Redmine

I am also having the very same issue. This needs urgent attention

I am running Redmine on a Opteron 64-bit CPU with Ubuntu 8.0.4 LTS, libruby1.8 and ruby1.8, using Passenger and Redmine 0.8.4

The ruby process will run fine for about 30 minutes and then start to use 100% of the CPU and eventually overload the server completely!

Im also running a small Redmine app with (up until now only private projects). When i stop using the app for something like 1 hour and then come back the ruby-process produces an average cpu-load at most 100%.

the site isnt responding for about 1/2 minute (while it has 100% load) and when it finally responds it works like a charm with nearly no cpu load!

System:

- Debian
- apache + mod_passenger
- mysql
- ruby 1.8

I haven't seen any memory leaks with my Redmine and my version of passenger. Passenger recently released a bugfix for a memory leak that might be the problem. http://blog.phusion.nl/2009/06/21/phusion-passenger-224-released/

Eric

More information on this problem :
I've switched of source browsing for Anonymous since my first post. Since that day, I was forced to Kill Rail process only once (before, it was once a day!).

And I don't think that is a Passenger problem because I had exactly the same bug with Mongrel...

I have the same problem with Mongrel + Debian + MySQL and Redmine 0.8.1 but haven't managed to find out what the problem is until now. How can I track this better with RoR?

There seems to be race conditions on the writing and reading of logs... We are disabling the logs right now to see if it continues to fail.

Yannick Warnier wrote:

There seems to be race conditions on the writing and reading of logs... We are disabling the logs right now to see if it continues to fail.

Did you have any luck with that?

Sorry,
I didn't test because a real leak (another one...) of time...

Please just tell me how to disable logs.
I've tried to set "config.logger = nil" into my production.rb file, but without success.

Thanks

Alexander Pánek wrote:

Yannick Warnier wrote:

There seems to be race conditions on the writing and reading of logs... We are disabling the logs right now to see if it continues to fail.

Did you have any luck with that?

No, and actually we didn't find the problem. In the end we removed a series of options whitout which Redmine still runs fine, and it stopped happening. We have no internal experience in RoR debugging, which makes it very difficult for us to know exactly what we are looking for.

I still believe that recovering details from several large SVN or Hg repositories might have been the main reason, but we failed to prove it as well.

Yannick Warnier wrote:

Alexander Pánek wrote:

Yannick Warnier wrote:

There seems to be race conditions on the writing and reading of logs... We are disabling the logs right now to see if it continues to fail.

Did you have any luck with that?

No, and actually we didn't find the problem. In the end we removed a series of options whitout which Redmine still runs fine, and it stopped happening. We have no internal experience in RoR debugging, which makes it very difficult for us to know exactly what we are looking for.

What options did you change/remove, if I may ask?

I still believe that recovering details from several large SVN or Hg repositories might have been the main reason, but we failed to prove it as well.

Yea, that could be a reason.. did you have a cron job set up for that?

Anyways, thanks for the info!

Cédric RICARD wrote:

Please just tell me how to disable logs

I honestly don't know how to do that. I've never seen the need for that, nor do I really recommend it.

Same issue here...
It even now leads to a crash of mysql since memory is so out.

Here are my details:
Ruby version 1.8.7 (i486-linux)
RubyGems version 1.3.5
Rack version 1.0
Rails version 2.3.4
Active Record version 2.3.4
Active Resource version 2.3.4
Action Mailer version 2.3.4
Active Support version 2.3.4
Application root /var/www/qw-dev.net/www
Environment production
Database adapter mysql
Database schema version 20091108092559

I am running Redmine trunk on Apache + Passenger:
www-servers/apache-2.2.14-r1
www-apache/passenger-2.2.5
dev-db/mysql-5.0.84-r1

• * * * * root ruby /var/www/qw-dev.net/www/script/runner "Repository.fetch_changesets" -e production

(+ a reposman cron every 10 minutes to autocreate new repositories)

HELP please!!

Just migrated my redmine from CentOS 5.2 to Debian 5.0.5
On Debian Redmine's SVN module crashes everytime becouse of memory leaks.

On CentOS everything was very stable.
Any ideas?