Bottom-Up Database Hardware Benchmarking

Greg Smith

2ndQuadrant US

04/13/2011

Greg Smith Bottom-Up Database Hardware Benchmarking

About this presentation

I The master source for these slides is:

http://projects.2ndquadrant.com
I Source code to automate testing available there too
I Slides are released under the Creative Commons Attribution
3.0 United States License:
http://creativecommons.org/licenses/by/3.0/us

Greg Smith Bottom-Up Database Hardware Benchmarking

Why should you always benchmark your hardware?

I Many useful tests will only run when the server isn’t being
used yet
I Software stacks are complicated
I Spending money on upgrades only helps if you upgrade the
right thing usefully
I Vendors lie

Greg Smith Bottom-Up Database Hardware Benchmarking

Systematic Benchmarking

I Memory
I CPU
I Disk
I Database server
I Application

Greg Smith Bottom-Up Database Hardware Benchmarking

Databases and the CPU

I MySQL and PostgreSQL use only a single CPU per query

I Queries executing against cached data will bottleneck on CPU
I Both CPU and memory need to be fast for individual queries
to be fast

Greg Smith Bottom-Up Database Hardware Benchmarking

STREAM Benchmarking

http://www.advancedclustering.com/company-blog/

Greg Smith Bottom-Up Database Hardware Benchmarking

Oracle Calling Center OLTP Benchmark

http://it.anandtech.com/IT/showdoc.aspx?i=3769&p=4

Greg Smith Bottom-Up Database Hardware Benchmarking

stream-scaling memory read test

git clone git://github.com/gregs1104/stream-scaling.git

cd stream-scaling
./stream-scaling

Greg Smith Bottom-Up Database Hardware Benchmarking

Memory Speeds - DDR2 Era

Greg Smith Bottom-Up Database Hardware Benchmarking

Memory Speeds - DDR3 Era

Greg Smith Bottom-Up Database Hardware Benchmarking

Sources for slow memory results

I Single channel RAM/slot mistakes

I Incorrect SPD/timing/voltage
I Bad RAM/CPU multiplier combination
I Poor quality RAM
I BIOS setup error

Greg Smith Bottom-Up Database Hardware Benchmarking

CPU Tests

I Synthetic CPU benchmarks don’t work anymore

I Use an in-memory, CPU intensive database test instead
I Heavy trivial SELECT statements work well

Greg Smith Bottom-Up Database Hardware Benchmarking

Sources for slow CPU results

I Slow memory
I Power management throttling
I Linux: /proc/cpuinfo shows 1000MHz suggests you need to
adjust the CPUFreq Governor to “performance”

Greg Smith Bottom-Up Database Hardware Benchmarking

Disk Tests

I Sequential write: INSERT, Bulk loading (when not CPU

limited)
I Sequential read: SELECT * FROM and similar table
sequential scans
I Seeks: SELECT using index, UPDATE
I Commit fsync rate: INSERT, UPDATE

Greg Smith Bottom-Up Database Hardware Benchmarking

dd test

I Compute 2X the size of your RAM in 8KB blocks

I blocks = 250,000 * gigabytes of RAM

time sh -c "dd if=/dev/zero of=bigfile bs=8k count=X &&

sync"
time dd if=bigfile of=/dev/null bs=8k

I Watch vmstat and/or iostat during disk tests

I vmstat’s bi and bo will match current read/write rate
I Note the CPU percentage required to reach the peak rate

Greg Smith Bottom-Up Database Hardware Benchmarking

bonnie++

./bonnie++
bon csv2html

I Ignore the per-character and create results, look at the block

output/input ones
I Random Seeks:
I The test runs SeekProcCount processes (default 3) in parallel,
doing a total of 8000 random seek reads to locations in the
file. In 10% of cases, the block read is changed and written
back.

Greg Smith Bottom-Up Database Hardware Benchmarking

bonnie++ ZCAV

./zcav -f/dev/sda > t500

I Must get a recent version of bonnie++ for ZCAV to scale

properly for TB drives (1.03e works)
I ZCAV on experimental branch (1.96) gave useless results for
me
I Download somewhat broken gnuplot script sample and typical
results from:

http://www.coker.com.au/bonnie++/zcav/results.html

Greg Smith Bottom-Up Database Hardware Benchmarking

Improved bonnie++ ZCAV gnuplot script

unset autoscale x
set autoscale xmax
unset autoscale y
set autoscale ymax
set xlabel "Position GB"
set ylabel "MB/s"
set key right bottom
set terminal png
set output "zcav.png"
plot "raid0" title "7200RPM RAID 0 3 Spindles",
"single" title "7200RPM Single Drive"

Greg Smith Bottom-Up Database Hardware Benchmarking

bonnie++ ZCAV: Laptop 7200RPM Disk

Greg Smith Bottom-Up Database Hardware Benchmarking

Scaling of 3-Disk RAID0 with 7200RPM SATA Disks

Greg Smith Bottom-Up Database Hardware Benchmarking

3-Disk Short-Stroked RAID0, larger 7200RPM SATA

Greg Smith Bottom-Up Database Hardware Benchmarking

Read seeks/second - sysbench

THREADS=1
GB=10
MODE=rndrd
OPTIONS="--test=fileio --num-threads=$THREADS
--file-block-size=8K --file-test-mode=$MODE
--file-num=$GB --file-total-size=${GB}G
--file-fsync-freq=0 --file-fsync-end=no"
sysbench prepare $OPTIONS
sysbench run --max-time=60 $OPTIONS
sysbench cleanup $OPTIONS

Greg Smith Bottom-Up Database Hardware Benchmarking

Sample sysbench random read results

Read 78.125Mb Written 0b

Total transferred 78.125Mb (1.0059Mb/sec)
128.75 Requests/sec executed

I That’s 128.75 seeks/second over 10GB, resulting in a net

throughput of 128.75 * 8KB/s = 1.01MB/s
I Consider both the size of the disk used and the number of
clients doing seeks

Greg Smith Bottom-Up Database Hardware Benchmarking

More customizable seek tests

I bonnie++ experimental (currently at 1.96)

I iozone
I fio
I Windows: HD Tune does everything but commit rate

Greg Smith Bottom-Up Database Hardware Benchmarking

Sources for slow disk results

I Poor mapping to underlying hardware

I Buggy driver
I Insufficient bandwidth to storage
I Bottlenecking at CPU/memory limits
I Bad performing filesystem or filesystem misaligned with stripe
sizes
I Writes faster than reads? Probably low read-ahead settings
somewhere.
I Vibration: don’t shout at your JBODs! They don’t like it!
http://it.toolbox.com/blogs/database-soup/the-problem-with-iscsi-30602
http://blog.endpoint.com/2008/09/filesystem-io-what-we-presented.html
http://www.youtube.com/watch?v=tDacjrSCeq4

Greg Smith Bottom-Up Database Hardware Benchmarking

fsync tests

sysbench --test=fileio --file-fsync-freq=1 --file-num=1

--file-total-size=16384 --file-test-mode=rndwr run
| grep "Requests/sec"

I Database insert-only tests

Greg Smith Bottom-Up Database Hardware Benchmarking

Solid State Drives

I Writes are batched to block size by caching small ones

I There must be a write cache for good speed and to reduce
wear
I Look for the battery, capacitor, or super-capacitor to allow
flushing writes when power is lost
I Ask what happens when someone trips over the power cord
I Manufacturer doesn’t say? Assume your data is toast.

Greg Smith Bottom-Up Database Hardware Benchmarking

Good and bad drives

I Bad: Intel X25-M, X25-E, and most cheap consumer drives

I Good: OCZ Vertex 2 Pro, Intel 320 series
I Enterprise SSD models usually get this right, sometimes with
weird downsides
I Run diskchecker.pl and pull the plug yourself:
http://brad.livejournal.com/2116715.html
I Pull the plug on write-heavy database tests, too

Greg Smith Bottom-Up Database Hardware Benchmarking

Sample laptop disk specification

I ST9320423AS Momentus 7200.4 320GB

I 7200 RPM
I 16MB Cache
I Average seek: 11ms read/13ms write
I Average rotational latency: 4.17ms

Greg Smith Bottom-Up Database Hardware Benchmarking

Computed parameters

I Rotational latency = 1 / RPM / 60 / 2

I IOPS=1/(latency+seek)
I IOPS = 1/(((1/(RPM/60))/2) + S)
I IOPS = 1/(4.17ms + 11ms) = 65.9 IOPS

Greg Smith Bottom-Up Database Hardware Benchmarking

IOPS Calculators and Info

http://www.wmarow.com/strcalc/
http://www.dbasupport.com/oracle/ora10g/disk_IO_02.shtml
http://storageadvisors.adaptec.com/2007/03/20/sata-iops-measurement/

Greg Smith Bottom-Up Database Hardware Benchmarking

Sample disk results

Disk Seq Seq bonnie++ Read-only Commit Drive

Count Rd Wr seeks seeks Rate Model

1 71 58 232 @ 4GB 194 @ 4GB 105/s 7200.4

1 59 54 177 @ 16GB 56 @ 100GB 10212/s WD160

3 125 119 371 @ 16GB 60 @ 100GB 10855/s RAID0

1 254 147 3935@ 32GB 3417@100GB 5005/s SSD

I Commit rate for 7200.4 laptop drive is 1048/s with unsafe
volatile write cache
I Non-laptop spinning drives include a 256MB battery-backed
write cache, Linux SW RAID

Greg Smith Bottom-Up Database Hardware Benchmarking

Using sysbench for database comparisons

I Originally targeted MySQL database testing

I Use current source code from development repo:
https://launchpad.net/sysbench
I Now tests PostgreSQL fairly for OLTP read-only transactions
I Standard OLTP tests quietly fail on PostgreSQL due to
transaction isolation differences

Greg Smith Bottom-Up Database Hardware Benchmarking

sysbench compilation

apt-get install bzr

bzr checkout https://code.launchpad.net/\
sysbench-developers/sysbench/0.4
cd 0.4
./autogen.sh
./configure --with-pgsql
make

Greg Smith Bottom-Up Database Hardware Benchmarking

Server configuration for sysbench results

I Quad-Core Intel i870, 8 Hyper-Threaded Cores

I 16GB DDR3-1600 RAM
I Areca ARC-1210 SATA II PCI-e x8 RAID controller, 256MB
write cache
I DB: 3x640GB Western Digital SATA disks, short-stroked,
Linux software RAID-0
I WAL: 160GB Western Digital SATA disk
I Ubuntu 10.04, Linux Kernel 2.6.32-26-generic x86 64
I OS on separate disk
I XFS filesystem
I Default database configurations

Greg Smith Bottom-Up Database Hardware Benchmarking

sysbench with MySQL

echo "create database sysbench;" | mysql -h localhost -u

root
sysbench --mysql-user=root --db-driver=mysql
--mysql-table-engine=innodb --mysql-db=sysbench
--test=oltp prepare
sysbench ... --oltp-read-only=on --oltp-test-mode=simple
--init-rng --max-requests=0
--max-time=$TIME --num-threads=$THREADS run
sysbench ... cleanup

Greg Smith Bottom-Up Database Hardware Benchmarking

sysbench with PostgreSQL

sudo su - postgres -c "createdb sysbench"

sudo su - postgres -c "psql -c \"alter user postgres with
password ’password’;\""
sysbench --pgsql-user=postgres --pgsql-password=password
--pgsql-db=sysbench --pgsql-host=localhost
--db-driver=pgsql --test=oltp prepare
sysbench ... --oltp-read-only=on --oltp-test-mode=simple
--init-rng --max-requests=0
--max-time=$TIME --num-threads=$THREADS run
sysbench ... cleanup

Greg Smith Bottom-Up Database Hardware Benchmarking

sysbench read-only size scaling, 10,000 rows

Greg Smith Bottom-Up Database Hardware Benchmarking

Simple PostgreSQL Configuration - 2GB or more of RAM

I shared buffers = 512MB

I checkpoint segments = 32
I wal buffers = 16MB
I http://wiki.postgresql.org/wiki/Tuning_Your_
PostgreSQL_Server
I To set these parameters and the also important work mem:
I apt-get install pgtune
I https://github.com/gregs1104/pgtune

Greg Smith Bottom-Up Database Hardware Benchmarking

Serious database application tests

I Include read and write transactions

I Track latency as well as transactions/second throughput
I Note size of database relative to RAM
I Make sure load generator isn’t the bottleneck
I Only real way to test subtle tuning like I/O scheduling

Greg Smith Bottom-Up Database Hardware Benchmarking

What should you do?

I Trust no one
I Don’t start on application benchmarks until you’ve proven
basic performance
I Don’t spend too long on basic performance if you can switch
to application benchmarks
I Vendors alternate among lying, misunderstanding what you
want, and trying to make you feel dumb
I Use simple, standard tools whenever possible to minimize
vendor disputes
I Be prepared to translate to your vendor’s language and
subvert their agenda
I Never spend real money on hardware unless you can return it
if it sucks

Greg Smith Bottom-Up Database Hardware Benchmarking

For more information...

I Performance tuning of PostgreSQL 8.1 through 9.0, from

hardware to scaling via replication
I And lots of hardware, OS tuning, and monitoring
Greg Smith Bottom-Up Database Hardware Benchmarking
Questions?

I The BOFs await...

Greg Smith Bottom-Up Database Hardware Benchmarking