MegaRAID Performance Tuning & Benchmark Tips

MegaRAID Performance Tuning & Benchmark Tips

English
30 Pages
Read
Download
Downloading requires you to have access to the YouScribe library
Learn all about the services we offer

Description

MegaRAID Benchmark Tips
January
27, 2010 Benchmark Review Requirements

Providing
the following information will improve our ability to support product
evaluations

Summary of test cases
you
will
be
performing

System details and benchmark
parameters listed
on
slide
3

Questions regarding configuration and
benchmark results
should be
addressed to:

L
SI Technical
Support

support@lsi.com

(
For Product Evaluations)

C
raig McCombs

craig.mccombs@lsi.com

J
ames Evans

james.evans@lsi.com
LSI Proprietary
2 Elements that Effect Performance

System

M
otherboard,
Chip
Set, BIOS, Processor,
Memory

S
ystem chip set and memory s
peed can impact benchmark performance

R
ecommend
8-wide (x8) PCIe
Generation-2

slot for all 6 Gb/s SAS
benchmarks

O
perating System with latest
Service
Pack and Updates

RAID controller

Firmware, BIOS,
driver version

D
isk
write cache policy
setting
–R
A
I
D
l
e
v
e
l

Stripe size

R
ead and
Write
Caching policies

Drives

T
otal
number of drives

N
umber of
drives per channel

Enclosure model
+
firmware

D
rive interface
(SATA, SAS) and
Speed (i.e.
1.5Gb/s,
3Gb/s,
6Gb/s)

M
ake and model of drive

Firmware version

C
lass of
the
drive
(e.g.,
HDD disk
drive,
SSD solid
state,
etc.)

Benchmark Tool

Test profile ( Block
Size, Sequential
R
ead or Write,
Random Read or Write)
LSI Proprietary
3





Tips
Use the latest firmware for testing the RAID ...

Subjects

Informations

Published by
Reads 8 769
Language English
Document size 1 MB
Report a problem
MegaRAID Benchmark TipsJanuary 27, 2010
Benchmark Review Requirements•Providing the following information will improve our ability to support product evaluations–Summary of test cases you will be performing–System details and benchmark parameters listed on slide 3•Questions regarding configuration and benchmark results should be addressed to:–LSI Technical Support•support@lsi.com–(For Product Evaluations)–Craig McCombs•craig.mccombs@lsi.com–James Evans•james.evans@lsi.comLSI Proprietary2
Elements that Effect Performance•System–Motherboard, Chip Set, BIOS, Processor, Memory•System chip set and memory speed can impact benchmark performance•Recommend 8-wide (x8) PCIeGeneration-2 slot for all 6 Gb/s SAS benchmarks–Operating System with latest Service Pack and Updates•RAID controller–Firmware, BIOS, driver version–Disk write cache policy setting–RAID level–Stripe size–Read and Write Caching policies•Drives–Total number of drives•Number of drives per channel –Enclosure model + firmware–Drive interface (SATA, SAS) and Speed (i.e. 1.5Gb/s, 3Gb/s, 6Gb/s)–Make and model of drive–Firmware version–Class of the drive (e.g., HDD disk drive, SSD solid state, etc.)•Benchmark Tool–Test profile ( Block Size, Sequential Read or Write, Random Read or Write)LSI Proprietary3
spiTƒUse the latest firmware for testing the RAID controller–How to update (flash) RAID on controller FW?ƒMegaRAID Storage Management (MSM)ƒMegaRAID Command Line Interface (MegaCLI)ƒReboot systemƒDownload latest FW from MegaRAID DownloadsƒTo configure the RAID Adapter and create logical arrays use either–CTRL-H utility during BIOS POST–Use MegaRAID Storage Management (MSM) running from OS•Use the latest OS drivers (use Storportdriver when possible)–Can be downloaded from MegaRAID Downloads•To achieve maximum sustained RAID controller throughput–For maximum bandwidth (MB/s) use 64KB to 1MB sequential access patterns–For maximum operations (IO/s) use 0.5KB to 8KB random access patters–Connect enough drives to saturate the RAID controller to achieve maximum performance. •For example use 20 or more 15K SAS drives to achieve maximum bandwidth performance–Use all SAS channels connected to drives or drive enclosures•With multiple drive enclosures make sure drives evenly distributed across enclosures–Drive technology and cache size can significantly impact performance–Benchmark IO request queue depth will impact performance.  Recommend IO request queue depth of 16 or greaterLSI Proprietary4
NOStandard Benchmark Tool•IOmeteris a I/O performance analysis tool for servers and workstations•Obtain latest IOmeterfrom Source Forge–IOmeterUser’s Guide IOmeter.org–Windows 32-bit version (2006.07.27) satisfactory for 32-bit or 64-bit OS•Throughput is measured in IO/sec and MB/sec–I/Os per second for random/transactional workloads–Megabytes per second for streaming/sequential workloads•(See Appendix for IOmeterworkload configuration examples)TE: LSI canp rovideI OmeterLSI Proprietaryperformancet ests criptst og enerater esults5
yrateirporP ISL.ecnamroMegaRAID Settings for Maximum HDD Performancef•–RSAtIriDp e0 ,S i1z e:&  10––RWeriated  PPoolliiccyy::  r–IO Policy: –Disk Cache Policy:•RAID 5 & 6–Stripe Size: –Read Policy: –Write Policy: –IO Policy: –Disk Cache Policy:–Init State:eBK652Always Read AheadWrite Thru (For streaming sequential performance)Write Back (For transactional/random tests)Direct IOEnabled (1)p6 rehgih ecudorp semitemos nac tnemtsujda retemarap gnihcac lanoitidda ,daolkrow dna noitarugifnoc cificeps a roF .sdaolkrow dna snoitarugifnoc nommoc ssorca stluser lamitpo ecudorp snoitadnemmocer evoba eht eton esaelP)2(ecnamrofrep detimil evird fo secnatsni gniziminiM  .sesoprup gnitset rof ecnamrofrep evird secnahnE)1()emulov eritne ssorca atad etirw ro( noitazilaitini lluF)1( delbanEOI tceriDkcaB etirWdaehA daeR syawlABK652
arap gnihcac lanoitidda ,daolkrow dna noitarugifnoc cificeps a roF .sdaolkrow dna snoitarugifnoc nommoc ssorca stluser lamitpo ecudorp snoitadnemmocer evoba eht eton esaelP)3(sDSS rof noitacilppa lacipyt si hcihw ecnamrofrep modnar/lanoitcasnart rof ezimitpO)2(ecnamrofrep detimil evird fo secnatsni gniziminiM  .sesoprup gnitset rof ecnamrofrep evird secnahnE)1()emulov eritne ssorca atad etirw ro( noitazilaitini lluF)1( delbanEOI tceriD)2(kcaB etirW)2(daehA daeR oNBK652MegaRAID Settings for Maximum SSD Performance•RAID 0, 1 & 10–Stripe Size: –Read Policy: –Write Policy: –IO Policy: –Disk Cache Policy:•RAID 5 & 6–Stripe Size: –Read Policy: –Write Policy: –IO Policy: –Disk Cache Policy:–Init State:BK652No Read Ahead(2)Write Thru(2)Direct IOEnabled (1)7yrateirporP ISL.ecnamrofrep rehgih ecudorp semitemos nac tnemtsujda retem
Read Policies•Adaptive Read Ahead–This specifies that the controller uses read-ahead if the two most recent disk accesses occurred in sequential sectors. If all read requests are random, the algorithm reverts to No Read Ahead, however all requests are still evaluated for possible sequential operation.–Recommended for all HDD configurations•Always Read Ahead–The controller reads ahead all the data until the end of the stripe from the disk•No Read Ahead –Only the requested data is read and the controller does not read ahead any data–Recommended for all SDD configurationsLSI Proprietary8
Caching Policies•Direct IO–All read data is transferred directly to host memory bypassing RAID controller cache. Any Read Ahead data is cached.–All write data is transferred directly from host memory bypassing RAID controller cache if Write-Through cache mode is set–Recommended for all configurations•Cached IO–All read and write data passes through controller cache memory on its way to or from host memory. Includes write data in Write-Through .edom•This legacy option is being eliminated in a future releaseLSI Proprietary9
Write Policies•Write-Through–Caching strategy where data is written to disks before a completion status is returned to the host operating system –Considered more secure, since a power failure will be less likely to cause undetected drive write data loss with no battery-backed cache present–Recommend Write-Through for RAID 0, 1 and 10 and all SSD configurations•Provides optimum performance for streaming/sequential benchmarks•iSnitnoc cea dcahtea  iwsh imcho vceadn  diimrepcrtolyv efr oovme rthalel  pheorsfto trom tahnec dei fsokrs ,s titr eaavmoiidnsg  cwooprykilnoga tdhs ei f dDaitrae icnt tIeOr mmeodidaer iys  .tes•Write-Back–A caching strategy where write operations result in a completion status being sent to the host operating system as soon as data is in written to the RAID cache. Data is written to the disk when it is forced out of controller cache memory.–Write-Back is more efficient if the temporal and/or spatial locality of the requests is smaller than the controller cache size. –Write-Back is more efficient in environments with “bursty” write activity.–Battery backed cache can be used to protect against data loss as a result of a power failure or system crash.–Recommend Write-Back for RAID 0, 1, and 10 configurations•Provides optimum performance for transactional(random real world) benchmarks–Recommend Write-Back for RAID 5 and 6 configurations•Provides best overall performance•Improves performance of RAID-5 and 6 data redundancy generationLSI Proprietary01
Cache Variables•RAID Controller Cache–Various cache sizes supported on RAID adapters–Faster cache speed improves sequential RAID 5/6 write performance:•Example: DDR vsDDRII•Example: 533 MHz vs. 667 MHz vs. 800 MHz–Memory bit width will affect RAID 5/6 Seq. write performance•72-bit has up to an 18% advantage over 40-bit–Recommend high speed and large size RAID cache memory•System Cache–Performance impacted by system chip set and memory speed for many IO workloads. However run IOmeterbenchmark on unformatted (raw) device. This bypasses system cache for accurate storage performance measurements.•Drive Cache–Recall performance impacted by drive cache size for many IO workloadsLSI Proprietary11