Comments on: MySQL Performance on Memory Appliance

By: pave

pave — Sun, 15 Feb 2009 16:39:54 +0000

sucks for device i guess

But i’m just showing what the difference is between DELL’s md3000 with 15K rpms HDs with raid 10 VS. ram based solution

By: peter

peter — Sun, 15 Feb 2009 06:25:51 +0000

I see. 1394.15 Sysbench transactions per second is not a lot for device with 600k IOs/sec capacity.

By: Pavel

Pavel — Fri, 13 Feb 2009 20:16:37 +0000

if u’re talking about IOPS – that would be 600k

By: peter

peter — Thu, 12 Feb 2009 20:01:22 +0000

Pavel,

What is about raw IO capacity ?

Also in your test I see you’re using just 1M table size – if Innodb is well configured this should be in memory workload.

By: Pavel

Pavel — Wed, 11 Feb 2009 22:19:16 +0000

I’m sorry Kevin, but unfortunately I cannot disclose this info since we do not own the device and I would have to ask permission from the manufacturer :-/

By: Kevin Burton

Kevin Burton — Wed, 11 Feb 2009 21:41:10 +0000

Hey Pavel.

Thanks for the numbers.

Can you say what memory device it is?

By: Pavel

Pavel — Wed, 11 Feb 2009 20:44:04 +0000

and the same test on MD3000 15K RPM HDs RAID 10

sysbench –test=oltp –mysql-table-engine=innodb –oltp-table-size=1000000 –mysql-socket=/tmp/mysql.sock –mysql-db=test –mysql-user=root –mysql-password= prepare
sysbench –num-threads=16 –max-requests=100000 –test=oltp –mysql-table-engine=innodb –oltp-table-size=1000000 –mysql-socket=/tmp/mysql.sock –mysql-db=test –mysql-user=root –mysql-password= run

Running the test with following options:
Number of threads: 16

Doing OLTP test.
Running mixed OLTP test
Using Special distribution (12 iterations, 1 pct of values are returned in 75 pct cases)
Using “BEGIN” for starting transactions
Using auto_inc on the id column
Maximum number of requests for OLTP test is limited to 100000
Threads started!
Done.

OLTP test statistics:
queries performed:
read: 1400014
write: 500005
other: 200002
total: 2100021
transactions: 100001 (332.53 per sec.)
deadlocks: 0 (0.00 per sec.)
read/write requests: 1900019 (6318.06 per sec.)
other operations: 200002 (665.06 per sec.)

Test execution summary:
total time: 300.7284s
total number of events: 100001
total time taken by event execution: 4809.0591
per-request statistics:
min: 0.0035s
avg: 0.0481s
max: 1.0817s
approx. 95 percentile: 0.1054s

Threads fairness:
events (avg/stddev): 6250.0625/90.30
execution time (avg/stddev): 300.5662/0.01

SEE THE DIFFERENCE ???

By: Pavel

Pavel — Wed, 11 Feb 2009 20:42:39 +0000

Here is OLTP test run on another memory storage solution:

sysbench –test=oltp –oltp-table-size=1000000 –db-driver=mysql –mysql-table-engine=innodb –mysql-socket=/var/lib/mysql/mysql.sock –mysql-user=root prepare

sysbench –num-threads=16 –max-requests=100000 –test=oltp –mysql-table-engine=innodb –db-driver=mysql –mysql-socket=/var/lib/mysql/mysql.sock –mysql-user=root run

OLTP test statistics:
queries performed:
read: 1400224
write: 500046
other: 200018
total: 2100288
transactions: 100002 (1394.15 per sec.)
deadlocks: 14 (0.20 per sec.)
read/write requests: 1900270 (26491.99 per sec.)
other operations: 200018 (2788.49 per sec.)

Test execution summary:
total time: 71.7300s
total number of events: 100002
total time taken by event execution: 1145.6317
per-request statistics:
min: 0.0027s
avg: 0.0115s
max: 0.0979s
approx. 95 percentile: 0.0313s

Threads fairness:
events (avg/stddev): 6250.1250/66.05
execution time (avg/stddev): 71.6020/0.01

By: francesco

francesco — Sat, 27 Sep 2008 07:34:17 +0000

I am very interested in MySQL on Memory Appiance. We have a database with 50 millions of records (emails), and 50 columns (100byte foreach). We need to be fast (50k-100k request/sec; 20k update/sec). With mySql and a memory applance is it possible?

By: Chris Nigh

Chris Nigh — Thu, 19 Jun 2008 22:43:19 +0000

5.0.51a
sorry..