Dell | PowerVault MD1000 SAS Solution | User`s guide | Dell PowerVault MD1000 SAS Solution User`s guide

Dell PowerVault MD1000 SAS Solution User`s guide
A Reference Guide for Optimizing the
Dell™ PowerVault™ MD1000 SAS Solution
Authored By:
SAS Solution Team
April, 2006 rev A01
THIS WHITE PAPER IS FOR INFORMATION PURPOSES ONLY, AND MAY CONTAIN TYPOGRAPHICAL ERRORS AND TECHNICAL
INACCURACIES. THE CONTENT IS PROVIDED AS IS WITHOUT EXPRESS OR IMPLIED WARRANTIES OF ANY KIND.
____________________
Information in this document is subject to change without notice.
© Copyright 2006 Dell Inc. All rights reserved.
Reproduction in any manner whatsoever without the written permission of Dell Inc. is strictly forbidden.
Dell, the Dell Logo, PowerVault, PowerEdge and OpenManage are trademarks of Dell Inc. Microsoft is a
registered trademark of Microsoft Corporation. Other trademarks and trade names may be used in this
document to refer to either the entities claiming the marks and names or their products. Dell disclaims
proprietary interest in the marks and names of others.
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
TABLE OF CONTENTS
1.
INTRODUCTION ...................................................................................................................................................................4
Serial Attached SCSI (SAS)....................................................................................................................................................5
2. RECOMMENDED MAINTENANCE FOR YOUR DELL POWERVAULT MD1000 SAS SOLUTION .........................................6
3. DELL’S SAS PRODUCT OFFERING .....................................................................................................................................7
SAS Cabling............................................................................................................................................................................8
Physical Drive Support ..........................................................................................................................................................8
3rd Party Hardware support...................................................................................................................................................9
Server Support ........................................................................................................................................................................9
4. POWERVAULT MD1000 STORAGE ENCLOSURE DETAILS ................................................................................................9
Enclosure Modes ....................................................................................................................................................................9
Unified Topology....................................................................................................................................................................9
Split Topology.......................................................................................................................................................................10
Daisy-Chaining Multiple PowerVault MD1000 Storage Enclosures................................................................................10
PowerVault MD1000 – Common Features and Capabilities............................................................................................11
5. PERC 5/E CONTROLLER - DETAILED FEATURES ............................................................................................................15
Considerations when implementing a RAID Solution ........................................................................................................15
Controller Features ................................................................................................................................................................................15
Capacity Requirements...........................................................................................................................................................................17
6.
DELL OPENMANAGE SERVER ADMINISTRATOR .............................................................................................................19
Managing the PERC 5/E Controller and the PowerVault MD1000 Storage Enclosure..................................................19
Using the Create Virtual Disk Wizards ...............................................................................................................................19
Reconfiguring Virtual Disks -- RAID Migration and Capacity Expansion.......................................................................20
Moving Physical Disks and Virtual Disk Configurations – Disk Roaming and Disk Migration .....................................21
Importing Migrated Virtual Disks .......................................................................................................................................22
Upgrading from Dell OpenManage Array Manager .........................................................................................................24
7. STORAGE APPLICATIONS AND COMPONENTS..................................................................................................................25
Identify customer usage model and needs...........................................................................................................................25
Picking a solution based on the application profile............................................................................................................26
Application specific guidelines ............................................................................................................................................27
E-Mail Servers ........................................................................................................................................................................................27
Web Servers.............................................................................................................................................................................................28
Database or Online Transaction Processing Servers...........................................................................................................................29
File Servers..............................................................................................................................................................................................30
Streaming Media Servers........................................................................................................................................................................32
Hot Spares.............................................................................................................................................................................33
8. SUMMARY AND CONCLUSIONS ........................................................................................................................................33
9. APPENDIX – A: REFERENCES............................................................................................................................................35
10.
APPENDIX – B: GLOSSARY...........................................................................................................................................35
11.
APPENDIX – D: REVISION HISTORY .............................................................................................................................35
PAGE 3
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
1. Introduction
A complete enterprise storage solution involves various software and hardware components. A
given storage project can be addressed using several approaches. Therefore, it is essential to
identify optimum storage solutions for different applications.
Selecting the right storage components not only impacts the effectiveness of a given application
but may also improve the performance of the entire IT infrastructure which relies on that storage.
See table 1 for examples of key storage components.
Table 1: Examples of the components in a typical storage solution
Storage Solution Components
Examples
Physical Storage Device Technology
Storage Expansion Enclosure (e.g., PowerVault
MD1000)
Interconnect Technology
3 Gbps Serial Attached SCSI (SAS)
Host System Interface Technology
Non-RAID (HBA) or RAID controllers (e.g. PowerEdge
Expandable Raid Controller)
Management Capabilities
Dell OpenManage Server Administrator
As with many IT solutions, storage deployments can be assessed using numerous criteria, each
with varying levels of importance. In general, the effectiveness of a given storage solution should
be measured using the following attributes:
•
•
•
•
•
•
•
•
Reliability – The solution remains functional and ensures that data is unaltered
Availability – Data remains accessible by the application.
Serviceability – The hardware has easy access to assemblies/components, allowing easy
maintenance and replacement.
Redundancy – The hardware provides redundancy for both the data storage, as well as
for the hardware management capabilities.
Performance
o Bandwidth Utilization – Data throughput / available Bandwidth
o IOPS – I/Os per second
Data Protection – Data is protected against catastrophic system failure or components
failure.
Scalability – Storage capacity can be expanded as the application necessitates.
Cost – Measured in total lifetime cost, acquisition cost, or $ per Gigabyte.
An optimal storage solution addresses all key attributes to a greater or lesser extent depending on
their relative importance. This paper presents a study of the technological, functional and
behavioral differences of the storage solution components and measures them against the
attributes listed above. Understanding each component with respect to these attributes will help in
making informed decisions about the optimal solution for a given storage application.
PAGE 4
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
The paper presumes that the reader has a basic understanding of different RAID levels and
interconnect technologies – such as Serial Attached SCSI (SAS) and PCI Express. In this paper,
we will focus on Serial Attached SCSI storage technology, specifically for Dell’s PowerVault
MD1000 storage expansion enclosure and Dell’s PERC 5/E RAID controller. Other storage
technologies such as SCSI or Fibre Channel, as well as topologies such as Network Attached
Storage (NAS) or Storage Area Network (SAN) are outside the scope of this document. All storage
applications will be treated independent of any specific host system or server.
Serial Attached SCSI (SAS)
The SCSI (Small Computing System Interface) protocol originated more than 20 years ago and is
used in the majority of server storage needs (e.g. storage that is internal to a server or physically
connected to a server in an external disk expansion enclosure). The market prevalence of SCSI
has created economies of scale making it one of the most affordable storage interconnect
technologies available today. Many businesses rely on SCSI physical disks to help deliver
extremely cost-effective and reliable storage. In fact, SCSI physical disks have advanced over the
years offering increased performance and more sophisticated features with each iteration.
However, given the inherent limitations of its parallel architecture, the current version of SCSI
physical disks, U320 or Ultra320, will be the last version of the traditional SCSI technology.
Serial Attached SCSI (SAS) is a major iteration of SCSI technology and introduces a host of
benefits when compared to older SCSI technology including,
Dramatic improvement in bandwidth performance
• The first generation of SAS will support 3.0 Gbps (or 300 MB/sec) per “lane”
• Ultra320 SCSI is the shared-bus architecture. SAS introduces a point to point topology
enabling systems with many physical disks to scale bandwidth far beyond the bandwidth
capacity of Ultra320 SCSI.
Figure 1: Parallel U320 SCSI bus versus point to point Serial Attached SCSI connection
Up to320 MBps
Up to 3.0 Gbps per link
•
PAGE 5
SAS introduces the concept of port aggregation to storage interconnects. External storage
devices can connect via a x4 (“by four”)wide-lane. Cabling aggregates four SAS lanes, each
operating at 3.0 Gbps, for a full external connection bandwidth of up to 12.0 Gbps on a single
connector.
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
Figure 2: Bandwidth improvement with SAS port aggregation
One parallel SCSI cable
Up to 320MB/s
performance
One x4 SAS cable
Up to 12.0 Gbps performance
Improved capacity scalability
• SAS allows for port expansion, enabling the ability to “daisy-chain” storage enclosures for
dramatic improvements in capacity scalability. With Ultra320 SCSI, systems were limited to a
maximum of 16 devices on a single SCSI channel. This typically translated to maximum of 14
physical disks connected to a single SCSI channel on a RAID controller. With SAS expansion,
a single SAS port can be expanded out to multiple SAS devices (similar to how Ethernet hubs
can expand a single Ethernet connection into multiple devices).
Other important changes to keep in mind about SAS:
• SAS continues to use the SCSI command sets for driver & software compatibility.
• SAS is not backwards compatible with Parallel SCSI.
2. Recommended Maintenance for your Dell PowerVault MD1000 SAS Solution
Dell strives to provide the best possible customer experience for all of our enterprise server and
storage products. Proper maintenance of your solution is essential to help ensure that the
equipment remains in optimal working condition. Dell recommends the following suggested
maintenance activities.
Keep firmware/drivers/utilities up to date
In general, Dell posts planned maintenance updates for firmware twice a year. To stay current
with the latest updates, customers can subscribe to Dell notification tools. Click the following link
to sign up for technical updates.
PAGE 6
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
http://support.dell.com/support/topics/global.aspx/support/notifications/en/index?c=us&l=en&s
=gen
Enable Background Read Patrol
Background Read Patrol, a function of the RAID controller, is an automatically scheduled
background task that scans and helps resolve possible physical disk errors that may lead to
physical disk failure and result in data integrity issues. Dell recommends implementing Background
Read Patrol in all configurations.
See section 5 for details.
Schedule and run periodic consistency checks / Background Initialization (BGI)
Consistency checks / BGI are performed by many RAID controllers to verify that all stripes in a
redundant virtual disk are consistent and to automatically correct stripes where inconsistencies are
found. BGI is a consistency check that automatically runs on all redundant arrays. Depending on
workload, Dell recommends that consistency checks should be run manually on the virtual disks at
least once every month to help ensure data integrity. Consistency checks and Background Read
Patrol are complementary and both should be used.
See section 5 for details.
Enable Hot Spares
Hot spare functionality provides extra security, availability and redundancy by automatically
replacing a failed disk in a RAID group and allowing the rebuild of the degraded array to begin
immediately. Dell PERC controllers support Hot Spare functionality and it is recommended
whenever possible.
See section 5 for details.
Document your configuration and keep the documentation current
Documenting your configuration facilitates faster and easier recovery should a failure occur. Make
sure to include RAID levels, physical disks included in your virtual disks, firmware and driver
versions and usage of hot spares in your documentation.
Perform periodic physical inspection of your connections and cabling
Damaged cables and loose connections may cause functional issues with your storage solution.
Dell recommends a physical equipment audit two times a year or anytime after re-cabling or redeploying your storage hardware.
3. Dell’s SAS Product Offering
The PowerVault MD1000 is a feature-optimized storage expansion enclosure which utilizes the
latest industry standard storage interconnect technology, Serial Attached SCSI (SAS). This 3U
enclosure supports between 2 and 15 physical SAS drives. Along with Dell’s host-based RAID
controller, PERC 5/E, the PowerVault MD1000 storage enclosure introduces new levels of
performance, capacity, and scalability. Major features of the PowerVault MD1000 and PERC 5/E
solution include:
PAGE 7
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
Performance
• Support for 3.0 Gbps SAS physical disks (both 10,000 and 15,000 RPM)
• External cable supports a SAS x4 wide-lane to achieve a total interface bandwidth of up to
12.0 Gbps connection to the host controller
• Host interconnect with the PERC 5/E host-based controller
• Support for RAID level 0
Availability
• Support for RAID levels 1, 5, 10 and 50
• Redundant, hot pluggable power supplies with integrated cooling fan modules
• Redundant enclosure management modules (EMMs)
• Hot pluggable physical disks
Scalability/Expandability
• Support for daisy-chaining up to three PowerVault MD1000 expansion enclosures per PERC
5/E connection
• Supports between 2 and 15 physical SAS drives
• Supports 36GB, 73GB, 146GB, and 300 GB 3.5-inch SAS physical disks
Ease-of-Use
• Common management software with PowerEdge™ servers through Dell OpenManage Server
Administrator
PERC 5/E is the RAID controller used to connect the PowerVault MD1000 storage enclosure to
one or two PowerEdge servers. It supports 3 Gbps Serial Attached SCSI (SAS) as the storage
interconnect technology and PCI Express (PCI-E) as the host-based interconnect technology. The
“E” in PERC 5/E stands for “External” because the PERC 5/E controller supports only physical
disks that attach to a server externally.
The PERC 5/E controller offers:
• Intel IOP333 I/O processor based on Intel XScale Microarchitecture
• 256MB of customized DDR2 400MHz, Error-Correcting Code (ECC) cache memory
• 3 Gbps maximum speed for each SAS lane
• Two external x4 (“by four”) SAS ports, each aggregating 4 SAS lanes for a total bandwidth per
port of 12.0 Gbps
• x8 PCI E host interface for a total bandwidth of 32.0 Gbps
• Up to 72 hours of intelligent, transportable, battery-backed, cache memory
For detailed information on the PERC 5/E controller, see section 6.
SAS Cabling
The PowerVault MD1000 storage enclosure and the PERC 5/E controller both support industry
standard 3.0 Gbps SAS cables in 1m, 2m, and 4m cable lengths. Each cable supports x4 SAS
connectors to deliver up to 12.0 Gbps of total bandwidth from the host to the enclosure. Only Dellqualified cables are supported for connecting the PowerVault MD1000 storage enclosure to the
PERC 5/E controller.
Physical Drive Support
The PowerVault MD1000 storage expansion enclosure supports the latest SAS physical disks,
currently specified as 3.0 Gbps SAS. SAS solutions from Dell include both 10,000 RPM and
15,000 RPM spindle speeds and a variety of capacities to allow customers the maximum flexibility.
PAGE 8
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
For the latest information on support for specific physical disks, see the Dell Storage web site at:
http://www.us.dell.com/storage/
3rd Party Hardware support
To help provide the best possible customer experience, Dell supports only hardware purchased
direct from Dell for the PowerVault MD1000 and PERC 5/E controller. Dell-developed server and
storage solutions include many thousands of person-hours of engineering testing and validation to
help ensure the highest product quality. For example, physical disks purchased direct from Dell
may include features not generally available to the rest of the industry. As such, Dell supports only
Dell tested and qualified hardware. Hardware purchased direct from Dell carries the full Dell
1
rd
system limited warranty and helps ensure full compatibility. Materials purchased from 3 party
vendors for use in Dell systems are not covered under Dell’s limited warranty and may not deliver
an acceptable customer experience.
Server Support
The PowerVault MD1000 storage enclosure and PERC 5/E solution are supported on the
PowerEdge 1850, 2850, 6800, and 6850 servers.
4. PowerVault MD1000 Storage Enclosure Details
Enclosure Modes
The enclosure can be set to either unified or split mode through the use of a two-position switch
located on the front panel of the enclosure. The enclosure can be configured into the selected
mode only at power up. Any changes to the mode switch made while the enclosure is powered on
will not take effect until the enclosure is powered off and back on again.
Note: Any time you power on the PowerVault MD1000 storage enclosure, ensure the mode switch
is set in the desired position. If the position of the switch changes inadvertently, data loss could
occur as the controller may recognize an inaccurate number of attached physical disks versus the
previous mode. In order to cycle power on your MD1000, all attached servers, must be turned off.
Unified Topology
In unified mode, a single server has access to all fifteen physical disks in the PowerVault MD1000
storage enclosure through a single SAS cable. The unified mode can also provide a single server
access to a maximum of forty five physical disks by daisy-chaining up to three enclosures from a
single host port. The main advantage of unified mode is that it provides a high degree of physical
disk scalability.
1
PAGE 9
For a copy of our guarantees or limited warranties, please write Dell USA L.P., Attn: Warranties, One Dell Way, Round
Rock, TX 78682. For more information, visit www.dell.com/warranty.
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
Split Topology
In split mode, the enclosure is divided into two logical enclosures; with each host connection
accessing its own set of physical disks. The primary Enclosure Management Module (EMM)
(located on the left side of the enclosure when viewed from the rear), accesses physical disk slots
7 through 14. The secondary EMM accesses physical disk slots 0 through 6. To connect a server
to an enclosure in split mode, the In port of each EMM should be attached to a connector of a
PERC 5/E controller. The Out port of an EMM in split mode is disabled, since enclosure daisychaining is not supported in this mode.
Split mode is useful in deployments where a single PowerVault MD1000 storage enclosure help
provide limited amounts of storage each for two different servers.
Note: In split mode, each host server can see its own physical disks (e.g. drive., physical disk 0 – 6
or 7 – 14) but not the other server’s physical disks.
Note: Split mode does not support daisy-chaining.
Note: A system with a single EMM set in split mode will function properly, but the server will only
be able to see the physical disks that are connected to the EMM.
Dual Hosts
Figure 3: Split mode configuration
Daisy-Chaining Multiple PowerVault MD1000 Storage Enclosures
In unified mode, up to three enclosures may be daisy-chained together from a single PERC 5/E
controller port. A maximum of six PowerVault MD1000 storage enclosures (3 enclosures per port)
may be connected to a single PERC 5/E controller. (Note: Although this is a supported
configuration, one should consider total solution availability before attaching this quantity of storage
to a single-controller solution.). When connecting multiple enclosures, the first enclosure is
connected to the PERC 5/E controller via the In port of the primary EMM. The second enclosure is
connected from the Out port of the primary EMM on the first enclosure to the In port of the primary
EMM on the second enclosure. The connection is made with a standard SAS cable. The third
enclosure is connected in the same manner as the second enclosure.
PAGE 10
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
Figure 4: Max configuration on a single PERC 5/E controller port
Figure 5: Max configuration on a single PERC 5/E controller (both ports connected)
PowerVault MD1000 – Common Features and Capabilities
Regardless of the topology, the PowerVault MD1000 storage enclosure provides a set of common
features and capabilities.
World Wide Name (WWN) Assignment
The SAS standard requires that all devices on a SAS domain have a unique address, called a
World Wide Name or WWN. The PowerVault MD1000 storage enclosure complies with the SAS
standard. For more information on WWN assignment, please refer to the SAS Specification
referenced in Appendix A.
PAGE 11
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
Mixing Physical Disk Capacities
By supporting the mixing of physical disk of different sizes within an enclosure, the PowerVault
MD1000 storage enclosure enables optimal matching of physical disk type to the requirements of
multiple applications. Mixing of physical disk sizes is supported but not recommended within a
single RAID group. The PowerVault MD1000 storage enclosure supports 36 GB, 73 GB, 146 GB,
and 300 GB capacities for SAS. When operating in a RAID configuration, the lowest capacity
physical disk defines the RAID virtual disk size. For example, a RAID 1 configuration with two
different sized physical disks, one 36 GB and another 146 GB physical disk, will be limited to 36
GB. If the same two physical disks were configured in RAID 0 configuration, the virtual disk size
would be limited to 73 GB (the equivalent of two 36 GB physical disks in RAID 0 configuration).
Mixing Physical Disk Spindle Speeds
Mixing different physical disk spindle speeds is supported, but not recommended within a single
RAID group. The PowerVault MD1000 storage enclosure supports 10,000 and 15,000 RPM
spindle speeds. Depending upon the RAID configuration, the lowest spindle speed may dictate the
maximum performance of the RAID group.
SCSI Enclosure Services (SES)
The EMM continuously monitors environmental variables such as temperature and voltage. The
PowerVault MD1000 storage enclosure works in conjunction with Dell OpenManage Server
Administrator to set trigger thresholds that warn the user when these variables exceed safe
operating levels. The enclosure will automatically power down when these environmental variables
exceed a critical threshold level in order to avoid potential equipment damage. Both local visual
indicators (e.g. LEDs) and audio alerts are triggered when faults are detected. The audio alerts on
the enclosure are disabled by default, but can be enabled using Server Administrator. The
following SES functions are provided by the EMMs in the PowerVault MD1000 storage enclosure:
•
•
•
•
•
•
•
Enclosure shelf faults detection
Physical disk Power/Fault/Offline status
Power supply status and shutdown capabilities
Fan speed status and control
Temperature monitoring
Over/under temperature shutdown capabilities
Alarms (local audible and visual LED indicators)
Hot Plugging
•
•
The PowerVault MD1000 storage enclosure is designed for hot plugging certain
components to help prevent damage when pulling or replacing. Hot plugging is a concept
which allows for the safe removal and replacement of components while the power is still
on. Hot plugging allows server applications and disk activities to continue uninterrupted,
while maintenance or repair actions take place. Additional steps may be required for
physical disks to become usable. See the RAID controller User’s Guide for more
information.
Hot Pluggable Devices
o Physical disk (all Disk I/O activities must be halted first)
o Power supply module/ cooling fan module
Note: Hot plugging of SAS cables is not supported.
PAGE 12
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
External SAS Cables
Depending on the enclosure topology, one or two external SAS cables may be attached between
the host system(s) and the PowerVault MD1000 storage enclosure. SAS cables use four active 3
Gbps SAS lanes in a single cable to deliver an aggregate of 12.0 Gbps connection. All PowerVault
MD1000 SAS cables purchased from Dell have been tested and comply with the ANSI SAS 1.0
standard. Dell supports using only Dell-provided cables.
External SAS Connector Types
SAS uses a 25-pin connector whose specification is governed by the SFF-8470 specification. For
more information, refer to the specification referenced in Appendix A.
Regardless of the topology, the PowerVault MD1000 storage enclosure also provides some
common capabilities as listed below:
Redundancy Capabilities
•
Management Redundancy
With two EMMs installed, the PowerVault MD1000 storage enclosure’s internal enclosure
functions on the PowerVault MD1000, such as temperature regulation, LED control, and
alarm control, are maintained in the event of an EMM failure. However, even with dual
EMMs installed, enclosure status will not be reported to the host in the event of an EMM
failure.
•
Power Supply and Cooling Redundancy
The PowerVault MD1000 storage enclosure requires that both power supplies be installed
for normal operation. The system can operate using a single power supply for a limited
time in the event a power supply fails or is removed for service.
The PowerVault MD1000 storage enclosure provides four individually controlled fans, two
installed in each power supply. The unit is thermally designed to continue operating in the
event of one fan blower failure (3+1 cooling redundancy). In the event of a fan failure, the
power supply containing the failed fan must be replaced.
PAGE 13
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
Reliability Capabilities
•
Thermal Reliability
A fully-redundant PowerVault MD1000 storage enclosure contains four digital temperature
sensors, two on the backplane and one on each EMM. In conjunction with the SES
processor, the temperature sensors provide device protection against adverse thermal
conditions. The PowerVault MD1000 storage enclosure is designed to operate within an
ambient temperature range of 10°C (50°F) to 35°C (95°F). When either of the backplane
sensors detects temperature below 3°C (37.4°F) or above 55°C (131°F), the SES
processor notifies the host of a critical temperature situation. This triggers the host to
perform a graceful shutdown of the enclosure within three minutes. In the event that the
host is unable to initiate the enclosure shutdown process or a host is not connected to the
enclosure, the PowerVault MD1000 storage enclosure has an embedded thermal
shutdown feature. This is activated by the SES processor in either of the following cases:
o
Either of the backplane temperature sensors detects temperature below -2°C
(28.4°F) or above 60°C (140°F).
o
Temperature sensor on an EMM detects temperatures below -2°C (28.4°F) or
above 65°C (149°F).
The PowerVault MD1000 storage enclosure shuts down within 5 seconds of the activation
of an embedded thermal shutdown.
PowerVault MD1000 Serviceability Capabilities
PAGE 14
•
All the major PowerVault MD1000 components are easily accessible and replaceable
even when in a rack.
•
The PowerVault MD1000 storage enclosure is equipped with LED indicators for visual
status reporting on the status of internal components.
•
The PowerVault MD1000 storage enclosure’s audible alarm notifies users of critical/noncritical status (for details, see PowerVault MD1000 Hardware Owners Manual). Status is
also reported to the host system’s management tool, Server Administrator (for details, see
the OpenManage User’s Guide.)
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
5. PERC 5/E Controller - Detailed Features
PERC 5/E is the RAID controller used to connect the PowerVault MD1000 storage enclosure with
PowerEdge servers. It supports 3 Gbps Serial Attached SCSI (SAS) as the storage interconnect
technology and PCI -E as the host-based interconnect technology. PERC 5/E controller supports
RAID levels 0, 1, 5, 10 and 50.
Considerations when implementing a RAID Solution
The selection of a RAID solution in any environment is governed by several aspects including
controller and capacity requirements. Each of these factors is discussed in detail in the following
sections.
Controller Features
The Server Administrator is the recommended application for configuring and managing the
PowerVault MD1000 storage enclosure. Apart from the Server Administrator software, the BIOS
configuration utility (<CTRL> <R>) may be used to perform pre-OS configuration and fault
recovery. Some of the key features of the PERC 5/E controller are listed here:
•
Virtual Disk Initialization
PERC 5/E controller is capable of performing initialization on the virtual disk(s) by either
writing 0’s to a portion of the physical disk (fast initialization) or to the entire physical disk
(full initialization). Full initialization guarantees that any legacy data is cleared and all
sectors are healthy.
•
Background Patrol Read
Background Patrol Read is an automatically scheduled background task that scans and
resolves possible physical disk errors that may lead to physical disk failure and result in
data integrity issues. Dell recommends implementing Background Patrol Read in all the
configurations.
•
Consistency Check/Background Initialization (BGI)
Consistency checks / BGI are performed by PERC 5/E controller to verify that all stripes in
a redundant virtual disk are consistent and to automatically correct stripes where
inconsistencies are found. BGI is a consistency check that automatically runs on all
redundant arrays. Depending on workload, Dell recommends that consistency checks are
run manually on virtual disks at least once every month to ensure data integrity. Running a
consistency check is complement to running Patrol Read and both should be used.
PAGE 15
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
•
VER A00
RAID Level Migration
Changing the RAID level of a virtual disk from the existing level (e.g. RAID 1) to another
level (e.g. RAID 5) is called RAID level migration. Dell recommends that data in the virtual
disk be backed up before a RAID level migration is initiated. Dell also strongly
recommends that the size of the physical disks to which the RAID level is to be migrated is
greater than or equal to the size of the physical disks in the existing virtual disk. Dell
supports RAID level migration as follows:
RAID 1 to RAID 0
RAID 0 to RAID 1
RAID 5 to RAID 0
RAID 0 to RAID 5
RAID 1 to RAID 5
•
Capacity Expansion
Expanding the capacity of a virtual disk by adding another physical disk is called
capacity expansion. Dell supports capacity expansion on all the supported RAID levels
except RAID 1. Dell recommends that the data in the virtual disk be backed up before
capacity expansion is initiated. For more details on capacity expansion, see the Dell
PERC 5/E User’s Guide.
•
Controller Write Cache Policy Options
Write–Back Caching
In write-back caching, data transfer is completed when the controller cache receives all
data from host for the write transaction. Write-back caching is faster than write through
caching. The main advantage of write-back caching is improved performance since the
controller does not have to wait for an acknowledgement from the physical disk before
proceeding. The battery backup unit for the controller is required to enable write-back
caching. For more details on write-back caching, see the Dell PERC 5/E User’s Guide.
Write-Through Caching
In write-through caching, a data transfer is completed when the disk subsystem receives
all of the data from the host. Write-through caching provides additional security because
data must be committed to the physical disk before proceeding. There may be a minimal
performance impact since the controller must wait for the physical disk to return a good
status to the controller before proceeding to the next operation.
•
Hot Spare Support
When a physical disk is assigned to a virtual disk as a hot spare, the PERC 5/E controller
will automatically rebuild the virtual disk using the hot spare in the event of a failed physical
disk. This restores the virtual disk to optimal redundancy. The hot spare disk must be
equal to or greater than the largest physical disk size within any of the virtual disks. Dell
recommends assigning hot spares to redundant virtual disks only, (i.e. RAID 1, RAID 5,
RAID 10 and RAID 50.)
Global Hot Spare
PAGE 16
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
A global hot spare is a backup physical disk capable of replacing a failed physical disk in
any virtual disk on PERC 5/E controller. The capacity of a global hot spare should be
greater than or equal to the largest physical disk in all the disk groups.
Dedicated Hot Spare
A dedicated hot spare is a backup physical disk capable of replacing a failed physical disk
in a specific PERC 5/E virtual disk. The capacity of a dedicated hot spare should be
greater than or equal to the largest physical disk in the virtual disk to which it is assigned.
Dell recommends assigning dedicated hot spares to all business critical virtual disks.
•
Disk Roaming
Disk roaming is initiated when the physical disks are changed to different ports on the
same controller. The physical disk roaming feature will detect the RAID configuration
from the configuration metadata residing on the physical disk.
•
Rebuild
The process of restoring the redundancy of a RAID 1, 5, 10 or 50 virtual disk is called
rebuilding. The rebuilding process can be initiated manually or automatically. Dell
recommends choosing the automatic rebuild option to prevent running in the degraded
mode for an extended period of time.
•
Mixed-Size Physical Disk Within RAID 10 or RAID 50
When mixing physical disk sizes in RAID 10 or RAID 50, the controller will not coerce the
larger virtual disks to match the size of the smallest virtual disk. PERC 5/E controller will
stripe across all disks until the smallest one is full, then it will continue to stripe across the
remaining disks until the next one is full, and so on until all the available space is used.
This process can impact performance as the array fills up, since the numbers of stripes
decreases as disks fill up.
Capacity Requirements
The PERC 5/E controller supports up to sixty four virtual disks simultaneously of which each
virtual disk can contain up to thirty two physical disks. The maximum number of total virtual
disks that can be supported on each physical disk is sixteen. PERC 5 controllers support
virtual disks greater than 2 TB in size, but Dell recommends checking the capacity of the
operating systems to support boot volumes of that capacity.
http://www.microsoft.com/windowsserversystem/storage/getstorfacts.mspx
http://www.dell.com/downloads/global/vectors/2004_2tblun.pdf
PAGE 17
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
Table 2: PERC 5/E controller features
Feature
Comment
SAS Technology
3Gbps
Host Bus Interface
x8 PCI Express
Internal connectors
N/A
External Connectors
2
RAID Levels Supported
0, 1, 5, 10, 50
Cache Size (DIMM)
256MB DDR2
Write Cache Options
Write-Through, Write-Back
Read Cache Options
Read-Ahead, No Read Ahead
BBU Battery Life
Up to 72 hours
Maximum Number of Virtual
Disks/Controller
Maximum Number of
Physical Drives/Virtual Disk
Form Factor
I/O Processor (IOP)
PAGE 18
64
32
Full Height, Half-Length PCI
Adapter card
Intel IOP333
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
6. Dell OpenManage Server Administrator
Dell OpenManage Server Administrator is a management application for configuring PowerEdge
servers. The Server Administrator’s configuration abilities for host-based RAID and non-RAID disk
storage are of particular interest for this paper. The Server Administrator enables you to perform
controller and enclosure functions from a single, graphical or command line interface. The
graphical user interface (GUI) is wizard-driven with features for novice and advanced users and
detailed online help. The command line interface is fully-featured and scriptable.
The Server Administrator provides SAS support with the PERC 5/E controller and the PowerVault
MD1000 storage enclosure. Server Administrator also supports the SCSI and ATA technologies.
Managing the PERC 5/E Controller and the PowerVault MD1000 Storage Enclosure
After installing Server Administrator, the user can expand the storage object in the tree view to
display the storage components attached to the system. These components include supported
controllers and enclosures, virtual disks, physical disks, and the connectors (SAS ports and/or
SCSI channels) that comprise the topography of all attached storage.
The PERC 5/E controller and the PowerVault MD1000 storage enclosure are represented in
Server Administrator’s tree view. You can select the PERC 5/E object and the PowerVault MD1000
object to display the status and to provide access to PERC 5/E controller and PowerVault storage
enclosure tasks.
Using the Create Virtual Disk Wizards
You can create virtual disks using either Server Administrator’s Create Virtual Disk Express or the
Create Virtual Disk Advanced wizard. These wizards enable you to quickly configure virtual disk
attributes, such as the RAID level, size, and physical disk selection.
The Express wizard displays the RAID levels supported by the controller and enables you to select
the desired RAID level and the disk size for your virtual disk. After you select the RAID level, the
Express wizard selects an appropriate physical disk given the size of the physical disks and the
available disk space. The Express wizard displays the physical disks it has selected so that you
can confirm the selection before creating the virtual disk.
The Advanced wizard allows you to select physical disks and specify additional virtual disk
properties such as the read, write, and cache policies.
PAGE 19
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
Figure 6: Screenshot of Create Virtual Disk Advanced Wizard with Disk Selection Displayed
The Create Virtual Disk Express and Create Virtual Disk Advanced wizards set the virtual disk
initialization to “Fast Initialize” by default. After completing the creation of a virtual disk using the
Express or Advanced wizard, the PERC 5/E controller implements the virtual disk configuration on
the selected physical disks and completes a fast initialize.
Reconfiguring Virtual Disks -- RAID Migration and Capacity Expansion
Server Administrator supports RAID migration and capacity expansion with the Reconfigure Virtual
Disk wizard. This wizard enables you to change a virtual disk’s RAID level or increase its size by
adding one or more physical disks.
The possibilities for reconfiguring a virtual disk depend on the capabilities of the RAID controller,
the existing RAID level, and the available physical disks. The following table describes the possible
reconfiguration scenarios on the PERC 5/E controller.
PAGE 20
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
Table 3: Virtual disk reconfiguration on the PERC 5/E RAID controller
PERC 5/Controller
Starting RAID Level
Target RAID Level Disk Requirements
RAID 0
RAID 0
Add at least one additional disk
RAID 0 (on a single disk) RAID 1
Add a single disk
RAID 0
RAID 5
Add at least one additional disk
RAID 1
RAID 0
With or without adding additional
disks
RAID 1
RAID 5
Add additional disks
RAID 5
RAID 0
With or without adding additional
disks
RAID 5
RAID 5
Add additional disks
Other considerations may apply to virtual disk reconfiguration. For example, you cannot create
more than sixty four virtual disks on the PERC 5/E controller. After reaching this limit, you will not
be able to reconfigure any of the virtual disks on the controller.
The Reconfigure Virtual Disk wizard displays the available target RAID levels and the physical
disks. After you have selected the new RAID level and the physical disks, the Reconfigure Virtual
Disk wizard displays the previous and the new configuration so that you can confirm or reject the
changes.
Moving Physical Disks and Virtual Disk Configurations – Disk Roaming and Disk
Migration
Disk roaming refers to physically moving a disk from one cable connection to another or from one
backplane or enclosure slot to another on the same controller. The PERC 5/E controller recognizes
the relocated physical disks and logically restores the physical disk and its data to the proper virtual
disk. Disk roaming can be performed only when the system is shut down.
Disk migration refers to moving physical disks from one controller to another. Virtual disks residing
on the relocated physical disks are identified as a foreign configuration on the receiving PERC 5/E
controller. The Server Administrator enables you to import the foreign configuration so that the
migrated virtual disks can be managed on the receiving PERC 5/E controller.
The following conditions must be met to successfully migrate the physical disks and their resident
virtual disks:
PAGE 21
•
Virtual disks on a SAS controller can only be migrated to another SAS controller.
•
When moving an enclosure, power down the enclosure and the server before moving the
enclosure. The enclosure can be moved to any connector number on the receiving
controller.
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
The following procedure describes how to migrate the physical disks from one controller to another:
1.
Shut down the system from which the physical disks are being moved.
2.
Shut down the server of the receiving controller if it does not have a preexisting virtual disk
configuration.
3.
Place the physical disks into the new enclosure.
4.
Start up the system connected to the receiving controller.
5.
Use Server Administrator’s Import Foreign Configuration wizard to import the migrated virtual
disks on the receiving controller.
Importing Migrated Virtual Disks
After migrating the physical disks, Server Administrator enables you to import the virtual disks or
“foreign” configuration that reside on the relocated physical disks.
To be imported, the foreign configuration must contain a virtual disk that is in either “Ready” or
“Degraded” state. For proper completion of this task, all of the virtual disk data must be present. If
the virtual disk is using a redundant RAID level, the additional redundant data does not need to be
present. For example, if the foreign configuration contains one side of a mirror in a RAID 1 virtual
disk, then the virtual disk is in a degraded state and can be imported. On the other hand, if the
foreign configuration contains only one physical disk that was originally configured as a RAID 5
using three physical disks, then the RAID 5 virtual disk cannot be imported.
In addition to the virtual disks, a foreign configuration may consist of a physical disk that was
assigned as a hot spare on one controller and then moved to another controller. The Import
Foreign Configuration wizard imports the new physical disk as a hot spare. If the physical disk was
a dedicated hot spare on the previous controller, but the virtual disk to which the hot spare was
assigned is no longer present in the foreign configuration, then the physical disk is imported as a
global hot spare.
If a physical disk contains all or some portion of a foreign configuration, then Server Administrator
displays the physical disk state as “Foreign.”
PAGE 22
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
Figure 7: Physical Disk with a Foreign State
The Server Administrator displays the Import Foreign Configuration task when the controller
detects a foreign configuration. If you have an incomplete foreign configuration which cannot be
imported, using the Clear Foreign Configuration task, erases the foreign data on the physical disks.
PAGE 23
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
Figure 8: Controller Tasks Drop-down Menu With the Import Foreign Configuration and Clear
Foreign Configuration Tasks Displayed
Upgrading from Dell OpenManage Array Manager
Dell has discontinued the release of Dell OpenManage Array Manager. Server Administrator is a
replacement for Array Manager. Customers wishing to take advantage of SAS technology must
upgrade their Array Manager installation to Server Administrator. Server Administrator also
provides Linux support which was not offered with Array Manager.
Note: OMSA: Server Administrator does not support NetWare. Existing NetWare customers may
continue to use their Array Manager licenses for existing systems, but additional licenses will not
be granted for new systems.
Note: Server Administrator does not provide support for Fibre channel arrays.
The following sections describe the considerations for upgrading to Server Administrator.
Virtual Disk Preservation
You can preserve assigned virtual disk names when migrating from Array Manager to Server
Administrator. To do so, you must not uninstall Array Manager prior to installing Server
Administrator. If Array Manager is uninstalled prior to installing Server Administrator, the virtual
disks are renamed. Irrespective of whether Array Manager is uninstalled, Server Administrator
identifies and manages the virtual disks created with Array Manager.
PAGE 24
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
SNMP Traps
The architecture for handling SNMP traps and the Management Information Base (MIB) is different
in Server Administrator than in Array Manager. You may need to modify the applications that are
customized to receive SNMP traps from Array Manager.
Event Numbering
The numbering scheme for Storage Management alerts or events is different than the numbers
used for the corresponding Array Manager events. See the Alert Messages chapter in the Storage
Management online help for more information.
7. Storage Applications and Components
Identify customer usage model and needs
In order to select the correct storage solution, it is important to understand the application and user
requirements. A good starting point is to use the following basic storage profiling considerations.
Table 4: Storage Profiling Considerations
PAGE 25
Characteristic
Performance
Values
• Bandwidth (MB per
sec.)
• I/O size (KB/MB)
• I/O Profile (read/write
and random/sequential
access mix)
• Latency
Storage
Capacity
Needs
Gigabytes
Storage
Growth Rate
Percent increase per year
Criticality
Low, Medium, High, Very
High
Description
Performance is the overall ability of the
solution to read and write data to the disk.
The performance requirements are
usually determined by the type of
application being utilized. Different
applications have different performance
requirements. For example, a database or
e-mail server has mostly random disk
access operations while a streaming
media server would have mostly
sequential disk access.
Storage capacity is the amount of storage
space required by the application and user
data. For example, e-mail storage for 100
users would require much less storage
capacity than an e-mail store for 1000
users.
Storage growth is the expected increase
in the amount of the capacity that will be
required as the usage increases. This can
be estimated by forecasting the number of
users or clients expected to access the
application in the future.
Criticality defines the impact to business
needs if the storage is offline. This
characteristic is important for choosing the
right RAID level.
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
Picking a solution based on the application profile
One of the most important factors to consider when selecting a storage solution is the type of
application that will be utilizing the storage solution. This defines the overall purpose of the
server/storage solution and will determine what RAID configuration will be optimal for the
application. Table 5 outlines the recommended RAID solution based on the usage of the server.
Table 5: RAID Configuration Recommendations
Application
RAID Level
Concatenated
0
1
10
5
50
Email
Web
Database
or On-line
transaction
processing
Streaming
Media
File –
Archival
File – User
File Stores
Recommended
Not Recommended
Possible
General RAID Configuration Guidelines
RAID 0
RAID 0 is generally not a recommended solution due to lack of data redundancy and protection.
However, it may be utilized in situations where this is not required and maximum storage capacity
and performance are essential.
RAID 1
A RAID 1 solution, while not ideal, could be utilized for small workgroup servers or servers with low
storage capacity and growth requirements. Any server beyond a small workgroup server is likely to
need more storage and consequently, a more efficient RAID solution.
PAGE 26
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
RAID 10
RAID 10 has good I/O performance, excellent availability and redundancy. The biggest drawback
of RAID 10 is that it requires twice the number of physical disks needed for data. This RAID
configuration should be used in situations requiring maximum availability, redundancy, and
performance.
RAID 5
RAID 5 has good I/O performance, data protection and requires only one additional physical disk
than the number needed for data. RAID 5 should be used in situations where maximum storage
capacity is required along with a moderate amount of data protection.
RAID 50
A balance between RAID 5 and RAID 10, this solution offers good I/O performance, availability and
good storage capacity. This configuration offers slightly higher performance than RAID 5, but
requires multiple additional physical disks than the number needed for data. It does however
provide greater storage capacity than a RAID 10 solution with a decrease in performance and data
protection. This should be used in situations where greater redundancy and data protection is
required as well as a reasonable storage capacity.
Concatenated
This solution is not recommended due to lack of data protection and redundancy and no
performance gain.
Note: For more details on RAID configurations see the Dell ‘Getting Started with RAID’ document
(http://support.dell.com/support/edocs/storage/RAID/RAIDbk0.pdf)
Application specific guidelines
E-Mail Servers
The storage requirements for e-mail servers can vary depending on the size, the amount, and the
type of users. While small departmental e-mail servers may work well with a small amount of
storage and limited features; large corporate e-mail servers normally require greater storage
capacity, very high availability, performance, and scalability. I/O profiles will vary depending on the
number of users and type of mail and attachments sent.
Table 6: Email Server General I/O profile
PAGE 27
I/O Profile
I/O Profile
(Read/Write)
(Sequential/Random)
60/40
Random
Bandwidth
I/O
Size
Heavy
4k
Latency
Sensitivity
High
Growth
Rate
High
Criticality
High
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
Table 7: Email Sever RAID Guidelines
RAID Level
Concatenated
Application
0
1
10
5
50
E-Mail
Recommended
Not Recommended
Possible
Recommended:
•
RAID 10 – Recommended for mission-critical e-mail servers where availability and
redundancy and/or performance are the highest priority and where reduced storage
capacity is acceptable.
•
RAID 5 – Recommended for e-mail servers that require maximum storage capacity and
moderate data protection and performance.
•
RAID 50 – Recommended for those solutions that require greater redundancy and data
protection and a balance between storage capacity and performance.
Possible:
•
RAID 1 – Possible solution for small e-mail servers which do not require high storage
capacity.
Not recommended:
•
RAID 0, Concatenated – Not recommended due to lack of redundancy and data
protection.
Note: While these configurations are not recommended, they can be configured and
utilized.
Web Servers
Web severs are usually high traffic systems where read operations are the most common disk
activity as web pages are requested by users. They can be intranet sites with minimal internal
company traffic or global internet portal sites that receive hundreds of thousands of users daily.
Performance is generally the main concern with redundancy and data protection being less of a
concern. For large web sites, there are usually several web servers responding to client requests
and as a result the need for redundancy and data protection is not as important. Also web servers
may not require as much scalability or disk capacity since website content is relatively static and
changes are usually minor content modifications.
Table 8: Web General IO profile
PAGE 28
I/O Profile
I/O Profile
(Read/Write)
(Sequential/Random)
95/5
Random
Bandwidth
I/O
Size
Latency
Sensitivity
Growt
h Rate
Criticality
Moderate
< 64k
Moderate
Low
High
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
Table 9: Web RAID Guidelines
RAID Level
Concatenated
Application
0
1
10
5
50
Database
Recommended
Not Recommended
Possible
Recommended:
•
RAID 10 – Recommended for enterprise web server solutions where availability and
redundancy and performance are the highest priority, usually for stand-alone portal sites
that are critical to the organization’s business.
•
RAID 5 – Recommended for web servers that require maximum storage capacity and
moderate data protection and performance.
•
RAID 50 – Recommended for those solutions that require a balance between storage
capacity and performance.
•
RAID 0 – Recommended for solutions where the web server will be part of a group of web
servers that service a large internet portal and performance is the highest concern. In this
situation, availability and redundancy are addressed by the cluster group and do not need
to occur at the disk level.
•
RAID 1 – Good solution for small websites which do not require high storage capacity.
Not recommended:
•
Concatenated - This solution is not recommended due to lack of redundancy and data
protection
Note: While this configuration is not recommended, it can be configured and utilized.
Database or Online Transaction Processing Servers
®
Database servers can range from simple workgroup databases like Microsoft Access with a few
hundred users to mission-critical enterprise databases like Oracle or SQL Server with thousands of
users. Database applications will always benefit from some data protection while other
requirements such as performance and availability will vary. As a general rule, the more critical the
database, greater is the need for data protection. Additionally, the performance requirements
increase relative to the number of users accessing the database.
Online Transaction Processing (OLTP) oriented servers are used in a number of industries for the
entry and retrieval of transactions. For example, OLTP is common in banking, airlines, mail-order,
and supermarkets. These servers are generally mission-critical and require maximum availability
and redundancy possible.
Table 10: Database or OLTP General IO profile
PAGE 29
I/O Profile
I/O Profile
(Read/Write)
(Sequential/Random)
80/20
Random
Bandwidth
I/O
Size
Latency
Sensitivity
Growt
h Rate
Criticality
Moderate
8k
Moderate
Low
High
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
Table 11: OLTP RAID Guidelines
Application
RAID Level
Concatenated
0
1
10
5
50
OLTP
Recommended
Not Recommended
Possible
Recommended:
•
RAID 10 – Since OLTP systems are critical for most business, this RAID configuration is
highly recommended due to the high availability and redundancy and good performance.
•
RAID 5 – Recommended for OLTP servers that require maximum storage capacity and
moderate data protection and performance.
•
RAID 50 – Recommended for those solutions that require a balance between storage
capacity and performance.
Possible:
•
RAID 1 – Possible solution for situations which do not require high storage capacity.
Not recommended:
•
RAID 0, Concatenated - These are not recommended due to lack of redundancy and data
protection.
Note: While these configurations are not recommended, they can be configured and utilized
File Servers
File servers can be used for file archival or more dynamic storage where files are changed, added
and deleted on a daily basis. These servers range from workgroup to the corporate level. Storage
capacity is a key attribute of these servers as users add more and more files. File servers are
generally not mission critical systems so lower levels of availability and redundancy are acceptable
as data is usually backed up and can be restored in a matter of hours.
Archival file server characteristics and recommendations:
Table 12: Archival File Server General I/O profile
PAGE 30
I/O Profile
I/O Profile
(Read/Write)
(Sequential/Random)
90/10
Sequential
Bandwidth
I/O
Size
Latency
Sensitivity
Growt
h Rate
Criticality
Moderate
>64k
High
Varies
Low
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
Table13: Archival File Server RAID Guidelines
Application
RAID Level
Concatenated
0
1
10
5
50
File Archival
Recommended
Not Recommended
Possible
Recommended:
•
RAID 10 – Recommended due to high availability, redundancy, and good performance.
•
RAID 5 – Recommended for file servers that require maximum storage capacity and
moderate data protection and performance.
•
RAID 50 – Recommended for those solutions that require a balance between storage
capacity and performance.
Possible:
•
RAID 1 – Possible solution in situations which do not require high storage capacity.
Not recommended:
•
RAID 0, Concatenated - These are not recommended due to lack of redundancy and data
protection. Even for long term archival storage which is intended to be a backup of
important files, it is not a recommended solution.
Note: While these configurations are not recommended, they can be configured and utilized.
User file store characteristics and recommendations
Table 14: User Store File Server General IO profile
I/O Profile
I/O Profile
(Read/Write)
(Sequential/Random)
80/20
Sequential
Bandwidth
I/O
Size
Latency
Sensitivity
Growt
h Rate
Criticality
Heavy
>64k
High
Varies
Moderate
Table 15: User Store File Server RAID Guidelines
Application
RAID Level
Concatenated
0
1
10
5
50
File – User
file stores
Recommended
PAGE 31
Not Recommended
Possible
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
Recommended:
•
RAID 5 – Recommended for file servers that require maximum storage capacity and
moderate data protection and performance.
•
RAID 50 – Recommended for those solutions that require a balance between storage
capacity and performance.
Possible:
•
RAID 1 – Possible solution for situations which do not require high storage capacity.
•
RAID 10 – Possible solution for situations that require greater availability and redundancy
but not as much disk capacity.
•
RAID 0 – While this solution is not redundant, RAID 0 may be useful for file servers that
require better performance and maximum storage capacity, for example, where backups
are expected to maintain data in the event of a failure.
•
Concatenated – This solution could be used for situations that require maximum scalability
as physical disks could be added as storage needs grow. For file servers that can afford
some downtime and rely on backups to restore data, this could be a solution.
Streaming Media Servers
Streaming media servers are systems that provide web-casting, video conferencing, internet
entertainment (e.g. Internet TV or Internet radio), and multimedia services. These systems
generally require a balance between storage capacity, availability, redundancy, and performance.
Like web servers, they may also be part of a group of systems that work together to provide
content.
Table 16: Streaming Media IO profile
I/O Profile
I/O Profile
(Read/Write)
(Sequential/Random)
98/2
Sequential
Bandwidth
I/O
Size
Latency
Sensitivity
Growt
h Rate
Criticality
Heavy
>64k
High
High
Varies
Table 17: Streaming Media RAID Guidelines
Application
RAID Level
Concatenated
0
1
10
5
50
Streaming
Media
Recommended
Not Recommended
Possible
Recommended:
•
PAGE 32
RAID 5 – Recommended for streaming media servers that require maximum storage
capacity and moderate protection and performance. This is the recommended RAID
configuration for stand alone streaming media servers.
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
•
RAID 50 – Recommended for those solutions that require a balance between storage
capacity and performance. This solution provides greater availability and redundancy but
less storage capacity.
•
RAID 0 – While not redundant, this solution may be useful for streaming media servers
that require better performance and are members of a cluster of servers where availability
and redundancy are handled at a higher level.
Possible:
•
RAID 1, RAID 10 – Possible in situations which do not require high storage capacity. In
general servers utilized for streaming media require large amounts of storage capacity for
content, especially uncompressed video.
Not Recommended:
•
Concatenated – This solution is not recommended due to lack of redundancy and limited
performance.
Note: While these configurations are not recommended, they can be configured and utilized.
Hot Spares
Hot Spare functionality provides extra security, availability and redundancy by automatically
replacing a failed physical disk in a RAID group and allowing the rebuild of the degraded array to
begin immediately. PERC 5/E controllers support Hot Spare functionality and it is recommended
whenever possible.
8. Summary and Conclusions
It is important to understand the characteristics of each storage component in order to assess the
performance of the overall application. It is also necessary to understand how the storage
component is configured since this has a direct impact on the effectiveness of an application. Table
8-1 summarizes typical storage recommendations for various applications. Recommendations
listed in the table are general conclusions and the actual storage solution may have additional
factors which need to be considered before determining an optimal solution. Additional factors
could include:
• Data and power back-up strategies of an Enterprise IT
• Storage deployment environment:
o Number of active and passive users
o Type of data – mission critical or non-mission critical
• Storage deployment budget
• Long term storage consolidation strategy
Table 18: Typical Storage Recommendation for Different Storage Applications
Application
Key Parameters
RAID Level
Email
Availability
Performance
Scalability
Availability
RAID 10 provides
both availability
and performance
RAID 10 provides
Database
PAGE 33
Remarks
Storage Recommendations
Servers
PowerVault
MD1000
Daisy-chaining
provides capacity
scalability
Daisy-chaining
Host Controller
For added
performance enable
Write back cache
For added
Recommendation targets
a typical enterprise e-mail
Server
Performance
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
Performance
both availability
and performance
provides capacity
scalability
performance enable
Write back cache
Web
Performance
RAID 0 in Web
Server farms
provides maximum
performance of all
RAID levels
Enable read ahead
capabilities for
added performance
boost
On-line
Transaction
High Availability
RAID 10
provides highest
availability of all
RAID levels
Split Topology
spanning RAID 0
across channels
provides better
performance if
daisy-chaining for
additional capacity
is not required
Daisy-chaining
provides capacity
scalability
File
Scalability
RAID 5
maximizing
storage with some
data protection
Daisy-chaining
provides capacity
scalability
Enable Global Hot
Spare for improved
availability for RAID
5
Streaming
Media
Performance
RAID 0
provides maximum
performance of all
RAID levels
Split Topology
spanning RAID 0
across channels
provides better
performance if
daisy-chaining for
additional capacity
is not required
Enable read ahead
capabilities for
added performance
boost
PAGE 34
For added
performance enable
Write back cache
VER A00
requirements for database
servers depend on
number of users
Web server farms, usually
have multiple servers
providing replicated web
data. Thus, data
availability is handled at
server level
Recommendation targets
Enterprise applications
which require higher
availability than
performance.
In general for all OLTP,
data protection is of
utmost importance
These servers usually
have data backed up
periodically and hence,
require minimum data
protection from storage
components
Performance is of utmost
importance for Streaming
servers. Data on these
servers is usually backed
up on external media such
as tape libraries or optical
media - CD/DVD
5/06/2005
A REFERENCE GUIDE FOR OPTIMIZING DELL™ MD1000 SAS SOLUTIONS
VER A00
9. Appendix – A: References
1.
SAS Protocol Specification
(http://www.t10.org/drafts.htm#SCSI3_SAS)
2.
SFF-8470 Connector Specification
(http://www.sffcommittee.com/ns/index.html)
3.
Beyond The 2-TB SCSI Logical Unit
(http://www.dell.com/downloads/global/vectors/2004_2tblun.pdf)
10. Appendix – B: Glossary
Array – Collection of multiple physical disks to form a single logical volume
BBU – Battery Backup Unit
BIOS – Basic Input Output System
Cluster – Multi-initiator Environment with MSCS operating environment
HBA – Host Bus Adapter (non-RAID controllers)
Hot Plug – Insertion or removal of device without the need to quiescent I/Os
PERC – PowerEdge Expandable RAID Controller
RAID – Redundant Array of Inexpensive Disks or Redundant Array of Independent Disks
SAS – Serial Attached SCSI
SCSI – Small Computer System Interface
U320 – Ultra 320 (320MB/sec through put)
11. Appendix – D: Revision History
PAGE 35
Revision
Description
Date
Modified By
A00
First Release
4/17/2006
SAS Solution
Core Team
5/06/2005
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

advertising