[rt-users] Please help - rt4 slow after upgrade

Chris Hall hiro24 at gmail.com
Thu May 19 10:33:00 EDT 2011


Well, resolved the below a bit by adding to RT_SiteConfig.pm:

Set($WebSessionClass, "Apache::Session::File");

However, CPU usage still peaks out on apache2 processes, although it appears
to be not AS frequent.  Anybody have any other suggestions on areas to look
into?  Some are saying the slowdown seems to occur most frequently when
updating tickets.

On Wed, May 18, 2011 at 8:36 PM, Chris Hall <hiro24 at gmail.com> wrote:

> I actually saw something about that in a message to the mailing list, and
> already turned it off as well, as a precaution.   After searching around
> some and trying out some tests, it seems the delay happens after the GET
> request to the apache server, and watching the logs, nothing appears in the
> log until about 20 seconds after the request to say the user asked for it..
> so there's about 20 - 30 seconds of dead time from the GET to when the log
> shows a request was made.  On a forum recommendation, I enabled slow mysql
> logging and noticed several entries in the log.  Here is a sample:
>
> # User at Host: rt_user[rt_user] @ localhost []
> # Query_time: 3246  Lock_time: 0  Rows_sent: 1  Rows_examined: 0
> SELECT GET_LOCK('Apache-Session-8a63b00ecb1829c65682fc75db1c05be', 3600);
> # User at Host: rt_user[rt_user] @ localhost []
> # Query_time: 3204  Lock_time: 0  Rows_sent: 1  Rows_examined: 0
> SELECT GET_LOCK('Apache-Session-8a63b00ecb1829c65682fc75db1c05be', 3600);
> # User at Host: rt_user[rt_user] @ localhost []
> # Query_time: 2897  Lock_time: 0  Rows_sent: 1  Rows_examined: 0
> SELECT GET_LOCK('Apache-Session-8a63b00ecb1829c65682fc75db1c05be', 3600);
> # User at Host: rt_user[rt_user] @ localhost []
> # Query_time: 2885  Lock_time: 0  Rows_sent: 1  Rows_examined: 0
> SELECT GET_LOCK('Apache-Session-8a63b00ecb1829c65682fc75db1c05be', 3600);
> # Time: 110518 19:51:37
> # User at Host: rt_user[rt_user] @ localhost []
> # Query_time: 1660  Lock_time: 0  Rows_sent: 1  Rows_examined: 0
> SELECT GET_LOCK('Apache-Session-8a63b00ecb1829c65682fc75db1c05be', 3600);
> # Time: 110518 19:56:44
> # User at Host: rt_user[rt_user] @ localhost []
> # Query_time: 14  Lock_time: 0  Rows_sent: 1  Rows_examined: 0
> SELECT GET_LOCK('Apache-Session-29fd7421a2e8981d99740258bbb9aea3', 3600);
>
>
> Dunno if they are of any help, but to me, it seems to indicate several
> SELECT queries that are running long for... whatever reason.  Do these logs
> make sense to anybody?
>
> On Wed, May 18, 2011 at 6:33 PM, Jonah Hirsch <jwh83 at nau.edu> wrote:
>
>> I was having this issue, same exact versions and problems with apache2. I
>> turned off mod_deflate and that helped a bit, but I'm still getting some
>> slowdowns, just not as much as before.
>> -----------------------
>>  Jonah Hirsch
>>
>>
>>
>> On Wed, May 18, 2011 at 11:31 AM, Chris Hall <hiro24 at gmail.com> wrote:
>>
>>> Not sure where to start here, but trying to diagnose why after an upgrade
>>> from 3.8.8 to 4.0.0 things have slowed down tremendously.  Doing even
>>> pedestrian tasks such as loading the home screen or opening the ticket
>>> search will about 30% of the time decide it wants to max out the cpu with
>>> the apache2 process for roughly 20 - 30 seconds.  Sometimes, loading things
>>> happens almost instantaneously.  Other times, there's this weird delay.  Not
>>> sure where to start looking, are there any known issues that I might need to
>>> be aware of?
>>
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.bestpractical.com/pipermail/rt-users/attachments/20110519/5d1ab067/attachment.htm>


More information about the rt-users mailing list