Is the SMP fix still relevant now that 5.1.28 is out?

[Matt]

Hi...

From what I read, there are several changes that have landed in MySQL
5.1.28 that address the same issues that your patches address. I'm
mostly interested in the SMP fix - is that patch still relevant?

Cheers,

Matt

[MARK CALLAGHAN]

What changes in 5.1.28 are you talking about?

>
> >
>

--
Mark Callaghan
mdca...@gmail.com

[Matt]

On Sep 17, 3:03 am, "MARK CALLAGHAN" <mdcal...@gmail.com> wrote:

> What changes in 5.1.28 are you talking about?

This one:

http://bugs.mysql.com/bug.php?id=34409

and this one:

http://bugs.mysql.com/bug.php?id=38185

They are both related to InnoDB, and my impression was that most of
the performance enhancements in the Google patches, specifically when
talking about many-CPU systems, were dealing with these same issues.
I must confess that I don't possess the knowledge to actually look at
the patches and figure out whether I'm right or not, I'm only looking
at what has been written on blogs etc.

I'm not talking about any of the patches that Google has release
relating to added/changed functionality, I'm only talking about
performance.

My company is about to buy an 8-CPU server, and I am trying to make
sense of all of the various versions of MySQL available (5.0, 5.1,
Percona, Google, etc), and which will best serve our needs.

Thanks,

Matt

[MARK CALLAGHAN]

On Tue, Sep 16, 2008 at 4:21 PM, Matt <matt.p...@gmail.com> wrote:
>
> On Sep 17, 3:03 am, "MARK CALLAGHAN" <mdcal...@gmail.com> wrote:
>
>> What changes in 5.1.28 are you talking about?
>
> This one:
>
> http://bugs.mysql.com/bug.php?id=34409

This fixes a bug (334409) introduced in MySQL 5.1.24

>
> and this one:
>
> http://bugs.mysql.com/bug.php?id=38185
>
> They are both related to InnoDB, and my impression was that most of
> the performance enhancements in the Google patches, specifically when
> talking about many-CPU systems, were dealing with these same issues.
> I must confess that I don't possess the knowledge to actually look at
> the patches and figure out whether I'm right or not, I'm only looking
> at what has been written on blogs etc.

I was confused at first from the announcement for the 5.1.28 patch. I
thought perhaps a lot more was fixed.

>
> I'm not talking about any of the patches that Google has release
> relating to added/changed functionality, I'm only talking about
> performance.
>
> My company is about to buy an 8-CPU server, and I am trying to make
> sense of all of the various versions of MySQL available (5.0, 5.1,
> Percona, Google, etc), and which will best serve our needs.

These bug fixes are nice to have but don't fix the SMP problems. The
Google SMP fix patch also doesn't fix all of the SMP problems but
fixes much more of the problem than the bug fixes in 5.1.28. The only
alternative to the Google SMP fix patch at this time is the Percona
branch and I highly recommend it. Frankly, I think that everyone
should use the Percona branch rather than the SMP fix directly and
then we can lobby Percona to include the appropriate bits and pieces
from the Google patch into their branch.

>
> Thanks,
>
> Matt
>
> >
>

--
Mark Callaghan
mdca...@gmail.com

[Matt]

On Sep 17, 12:33 pm, "MARK CALLAGHAN" <mdcal...@gmail.com> wrote:

> These bug fixes are nice to have but don't fix the SMP problems. The
> Google SMP fix patch also doesn't fix all of the SMP problems but
> fixes much more of the problem than the bug fixes in 5.1.28. The only
> alternative to the Google SMP fix patch at this time is the Percona
> branch and I highly recommend it. Frankly, I think that everyone
> should use the Percona branch rather than the SMP fix directly and
> then we can lobby Percona to include the appropriate bits and pieces
> from the Google patch into their branch.

Cool, thanks so much for that - it's good to hear.

Cheers,

Matt