Re: document for mysql performance improvement

2010-09-23 Thread Johnny Withers
If this is a dedicated MySQL box, i would increase the InnoDB buffer pool setting in my.cnf to at least 8GB, 12 would be even better (you did say you have 16GB of ram in the machine I believe). Also, what is the output of: show status like '%tmp%'; JW On Wed, Sep 22, 2010 at 8:01 PM, Vokern wr

Re: document for mysql performance improvement

2010-09-22 Thread Vokern
2010/9/23 Johnny Withers > > Can you show us the output of: show status like '%innodb%' > JW > Sure. mysql> show status like '%innodb%'; +---++ | Variable_name | Value | +---+---

Re: document for mysql performance improvement

2010-09-22 Thread Johnny Withers
Can you show us the output of: show status like '%innodb%' JW On Tue, Sep 21, 2010 at 10:11 PM, vokern wrote: > And this is the innodb file size, does this matter for performance? > > $ du -h ibdata* > 11G ibdata1 > 11G ibdata2 > 11G ibdata3 > 59G ibdata4 > > > > 2010/9/22 voke

Re: document for mysql performance improvement

2010-09-22 Thread Baron Schwartz
Hi, > If its an IO problem the first and easiest thing to do is (probably) look at > your disk subsystem. You can easily achieve higher disk IO by increasing the > number of disks and implementing something like RAID1+0. Or you can be logical about it and try to determine whether the IO performan

Re: document for mysql performance improvement

2010-09-21 Thread vokern
And this is the innodb file size, does this matter for performance? $ du -h ibdata* 11G ibdata1 11G ibdata2 11G ibdata3 59G ibdata4 2010/9/22 vokern : > This is piece of the setting in my.cnf: > > set-variable = innodb_buffer_pool_size=4G > set-variable = innodb_additional_mem_p

Re: document for mysql performance improvement

2010-09-21 Thread vokern
This is piece of the setting in my.cnf: set-variable = innodb_buffer_pool_size=4G set-variable = innodb_additional_mem_pool_size=20M set-variable = innodb_flush_log_at_trx_commit=2 set-variable = innodb_data_file_path=ibdata1:10G;ibdata2:10G;ibdata3:10G;ibdata4:10G:autoextend key_buffer

Re: document for mysql performance improvement

2010-09-21 Thread petya
Hi, Check your slow queries first. Large full scans can cause unwanted disk io. Do you use MyISAM or InnoDB? From your status, you seem to have intensive MyISAM locking. Peter On 09/21/2010 04:10 PM, Johan De Meersman wrote: Also, mailing list doesn't want to distribute attachments :-) Here

Re: document for mysql performance improvement

2010-09-21 Thread Johan De Meersman
Also, mailing list doesn't want to distribute attachments :-) Here's a link to the metrics view I was on about earlier: http://www.tuxera.be/mysqlstats.zip On Tue, Sep 21, 2010 at 4:08 PM, Johan De Meersman wrote: > > > On Tue, Sep 21, 2010 at 4:03 PM, wrote: > >> Quoting Johan De Meersman : >>

Re: document for mysql performance improvement

2010-09-21 Thread Johan De Meersman
On Tue, Sep 21, 2010 at 4:03 PM, wrote: > Quoting Johan De Meersman : > >> Your raid controller is "lying" to you - you can't have RAID10 with just >> two >> disks :-p Don't worry about that, though - it's a good enough config. >> > > Good enough? If he is genuinely saturating the disk with IO (a

Re: document for mysql performance improvement

2010-09-21 Thread a . smith
Quoting Johan De Meersman : Your raid controller is "lying" to you - you can't have RAID10 with just two disks :-p Don't worry about that, though - it's a good enough config. Good enough? If he is genuinely saturating the disk with IO (as he states the problem is IO) then it isnt good enoug

Re: document for mysql performance improvement

2010-09-21 Thread Bruce Ferrell
Mysql tuner is a very useful tool to pull metrics http://blog.mysqltuner.com/ On 09/21/2010 05:48 AM, Jangita wrote: > I find this quite good > > http://www.mysqlperformanceblog.com/ > > Send your my.cnf and maybe we could look at it and pick anything that would > help. > > > Jangita | +254 76

Re: document for mysql performance improvement

2010-09-21 Thread Giles Coochey
>>> >>> The disk is exactly Raid10. >>> The CPU is two 2.5G*4, totally 16G memory. >>> > > Two disks with SAS driver, 15K rpm. > RAID-10 with 2 disks? or do you mean RAID0 or RAID1??? Can't see how you would get RAID10, minimum of 6 disks for that, no? -- MySQL General Mailing List For list ar

Re: document for mysql performance improvement

2010-09-21 Thread a . smith
Quoting vokern : Two disks with SAS driver, 15K rpm. Ok so you have fast disks, but with only 2 disks it is normal you will be quite restricted by DISK IO. By adding more in multiples of 2 and stripping across all you achieve RAID1+0 and higher max IO... Andy. -- MySQL General Mail

Re: document for mysql performance improvement

2010-09-21 Thread vokern
2010/9/21 : > Quoting vokern : > >> >> The disk is exactly Raid10. >> The CPU is two 2.5G*4, totally 16G memory. >> > > And how many disks do you have, and what type (SATA/SAS/FC etc) what RPM? To > improve IO you can add more disks, or upgrade to faster disks. > > Two disks with SAS driver, 15K

Re: document for mysql performance improvement

2010-09-21 Thread vokern
Thank you all for the kind helps. I will check them and if still have problems I will come back. 2010/9/21 Machiel Richards : > > http://www.mysqlperformanceblog.com/2006/09/29/what-to-tune-in-mysql-server-after-installation/ > > > http://www.mysql.com/why-mysql/performance/ > > http://www.debian

Re: document for mysql performance improvement

2010-09-21 Thread a . smith
Quoting vokern : The disk is exactly Raid10. The CPU is two 2.5G*4, totally 16G memory. And how many disks do you have, and what type (SATA/SAS/FC etc) what RPM? To improve IO you can add more disks, or upgrade to faster disks. -- MySQL General Mailing List For list archives: http://l

RE: document for mysql performance improvement

2010-09-21 Thread Jangita
I find this quite good http://www.mysqlperformanceblog.com/ Send your my.cnf and maybe we could look at it and pick anything that would help. Jangita | +254 76 918383 | MSN & Y!: jang...@yahoo.com Skype: jangita | GTalk: jangita.nyag...@gmail.com -Original Message- From: vokern [mai

Re: document for mysql performance improvement

2010-09-21 Thread vokern
2010/9/21 : > If its an IO problem the first and easiest thing to do is (probably) look at > your disk subsystem. You can easily achieve higher disk IO by increasing the > number of disks and implementing something like RAID1+0. What is your > current disk configuration? > The disk is exactly Rai

Re: document for mysql performance improvement

2010-09-21 Thread a . smith
If its an IO problem the first and easiest thing to do is (probably) look at your disk subsystem. You can easily achieve higher disk IO by increasing the number of disks and implementing something like RAID1+0. What is your current disk configuration? Andy. -- MySQL General Mailing Li

Re: document for mysql performance improvement

2010-09-21 Thread Johan De Meersman
Yep. There's rather extensive documentation on http://www.mysql.com. You'll need to read it and compare to the metrics you're taking off your own server, draw conclusions and apply them to your setup. You *are* pulling metrics, aren't you, and not hoping for some magic wand to make it all happen ?

Re: document for mysql performance improvement

2010-09-21 Thread vokern
Yes we have changed some arguments in my.cnf like key_buffer_size, sort_buffer_size etc. 2010/9/21 Machiel Richards : > Good day > >   There is quite a lot of documentation available for MySQL performance > management. > >     However, a quick question on this matter. Have you changed any