Comments on: Returning to InnoDB scalability

By: Pool Table And Ping Pong

Pool Table And Ping Pong — Sun, 02 Mar 2008 09:57:11 +0000

Pool Table Lights Guide: How To Light Up Your Game

The proper pool table lights is critical to playing your best game of pool. The ideal pool table light fixture is attractive, unobtrusive, and provides just enough illumination without creating a glare. Choosing the right pool table lights for your hom…

By: Ich bin root

Ich bin root — Fri, 01 Feb 2008 18:42:57 +0000

MySQL-Käferjagd

Für einen neuen, dicken Datenbankserver (8 CPU-Kerne, 16GB RAM) habe ich eine passende MySQL-Version gesucht. Dabei bin ich auf einige - teilweise sehr beunruhigende - Bugs-Reports gestoßen.MySQL 5.0.32 CommunityEigentlich wollte ich das MySQL-Paket …

By: Vadim

Vadim — Fri, 29 Dec 2006 16:58:00 +0000

Alan,

We don’t expect that. This question is to MySQL developers, we are not.
afaik MySQL are not stress tested on HP-UX and other platforms, excluding Linux/Solaris.

By: alan

alan — Mon, 11 Dec 2006 04:00:59 +0000

I think it is already disappointing that mysql is only tuned for Linux, like the PTHREAD_MUTEX_ADAPTIVE_NP which only exists on Linux, but why don’t you just use pthread_mutex_trylock and pthread_mutex_lock ? how do you expect that HP-UX or others will perfom well without stress testing and improving it ?

By: peter

peter — Tue, 28 Nov 2006 09:06:19 +0000

Prem,

Please ask your question at out forum - forum.mysqlperformanceblog.com
Comments on unrelated topic is not great for discussion.

By: Prem

Prem — Tue, 28 Nov 2006 06:57:59 +0000

Peter,

Thanks for your immediate reply.

If the bottleneck is System (HP-UX), How to find out the system factors which affects the MySQL performance?

Please share your ideas….

Thanks in Advance,
Prem

By: peter

peter — Mon, 27 Nov 2006 09:07:55 +0000

Prem,

There could be some source differences but these are mostly to work-around missing or different functionality, most of things work same across all Unix platforms. All of the platforms scale differently however so I’m not surprized - even different Linux versions can have 20%+ performance difference. Solaris can perform quite differently from Linux etc.

By: Prem

Prem — Mon, 27 Nov 2006 07:05:31 +0000

Hi,

This is regarding MySQL5.0.20a performance difference between HP-UX and Linux. I ran Sysbench Read only test on both platform. For 4 thread, I got 280 transaction in Hp-UX and 710 transaction in RHEL. Also i ran Sysbench Read-Write OLTP test for 4 thread, i got 69 in HP-UX and 119 in RHEL.

Why this differences? Is this because of HP-UX specific codes in MySQL Source?

Thanks,
Prem

By: MySQL Performance Blog » Bug fix of InnoDB scalability problem

MySQL Performance Blog » Bug fix of InnoDB scalability problem — Tue, 14 Nov 2006 10:11:19 +0000

[...] I was pretty busy last month with project which will be annonced very soon (I hope), but I can’t miss bug fix of my favorite bug 15815. I wrote about this problem before and also investigated in my presentation. Finally bug fix was pushed into 5.0-bk tree and now I have it in my hands. [...]

By: peter

peter — Fri, 20 Oct 2006 15:28:29 +0000

Mikhail,

I would even add it has to be approved by Innobase/Oracle as it is their code.
This has slight chance as, as I know they work on their own version of the patch which will be included in some later MySQL release…