ourdelta-developers team mailing list archive

[Neil Katin <launchpad@xxxxxxxxxxx>]

> So Neil, if you can bear with us, we're prepping the 5.0.77 build
> right now with updated/new patches. If that version still has the
> problem, then at least we'll be able to toggle the patch that we
> suspect might be causing the problem, allowing you to catch the issue.
> Oki?

Thanks for the fast response.  Here's what we plan to do:

We had already planned on restarting the database tonight so we could
"take the downtime" when we wanted, instead of having the db lock up.
At that time we will turn on the --skip-name-resolve option and see
if that works-around the problem.  If so, then this is probably the
truncated hostname issue in client-statistics.  Hopefully that will
give you the feedback you need even sooner than the 5.0.77
release.

Sound OK to you?

-- 
Memory leak in mysqld?
https://bugs.launchpad.net/bugs/344447
You received this bug notification because you are a member of OurDelta-
developers, which is the registrant for OurDelta.

Status in OurDelta - Builds for MySQL: New

Bug description:

We recently upgraded to the ourdelta patch set (Centos5.2, x86_64, 5.0.67.d7-44.el5_2).  We were running the provenscaling version mysql).

Ever since the upgrade we seem to be having a memory leak in mysqld; the executable is growing by about 3gb a day.  Here's an rrd graph of our memory usage; you can see the "green" (apps) portion of memory growing right after we performed the upgrade (at the end of week 8).

http://www.sonic.net/~nk31/memory-month.png

The application didn't change at that time; we are running the exact same SQL queries and workload before and after the update.

I realize that I don't have enough info to allow you to fix any problems in this but report (but if you can think of something we could measure for you, I will).  But I did want to see if anyone else has seen any leakage issues in this version of mysqld.

Thanks.