mixedpuppy at gmail
Nov 6, 2009, 10:33 AM
Post #3 of 3
(338 views)
Permalink
|
On 05/11/09 7:59 AM, Scott McNab wrote:
> Hello.
> We are running several Trac projects with no problems, however, there is
> one particular ticket in one project that seems to hang Apache when
> viewed. The strange thing is, every other ticket in the same project is
> perfectly fine, its just one particular ticket that takes 9 minutes to load!
> Here is a capture of the Trac log at debug level when viewing this
> ticket (notice the time lapse between the last two lines):
> 2009-11-05 09:33:15,563 Trac[main] DEBUG: Dispatching <Request "GET
> u'/ticket/636'">
> 2009-11-05 09:33:15,750 Trac[svn_fs] DEBUG: Subversion bindings imported
> 2009-11-05 09:33:15,758 Trac[api] DEBUG: action controllers for ticket
> workflow: ['ConfigurableTicketWorkflow']
> 2009-11-05 09:33:16,307 Trac[default_workflow] DEBUG:
> render_ticket_action_control: action "leave"
> 2009-11-05 09:33:16,308 Trac[default_workflow] DEBUG:
> render_ticket_action_control: action "resolve"
> 2009-11-05 09:33:16,311 Trac[default_workflow] DEBUG:
> render_ticket_action_control: action "reassign"
> 2009-11-05 09:33:16,582 Trac[default_workflow] DEBUG:
> render_ticket_action_control: action "accept"
> 2009-11-05 09:33:16,585 Trac[session] DEBUG: Retrieving session for ID
> u'abcd'
> 2009-11-05 09:33:16,588 Trac[chrome] DEBUG: Prepare chrome data for request
> 2009-11-05 09:42:39,607 Trac[main] DEBUG: 587 unreachable objects found.
> The ticket itself is very small, with only a couple of lines of text as
> ticket content. I cannot see any reason why this ticket should take so
> long to load!
> Note: during the time the ticket is loading, Apache is consuming 100% CPU.
> Has anyone got any idea what might be going on here? We are using Trac
> 0.11.2.1
> Thanks
> Scott
First, I'd suggest upgrading to the latest trac and genshi and see if
the problem is resolved.
The last log entry is produced by the gc.collect call in main.py, which
affects every request. That takes some time, certainly NOT 9 minutes.
[I'm stating this by memory, been a while] I removed that call and
gained less than 1 second performance increase (but big enough to be
worth the removal, I think around 150ms)
The weird thing is you're saying this happens *only* on one specific
ticket? In that case I'd look at what the data is, perhaps try to
duplicate the ticket exactly into a new ticket and see if you have the
problem with the new ticket. Does the ticket data have any wiki macros,
unicode chars, etc., that might not be in any other tickets? Ah, you
say it's just a couple lines of simple text. weird.
The 100% cpu is normal, the bulk of that is genshi processing templates.
The trick is keeping the start-to-finish time of a request as low as
possible. Have you customized any templates? What plugins do you have
installed? If either is yes, remove them all, if that fixes is,
reinstate them one at a time until you find the culprit.
Shane
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "Trac Users" group.
To post to this group, send email to trac-users [at] googlegroups
To unsubscribe from this group, send email to trac-users+unsubscribe [at] googlegroups
For more options, visit this group at http://groups.google.com/group/trac-users?hl=en
-~----------~----~----~----~------~----~------~--~---
|
