#8145 closed enhancement (worksforme)
Memory leaks gone
Reported by: | Owned by: | ||
---|---|---|---|
Priority: | normal | Milestone: | |
Component: | general | Version: | 0.11.4rc1 |
Severity: | normal | Keywords: | memory |
Cc: | Branch: | ||
Release Notes: | |||
API Changes: | |||
Internal Changes: |
Description
Just wanted to show you the memory usage graph of my server to demonstrate the much better memory management in trac 0.11.4rc1.
I got rid of trac 0.10.3 the 19th. Before I had a cron to restart apache every day.
Nothing else changed on the server since the 19th.
Current configuration :
Trac: 0.11.4rc1 Python: 2.4.4 (#1, Oct 22 2008, 19:49:52) [GCC 4.1.2 20061115 (prerelease) (Debian 4.1.1-21)] setuptools: 0.6c9 SQLite: 3.3.8 pysqlite: 2.3.2 Genshi: 0.5.1 mod_python: 3.2.10 Pygments: 1.0 Subversion: 1.4.2 (r22196) jQuery: 1.2.6
Thanks for this improvement ;)
Attachments (0)
Change History (5)
comment:1 by , 16 years ago
Version: | none → 0.11.4rc1 |
---|
comment:2 by , 16 years ago
follow-up: 4 comment:3 by , 16 years ago
I really meant trac 0.10.3 (debian etch package).
The usage of the server is the same, idle :P
I have three different python version on my server. How can I tell to mod python which python interpretor I want it to use. I assume it's a directive in the virtualhost configuration.
For the others I'm stuck with debian etch packages.
comment:4 by , 16 years ago
Resolution: | → worksforme |
---|---|
Status: | new → closed |
Replying to Sylvain Rabot <sylvain@…>:
I really meant trac 0.10.3 (debian etch package).
Great! That's must be a world premiere then ;-) 0.11 used to be much more memory hungry that 0.10.x used to be, and all the reports so far have been more of the "why is 0.11 now eating all my memory?" kind…
I have three different python version on my server. How can I tell to mod python which python interpretor I want it to use. I assume it's a directive in the virtualhost configuration.
Well, you can't, mod_python is a C extension which is compiled for a specific Python version, therefore it's linked to one specific Python library (and attempting to cheat usually results in a crash ;-) ).
Thanks for the feedback.
comment:5 by , 15 years ago
Kind of off topic….but what program did you use to get the graph of memory consumption. Right now I have to use 'top' Thanks, Lucas
Replying to sylvain@…:
Thanks! Any good news on this front is welcomed ;-)
I assume you meant 0.11.3?
Same kind of usage? Did you investigate the access logs?
If those are the same, I'd be really interested in knowing which change between 0.11.3 and 0.11.4rc1 could have improved the memory leak that much. Unless you're really talking of 0.10.3 of course ;-)
Here are a few additional improvement suggestions:
2.6 should have better memory management for unicode strings
3.3.1 preferred
1.5.x preferred (known mem leak in versions < 1.5)