
6. When you are done, click Save Graph to display the graph showing the defined metrics.
7. You can edit the metrics displayed in a graph by clicking Edit Graph next to the graph's title.
Grouping performance metrics
How to save groups of performance graphs.
About this task
Groups of graphs can be saved using named presets. This allows you to display different groups of related
metrics graphs for analysis.
Procedure
1. Click Performance in the left hand pane.
2. Click Unsaved Preset at the top of the Performance Metrics page, then click Save As.
62
Using OpsCenter
3. In the Save Preset dialog, enter a name for the preset and click Save. The first preset you define is the
default.
4. If you modify the graphs displayed in the named preset, click the preset name at the top of the page
and click Save to update the preset, or Save As to define a new preset. You may also set the default
preset by selecting Make Default. Non-default presets can be deleted by selecting Delete.
5. To display another preset, click the current preset name at the top of the page and select the new
preset name.
Cluster performance metrics
Cluster metrics are aggregated across all nodes in the cluster. Cluster metrics are a good way to
monitor cluster performance at a high level. OpsCenter tracks a number of cluster-wide metrics for read
performance, write performance, memory and capacity.
Watching for variations in cluster performance can signal potential performance issues that may require
further investigation. For general performance monitoring, watching for spikes in read and write latency,
along with an accumulation of pending operations can signal issues that may require further investigation.
Drilling down on high-demand column families can further pinpoint the source of performance issues with
your application.
Write requests
The number of write requests per second on the coodinator nodes, analogous to client writes. Monitoring
the number of requests over a given time period can give you and idea of system write workload and
usage patterns.
Write request latency
The response time (in milliseconds) for successful write requests. The time period starts when a node
receives a client write request, and ends when the node responds back to the client. Optimal or acceptable
levels of write latency vary widely according to your hardware, your network, and the nature of your write
load. For example, the performance for a write load consisting largely of granular data at low consistency
levels would be evaluated differently from a load of large strings written at high consistency levels.
Read requests
The number of read requests per second on the coordinator nodes, analogous to client reads. Monitoring
the number of requests over a given time period can give you and idea of system read workload and usage
patterns.
Read request latency
The response time (in milliseconds) for successful read requests. The time period starts when a node
receives a client read request, and ends when the node responds back to the client. Optimal or acceptable
levels of read latency vary widely according to your hardware, your network, and the nature of your
application read patterns. For example, the use of secondary indexes, the size of the data being requested,
63
Using OpsCenter
and the consistency level required by the client can all impact read latency. An increase in read latency
can signal I/O contention. Reads can slow down when rows are fragmented across many SSTables and
compaction cannot keep up with the write load.
Cassandra JVM memory usage
The average amount of Java heap memory (in megabytes) being used by Cassandra processes.
Cassandra opens the JVM with a heap size that is half of available system memory by default, which still
allows an optimal amount of memory remaining for the OS disk cache. You may need to increase the
amount of heap memory if you have increased column family memtable or cache sizes and are getting
out-of-memory errors. If you monitor Cassandra Java processes with an OS tool such as top, you may
notice the total amount of memory in use exceeds the maximum amount specified for the Java heap. This
is because Java allocates memory for other things besides the heap. It is not unusual for the total memory
consumption of the JVM to exceed the maximum value of heap memory.
JVM CMS collection count
The number of concurrent mark-sweep (CMS) garbage collections performed by the JVM per second.
These are large, resource-intensive collections. Typically, the collections occur every 5 to 30 seconds.
JVM CMS collection time
The time spent collecting CMS garbage in milliseconds per second (ms/sec).
Note: A ms/sec unit defines the number of milliseconds for garbage collection for each second that
passes. For example, the percentage of time spent on garbage collection in one millisecond (.001
sec) is 0.1%.
JVM ParNew collection count
The number of parallel new-generation garbage collections performed by the JVM per second. These are
small and not resource intensive. Normally, these collections occur several times per second under load.
JVM ParNew Collection Time
The time spent performing ParNew garbage collections in ms/sec. The rest of the JVM is paused during
ParNew garbage collection. A serious performance hit can result from spending a significant fraction of
time on ParNew collections.
Data size
The size of column family data (in gigabytes) that has been loaded/inserted into Cassandra, including any
storage overhead and system metadata. DataStax recommends that data size not exceed 70 percent of
total disk capacity to allow free space for maintenance operations such as compaction and repair.
Total bytes compacted
The number of sstable data compacted in bytes per second.
Total compactions
The number of compactions (minor or major) performed per second.
Pending task metrics
Pending task metrics track requests that have been received by a node, but are waiting to be processed.
An accumulation of pending tasks on a node can indicate a potential bottleneck in performance and should
be investigated.
64
Using OpsCenter
Cassandra maintains distinct thread pools for different stages of execution. Each of these thread pools
provide granular statistics on the number of pending tasks for that particular process. If you see pending
tasks accumulating, it is indicative of a cluster that is not keeping up with the workload. Essentially,
pending tasks mean that things are backing up, which is usually caused by a lack of (or failure of) cluster
resources such as disk bandwidth, network bandwidth or memory.
Pending task metrics for writes
Pending tasks for the following metrics indicate that write requests are arriving faster than they can be
handled.
Flushes pending
The flush process flushes memtables to disk as SSTables. This metric shows the number of memtables
queued for the flush process. The optimal number of pending flushes is 0 (or at most a very small number).
A value greater than 0 indicates either I/O contention or degrading disk performance (see disk metrics such
as disk latency, disk throughput, and disk utilization for indications of disk health).
Flush sorter tasks pending
The flush sorter process performs the first step in the overall process of flushing memtables to disk as
SSTables.
memtable post flushers pending
The memtable post flush process performs the final step in the overall process of flushing memtables to
disk as SSTables.
Write requests pending
The memtable post flush process performs the final step in the overall process of flushing memtables to
disk as SSTables.
Replicate on write tasks pending
When an insert or update to a row is written, the affected row is replicated to all other nodes that manage a
replica for that row. This is called the ReplicateOnWriteStage. This metric tracks the pending tasks related
to this stage of the write process. During low or moderate write load, you should see 0 pending replicate on
write tasks (or at most a very low number). A continuous high number signals a need to investigate disk I/O
or network contention problems.
Pending task metrics for reads
Pending tasks for the following metrics indicate I/O contention, and can manifest in degrading read
performance.
Read requests pending
The number of read requests that have arrived into the cluster but are waiting to be handled. During low
or moderate read load, you should see 0 pending read operations (or at most a very low number). A
continuous high number of pending reads signals a need for more capacity in your cluster or to investigate
disk I/O contention. Pending reads can also indicate an application design that is not accessing data in the
most efficient way possible.
Read repair tasks pending
The number of read repair operations that are queued and waiting for system resources in order to run.
The optimal number of pending read repairs is 0 (or at most a very small number). A value greater than 0
indicates that read repair operations are in I/O contention with other operations. If this graph shows high
values for pending tasks, this may suggest the need to run a node repair to make nodes consistent. Or,
65
Using OpsCenter
for column families where your requirements can tolerate a certain degree of stale data, you can lower the
value of the column family parameter read_repair_chance.
Compactions pending
An upper bound of the number of compactions that are queued and waiting for system resources in order
to run. This is a worst-case estimate. The compactions pending metric is often misleading. An unrealistic,
high reading often occurs. The optimal number of pending compactions is 0 (or at most a very small
number). A value greater than 0 indicates that read operations are in I/O contention with compaction
operations, which usually manifests itself as declining read performance. This is usually caused by
applications that perform frequent small writes in combination with a steady stream of reads. If a node
or cluster frequently displays pending compactions, that is an indicator that you may need to increase
I/O capacity by adding nodes to the cluster. You can also try to reduce I/O contention by reducing the
number of insert/update requests (have your application batch writes for example), or reduce the number
of SSTables created by increasing the memtable size and flush frequency on your column families.
Pending task metrics for cluster operations
Pending tasks for the following metrics indicate a backup of cluster operational processes such as those
maintaining node consistency, system schemas, fault detection, and inter-node communications. Pending
tasks for resource-intensive operations (such as repair, bootstrap or decommission) are normal and
expected while that operation is in progress, but should continue decreasing at a steady rate in a healthy
cluster.
Manual repair tasks pending
The number of operations still to be completed when you run anti-entropy repair on a node. It will only
show values greater than 0 when a repair is in progress. Repair is a resource-intensive operation that is
executed in stages: comparing data between replicas, sending changed rows to the replicas that need to
be made consistent, deleting expired tombstones, and rebuilding row indexes and bloom filters. Tracking
the state of this metric can help you determine the progress of a repair operation. It is not unusual to
see a large number of pending tasks when a repair is running, but you should see the number of tasks
progressively decreasing.
Gossip tasks pending
Cassandra uses a protocol called gossip to discover location and state information about the other nodes
participating in a Cassandra cluster. In Cassandra, the gossip process runs once per second on each
node and exchanges state messages with up to three other nodes in the cluster. Gossip tasks pending
shows the number of gossip messages and acknowledgments queued and waiting to be sent or received.
The optimal number of pending gossip tasks is 0 (or at most a very small number). A value greater than 0
indicates possible network problems (see network traffic for indications of network health).
Hinted handoff pending
While a node is offline, other nodes in the cluster will save hints about rows that were updated during the
time the node was unavailable. When a node comes back online, its corresponding replicas will begin
streaming the missed writes to the node to catch it up. The hinted handoff pending metric tracks the
number of hints that are queued and waiting to be delivered once a failed node is back online again. High
numbers of pending hints are commonly seen when a node is brought back online after some down time.
Viewing this metric can help you determine when the recovering node has been made consistent again.
Hinted handoff is an optional feature of Cassandra. Hints are saved for a configurable period of time (an
hour by default) before they are dropped. This prevents a large accumulation of hints caused by extended
node outages.
Internal responses pending
The number of pending tasks from various internal tasks such as nodes joining and leaving the cluster.
66
Using OpsCenter
Migrations pending
The number of pending tasks from system methods that have modified the schema. Schema updates
have to be propagated to all nodes, so pending tasks for this metric can manifest in schema disagreement
errors.
Miscellaneous tasks pending
The number of pending tasks from other miscellaneous operations that are not ran frequently.
Request response pending
The progress of rows of data being streamed from the receiving node. Streaming of data between nodes
happens during operations such as bootstrap and decommission when one node sends large numbers of
rows to another node.
Streams pending
The progress of rows of data being streamed from the sending node. Streaming of data between nodes
happens during operations such as bootstrap and decommission when one node sends large numbers of
rows to another node.
Column family performance metrics
Column family metrics allow you to drill down and locate specific areas of your application workloads that
are the source of performance issues. If you notice a performance trend at the OS or cluster level, viewing
column family metrics can provide a more granular level of detail.
The metrics for KeyCache Hits, RowCache Hits and SSTable Size can only be viewed on a single column
family at a time. Otherwise, all column family metrics are available for specific column families as well as
for all column families on a node.
In addition to monitoring read latency, write latency and load on a column family, you should also monitor
the hit rates on the key and row caches for column families that rely on caching for performance. The more
requests that are served from the cache, the better response times will be.
OpsCenter 4.1 and later has been optimized to handle thousands of column families efficiently. If a column
family experiences a dramatic dip in performance, check the Pending Tasks metric for a back-up in queued
operations.
Viewing SSTable Size and SSTable Count for a specific column family (or counts for all families) can help
with compaction tuning.
Column family local writes
The write load on a column family measured in requests per second. This metric includes all writes to a
given column family, including write requests forwarded from other nodes. This metric can be useful for
tracking usage patterns of your application.
Column family local write latency
The response time in milliseconds for successful write requests on a column family. The time period starts
when nodes receive a write request, and ends when nodes respond. Optimal or acceptable levels of
write latency vary widely according to your hardware, your network, and the nature of your write load. For
example, the performance for a write load consisting largely of granular data at low consistency levels
would be evaluated differently from a load of large strings written at high consistency levels.
67
Using OpsCenter
Column family local reads
The read load on a column family measured in requests per second. This metric includes all reads to a
given column family, including read requests forwarded from other nodes. This metric can be useful for
tracking usage patterns of your application.
Column family local read latency
The response time in milliseconds for successful reads on a column family. The time period starts when
a node receives a read request, and ends when the node responds. Optimal or acceptable levels of
read latency vary widely according to your hardware, your network, and the nature of your application
read patterns. For example, the use of secondary indexes, the size of the data being requested, and the
consistency level required by the client can all impact read latency. An increase in read latency can signal
I/O contention. Reads can slow down when rows are fragmented across many SSTables and compaction
cannot keep up with the write load.
Column family key cache requests
The total number of read requests on the row key cache.
Column family key cache hits
The number of read requests that resulted in the requested row key being found in the key cache.
Column family key cache hit rate
The percentage of cache requests that resulted in a cache hit that indicates the effectiveness of the key
cache for a given column family. The key cache is used to find the exact location of a row on disk. If a row
is not in the key cache, a read operation will populate the key cache after accessing the row on disk so
subsequent reads of the row can benefit. Each hit on a key cache can save one disk seek per SSTable.
If the hits line tracks close to the requests line, the column family is benefiting from caching. If the hits fall
far below the request rate, this suggests that you could take actions to improve the performance benefit
provided by the key cache, such as adjusting the number of keys cached.
Column family row cache requests
The total number of read requests on the row cache. This metric is only meaningful for column families with
row caching configured (it is not enabled by default).
Column family row cache hits
The number of read requests that resulted in the read being satisfied from the row cache. This metric is
only meaningful for column families with row caching configured (it is not enabled by default).
Column family row cache hit rate
The percentage of cache requests that resulted in a cache hit that indicates the effectiveness of the row
cache for a given column family. This metric is only meaningful for column families with row caching
configured (it is not enabled by default). The graph tracks the number of read requests in relationship to the
number of row cache hits. If the hits line tracks close to the requests line, the column family is benefiting
from caching. If the hits fall far below the request rate, this suggests that you could take actions to improve
the performance benefit provided by the row cache, such as adjusting the number of rows cached or
modifying your data model to isolate high-demand rows.
Column family SSTable size
The current size of the SSTables for a column family. It is expected that SSTable size will grow over time
with your write load, as compaction processes continue doubling the size of SSTables. Using this metric
together with SSTable count, you can monitor the current state of compaction for a given column family.
68
Using OpsCenter
Viewing these patterns can be helpful if you are considering reconfiguring compaction settings to mitigate I/
O contention.
Column family SSTable count
The current number of SSTables for a column family. When column family memtables are persisted to disk
as SSTables, this metric increases to the configured maximum before the compaction cycle is repeated.
Using this metric together with SSTable size, you can monitor the current state of compaction for a given
column family. Viewing these patterns can be helpful if you are considering reconfiguring compaction
settings to mitigate I/O contention.
Column family pending reads and writes
The number of pending reads and writes on a column family. Pending operations are an indication that
Cassandra is not keeping up with the workload. A value of zero indicates healthy throughput. If out-ofmemory events become an issue in your Cassandra cluster, it may help to check cluster-wide pending
tasks for operations that may be clogging throughput.
Bloom filters are used to avoid going to disk to try to read rows that don't actual exist.
Column family bloom filter space used
The size of the bloom filter files on disk. This grows based on the number of rows in a column family and is
tunable through the per-CF attribute,bloom_filter_fp_chance; increasing the value of this attribute shrinks
the bloom filters at the expense of a higher number of false positives. Cassandra reads the bloom filter files
and stores them on the heap, so large bloom filters can be expensive in terms of memory consumption.
Note: Bloom filters are used to avoid going to disk to try to read rows that don't actual exist.
Column family bloom filter false positives
The number of false positives, which occur when the bloom filter said the row existed, but it actually did not
exist in absolute numbers.
Column family bloom filter false positive ratio
The fraction of all bloom filter checks resulting in a false positive. This should normally be at or below .01.
A higher reading indicates that the bloom filter is likely too small.
System performance metrics
As with any database system, Cassandra performance greatly depends on underlying systems on which
it is running. Tracking operating system metrics on your Cassandra nodes to watch for disk I/O, network,
memory and CPU utilization trends can help you identify and troubleshoot hardware-related performance
problems.
Monitoring Cassandra nodes for increasing disk and CPU utilization can help identify and remedy issues
before performance degrades to unacceptable levels. The graphs in OpsCenter provide a quick way to
view variations in OS metrics at a glance, and drill-down for specific data points. Especially in systems with
heavy write loads, monitoring disk space is also important. It allows for advanced expansion planning while
there is still adequate capacity to handle expansion and rebalancing operations.
Memory
Shows memory usage metrics in megabytes.
•
•
•
Linux—Shows how much total system memory is currently used, cached, buffered or free.
Windows—Shows the available physical memory, the cached operating system code, and the allocated
pool-paged-resident and pool-nonpaged memory.
Mac OS X—Shows free and used system memory.
69
Using OpsCenter
CPU
Shows average percentages for CPU utilization metrics, which is the percentage of time the CPU was idle
subtracted from 100 percent. CPU metrics can be useful for determining the origin of CPU performance
reduction.
•
•
Linux—Shows how much time the CPU devotes to system and user processes, to tasks stolen
by virtual operating systems, to waiting for I/O to complete, and to processing nice tasks. High
percentages of nice may indicate that other processes are crowding out Cassandra processes, while
high percentages of iowait may indicate I/O contention. On fully virtualized environments like Amazon
EC2, a Cassandra cluster under load may show high steal values while other virtual processors use the
available system resources.
Windows and Mac OS X—Shows how much time the CPU spends on user processes and system
processes.
Load
The amount of work that a computer system performs. An idle computer has a load number of 0 and each
process using or waiting for CPU time increments the load number by 1. Any value above one indicates
that the machine was temporarily overloaded and some processes were required to wait. Shows minimum,
average, and maximum OS load expressed as an integer.
Disk usage (GB)
Tracks growth or reduction in the amount of available disk space used. If this metric indicates a growth
trend leading to high or total disk space usage, consider strategies to relieve it, such as adding capacity
to the cluster. DataStax recommends leaving 30-50% free disk space for optimal repair and compaction
operations.
Disk usage (percentage)
The percentage of disk space that is being used by Cassandra at a given time. When Cassandra is reading
and writing heavily from disk, or building SSTables as the final product of compaction processes, disk
usage values may be temporarily higher than expected.
Disk throughput
The average disk throughput for read and write operations, measured in megabytes per second.
Exceptionally high disk throughput values may indicate I/O contention. This is typically caused by
numerous compaction processes competing with read operations. Reducing the frequency of memtable
flushing can relieve I/O contention.
Disk rates
•
•
Linux and Windows—Averaged disk speed for read and write operations.
Mac OS X—Not supported.
Disk latency
•
•
70
Linux and Windows—Measures the average time consumed by disk seeks in milliseconds. Disk latency
is among the higher-level metrics that may be useful to monitor on an ongoing basis by keeping this
graph posted on your OpsCenter performance console. Consistently high disk latency may be a
signal to investigate causes, such as I/O contention from compactions or read/write loads that call for
expanded capacity.
Mac OS X—Not supported.
Using OpsCenter
Disk request size
•
•
Linux and Windows—The average size in sectors of requests issued to the disk.
Mac OS X—Not supported.
Disk queue size
•
•
Linux and Windows—The average number of requests queued due to disk latency issues.
Mac OS X—Not supported.
Disk utilization
•
•
Linux and Windows—The percentage of CPU time consumed by disk I/O.
Mac OS X—Not supported.
Network traffic
The speed at which data is received and sent across the network, measured in kilobytes per second.
Alert metrics
From the Alerts area of OpsCenter Enterprise Edition, you can configure alert thresholds for a number of
Cassandra cluster-wide, column family, and operating system metrics. This proactive monitoring feature is
available only in OpsCenter Enterprise Edition.
Commonly watched alert metrics
Metric
Definition
Node down
When a node is not responding to requests, it is
marked as down.
Write requests
The number of write requests per second.
Monitoring the number of writes over a given time
period can give you and idea of system write
workload and usage patterns.
Write request latency
The response time (in milliseconds) for successful
write operations. The time period starts when a
node receives a client write request, and ends
when the node responds back to the client.
Read requests
The number of read requests per second.
Monitoring the number of reads over a given time
period can give you and idea of system read
workload and usage patterns.
Read request latency
The response time (in milliseconds) for successful
read operations. The time period starts when a
node receives a client read request, and ends when
the node responds back to the client.
CPU usage
The percentage of time that the CPU was busy,
which is calculated by subtracting the percentage of
time the CPU was idle from 100 percent.
Load
Load is a measure of the amount of work that a
computer system performs. An idle computer has a
71
Using OpsCenter
Metric
Definition
load number of 0 and each process using or waiting
for CPU time increments the load number by 1.
Advanced Cassandra alert metrics
72
Metric
Definition
Heap max
The maximum amount of shared memory allocated
to the JVM heap for Cassandra processes.
Heap used
The amount of shared memory in use by the JVM
heap for Cassandra processes.
JVM CMS collection count
The number of concurrent mark-sweep (CMS)
garbage collections performed by the JVM per
second.
JVM ParNew collection count
The number of parallel new-generation garbage
collections performed by the JVM per second.
JVM CMS collection time
The time spent collecting CMS garbage in
milliseconds per second (ms/sec).
JVM ParNew collection time
The time spent performing ParNew garbage
collections in ms/sec.
Data size
The size of column family data (in gigabytes) that
has been loaded/inserted into Cassandra, including
any storage overhead and system metadata.
Compactions pending
The number of compaction operations that are
queued and waiting for system resources in order
to run. The optimal number of pending compactions
is 0 (or at most a very small number). A value
greater than 0 indicates that read operations are
in I/O contention with compaction operations,
which usually manifests itself as declining read
performance.
Total bytes compacted
The number of sstable data compacted in bytes per
second.
Total compactions
The number of compactions (minor or major)
performed per second.
Flush sorter tasks pending
The flush sorter process performs the first step in
the overall process of flushing memtables to disk as
SSTables. The optimal number of pending flushes
is 0 (or at most a very small number).
Flushes pending
The flush process flushes memtables to disk
as SSTables. This metric shows the number of
memtables queued for the flush process. The
optimal number of pending flushes is 0 (or at most
a very small number).
Gossip tasks pending
Cassandra uses a protocol called gossip to
discover location and state information about the
other nodes participating in a Cassandra cluster.
In Cassandra, the gossip process runs once
Using OpsCenter
Metric
Definition
per second on each node and exchanges state
messages with up to three other nodes in the
cluster. Gossip tasks pending shows the number of
gossip messages and acknowledgments queued
and waiting to be sent or received. The optimal
number of pending gossip tasks is 0 (or at most a
very small number).
Hinted handoff pending
While a node is offline, other nodes in the cluster
will save hints about rows that were updated during
the time the node was unavailable. When a node
comes back online, its corresponding replicas will
begin streaming the missed writes to the node
to catch it up. The hinted handoff pending metric
tracks the number of hints that are queued and
waiting to be delivered once a failed node is back
online again. High numbers of pending hints are
commonly seen when a node is brought back
online after some down time. Viewing this metric
can help you determine when the recovering node
has been made consistent again.
Internal response pending
The number of pending tasks from various internal
tasks such as nodes joining and leaving the cluster.
Manual repair tasks pending
The number of operations still to be completed
when you run anti-entropy repair on a node. It will
only show values greater than 0 when a repair is in
progress. It is not unusual to see a large number
of pending tasks when a repair is running, but
you should see the number of tasks progressively
decreasing.
Memtable postflushers pending
The memtable post flush process performs the final
step in the overall process of flushing memtables to
disk as SSTables. The optimal number of pending
flushes is 0 (or at most a very small number).
Migrations pending
The number of pending tasks from system methods
that have modified the schema. Schema updates
have to be propagated to all nodes, so pending
tasks for this metric can manifest in schema
disagreement errors.
Miscellaneous tasks pending
The number of pending tasks from other
miscellaneous operations that are not ran
frequently.
Read requests pending
The number of read requests that have arrived
into the cluster but are waiting to be handled.
During low or moderate read load, you should see
0 pending read operations (or at most a very low
number).
Read repair tasks pending
The number of read repair operations that are
queued and waiting for system resources in order
to run. The optimal number of pending read repairs
is 0 (or at most a very small number). A value
73
Using OpsCenter
Metric
Definition
greater than 0 indicates that read repair operations
are in I/O contention with other operations.
Replicate on write tasks pending
When an insert or update to a row is written, the
affected row is replicated to all other nodes that
manage a replica for that row. This is called the
ReplicateOnWriteStage. This metric tracks the
pending tasks related to this stage of the write
process. During low or moderate write load, you
should see 0 pending replicate on write tasks (or at
most a very low number).
Request response pending
Streaming of data between nodes happens during
operations such as bootstrap and decommission
when one node sends large numbers of rows to
another node. The metric tracks the progress of the
streamed rows from the receiving node.
Streams pending
Streaming of data between nodes happens during
operations such as bootstrap and decommission
when one node sends large numbers of rows to
another node. The metric tracks the progress of the
streamed rows from the sending node.
Write requests pending
The number of write requests that have arrived
into the cluster but are waiting to be handled.
During low or moderate write load, you should see
0 pending write operations (or at most a very low
number).
Advanced column family alert metrics
74
Metric
Definition
Local writes
The write load on a column family measured in
operations per second. This metric includes all
writes to a given column family, including write
requests forwarded from other nodes.
Local write latency
The response time in milliseconds for successful
write operations on a column family. The time
period starts when nodes receive a write request,
and ends when nodes respond.
Local reads
The read load on a column family measured in
operations per second. This metric includes all
reads to a given column family, including read
requests forwarded from other nodes.
Local read latency
The response time in microseconds for successful
read operations on a column family. The time
period starts when a node receives a read request,
and ends when the node responds.
Column family key cache hits
The number of read requests that resulted in the
requested row key being found in the key cache.
Using OpsCenter
Metric
Definition
Column family key cache requests
The total number of read requests on the row key
cache.
Column family key cache hit rate
The key cache hit rate indicates the effectiveness
of the key cache for a given column family by giving
the percentage of cache requests that resulted in a
cache hit.
Column family row cache hits
The number of read requests that resulted in the
read being satisfied from the row cache.
Column family row cache requests
The total number of read requests on the row
cache.
Column family row cache hit rate
The key cache hit rate indicates the effectiveness
of the row cache for a given column family by giving
the percentage of cache requests that resulted in a
cache hit.
Column family bloom filter space used
The size of the bloom filter files on disk.
Column family bloom filter false positives
The number of false positives, which occur when
the bloom filter said the row existed, but it actually
did not exist in absolute numbers.
Column family bloom filter false positive ratio
The fraction of all bloom filter checks resulting in a
false positive.
Live disk used
The current size of live SSTables for a column
family. It is expected that SSTable size will grow
over time with your write load, as compaction
processes continue doubling the size of SSTables.
Using this metric together with SSTable count, you
can monitor the current state of compaction for a
given column family.
Total disk used
The current size of the data directories for the
column family including space not reclaimed by
obsolete objects.
SSTable count
The current number of SSTables for a column
family. When column family memtables are
persisted to disk as SSTables, this metric increases
to the configured maximum before the compaction
cycle is repeated. Using this metric together with
live disk used, you can monitor the current state of
compaction for a given column family.
Pending reads and writes
The number of pending reads and writes on a
column family. Pending operations are an indication
that Cassandra is not keeping up with the workload.
A value of zero indicates healthy throughput.
Advanced system alert metrics
As with any database system, Cassandra performance greatly depends on underlying systems on which
it is running. To configure advanced system metric alerts, you should first have an understanding of the
baseline performance of your hardware and the averages of these system metrics when the system is
handling a typical workload.
75
Using OpsCenter
Linux memory metrics
Metric
Definition
Memory free
System memory that is not being used.
Memory used
System memory used by application processes.
Memory buffered
System memory used for caching file system
metadata and tracking in-flight pages.
Memory shared
System memory that is accessible to CPUs.
Memory cached
System memory used by the OS disk cache.
Linux CPU metrics
Metric
Definition
Idle
Percentage of time the CPU is idle.
Iowait
Percentage of time the CPU is idle and there is a
pending disk I/O request.
Nice
Percentage of time spent processing prioritized
tasks. Niced tasks are also counted in system and
user time.
Steal
Percentage of time a virtual CPU waits for a real
CPU while the hypervisor services another virtual
processor.
System
Percentage of time allocated to system processes.
User
Percentage of time allocated to user processes.
Linux Disk metrics
76
Metric
Definition
Disk usage
Percentage of disk space Cassandra uses at a
given time.
Free disk space
Available disk space in GB.
Used disk space
Used disk space in GB.
Disk read throughput
Average disk throughput for read operations in
megabytes per second. Exceptionally high disk
throughput values may indicate I/O contention.
Disk write throughput
Average disk throughput for write operations in
megabytes per second.
Disk read rate
Averaged disk speed for read operations.
Disk write rate
Averaged disk speed for write operations.
Disk latency
Average time consumed by disk seeks in
milliseconds.
Disk request size
Average size in sectors of requests issued to the
disk.
Using OpsCenter
Metric
Definition
Disk queue size
Average number of requests queued due to disk
latency.
Disk utilization
Percentage of CPU time consumed by disk I/O.
Windows memory metrics
Metric
Definition
Available memory
Physical memory that is not being used.
Pool nonpaged
Physical memory that stores the kernel and other
system data structures.
Pool paged resident
Physical memory allocated to unused objects that
can be written to disk to free memory for reuse.
System cache resident
Physical pages of operating system code in the file
system cache.
Windows CPU metrics
Metric
Definition
Idle
Percentage of time the CPU is idle.
Privileged
Percentage of time the CPU spends executing
kernel commands.
User
Percentage of time allocated to user processes.
Windows Disk metrics
Metric
Definition
Disk usage
Percentage of disk space Cassandra uses at a
given time.
Free disk space
Available disk space in GB.
Used disk space
Used disk space in GB.
Disk read throughput
Average disk throughput for read operations in
megabytes per second. Exceptionally high disk
throughput values may indicate I/O contention.
Disk write throughput
Average disk throughput for write operations in
megabytes per second.
Disk read rate
Averaged disk speed for read operations.
Disk write rate
Averaged disk speed for write operations.
Disk latency
Average time consumed by disk seeks in
milliseconds.
Disk request size
Average size in sectors of requests issued to the
disk.
Disk queue size
Average number of requests queued due to disk
latency.
77
Using OpsCenter
Metric
Definition
Disk utilization
Percentage of CPU time consumed by disk I/O.
Mac OS X memory metrics
Metric
Definition
Free memory
System memory that is not being used.
Used memory
System memory that is being used by application
processes.
Mac OS X CPU metrics
Metric
Definition
Idle
Percentage of time the CPU is idle.
System
Percentage of time allocated to system processes.
User
Percentage of time allocated to user processes.
Mac OS X disk metrics
Metric
Definition
Disk usage
Percentage of disk space Cassandra uses at a
given time.
Free space
Available disk space in GB.
Used disk space
Used disk space in GB.
Disk throughput
Average disk throughput for read/write operations
in megabytes per second. Exceptionally high disk
throughput values may indicate I/O contention.
Managing DataStax Enterprise Management Services
DSE comes with some enterprise management services which you can configure and run using
OpsCenter.
Repair Service
The Repair Service is configured to run continuously and perform repair operations across a DSE cluster in
a minimally impactful way.
The Repair Service runs in the background, constantly repairing small chunks of your cluster in order to
alleviate the pressure and potential performance impact of having to periodically run repair on entire nodes.
When the entire cluster has been repaired, the Repair Service re-calculates the list of subranges to repair
and starts over. Repairing the entire cluster one time is a cycle.
Note: DSE 3.0 or greater is required.
How Repair Service works
The Repair Service works by repairing small chunks of your cluster in the background. The service takes
a single parameter, time_to_completion, which is the maximum amount of time it takes to repair the entire
cluster once. Typically, you set this to a value lower than your lowest gc_grace_seconds setting (the
78
Using OpsCenter
default for gc_grace_seconds is 10 days). The service may run multiple repairs in parallel, but will run as
few as needed in order to complete within the amount of time specified, and will always avoid running more
than one repair in a single replica set.
The Repair Service uses an average of the throughput of recent repairs to calculate how many parallel
repairs can be completed in the current cycle.
Restarting opscenterd
The current state of the Repair Service is persisted locally on the opscenterd server every five minutes
by default. If opscenterd is restarted, the Repair Service resumes where it left off.
Known limitations
If a cluster is data center aware and has keyspaces using SimpleStrategy, the repair service will fail to
start
Changes in cluster topology
If a change in cluster topology occurs, the Repair Service stops its current cycle and waits for the ring to
stabilize before starting a new cycle. This check occurs every five minutes.
Topology changes:
•
•
•
•
Nodes moving
Nodes joining
Nodes leaving
Nodes going up/down
Changes in schemas
•
•
•
Keyspaces added while the repair service is running are repaired when the next subrange repair is
started.
Column families added to existing keyspaces are repaired immediately during the current cycle of the
Repair Service.
Keyspace or column family can be removed while the Repair Service is running without causing any
issues.
Starting and stopping the Repair Service
Starting and stopping the Repair Service
About this task
The service takes a single parameter, time_to_completion, which is the maximum amount of time it takes
to repair the entire cluster once. Typically, you set this to a value lower than your lowest gc_grace_seconds
setting (the default for gc_grace_seconds is 10 days).
Procedure
1. In the Services section of OpsCenter, toggle the Repair Service > Status button to On.
79
Using OpsCenter
2. In the Start Repair Service dialog, you enter a value for Time to completion field. (The default is 9
days.)
3. Select Start Repair Service.
The Repair Service status updates after one minute.
Stopping the Repair Service
About this task
Stopping the Repair Service is realtively straightforward.
Procedure
Toggle Services > Repair Service > Status from On to Off.
Any repairs that are running when the repair service is stopped run until completion. They are not
cancelled.
Checking progress
Progress for the current cycle can be seen in the Services section of the OpsCenter UI.
The Activity column
The Activity column in Services > Repair Service displays the progress of the current cycle as a
percentage of subranges that have been fully repaired. The Activity column updates every minute, or
when you click on the Services section in the left-hand navigation.
80
Using OpsCenter
Logging
All Repair Service activity is logged to:
Package install location
/var/log/opscenter/repair_service/<cluster_name>.log
Tarball install location
<install_location>/log/repair_service/<cluster_name>.log
The log file is automatically rotated at ~9.5MB, keeping up to five rotated logs. This is not currently
configurable.
Repair Service error handling
The Repair Service handles the following errors.
Error of a single range repair
When a single range repair fails, the repair is skipped temporarily and added to the end of the queue of
repairs and retried later. If a single range fails ten times, the Repair Service shuts down and fires an alert.
Too many errors in a single run
After a total of 100 errors during a single run, the Repair Service shuts down and fires an ALERT.
Time-outs
The Repair Service times out a single repair command after one hour. This counts towards an error for that
repair command and it is placed at the end of the queue of repairs and retried later.
Too many repairs in parallel
The Repair Service errors and shuts down if it has to run too many repairs in parallel. By default, this happens
if it estimates that it needs to run more than one repair in a single replica set to complete on time.
Note: These are all configurable.
Advanced configuration
The following is currently configurable by adding a [repair_service] section to the
opscenterd.conf file to apply to all clusters, or per-cluster by adding the section to the <cluster
name>.conf file. Settings in <cluster name>.conf override any settings in opscenterd.conf.
log_directory = /var/log/opscenter/repair_service/
The directory to log the repair service log to.
log_length = 10485760
The maximum size (in bytes) the repair service logs can grow before rotation (default 10485760 bytes)
max_rotate = 10
The maximum number of rotated repair service logs to keep (default 10)
persist_directory = /var/lib/opscenter/repair_service/
The directory to persist the repair service status to.
persist_period = 300
81
Using OpsCenter
How often to persist the repair service status (in seconds).
restart_period = 300
How often to check for a stable ring during topology changes (in seconds).
single_task_err_threshold = 10
How many times a single repair command must fail before the service shuts down.
max_err_threshold = 100
The maximum number of errors we can see during a single cycle of the repair service before shutting down.
max_parallel_repairs = 0
A cap on the maximum number of parallel repairs we will perform, taking precedence over our calculation
based on number of nodes/RF.
min_throughput = 512
The minimum number of bytes a single range repair must contain in order to be included in calculations to
determine how many parallel repairs to run (default: 512).
single_repair_timeout = 3600
The timeout placed on each repair command (in seconds).
min_repair_time = 5
The minimum amount of time we expect each repair command to take. If commands run faster, we will
sleep the difference. (in seconds).
snapshot_override
Enable or disable snapshot repairs by setting to true or false, respectively. This option is left unset by
default. If snapshot_override is unset, snapshot repairs will be enabled automatically if the version of
DataStax Enterprise or Cassandra supports them.
num_recent_throughputs
The number of recent throughputs used to calculate the average throughput, which is then used to determine
how many parallel repairs are needed. The default value is 20.
Capacity Service
Using trend analysis and forecasting, the Capacity Service helps users understand how their cluster is
performing within its current environment and workload, and gain a better sense of how time affects those
trends, both past and future.
Several types of metrics are collected, including Cassandra specific and platform specific (for example,
disk metrics, network metrics), at both the node and column-family level (where applicable). These metrics
are stored in Cassandra, on the cluster being managed by default. That data can be stored on a separate,
dedicated cluster as well.
Trend Analysis
The Trend Analysis component of the Capacity Service allows users to view historical metrics for any node
or column family, as well as aggregates across the entire cluster.
Forcasting
Using the Forecast feature, users are able to see a predicted trend for any metric-based on historical data
that has been collected.
Using forecasting
Use forecasting to predict trends in graphs based on past performance.
About this task
82
Using OpsCenter
Procedure
1. In the Performance section of the OpsCenter web interface, locate the graph you'd like to forecast and
click the Forecast icon.
2. In the Forecast dialog, enter the date and time you'd like to use for the end of the trend prediction.
The end date and time must be a minimum of two days into the future and a maximum of one year.
3. Click the Forecast button.
A predicted trend using polynomial curve fitting against historical data displays.
The forecast above shows that our node will be at 60% disk usage in a month, so we should probably
start thinking about adding capacity now.
Advanced configuration of forecasting
To configure forecasting.
The following is currently configurable by adding a [forecasting] section to the opscenterd.conf
file.
polyfit_degree = 2
Which degree polynomial equation to use in calculating the forecasting. (The default is three.)
83
Using OpsCenter
range_multiplier = 3
In order to generate a meaningful prediction, historical data is analyzed for a period larger than the range
being forecasted. (The default is three times larger than the range being forecasted.) For example, to
forecast one month into the future, three months of data is analyzed.
required_data_percentage = 0.5
The percentage of the historical data required to be available. (The default is 50%.)
Managing backups and restoring from backups
Using OpsCenter Enterprise Edition, you can take, schedule, and manage backups across all registered
clusters.
A backup is a snapshot of all on-disk data files (SSTable files) stored in the data directory. Backups are
taken per keyspace and while the system is online. A backup first flushes all in-memory writes to disk, then
makes a hard link of the SSTable files for each keyspace. Backups are stored in the snapshots directory
of the column family that's being snapshotted. For example, /var/lib/cassandra/data/OpsCenter/
settings/snapshots.
You must have enough free disk space on the node to accommodate making snapshots of your data files.
A single snapshot requires little disk space. However, snapshots will cause your disk usage to grow more
quickly over time because a snapshot prevents old obsolete data files from being deleted. OpsCenter Data
Backups allows you to specify a schedule to remove old backups and prevent backups from being taken
when disk space falls below a specified level.
Note: OpsCenter Data Backups does not show or manage manual snapshots taken using the
nodetool snapshot command.
Scheduling a backup
To schedule a backup:
Procedure
1. In the OpsCenter Dashboard, click Data Backups.
2. Click Schedule Backup.
3. In Add Backup, select the backup parameters:
a) Select a Keyspace to backup—Select the keyspace that you want to back up.
b) Schedule—Select a frequency and timezone for your backup. GMT is the default timezone.
c) Cleanup—Choose a frequency to remove old backups. (If not specified, you should manually
cleanup snapshots.)
d) Run Script Before—Optionally, enter the name of a script that will be run before the backup is
started. See Using custom scripts before and after backups for details.
e) Run Script After—Optionally, add the name of a script that will be run after the backup is
completed. See Using custom scripts before and after backups for details.
4. Click Save.
5. To set the percentage of free disk space at which backups are prevented, click Configure and then
enter the appropriate information.
The percentage of free disk space that you set applies to all nodes in the cluster. Detailed information
about the backup is recorded in the Event Log.
Restoring from a backup
You can restore from any local backups that have been run by OpsCenter, but not from snapshots run from
nodetool. You can pick any subset of column families that exist in the snapshot to restore.
84
Using OpsCenter
About this task
To restore a backup:
Procedure
1.
2.
3.
4.
5.
6.
7.
Click Data Backups.
Find the backup you wish to restore in the list of backups.
Click Restore.
Choose which keyspaces to restore.
(Optional) Choose individual column families within a chosen keyspace to restore.
(Optional) Click the Truncate/delete existing data before restore checkbox.
(Optional) Click the Throttle stream throughput at ____ MB chekbox and enter the desired throttling
value.
8. Click Restore.
Using custom scripts before and after backups
You can configure custom scripts will run before or after a backup.
Scheduled backups can be configured to run custom scripts before and after the backup is performed.
These custom scripts need to be located in /usr/share/datastax-agent/bin/backup-scripts
for package installations, or in <install_location>/bin/backup-scripts for tarball installations.
This directory also contains example scripts. The scripts need to be executable, and are run as the
DataStax agent user (by default datastax-agent). Any custom scripts should exit with a status of 0 if all
operations completed successfully. Otherwise, it should exit with a non-zero status to indicate a failure.
Post-backup scripts are sent a list of files in the backup to stdin, one file per line, but will not have any
arguments passed to them.
Data modeling
Keyspaces
You can create a new keyspace or manage keyspaces.
Selecting the Data Modeling area in the OpsCenter console lists the keyspaces in the cluster that you are
monitoring.
When you create a keyspace, you give it a name, choose a replica placement strategy, the total number of
replicas you want, and how those replicas are divided across your data centers (if you have a multiple data
center cluster).
Creating a keyspace
About this task
To create a new keyspace.
Procedure
1. Select Schema in the OpsCenter console.
2. Select Add in the Schema section of OpsCenter.
3. Give the keyspace a name. Keyspace names should not contain spaces or special characters or
exceed the filename size limit of your operating system (for example, 255 bytes on most Linux file
systems). Keyspace names are case sensitive.
85
Using OpsCenter
4. Set the replica placement strategy. The replica placement strategy (along with the cluster-configured
snitch) determines how replicas are placed on nodes throughout the cluster. Use one of three built-in
replica placement strategies:
a) SimpleStrategy—Single data center, rack unaware replica placement. This is the default strategy.
b) NetworkTopologyStrategy—Single or multiple data center, rack aware replica placement. This is
the recommended strategy.
c) OldNetworkTopologyStrategy—Two data centers only, rack aware replica placement. This
strategy is deprecated.
5. Choose how many total copies that you want of your keyspace data (replication factor). The
NetworkTopologyStrategy requires you to configure how many replicas you want per data center. The
data center name you enter should match the data center name used by your cluster-configured snitch.
Make sure to name the data center(s) correctly according to your snitch configuration.
6. If you do not want to start defining column families within your new keyspace right away, uncheck the I
would like to create a Column Family checkbox.
7. Click Save Keyspace.
Managing a keyspace
About this task
To manage a keyspace:
Procedure
1. Select Schema in the OpsCenter console.
2. From the list of keyspaces, select one of the keyspaces.
In Keyspace Settings, the replica placement strategy options for the keyspace appear.
3. From the list of column families (below Settings), select a column family to view its properties and to
view or change performance tuning metrics.
4. Click Add in the Column Family to add a column family to the keyspace.
5. Click Delete to delete the keyspace.
Managing column families
When you create a column family in Cassandra using an application, the CLI, or CQL 2 or earlier, the
column family appears in OpsCenter. You can use Schema to manage the column family.
You can also create one type of column family: the dynamic column family. Dynamic column families
are those that do not specify column names or values when the column family is created. An application
typically supplies this metadata. CQL 3, the default query language in Cassandra, does not support
dynamic column families. Earlier versions of CQL and the CLI support dynamic column families.
This version of OpsCenter does not support defining static column families (per-column meta data), row
key data types, or schema information for super column sub-columns described in Cassandra 1.0, or
earlier.
Creating a dynamic column family
About this task
To create a new dynamic column family:
Procedure
1. Select Schema in the OpsCenter console.
2. From the list of keyspaces, select the keyspace to contain the column family.
86
Using OpsCenter
3. Give the column family a name.
Column family names should not contain spaces or special characters and cannot exceed the filename
size limit of your operating system (255 bytes on most Linux file systems).
By default, column families are created with standard columns (column_type: Standard). If you want a
column family containing super columns choose column_type: Super.
4. Use compare_with to set the default data type for column names (or super column names). Setting
the default data type also sets the column sort order for the column family. For example, choosing
LongType would sort the columns within a row in numerical order. The sort order cannot be changed
after a column family is created, so choose wisely.
5. Use default_validation_class to set the default data type for column values (or super column subcolumn values). Always set this for dynamic column families.
6. Click Save Column Family.
Managing a column family
About this task
To manage a column family:
Procedure
1. Select Schema in the OpsCenter console.
2. From the list of keyspaces, select a keyspace.
The #CFs columns shows how many column families each keyspace contains.
3. From the list of the column families, select a column family.
Click one of the following buttons:
•
•
Add—See above.
Delete—Completely removes the column family from the keyspace. You may select more than one
column family in a keyspace to delete.
• View Metrics—Presents metrics for a column family. In the Metric Options dialog, select a column
family (CF) metric to view. To aggregate measurements across the entire cluster, all nodes in the
data center, or in a particular node, select Cluster Wide, All Nodes, or the IP address of a node. At
this point, you can add a graph of the measurements to the Performance Metrics area, or choose a
different column family to measure.
• Truncate—Deletes all data from the column family but does not delete the column family itself.
Removal of the data is irreversible.
4. When you select a column family, you see a list of manageable attributes: Properties (fully editable),
Metadata (Add or Delete), and Secondary Indexes (Add or Delete).
Browsing data
Browsing a Cassandra database
About this task
To browse a Cassandra database
Procedure
1. Select Data Explorer in the OpsCenter console.
A list of keyspaces in the cluster appears. By default, the list includes the OpsCenter keyspace, which
contains column families of data in the cluster.
2. Click one of the keyspaces. For example, click the OpsCenter keyspace.
87
Using OpsCenter
The list of column families appears.
3. Click a column family. For example, click events_timeline and expand the OpsCenter console
window, so you can see more values.
A row keys, columns, and data values of the events_timeline column family appear in tabular
format. You may notice that some data values are hex-encoded, not human-readable values, and other
values are perfectly readable. Internally, Cassandra stores row key names, column names and column
values as hex byte arrays. If possible, OpsCenter converts data to text you can read.
4. Click the heading row of the table to see the sort control (the arrow on the right). Toggle the sort control
to sort the rows in ascending or descending order.
5. To browse through the data, use the scroll bar or click Next to page through the data.
Examining row data
About this task
To zoom in on a particular row:
Procedure
1. Assuming the events_timeline column family is still selected, type the key for a row in the Key text entry
box.
For example, type the row key timeline_start. If you set up a secondary index on the column family, the
Key drop-down lists key choices.
2. Click the search control (the hourglass icon).
The columns and values of the row appear.
Removing all traces of Cassandra or DSE packages
After a failed attempt to provision a node or a cluster, you must remove all traces of Cassandra or DSE
packages from the affected nodes.
Procedure
1. Log into each node.
2. Run the following BASH script on each node using sudo.
# Stop services
/etc/init.d/cassandra stop
/etc/init.d/dse stop
/etc/init.d/datastax-agent stop
# Remove packages
PACKAGES=(dsc dsc1.1 dsc12 cassandra apache-cassandra1 dsc-demos \
dse dse-libhadoop-native dse-libhadoop dse-libcassandra dse-hive dselibhive dse-pig \
dse-libpig dse-demos dse-libsqoop dse-libtomcat dse-liblog4j dse-libsolr
dse-libmahout dse-full)
DEB_PACKAGES=(python-cql python-thrift-basic)
RPM_PACKAGES=(python26-cql python26-thrift)
if [ `which dpkg` ]; then
PLIST=(${PACKAGES[@]} ${DEB_PACKAGES[@]})
dpkg -P ${PLIST[*]}
rm -rf /etc/apt/sources.list.d/datastax.list
else
PLIST=(${PACKAGES[@]} ${RPM_PACKAGES[@]})
yum -y remove ${PLIST[*]}
88
Using OpsCenter
rm -rf
fi
/etc/yum.repos.d/datastax.repo
# Cleanup log and configuration files
rm -rf /var/lib/cassandra/* /var/log/{cassandra,hadoop,hive,pig}/* /etc/
{cassandra,dse}/* \
/usr/share/{dse,dse-demos} /etc/default/{dse,cassandra}
•
•
copy, paste, and run the script from the command line
save the script to a local file and copy to each node to run from the command line
89
Troubleshooting
Troubleshooting
This section lists some common problems experienced with OpsCenter and solutions or workarounds.
Internet Explorer web browser not supported
If you try to load the OpsCenter client in Microsoft Internet Explorer, a dialog displays informing you that it
is not supported.
OpsCenter is only supported on the latest versions of:
•
•
•
Apple Safari
Google Chrome
Mozilla Firefox
The SSTables in this snapshot '<tag>' are not compatible
If you receive an error message that includes "The SSTables in this snapshot '<tag>' are not compatible
with the current version of Cassandra", it means you must upgrade your snapshot to the current major
version of Cassandra or DSE.
How to
1. Log in to each node.
2. Run the sstableupgrade script for every keyspace and column family you wish to restore, passing it the
keyspace, column family, and OpsCenter snapshot tag that you got from the error message.
How you run the script depends on how you installed Cassandra or DSE.
3. Retry the restore from OpsCenter.
OpsCenter data growing too large
A bug, which has been fixed in 3.2.1, was not setting a TTL on metrics data being collected for a managed
cluster. Depending on your enviornment, this could cause some column families in the OpsCenter
keyspace to grow too large. The most common offenders will be the pdps (raw data points) and rollups60
(1m data points) column families.
If any of the column families have grown too large, you can truncate them to reclaim the space. If you
are not comfortable losing historical data for that granularity (for example, 1m), please contact DataStax
support.
Cannot create a keyspace
Due to a Python 2.6 or earlier bug, some users experience a problem adding a keyspace using Data
Modeling OpsCenter features. OpsCenter cannot save a newly created keyspace.
Upgrading Python generally fixes this problem.
Error exceptions.ImportError:libssl.so.0.9.8
Occurs when OpenSSL 1.0.0 is installed on RHEL 5.x, CentOS 5.x, OEL 5.5, Debian, or Ubuntu systems:
message.exceptions.ImportError: libssl.so.0.9.8
90
Troubleshooting
To fix, you can do either of the following:
•
Install OpenSSL 0.9.8:
•
1. RHEL-based systems: sudo yum install openssl098
2. Debian-based systems: sudo apt-get install libssl0.9.8
Install Python libssl 1.x:
1. Remove all OpenSSL modules from the OpsCenter installation lib directory. You do not have to
remove them globally.
Packaged installs: /usr/share/opscenter/content/lib
Tarball installs: install_location/lib
You can easily find these directories using:
find . -name OpenSSL -type d | xargs rm -rf
2. Install the latest pyOpenSSL module:
Debian-based systems: apt-get install python-openssl
RHEL-based systems: yum install python-openssl
3. Start or restart opscenterd.
Python used to run OpsCenter not built with SSL
In order to protect your AWS credentials when launching EC2 instances, OpsCenter needs to use HTTPS,
but if the version of Python that is running opscenterd was not compiled with SSL support OpsCenter will
not run even if SSL has been disabled in the configuration file.
To resolve the issue, first ensure that OpenSSL 0.9.8 is installed on your system. If you have compiled
Python manually, it is recommended that you install Python 2.6+ through your package manager. On
CentOS and RedHat Enterprise Linux, this is most easily done through EPEL packages.
If you must compile Python manually, make sure that SSL support is enabled. This blog post explains the
process for Python 2.5, but the same steps will work for Python 2.6 or 2.7.
DataStax agent port setting conflict
If you have a problem with OpsCenter, check for conflicts in port settings. The DataStax Agent uses port
7199 by default. If you have not changed the default port, check that Cassandra or another process on the
node, is not set up to use this port.
If you set the DataStax Agent port to a host name instead of an IP address, the DNS provider must
be online to resolve the host name. If the DNS provider is not online, intermittent problems should be
expected.
Limiting the metrics collected by OpsCenter
If you have many column families, the number of metrics OpsCenter collects can become quite large.
For information about how to reduce the number keyspaces or column families that are monitored, see
Controlling data collection.
91
Troubleshooting
Java not installed or JAVA_HOME environment variable not set
If Java is not installed or if OpsCenter cannot find JAVA_HOME, you may see an error such as:
/usr/share/datastax-agent/bin/datastax-agent: line 98:exec: -X: invalid
option
exec: usage: exec [-cl ] [-a name ] [ command [arguments ... ]]
[redirection ... ]
To correct this problem, install Java or set JAVA_HOME:export JAVA_HOME =<path_to_java>
Insufficient user resource limits errors
Insufficient resource limits may result in an insufficient nofiles error:
2012-08-13 11:22:51-0400 [] INFO: Could not accept new connection
(EMFILE )
See Recommended settings under Insufficient user resource limits errors in the Cassandra documentation.
Installing EPEL on CentOS 5.x or RHEL 5.x
Before installing OpsCenter on CentOS 5.x or RHEL 5.x, you must install EPEL (Extra Packages for
Enterprise Linux). EPEL contains dependent packages, such as Python 2.6+.
To install for both 32- and 64-bit systems:
$ sudo rpm -Uvh http://dl.fedoraproject.org/pub/epel/5/i386/epelrelease-5-4.noarch.rpm
Note: You do not have to install EPEL on other other machines.
Problems with provisioning
General troubleshooting steps
Ensure firewalls are properly opened between the opscenterd machine and each node.
Check the following files (on any of the nodes having problems) for any errors:
•
•
/var/log/cassandra/system.log
/var/log/datastax-agent/agent.log
Verify that Cassandra (or DSE) was not previously installed on any of the machines; if it was, all binaries,
configuration files, logs, etc. must be cleared out first.
Invalid repository credentials
Invalid repository credentials
Debian
When installing DSE, if you enter invalid values for DataStax Credentials, an error dialog displays text
along the lines of:
Installation stage failed: Installation failed on node 172.16.1.2: 'apt-get
update' failed.
92
Troubleshooting
Clicking Details displays the entire output from both stdout and stderr. If, in this output, you see the "401
Unauthorized", it means that the credentials entered were invalid.
RHEL
When installing DSE, if you enter invalid values for DataStax Credentials, an error dialog displays text
along the lines of:
Installation failed on node 172.16.1.2: 'yum install' failed.
Clicking Details displays the entire output from both stdout and stderr. If, in this output, you see "The
requested URL returned error: 401", it means that the credentials used were invalid.
Timed out waiting for Cassandra to start
If you receive this error, it most likely means that the Cassandra process failed to start on one or more
nodes. You should look in /var/log/cassandra/system.log for any errors that may need to be
resolved.
The following packages are already installed
If you receive an error that starts with this message, it means Cassandra (or DSE) is already installed on
the system. OpsCenter provisioning requires that any instances of Cassandra (or DSE) be completely
removed or purged before provisioning a new cluster.
Agents cannot connect to opscenterd
If you receive an error message that includes "The installed agent doesn't seem to be responding", there is
most likely a firewall issue preventing the installed agent from connecting to the opscenterd machine. You
should ensure that port 61620 is open on the opscenterd machine and check the agent logs.
Removing all Cassandra or DSE files after failed provisioning
If you provision a new cluster or add a node to an existing cluster via the OpsCenter UI and an error
occurs, you can retry the request, but before you retry you must remove all traces of Cassandra or DSE
packages from the nodes in the cluster. You can also remove simply remove the files if you do not wish to
retry the provisioning.
Running sstableloader results in broken data distribution
Running sstableloader on Cassandra versions prior to 1.2.12 or DataStax Enterprise 3.1.4 or earlier
results in broken data distribution. For more information, see CASSANDRA-6272.
We recommend upgrading to DataStax Enterprise 3.1.5 or higher, or Cassandra 1.2.12 or higher.
Cassandra 2.0.x is unaffected.
Timeout connecting to Cassandra 2.0 clusters
OpsCenter will timeout when connecting to Cassandra 2.0 clusters when rpc_server_type=hsha in
cassandra.yaml.
Due to a bug in Cassandra 2.0 (CASSANDRA-6373), the connection from opscenterd over thrift will
hang, causing a timeout in OpsCenter. The workaround is to change the rpc_server_type setting to
rpc_server_type=sync in cassandra.yaml.
93
Troubleshooting
Sophos Web Protection breaks browser access to OpsCenter on Windows
Sophos Web Protection prevents the OpsCenter UI from loading. The JavaScript console for the web
browser will show 404 errors, and the interface will not load. Adding a firewall rule in Web Protection
allowing access to port 8888 or disabling Web Protection fixes this issue.
94
OpsCenter API reference
OpsCenter API reference
The OpsCenter API facilitates the development of websites and programs to retrieve data and perform
Cassandra administrative actions. The OpsCenter API includes RESTful requests for programmatically
performing the same set of operations as the OpsCenter GUI.
95
Release Notes
Release Notes
Information about new features and resolved issues in the following releases.
4.1.4
Resolved issues
•
•
Latency, key and row cache hit rate, and network metrics are now displayed with the correct unit.
(OPSC-2976)
Fixed an issue with storing metrics in a separate cluster. (OPSC-2977)
4.1.3
New features
•
•
•
•
Optimized performance of loading graphs
Keyspaces added while Repair Service is running now included in current cycle
Option to drain nodes prior to stop/restart (OPSC-1250)
Option for opscenterd to only connect to seed hosts (OPSC-549)
Resolved issues
•
•
•
•
•
•
•
Race condition when opscenterd is restarted, but not the agents (OPSC-2485)
Post snapshot scripts not working when backing up "All Keyspaces" (OPSC-2576)
Obscure passwords in log files (OPSC-2554)
Errors in agent logs when compacting 0 bytes (OPSC-2546)
Agent not starting up when it can't resolve its own hostname (OPSC-2495)
Error when clicking the Fix link to install agents from the Overview tab (OPSC-2392)
Range calculation for repair service crossed tokens from other data centers (OPSC-2806)
Known issues
•
Graphs that display metric data that require conversion display unconverted data. The following metrics
display data that is off by a factor of 1000, in ms/op instead of Вµs/op:
•
•
•
•
write req latency
read req latency
cf write latency
cf read latency
The following metrics should display percentages, but only display values beteen 0 and 1:
•
•
key cache hit rate
row cache hit rate
The following metrics display kb/sec but their values are actually kb/min:
•
•
•
96
network received
network sent
total network traffic
Release Notes
•
The data is stored correctly and alerts work correctly, but the numbers on the graph are displayed
incorrectly. (OPSC-2976)
A missing option in the configuration file results in a problem storing metrics in a separate cluster.
(OPSC-2977)
4.1.2
New features
•
•
•
•
•
•
Significantly improved cassandra.yaml auto-detection on the agent, which no longer requires
manually configuring its location for tarball installs of DSE/Cassandra. (OPSC-2331)
DSE workload now displayed on node info popup. (OPSC-2388)
Several repair service options can now be configured per cluster. (OPSC-2254)
New streams indicator and dialog added to the Ring tab. (OPSC-2206)
Metrics can now be aggregated by datacenter. (OPSC-2179)
Rebalance proposal now displays rings separated by datacenter. (OPSC-2005)
Resolved issues
•
•
•
•
•
•
•
•
•
•
Fixed upgrade path from pre-4.0 installs of OpsCenter to 4.0 or greater. (OPSC-2435)
Definition updater no longer stops opscenterd from running if OpenSSL is not installed. (OPSC-2510)
Fixed repair service not running in multi-dc environments where all keyspaces exist across all
datacenters. (OPSC-2448)
Repair service not properly repairing CQL3 column families. (OPSC-2509)
Fixed choosing a node when a mismatch is found in Configure Cluster. (OPSC-2489)
Cluster wide aggregates for column family metrics are now properly displayed in the graph legend.
(OPSC-2470)
Fixed running backups on super column families in Cassandra 2.0. (OPSC-2381)
Fixed status updates for rolling restarts. (OPSC-2379)
Added an option to not force agent to use sudo for commands on agent. (OPSC-2122)
Fixed SSH connections to nodes failing because of SSH welcome message. (OPSC-579)
4.1.1
Resolved issues
•
•
Fixed issue with opscenterd breaking when updating definition files on platforms with Python 2.6.
A DataStax Agent now gracefully dies if it can't properly bind its HTTP server, rather than re-trying in a
loop.
Known issues
•
Starting OpsCenter after an upgrade on RHEL or CentOS sometimes fails with a pyOpenSSL error.
4.1.0
New features
•
•
•
New CF: Total Memtable Size metric for in-memory tables
New dialog when clicking a node, redesigned to display much more information
Custom date/time ranges are now supported in the Performance section
97
Release Notes
•
Graph configuration is now much more flexible
•
•
• Add multiple metrics to a single graph
• Add any subset of nodes and column families to a graph
Improved load time of Performance section
UI Navigation Changes
•
•
• Create Cluster and Add Cluster links have been merged into a single New Cluster link
• Ring View and List View have been merged into a single Nodes section
Definition files now automatically update. For information, see Configuring OpsCenter definition file
updates
OS X Mavericks support (dev only)
Resolved issues
•
•
•
Fixed issue with trying to grab too much metric data from Cassandra (symptom was "was too wide to
multiget" warnings)
API requests getting incorrectly cached by the browser
OpsCenter column families now use Cassandra's default compression settings
Known issues
•
Definition file updates don't work with Python 2.6
When an auto-update is performed on platforms with Python 2.6, an error occurs during the update
process which causes the current definition files to be removed automatically, preventing you from
managing and provisioning clusters, and preventing other OpsCenter operations from completing.
Workaround:
1. Add the following to opscenterd.conf to disable auto-update:
[definitions]
auto_update = False
2. Manually download the latest definition files.
a. In a terminal, go to the definitions directory. On package installations, this is /etc/
opscenter/definitions. On tarball installations, this is <opscenter-installlocation>/conf/definitions.
b. Download and install the latest definition files.
$ curl https://opscenter.datastax.com/definitions/4.1.0/
definition_files.tgz | tar xz
3. Restart opscenterd as described in Starting, restarting, and stopping OpsCenter.
4.0.3
New features
•
•
•
Provisioning support for latest DSE and DSC releases
Improved Repair Service throughput calculation, which is used to determine the number of parallel
repairs to run
Added repair service config option to force disabling snapshot repairs
Resolved issues
•
•
•
•
98
Repair service properly resumes after opscenterd restart
Fixed graphs using the "All Column Families" option
Fixed issues upgrading from opscenter-agent to datastax-agent
Fixed collecting Solr metrics in some environments
Release Notes
•
•
Fixed stack overflow in agent metric rollup code
Fixed permission errors when running backups on empty column families
4.0.2
New features
•
•
•
•
Ability to provision DataStax Community 2.0 nodes.
Status labels on nodes in ring view replaced by animation. Special statuses include moving, restarting,
and joining.
Log rotate properties for repair service logs are now configurable. See Advanced configuration for more
information.
The interface used by the opscenterd web server now defaults to 0.0.0.0 in opscenterd.conf.
Resolved issues
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Repair Service stopping because it requires an unnecessary number of parallel repairs.
Added the ability to provision DataStax Enterprise 3.2.x nodes.
Errors installing agents via the "Fix" link from tarball installations.
Installing agents via the "Fix" link on CentOS/RHEL distributions will no longer pull the latest version
from DataStax repositories.
Issue loading the Performance section when a preset group contains removed metrics.
Error starting up agent when local_interface is specified in address.yaml.
Exclude down nodes from Data Size calculations in ring view health summary.
Vertical scrollbar in Ring View no longer resets on update.
Passwords used for the nodetool and cassandra-cli commands are no longer logged by the
agents.
Repair Service status updates much sooner after starting within the UI.
Issue generating diagnostics tarball on nodes that had JMX/THRIFT authentication enabled.
Backup times incorrectly labeled as GMT.
Browser memory leaks in ring view.
Issue in agent.log when changing the log level. This fix includes requiring a restart of the agent when
changing the log level.
Diagnostics tarballs will no longer be allowed to fill up remaining disk space.
4.0.3
New features
•
•
•
Provisioning support for latest DSE and DSC releases
Improved Repair Service throughput calculation, which is used to determine the number of parallel
repairs to run
Added repair service config option to force disabling snapshot repairs
Resolved issues
•
•
•
•
•
•
Repair service properly resumes after opscenterd restart
Fixed graphs using the "All Column Families" option
Fixed issues upgrading from opscenter-agent to datastax-agent
Fixed collecting Solr metrics in some environments
Fixed stack overflow in agent metric rollup code
Fixed permission errors when running backups on empty column families
99
Release Notes
4.0.1
Resolved issues
•
Fixed issue with "Fix" link to deploy agents uninstalling DataStax Enterprise 3.2 RPM/Debian packages.
4.0
New features
•
Enterprise functionality now limited to DataStax Enterprise clusters.
•
•
•
•
Scheduled functionality such as alerts and backups previously configured on non-DSE
clusters now only run on DSE clusters.
• Merged the opscenter and opscenter-free packages into a single opscenter package.
Rebranded the OpsCenter agent as the DataStax agent.
SSL communication between opscenterd and the agents is now off by default.
Management support for new Enterprise services.
•
•
• Repair Service
• Capacity Service
Support for Cassandra 2.0
Redesigned some portions of the GUI:
•
•
•
• List View now has a bulk operations checkbox
• Ring View is now split up by datacenter
Re-added ability to add new nodes to an existing DSE cluster with vnodes enabled.
The Data Explorer is being deprecated in favor of DevCenter.
Ability to retry failed provisioning attempts.
Resolved issues
•
•
•
Ability to rebalance clusters that have some datacenters with vnodes enabled and others with vnodes
disabled.
Fixed issue with storing OpsCenter data in a separate cluster.
Fixed JMX authentication from the agent.
Known issues
•
•
•
100
Not able to provision new clusters on Cassandra 2.0 yet.
Adding existing clusters with rpc_server_type = hsha times out.
Clicking "Fix" when redeploying DataStax 3.2 agents uninstalls DataStax Enterprise 3.2 on the nodes
when DataStax Enterprise 3.2 is installed as an RPM or Debian package. The DataStax Enterprise 3.2
data on each node remains intact.
Tips for using DataStax documentation
Tips for using DataStax documentation
Navigating the documents
To navigate, use the table of contents or search in the left navigation bar. Additional controls are:
Hide or display the left navigation.
Go back or forward through the topics as listed in
the table of contents.
Toggle highlighting of search terms.
Print page.
See doc tweets and provide feedback.
Grab to adjust the size of the navigation pane.
Appears on headings for bookmarking. Right-click
the В¶ to get the link.
Toggles the legend for CQL statements and
nodetool options.
Other resources
You can find more information and help at:
•
•
•
•
•
•
Documentation home page
Datasheets
Webinars
Whitepapers
Developer blogs
Support
101
Send us feedback
Send us feedback
About this task
Thanks for using OpsCenter Enterprise Edition. Please take a moment to let us know what you think.
Procedure
1. Start OpsCenter and click Feedback at the top of the console.
The feedback form appears.
2. Enter your name, email address, and feedback.
3. Click Send Feedback.
We appreciate your comments.
102
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement