[rt-users] Searching on requestor in large databases

[Mat]

The main things changed in the my.cnf file (mysqls config file) are

set-variable    = key_buffer=128M
set-variable    = max_allowed_packet=1M
set-variable    = table_cache=256
set-variable    = sort_buffer=1M
set-variable    = record_buffer=1M
set-variable    = net_buffer_length=8K
set-variable    = myisam_sort_buffer_size=64M

originally I have key_buffer at something like 32mb, which meant mysql had
to continuelly copy the query results out to disk before it could get more
results. Massively slowing things down.

I think we could also play with the locking settings to speed up RT majorly,
as one of the big speed issues we are having now is queries waiting for
locks to be released before running.. But I'd rather not have to do that.

Regarding apache conf, I think that the issue with apache getting really big
is due to a bug in either rt2 or apache/mod_perl (or both) when tickets are
moved to a queue the user cant read, but thats in a different thread :)

Mat.

> -----Original Message-----
> From: Blair Zajac [mailto:blair at orcaware.com]
> Sent: Tuesday, 8 January 2002 6:40 PM
> To: Matthew Watson
> Cc: rt-users at lists.fsck.com
> Subject: Re: [rt-users] Searching on requestor in large databases
>
>
> Mat,
>
> Could you tell us what was poorly set up and what changes were made to
> have it work well?
>
> Best,
> Blair
>
> Matthew Watson wrote:
> >
> > Update on this..
> >
> >  After doing various things with indexes, I got our dba to have
> a looky at
> > it, and smartly he checked out the mysql cnf.. and rightly so,
> it was setup
> > pretty bad. After tweeking my.cnf and apache/mod_perl we have
> this system
> > running much better. Although it still barks a bit sometimes when a big
> > query has a lock on a table (so heaps of other queries bottle
> up behind it)
> >
> > Mat.
>
> --
> Blair Zajac <blair at orcaware.com>
> Web and OS performance plots - http://www.orcaware.com/orca/