Upgrading Cisco Elastic Services Controller. Cisco Elastic Services Controller
Reklam
Reklam
C H A P T E R
20
Upgrading Cisco Elastic Services Controller
Cisco Elastic Service Controller supports two type of upgrades:
• Backup and Restore Upgrade : This upgrade process involves stopping the ESC keepalive daemon (for
ESC HA), backing up the database, stopping and renaming (or deleting) the ESC instances, re-installing the ESC instances, and restore database. For information on the supported ESC versions for ESC 5.3
upgrade, see the table below.
• In-service upgrade : ESC supports in-service upgrade for Active/Standby high availability nodes with a minimum downtime.
You can upgrade the ESC instance as a standalone instance or as a high availability pair. The upgrade procedure is different for standalone and high availability pair.
This chapter lists separate procedures on how to upgrade ESC standalone and ESC High Availability instance.
You must review these instructions before you decide to upgrade the ESC instance. See the
for more information on the installation scenarios.
• ESC only support direct upgrade from previous two minor releases. For example, ESC 2.3 will support direct upgrade from ESC 2.1 and ESC 2.2. For any release older than the supported versions for direct upgrade, you need to perform the staged upgrade.
Note Example of an ESC major release, ESC 3.1
Example of an ESC minor release, ESC 3.1.1
Example of an ESC patch release, ESC 3.1.0.116.
• Upgrading ESC using RPM Package (referred to as RPM Upgrade in this chapter) applies only to the
ESC upgrade between ESC patch releases with the same release number. For Example, the upgrade from
ESC 3.1.0.116 to ESC 3.1.0.150.
• If you want to upgrade ESC between minor releases (for example, upgrade from ESC 2.3.1 to ESC
2.3.2) or major releases (for example, upgrade from ESC 3.0 to ESC 4.0 ), you can upgrade through
Backup and Restore upgrade process using qcow2 image.
• For ESC upgrade, you should be familiar with ESC installation process.
• For OpenStack, refer to the OpenStack installation procedures, see Chapter 4: Installing Cisco Elastic
Services Controller in OpenStack.
• For VMware, refer to the VMware Installation installation procedures, see Chapter 7: Installing
Cisco Elastic Services Controller in VMware vCenter.
Cisco Elastic Services Controller 5.4 Install and Upgrade Guide
141
Upgrading Cisco Elastic Services Controller
Upgrading Standalone ESC Instance
• For ESC HA, please refer to the ESC HA installation procedures, see Chapter 5 : Configuring High
Availability for OpenStack and Chapter 8: Configuring High Availability for VMware.
Table 6: Supported ESC Versions for Upgrading to ESC 5.4
Virtual Infrastructure
Manager
OpenStack
VMware
CSP
Supported Versions for Backup and
Restore Upgrade
Supported Versions for In-Service Upgrade
5.3, 5.2
5.3, 5.2
5.3, 5.2
5.3, 5.2
5.3, 5.2
5.3, 5.2
IMPORTANT NOTES
• After upgrading to the new ESC version, ESC service will manage the life cycle of all VNFs deployed in the previous release. To apply any new features (with new data models) to the existing VNFs, you must undeploy and redeploy these VNFs.
• Upgrade is supported for Active/Active HA only.
Note Do not change the VIP in ESC upgrade in ETSI deployment.
If you change ETSI's REST schema, from http to https in-flight, ESC stops sending recovery notification from
ESC core for any existing deployment.
•
Upgrading Standalone ESC Instance, on page 142
•
Upgrading ESC HA Active/Standby Instances, on page 144
•
In-Service Upgrade of the ESC HA Active/Standby Nodes in OpenStack, on page 146
•
In-Service Upgrade of ESC HA Active/Standby Nodes in Kernel-Based Virtual Machine (KVM), on page 150
•
In-Service Upgrade of the ESC HA Active/Standby Nodes in VMware, on page 154
•
In-Service Upgrade of the ESC HA Active/Standby Nodes in CSP, on page 158
Upgrading Standalone ESC Instance
To upgrade standalone ESC instance, perform the following tasks:
1.
Back up the ESC database. For more information, see
Backup the Database from the ESC Standalone
.
Note To backup any custom scripts used by the deployments, and for restoring them before restoring the database, you must re-install the scripts on the backup as well.
2.
Redeploy the ESC instance. For more information, see the below section, Deploy the ESC for Upgrade .
142
Cisco Elastic Services Controller 5.4 Install and Upgrade Guide
Upgrading Cisco Elastic Services Controller
Deploy the ESC for Upgrade
3.
Restore the ESC database on the new ESC instance. For more information, see the below section, Restoring the ESC Database .
Deploy the ESC for Upgrade
After backing up and shutting down of the old ESC VM, a new/upgraded (based on new ESC package) ESC
VM should be installed. All parameters for ESC installation should be the same as the old ESC VM deployment.
• For OpenStack, you need to register the new ESC qcow2 image using the Glance command with a new image name and then use new bootvm.py script and new image name to install ESC VM.
Note In OpenStack, if an old ESC VM was assigned with floating IP, the new ESC
VM should be associated with the same floating IP after the installation.
• For VMWare, you need to use the new ESC OVA file to install ESC VM. All other configurations and property values should be the same as the old VM.
Restoring the ESC Database
Restore the ESC database on the new ESC instance , using the following procedure:
Procedure
Step 1
Step 2
Step 3
Step 4
Step 5
Step 6
Connect to the new ESC instance using SSH.
$ ssh USERNAME@NEW_ESC_IP
Stop the ESC service.
$ sudo escadm stop
Check ESC service status to make sure all the services are stopped.
$ sudo escadm status
Restore the database files.
$ scp://<username>:<password>@<standby_ip>:<filename>
$ sudo escadm restore --file /path/where/file/scp-ed/to/db.tar.bz2
Start the ESC service:
$ sudo escadm start
After ESC service is started, the standalone ESC upgrade is complete. You can check the health of the new
ESC service by running
$ sudo escadm status in the new ESC VM.
In Openstack, after restoring the database successfully, delete the old ESC instance:
$ nova delete OLD_ESC_ID
Cisco Elastic Services Controller 5.4 Install and Upgrade Guide
143
Upgrading Cisco Elastic Services Controller
Important Notes:
Important Notes:
After upgrading to the new ESC version, ESC service will keep doing life cycle management of all VNFs deployed by the old version. However, to apply any new features (with new data models) to the VNFs deployed by the ESC with old version is not guaranteed. If you want to apply any new feature of the new ESC version to existing VNFs, you have to undeploy and redeploy those VNFs.
Upgrading ESC HA Active/Standby Instances
To upgrade ESC HA Active/Standby nodes, perform the following tasks:
1.
Back up the database from an old ESC HA Active/Standby active instance. For more information, see
Backup the Database from the ESC HA Active/Standby Instances .
Note To backup any custom scripts used by the deployments, and for restoring them before restoring the database, you must copy the scripts on the backup as well.
2.
Deploy new ESC HA Active/Standby nodes based on new ESC version. For more information, see the below section, Deploy the ESC HA Active/Standby nodes for Upgrade .
3.
Restore the database on active ESC instance (Standby ESC instance will sync with the active ESC instance
). For more information, see the below section, Restoring the ESC Database on New Active and Standby
Instances .
Deploying the ESC HA Active/Standby nodes for Upgrade
After backing up and shutting down the two old ESC VMs, based on new ESC package install the new ESC
VMs.
• For OpenStack, you need to register the new ESC qcow2 image using the Glance command with a new image name and then to use new bootvm.py script and new image name to install ESC VM. All other bootvm.py arguments should be the same as used to setup an old VMs.
• For VMWare, there are two steps to bring up HA Active/Standby pair in VMware: 1) setup two standalone instances 2) reconfigure each instance with HA Active/Standby info. All other configurations and property values should be the same as the old VMs.
• If VIP is used for Northbound access, keep VIP the same for the new deployment as used to reconfigure the old HA Active/Standby pair.
Restoring the ESC Database on New Active and Standby ESC Instances
Procedure
Step 1
Step 2
Shut down the Standby ESC instance .
Connect to the standby ESC instance using SSH.
$ ssh USERNAME@ESC_STANDBY_IP
:
Verify that the ESC instance is standby and note the name of the standby ESC HA Active/Standby instance
144
Cisco Elastic Services Controller 5.4 Install and Upgrade Guide
Upgrading Cisco Elastic Services Controller
Restoring the ESC Database on New Active and Standby ESC Instances
Step 3
Step 4
Step 5
Step 6
Step 7
Step 8
Step 9
Step 10
Step 11
$ sudo escadm status
If the output value shows "STANDBY", the node is the standby ESC node.
Note If a dynamic mapping file (dynamic_mapping.xml) is used by ESC service, the dynamic mapping file should be restored into the standby ESC VM. Before power off the standby ESC node, you need to copy the backup dynamic mapping file (dynamic_mapping.xml) to the path
/opt/cisco/esc/esc-dynamic-mapping/ .
Shutdown the standby ESC instance through OpenStack Kilo/Horizon using Nova command. For ESC VM instances based in VMware vSphere, shutdown the active instance through VMware client dashboard. An example of shutting down the standby ESC instance in OpenStack is shown below:
$ nova stop NEW_ESC_STANDBY_ID
Restore the database on the new Active ESC instance.
Connect to the active ESC instance using SSH.
$ ssh USERNAME@ESC_ACTIVE_IP
Verify that the ESC instance is active.
$ sudo escadm status
If the output value shows 'ACTIVE' , the node is the active ESC node.
Stop the ESC services on the active node and verify the status to ensure the services are stopped.
$ sudo escadm stop
$ sudo escadm status
Restore the database files.
$ sudo escadm restore --file /tmp/db.tar.bz2
$ scp://<username>:<password>@<standby_ip>:<filename>
Note If a dynamic mapping file (dynamic_mapping.xml) is used by ESC service, the dynamic mapping file should be restored into the ESC VM. Before starting the ESC node, you need to copy the backup dynamic mapping file (dynamic_mapping.xml) to the path /opt/cisco/esc-dynamic-mapping/ .
Reboot the VM to restart the full ESC service:
$ sudo escadm restart
Use the
$ sudo escadm status to check the status of the ESC service.
Start the standby ESC node.
Power on the standby ESC node through OpenStack Nova/Horizon or VMware client. After starting the standby node, ESC HA Active/Standby upgrade process should be complete.
Delete the old HA Active/Standby instance through OpenStack Nova/Horizon or VMware client. An example of deleting the VM on OpenStack is shown below:
$ nova delete OLD_ESC_ACTIVE_RENAMED OLD_ESC_STANDBY_RENAMED
Cisco Elastic Services Controller 5.4 Install and Upgrade Guide
145
Upgrading Cisco Elastic Services Controller
Upgrading VNF Monitoring Rules
Upgrading VNF Monitoring Rules
To upgrade the VNF monitoring rules, you must back up the dynamic_mappings.xml
file and then restore the file in the upgraded ESC VM. For more information, see the backup and restore procedures. For upgrade of
HA Active/Standby instance, see
Upgrading ESC HA Active/Standby Instances
. For upgrade of the standalone instance, see
Upgrading Standalone ESC Instance
.
In-Service Upgrade of the ESC HA Active/Standby Nodes in
OpenStack
In-Service upgrade in OpenStack using ESC RPM packages
Procedure
Step 1
Step 2
Step 3
Step 4
Step 5
Backup ESC database and log files.
a) Perform ESC database backup from active node. For more information on backing up the database, see
Backup the Database from the ESC HA Active/Standby Instances .
b) Collect and backup all logs from both active and standby VMs. To backup the log, use the following command:
# sudo escadm log collect
Note A timestamped file will be generated in: /var/tmp/esc_log-<timestamp>.tar.bz2
c) Copy the database backup file and logs files (generated in /tmp/esc_log-.tar.bz2)* out of ESC VMs.
Log into the ESC HA Active/Standby standby VM and stop the escadm service.
$ sudo escadm stop
Ensure the ESC VM is in STOP state. ESC may take some time to switch to the STOP state. If ESC status turns into STOP state, please note that it won't be the part of HA Active/Standby cluster and you will lose
HA Active/Standby function temporarily.
$ sudo escadm status
Expected output:
ESC status=0 ESC HA is stopped
Copy the RPM file for upgrade to the ESC VM and execute the rpm command for upgrade.
$ sudo rpm -Uvh /home/admin/cisco-esc-3.1.0-145.x86_64.rpm
Start the escadm service.
$ sudo escadm start
146
Cisco Elastic Services Controller 5.4 Install and Upgrade Guide
Upgrading Cisco Elastic Services Controller
In-Service upgrade in OpenStack using ESC qcow2 Image
Step 6
Step 7
Log into the ESC HA Active/Standby active VM and repeat step 2 to step 6 in active VM. Please note that after stop escadm service in active ESC VM, a failover will be triggered and the upgraded standby VM will take over the active role.
Check the ESC version on each instance to verify the version is upgraded correctly and make sure ESC service is running properly in new active VM.
# esc_version
# health.sh (in Active VM)
In-Service upgrade in OpenStack using ESC qcow2 Image
Procedure
Step 1
Step 2
Backup ESC database and log files.
a) Perform ESC database backup from active node. For more information on backing up the database, see
Backup the Database from the ESC HA Active/Standby Instances .
b) Collect and backup all logs from both active and standby VMs. To backup the log, use the following command:
# sudo escadm log collect
Note A timestamped file will be generated in: /var/tmp/esc_log-<timestamp>.tar.bz2
c) Copy the database backup file and logs files (generated in /tmp/esc_log-.tar.bz2)* out of ESC VMs.
Redeploy standby instance with the new version of ESC image, and wait for the data to be synchronized.
a) Delete the standby instance through Horizon Horizon Web UI or nova CLI. In OpenStack controller, running following command through nova client.
nova delete <secondary_vm_name> b) Register new ESC image into OpenStack Glance for redeployment usage.
glance image-create --name <image_name> --disk-format qcow2 --container-format bare
--file <esc_qcow2_file> c) Redeploy the standby ESC VM instance based on newer image version. Re-install the new standby instance by using the new ESC package (bootvm.py and new registered image). All other installation parameters should be the same as the former ESC VM deployment. For example, hostname, ip address, gateway_ip, ha_node_list, kad_vip, kad_vif have to use the same values. Once the new ESC instance with upgraded version is up, it will be in standby state.
d) Log into the new instance and run the following command to check the synchronization state of the new
ESC node.
$ drbdadm status
Wait until the output of drbdadm status show both nodes are "UpToDate" like the output below. It means the new ESC instance has completed the data synchronization from the active instance.
Cisco Elastic Services Controller 5.4 Install and Upgrade Guide
147
Upgrading Cisco Elastic Services Controller
In-Service upgrade in OpenStack using ESC qcow2 Image
Step 3
Step 4
Example for Active/Standby esc role:Secondary disk:UpToDate
101.1.0.119:7789 role:Primary peer-disk:UpToDate
Example for Active ESC esc role:Primary disk:UpToDate
101.1.0.120:7789 role:Secondary peer-disk:UpToDatee
Stop keepalived service on standby instance, Power off the active instance, and then start standby keepalived service.
a) To stop keepalived service, use the following command: escadm keepalived stop b) Log into the active instance, set ESC active node into maintenance mode.
$ sudo escadm op_mode set --mode=maintenance
Make sure there is no in-flight transaction ongoing before moving to the next step. To verify there are no in-flight transactions, use the following command:
For ESC 2.3:
$ sudo escadm ip_trans
For versions older than ESC 2.3, check escmanager log at (/var/log/esc/escmanager.log) and make sure there are no new transaction in escmanager log.
c) Log in to the upgraded standby instance and shut down the ESC service.
$ sudo escadm stop d) Power off the active instance through OpenStack Nova client/Horizon and make sure it is off. In OpenStack
Controller, run:
$ nova stop <active_vm_name>
$ nova list | grep <active_vm_name> e) Log into the previously upgraded standby instance which is in stopped state and restart the ESC service.
The standby ESC instance will take the active role (switchover will be triggered) and start providing services with new version.
$ sudo escadm start
Check the ESC version on the new active instance to verify the version is upgraded correctly.
$ sudo escadm status (check ha status)
Expected output:
0 ESC status=0 ESC HA Active Healthy
$ esc_version (check esc version) version : 3.x.x
148
Cisco Elastic Services Controller 5.4 Install and Upgrade Guide
Upgrading Cisco Elastic Services Controller
In-Service upgrade in OpenStack using ESC qcow2 Image
Step 5
Step 6 release : xxx
Re-deploy the old active instance with the new ESC image.
Delete the old active instance and redeploy it by using the new ESC package (bootvm.py and new registered image).
a) Log in to the new deployed instance and check ha status. The new instance should be in standby state:
$ sudo escadm status --v b) Run the following command to check the synchronization state of the new ESC standby node:
$ drbdadm status
Wait until the output of drbdadm status shown as UpToDate.
c) For the new ESC standby node, make sure the health check is passed and the ESC version are upgraded correctly.
$ sudo escadm status (check ha status)
Expected output:
0 ESC status=0 ESC HA Standby Healthy
$ esc_version (check esc version)version : 2.x.x
release : xxx
$ health.sh
Expected output:
ESC HEALTH PASSED
Go back in to the first upgraded active instance and check the health and keepalived state.
$ drbdadm status
Expected output:
1:esc/0 Connected Primary/Secondary UpToDate/UpToDate /opt/cisco/esc/esc_database ext4
2.9G 52M 2.7G 2%
$ sudo escadm status (check ha status)
Expected output:
0 ESC status=0 ESC HA Active Healthy
$ esc_version (check esc version) Expected output: version : 2.x.x
release : xxx
$ health.sh (check esc health)
Expected output:
ESC HEALTH PASSED
Cisco Elastic Services Controller 5.4 Install and Upgrade Guide
149
Upgrading Cisco Elastic Services Controller
In-Service Upgrade of ESC HA Active/Standby Nodes in Kernel-Based Virtual Machine (KVM)
Note Quick rollback: In case of an upgrade failure, shutdown the upgraded instance and start the old active instance to have a quick rollback.
Rollback Procedure for In-service Upgrade a.
Copy the database and log backup files to a location out of ESC VMs.
b.
Delete any remaining ESC instance and redeploy ESC HA Active/Standby VMs using qcow2 image with old version.
c.
Restore the database. Follow the procedures in the section, Upgrading ESC HA Active/Standby
Instance with Backup and Restore for HA Active/Standby database restore.
d.
After database restore, you should have ESC service back with the old version.
In-Service Upgrade of ESC HA Active/Standby Nodes in
Kernel-Based Virtual Machine (KVM)
In-Service Upgrade in KVM using ESC RPM packages
Use this procedure to upgrade ESC high-availability nodes with a minimum service interruption on a
Kernel-based virtual machine.
Procedure
Step 1
Step 2
Step 3
Step 4
Step 5
Backup ESC database and log files.
a) Perform ESC database backup from active node. For more information on backing up the database, see
Backup the Database from the ESC HA Active/Standby Instances .
b) Collect and backup all logs from both active and standby VMs. To backup the log, use the following command:
$ sudo escadm log collect
Note A timestamped log file will be generated in: /var/tmp/esc_log-<timestamp>.tar.bz2
c) Copy the database backup file and logs files (generated in /tmp/esc_log-.tar.bz2)* out of ESC VMs.
Log into the ESC HA Active/Standby standby VM and stop the ESC service.
$ sudo escadm stop
Make sure the standby ESC VM is in STOP state.
$ sudo escadm status --v
If ESC status=0 esc ha is stopped.
In standby VM, execute the rpm command for upgrade:
$ sudo rpm -Uvh /home/admin/cisco-esc-<latest rpm filename>.rpm
Log into the active instance, set ESC active node into maintenance mode.
$ sudo escadm op_mode set --mode=maintenance
150
Cisco Elastic Services Controller 5.4 Install and Upgrade Guide
Upgrading Cisco Elastic Services Controller
In-Service Upgrade in KVM using ESC OVA Image
Step 6
Step 7
Step 8
Step 9
Step 10
Make sure there are no in-flight transactions and no new transactions during the upgrade. Use following commands to check in-flight transactions.
$ sudo escadm ip_trans
For any build older than ESC 2.3, you may need to check escmanager log for transactions at
(/var/log/esc/escmanager.log).
Power off ESC active node and make sure it is completely shut down. In KVM ESC controller, execute the following commands:
$ virsh destroy <active_vm_name>
$ virsh list --all
Log in the upgraded ESC instance (previous standby one), start the ESC service. The upgraded VM will take over active role and provide ESC service.
$ sudo escadm restart
$ sudo escadm monitor start
Check the ESC version on the new active instance to verify the upgraded version is correct. Once it is in the active state, make sure ESC service is running properly in the new active VM.
$ sudo escadm status
Expected output:
0 ESC status=0 ESC HA Active Healthy
$ esc_version
$ health.sh
Expected output:
ESC HEALTH PASSED
Power on the old active instance. In KVM ESC controller, execute the following commands:
$ virsh start <active_vm_name>
Log into the VM which is still with old ESC version and repeat step 2, 3, 4, and 7 in the VM.
In-Service Upgrade in KVM using ESC OVA Image
Procedure
Step 1 Backup ESC database and log files.
a) Perform ESC database backup from active node. For more information on backing up the database, see
Backup the Database from the ESC HA Active/Standby Instances .
b) Collect and backup all logs from both active and standby VMs. To backup the log, use the following command:
$ sudo escadm log collect
Note A timestamped log file will be generated in: /var/tmp/esc_log-<timestamp>.tar.bz2
Cisco Elastic Services Controller 5.4 Install and Upgrade Guide
151
Upgrading Cisco Elastic Services Controller
In-Service Upgrade in KVM using ESC OVA Image
Step 2
Step 3
Step 4 c) Copy the database backup file and logs files (generated in /tmp/esc_log-.tar.bz2)* out of ESC VMs.
Redeploy standby ESC instance. Register new ESC image on the standby instance.
a) Delete the standby instance through lib vert Virsh commands. On KVM host, run the following command:
$ Virsh destroy the <secondary_vm_name>
$ Virsh undefine --remove-all-storage <secondary_vm_name> b) Copy the new ESC image into Kvm Host for redeployment usage: sshpass -p "host Password' scp /scratch/BUILD-2_x_x_x/BUILD-2_x_x_x/ESC-2_x_x_x.qcow2
root@HOSTIP: c) Redeploy the standby ESC VM instance based on newer image version. Re-install new the standby instance by using the new ESC package (bootvm.py and new registered image). All other installation parameters should be the same as the former ESC VM deployment. For example, hostname, ip address, gateway_ip, ha_node_list, kad_vip, kad_vif have to use the same values. Once the new ESC instance with upgraded version is up, it will be in standby state.
d) Log into the new instance and run the following command to check the synchronization state of the new
ESC node.
$ drbdadm status wait until the output of drbdadm status show both nodes are "UpToDate" like the output below. It means the new ESC instance has completed the data synchronization from the active instance.
esc/0 Connected Secondary/Primary UpToDate/UpToDate
Stop keepalived service on standby instance, Power off active instance, and then start standby keepalived service.
a) Log into the active instance, set ESC active node into maintenance mode.
$ sudo escadm op_mode set --mode=maintenance
Make sure there is no in-flight transaction ongoing before moving to the next step. To verify there are no in-flight transactions, use the following command:
$ sudo escadm ip_trans
Check escmanager log at (/var/log/esc/escmanager.log) and make sure there are no new transaction in escmanager log.
b) Log in to the upgraded standby instance and shut down the keepalived service.
$ sudo escadm stop c) Power off the active instance and make sure it has been completely turned off. In KVM ESC Controller, run:
$ virsh destroy <active_vm_name>
$ virsh list --all d) Log into the previously upgraded standby instance which is in stopped state and start the ESC service.
The standby ESC instance will take the active role (switchover will be triggered) and start providing services with new version.
$ sudo escadm restart
Check the ESC version on the new active instance to verify the version is upgraded correctly.
152
Cisco Elastic Services Controller 5.4 Install and Upgrade Guide
Upgrading Cisco Elastic Services Controller
In-Service Upgrade in KVM using ESC OVA Image
Step 5
Step 6
$ sudo escadm status (check ha status)
Expected output:
0 ESC status=0 ESC HA Active Healthy
$ esc_version (check esc version) version : 4.1.x
release : xxx
$ health.sh (check esc health)
Expected output:
ESC HEALTH PASSED
Re-deploy the old active instance with the new ESC image.
Delete the old active instance and redeploy it by using the new ESC package (bootvm.py and new registered image). All other installation parameters should be the same as the old ESC VM deployment. For example, hostname, ip address, gateway_ip, ha_node_list, kad_vip, kad_vif have to be the same values.
a) Log in to the new deployed instance and check ha status. The new instance should be in standby state:
$ sudo escadm status b) Run the following command to check the synchronization state of the new ESC standby node:
$ drbdadm status
Wait until the output of drbdadm status shown as UpToDate.
c) For the new ESC standby node, make sure the health check is passed and the ESC version are upgraded correctly.
$ sudo escadm status (check ha status)
Expected output:
0 ESC status=0 ESC HA Active Healthy
$ esc_version (check esc version)version : 4.1.x
release : xxx
$ health.sh
Expected output:
ESC HEALTH PASSED
Go back in to the first upgraded active instance and check the health and keepalived state.
$ drbdadm status
Expected output:
1:esc/0 Connected Primary/Secondary UpToDate/UpToDate /opt/cisco/esc/esc_database ext4
2.9G 52M 2.7G 2%
$ sudo escadm status (check ha status)
Expected output:
0 ESC status=0 ESC HA Active Healthy
$ esc_version (check esc version) Expected output: version : 2.x.x
release : xxx
$ health.sh (check esc health)
Expected output:
ESC HEALTH PASSED
Cisco Elastic Services Controller 5.4 Install and Upgrade Guide
153
Upgrading Cisco Elastic Services Controller
In-Service Upgrade of the ESC HA Active/Standby Nodes in VMware
Note Quick rollback: In case of an upgrade failure, shutdown the upgraded instance and start the old active instance to have a quick rollback.
Rollback Procedure for In-service Upgrade a.
Delete any remaining ESC instance and redeploy ESC HA Active/Standby VMs using qcow2 image with old version.
b.
Restore the database. Follow the procedures in the section, Upgrading ESC HA Active/Standby
Instance with Backup and Restore for HA Active/Standby database restore.
c.
After database restore, you should have ESC service back with the old version.
In-Service Upgrade of the ESC HA Active/Standby Nodes in
VMware
In-Service upgrade in VMware using ESC RPM packages
Use this procedure to upgrade the ESC high-availability nodes one node at a time with a minimum service interruption. This process leverages the ESC HA Active/Standby replication and failover capability to smoothly move ESC service to the new upgraded node without the manual database restore.
Procedure
Step 1
Step 2
Step 3
Step 4
Step 5
Backup ESC database and log files.
a) Perform ESC database backup from active node. For more information on backing up the database, see
Backup the Database from the ESC HA Active/Standby Instances .
b) Collect and backup all logs from both active and standby VMs. To backup the log, use the following command:
# sudo escadm log collect c) Copy the database backup file and logs files (generated in /tmp/esc_log-.tar.bz2)* out of ESC VMs.
Log into the ESC HA Active/Standby standby VM and stop the keepalived service.
$ sudo escadm stop
Make sure the standby ESC VM is in STOP state.
$ sudo escadm status --v
If ESC status=0 esc ha is stopped.
In the standby VM, execute the rpm command for upgrade:
$ sudo rpm -Uvh /home/admin/cisco-esc-2.2.9-50.rpm
Log into the active instance, set ESC active node into maintenance mode.
$ sudo escadm op_mode set --mode=maintenance
154
Cisco Elastic Services Controller 5.4 Install and Upgrade Guide
Upgrading Cisco Elastic Services Controller
In-Service upgrade in VMware using ESC qcow2 Image
Step 6
Step 7
Step 8
Step 9
Step 10
Make sure there are no in-flight transactions and no new transactions during the upgrade. From ESC 2.3, you may use following commands to check in-flight transactions.
$ sudo escadm ip_trans
For build older than ESC 2.3, you may need to check escmanager log and make sure no new transactions are recorded in this log file. The log file can be located at (/var/log/esc/escmanager.log).
Power off ESC active node. In VMware vSphare Client, select Home > Inventory > VMs and Templates , right click the active instance name from the left panel, and select Power > Power Off .
Log in to the upgraded ESC instance (previous standby one), and start the keepalived service. The upgraded
VM will take over active role and provide ESC service.
$ sudo escamd restart
Check the ESC version on the new active instance to verify the upgraded version is correct. Once it is in the active state, make sure ESC service is running properly in the new active VM.
$ sudo escadm status
Expected output:
0 ESC status=0 ESC HA Active Healthy
$ esc_version
$ health.sh
Expected output:
ESC HEALTH PASSED
Power on the old active instance. In VMware vSphare Client, select Home > Inventory > VMs and Templates , right click the active instance name from the left panel, then select Power > Power On .
Log into the VM which is still with old ESC version and repeat step 2, 3, 4, and 7 in the VM.
In-Service upgrade in VMware using ESC qcow2 Image
Procedure
Step 1
Step 2
Backup ESC database and log files.
a) Perform ESC database backup from the active node. For more information on backing up the database, see
Backup the Database from the ESC HA Active/Standby Instances
.
b) Collect and backup all logs from both active and standby VMs. To backup the log, use the following command:
# sudo escadm log collect
Note A timestamped log file will be generated in: /var/tmp/esc_log-<timestamp>.tar.bz2
c) Copy the database backup file and logs files (generated in /tmp/esc_log-.tar.bz2)* out of ESC VMs.
Redeploy standby ESC instance. Register new ESC image on the standby instance, and wait for the data to be synchronized.
Cisco Elastic Services Controller 5.4 Install and Upgrade Guide
155
Upgrading Cisco Elastic Services Controller
In-Service upgrade in VMware using ESC qcow2 Image
Step 3
Step 4 a) Delete the standby instance. To delete the standby ESC instance, you need to first "Power Off" the instance through vSphere Client and then use the Delete from Disk option. In VMware vSphare Client, select
Home > Inventory > VMs and Templates , right click the instance name from the left panel, then select
Power > Power Off . Now to delete the standby instance, select Home > Inventory > VMs and Templates , right click the instance name from the left panel, then select Delete from Disk .
b) Redeploy the standby ESC VM instance based on newer image version. Re-install new the standby instance by using the new ESC package (bootvm.py and new registered image). Once the new ESC instance with upgraded version is up, it will be in standby state.
c) Log into the new instance and run the following command to check the synchronization state of the new
ESC node.
$ drbdadm status
Wait until the output of drbdadm status show both nodes are "UpToDate" like the output below. It means the new ESC instance has completed the data synchronization from the active instance.
esc/0 Connected Secondary/Primary UpToDate/UpToDate
Stop keepalived service on standby instance, Power off active instance, and then start standby keepalived service.
a) Log into the active instance, set ESC active node into maintenance mode.
$ sudo escadm op_mode set --mode=maintenance
Make sure there is no in-flight transaction ongoing before moving to the next step. To verify there are no in-flight transactions, use the following command:
For ESC 2.3:
$ sudo escadm ip_trans
For versions older than ESC 2.3, check escmanager log at (/var/log/esc/escmanager.log) and make sure there are no new transaction in escmanager log.
b) Log in to the upgraded standby instance and shut down the keepalived service.
$ sudo escadm stop c) Power off the active instance and make sure the active instance has been powered off. In VMware vSphare
Client, select Home > Inventory > VMs and Templates , right click the instance name from the left panel, then select Power > Power Off .
d) Log into the previously upgraded standby instance which is in stopped state and start the keepalived service. The standby ESC instance will take the active role (switchover will be triggered) and start providing services with new version.
$ sudo escadm start
Check the ESC version on the new active instance to verify the version is upgraded correctly.
$ sudo escadm status --v(check ha status)
Expected output:
0 ESC status=0 ESC HA Active Healthy vimmanager (pgid 41908) is running monitor (pgid 42067) is running mona (pgid 42130) is running drbd (pgid 38488) is active
156
Cisco Elastic Services Controller 5.4 Install and Upgrade Guide
Upgrading Cisco Elastic Services Controller
In-Service upgrade in VMware using ESC qcow2 Image
Step 5
Step 6 snmp (pgid 2121) is running etsi (pgid 43247) is running pgsql (pgid 42566) is running keepalived (pgid 40281) is running portal (pgid 43307) is running confd (pgid 25644) is running filesystem (pgid 0) is running escmanager (pgid 42775) is running
$ esc_version (check esc version) version : 3.x.x
release : xxx
$ health.sh (check esc health)
Expected output:
ESC HEALTH PASSED
Re-deploy the old active instance with the new ESC image.
Delete the old active instance and redeploy it by using the new ESC package (bootvm.py and new registered image). All other installation parameters should be the same as the old ESC VM deployment. For example, hostname, ip address, gateway_ip, ha_node_list, kad_vip, kad_vif have to be the same values. To delete, in the VMware vSphare Client, access, Home > Inventory > VMs and Templates , right click the instance name from the left panel, then select Delete from Disk .
a) Log in to the new deployed instance and check ha status. The new instance should be in standby state:
$ sudo escadm status b) Run the following command to check the synchronization state of the new ESC standby node:
$ drbdadm status
Wait until the output of drbdadm status shown as UpToDate.
c) For the new ESC standby node, make sure the health check is passed and the ESC version are upgraded correctly.
$ sudo escadm status (check ha status)
Expected output:
0 ESC status=0 ESC HA Active Healthy
$ esc_version (check esc version)version : 3.x.x
release : xxx
$ health.sh
Expected output:
ESC HEALTH PASSED
Go back in to the first upgraded active instance and check the health and keepalived state.
$ drbdadm status
Expected output:
1:esc/0 Connected Primary/Secondary UpToDate/UpToDate /opt/cisco/esc/esc_database ext4
2.9G 52M 2.7G 2%
$ sudo escadm status (check ha status)
Expected output:
0 ESC status=0 ESC HA Active Healthy
$ esc_version (check esc version) Expected output:
Cisco Elastic Services Controller 5.4 Install and Upgrade Guide
157
Upgrading Cisco Elastic Services Controller
In-Service Upgrade of the ESC HA Active/Standby Nodes in CSP version : 3.x.x
release : xxx
$ health.sh (check esc health)
Expected output:
ESC HEALTH PASSED
Note Quick rollback: In case of an upgrade failure, shutdown the upgraded instance and start the old active instance to have a quick rollback.
Rollback Procedure for In-service Upgrade a.
Copy the database and log backup files to a location out of ESC VMs.
b.
Delete any remaining ESC instance and redeploy ESC HA Active/Standby VMs using qcow2 image with old version.
c.
Restore the database. Follow the procedures in the section, Upgrading ESC HA Active/Standby
Instance with Backup and Restore for HA Active/Standby database restore.
d.
After database restore, you should have ESC service back with the old version.
In-Service Upgrade of the ESC HA Active/Standby Nodes in CSP
Follow the steps to do the In Service upgrade of the ESC HA Active/Standby Nodes in CSP:
Before you begin
Verify that the ESC HA nodes running properly before the upgrade, by using the following command:
# escadm status
One node must be in active state and the other node in standby state. Verify the Active node by using the following command:
# health.sh
On a successful health check you receive the following:
ESC HEALTH PASSED
Procedure
Step 1
Step 2
Shutdown the Standby instance
Upgrade the standby ESC VM, before power off the VM. To upgrade, follow the steps: a.
Collect the logs from the standby ESC VM, and copy it to another machine, by using the following commands:
# collect_esc_log.sh
# scp /tmp/LOG_PACKAGE_NAME <username>@<backup_vm_ip>:<filepath> b.
Using CSP, power off the standby ESC.
Deploy the standby node again using the new ESC package
158
Cisco Elastic Services Controller 5.4 Install and Upgrade Guide
Upgrading Cisco Elastic Services Controller
In-Service Upgrade of the ESC HA Active/Standby Nodes in CSP
Step 3
Step 4
After power off the standby ESC VM, install new ESC VM for upgrading by using the new ESC package.
Except using different ESC packages from the former ESC VM, all other parameters for ESC installation must be the same as the previous ESC VM deployment.
Verify that the ESC node is in standby state. Use the following command to check the synchronization state of the new ESC standby node:
# drbdadm status
Wait until you get the following output, which shows that the new ESC VM has completed the data synchronization from the Active node and it is up to date.
[admin@esc-xyx-upgradetestha1-4-5-0-105 ~]$ drbdadm status esc role:Secondary disk:UpToDate
172.20.117.55:7789 role:Primary peer-disk:UpToDate
If a dynamic mapping file (dynamic_mapping.xm) is used by ESC service, it must be restored into the ESC
VM. Copy the Backup file to
/opt/cisco/esc-dynamic-mapping/ path.
Stop the Active node and trigger a switchover.
Power off the Active instance through CSP. Follows that, an HA switchover automatically triggers and the
Standby instance takes over the ESC service with the new ESC version. After new ESC instance becomes the
Active, verify that the new ESC Active node passes the health check.
Deploy new ESC node to replace old active
Install the new ESC VM for upgrading the previous active instance by using the new ESC package. Before powering off the ESC VM with the old version, collect the logs from the ESC VM, and copy it to another machine by using the following commands:
# collect_esc_log.sh
# scp /tmp/LOG_PACKAGE_NAME <username>@<backup_vm_ip>:<filepath>
Note If a dynamic mapping file is used by ESC service, the dynamic mapping file should be backed up at the same time with ESC logs. The default path of the dynamic mapping file is
/opt/cisco/esc/esc-dynamic-mapping/dynamic_mappings.xml.
Power off the ESC VM with old version through CSP.
Verify that the ESC node is in standby state. Use the following command to check the synchronization state of the new ESC standby node:
# drbdadm status
Wait until you get the following output, which shows that the new ESC VM has completed the data synchronization from the Active node and it is up to date.
[admin@esc-xyz-upgradetestha1-4-5-0-105 ~]$ drbdadm status esc role:Secondary disk:UpToDate
172.20.117.55:7789 role:Primary peer-disk:UpToDate
If a dynamic mapping file (dynamic_mapping.xm) is used by ESC service, it must be restored into the ESC
VM. Copy the Backup file to
/opt/cisco/esc-dynamic-mapping/ path.
Cisco Elastic Services Controller 5.4 Install and Upgrade Guide
159
Upgrading Cisco Elastic Services Controller
In-Service Upgrade of the ESC HA Active/Standby Nodes in CSP
Note If a dynamic mapping file is used by ESC service, the dynamic mapping file should be restored into the ESC VM. Copy the backup dynamic mapping file to the default path of the dynamic mapping file
/opt/cisco/esc/esc-dynamic-mapping/dynamic_mappings.xml.
Once you sucessfully complete the in-service upgrade, you can remove the old ESC instance.
Note After upgrading to the new ESC version, ESC service continues life cycle management of all VNFs deployed by the old version. If you want to apply any new feature of the new ESC version to existing
VNFs, undeploy those VNFs and do a new deployment.
160
Cisco Elastic Services Controller 5.4 Install and Upgrade Guide
P A R T
IX
Troubleshooting Cisco Elastic Services
Controller Installation
•
İndir
Herkese açık bağlantı güncellendi
Sohbetinize giden herkese açık bağlantı güncellendi. You may view and manage your chat links after registration here.
Belge herkese açık bağlantısı
Reklam