We've got a few internal tracs running and the performance has generally been pretty bad. Today the cheif ninja's here managed to track it down and page loads have gone from frequently 15+ seconds to a second or two.
We're running Apache 2 w/ Python 2.5, Trac 0.12multirepos-r7684 under mod_wsgi. Apache is doing basic auth with Kerberos against a windows AD Domain. The database is Postgresql. Using firebug it looked like a lot of time was spent loading CSS & JS files. In an effort to improve performance we had the /chrome location being handled outside of the wsgi process, directly by apache. However this really didn't help. What did help was a bit more tricky, but the difference is night and day. We use ZXTM (http://www.zeus.com/products/zxtm/) pretty heavily for load balancing, caching, etc. For demo purposes let's say one of the Tracs is hosted at intranet.example.com. We had zxtm reroute intranet.example.com/chrome to tracchrome.example.com/chrome - which is an entirely separate apache vhost with no authentication. This means that both zxtm and browsers will happily cache the heck out of it. The performance difference is incredible. Now this may not be a big win for other users - I think it's a particular combination of Kerberos + Basic Auth + probably something else. However if we can help just one other organization I thought I'd pass this tip along. --~--~---------~--~----~------------~-------~--~----~ 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@googlegroups.com To unsubscribe from this group, send email to trac-users+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/trac-users?hl=en -~----------~----~----~----~------~----~------~--~---
