'flush tables with read lock' causes mysql server to deadlock

[Liam Bedford]

Binary package hint: mysql-server-5.0

I'm working on a backup solution for systems that requires running "flush tables with read lock" followed by an lvm snapshot. In testing, I've had multiple instances where the flush tables will hang with no particular solution short of killing mysqld.
The distro currently includes 5.0.22, and lo and behold in the 5.0.23 changelog fixes:
"A compatibility issue with NPTL (Native POSIX Thread Library) on Linux could result in a deadlock with FLUSH TABLES WITH READ LOCK under some conditions. (Bug#20048)"

Can the patch be backported/fixed in 5.0.22 please?

[Mathias Gug (mathiaz)]

Thanks taking the time to report this bug and helping to make Ubuntu better. However, I am closing it because the bug has been fixed in the latest development version of Ubuntu - the Gutsy Gibbon.

If you need a fix for the bug in previous versions of Ubuntu, please follow the instructions for "How to request new packages" at [WWW] https://help.ubuntu.com/community/UbuntuBackports#request-new-packages .

[Scott Kitterman (kitterman)]

Please don't suggest Backports for bug fixes. Backports are for new
features. Since *-backports are not enabled by default, bugs needing to be
in released versions should be addressed through the SRU process. That way
all Ubuntu users can benifit.

[Liam Bedford (lbedford-deactivatedaccount)]

So where is the LTS for dapper? or is that only for some security updates? 5.0.22 of mysql is going to be pretty obsolete by the
time dapper has finished being supported in 2011. Or do we only get support if we pay for it or are willing to use backports?

[Mathias Gug (mathiaz)]

I've reopen the bug for dapper.

This bug may qualify for a SRU, which is described here: https://wiki.ubuntu.com/StableReleaseUpdates.

[Liam Bedford (lbedford-deactivatedaccount)]

okay, so where do things go from here? do you have to make the changes for it to be considered for SRU, or do I? it looks
like whoever is producing the gutsy one is?

[Scott Kitterman (kitterman)]

Here is the svn commit that fixed the problem:

http://lists.mysql.com/commits/6835

Someon who is more familiar that me (e.g. more than not at all) with the mysql code base should look this over and pull the relevant bits out for this bug (the commit fixed two bugs). It should be easy enough to produce an updated package from there.

[Soren Hansen (soren)]

On Tue, Jul 17, 2007 at 01:34:48PM -0000, Scott Kitterman wrote:
> Here is the svn commit that fixed the problem:
>
> http://lists.mysql.com/commits/6835
>
> Someon who is more familiar that me (e.g. more than not at all) with the
> mysql code base should look this over and pull the relevant bits out for
> this bug (the commit fixed two bugs). It should be easy enough to
> produce an updated package from there.

Just glancing at the patch, it seems that the two bugs were essentially
duplicates, so fixing one without fixing the other would have been
difficult. :) The patch applies just fine to the Dapper. I'll run a few
quick tests.

--
Soren Hansen
Ubuntu Server Team
http://www.ubuntu.com/

[Soren Hansen (soren)]

Can someone provide an easy way to reproduce this bug? I took a cleanly
installed mysql-server-5.0 instance, created a database, created a table
and dumped a stack of data into it, ran 'FLUSH TABLES WITH READ LOCK',
but no luck.

--
Soren Hansen
Ubuntu Server Team
http://www.ubuntu.com/

[Martin Pitt (pitti)]

Needs reproduction recipe, so it's not 'triaged'.

[Launchpad Janitor (janitor)]

[Expired for mysql-dfsg-5.0 (Ubuntu Dapper) because there has been no activity for 60 days.]

[Mathias Gug (mathiaz)]

Reopening

[JC Hulce (soaringsky)]

Thank you for taking the time to report this bug and helping to make Ubuntu better. The version of Ubuntu you're reporting this issue on is in End of Life status, and newer versions have fixed this issue. You can learn more about this at https://wiki.ubuntu.com/Releases