Resque workers not working on latest master branch

Created by: netdata-be

When starting the reque worker using the rake task nothing happens except this:

~# bundle exec rake environment resque:work QUEUE=post_receive,mailer,system_hook RAILS_ENV=production PIDFILE=tmp/pids/resque_worker.pid
Starting worker git1-0:7710:post_receive,mailer,system_hook
Registered signals
Running before_first_fork hooks with [#<Worker git1-0:7710:post_receive,mailer,system_hook>]
#~

There is no process running.

PS: I have already following the troubleshoot guide and it has nothing todo with it

Created by: hjanuschka

+1

By Administrator on 2013-01-09T06:59:59 (imported from GitLab project)

By Administrator on 2013-01-09T06:59:59 (imported from GitLab)

Created by: dzaporozhets

PID option is removed from resque version we use for current master.

Use

nohup bundle exec rake environment resque:work QUEUE=post_receive,mailer,system_hook RAILS_ENV=production  > /dev/null  2>&1 &

By Administrator on 2013-01-09T08:02:39 (imported from GitLab project)

By Administrator on 2013-01-09T08:02:39 (imported from GitLab)

Created by: hjanuschka

thx for reply :)

here is my output

gitlab@gitlab:/gitlab$ bundle exec rake environment resque:work QUEUE=post_receive,mailer,system_hook RAILS_ENV=production Connecting to database specified by database.yml Starting worker gitlab:19159:post_receive,mailer,system_hook Registered signals Running before_first_fork hooks with [#<Worker gitlab:19159:post_receive,mailer,system_hook>] gitlab@gitlab:/gitlab$

does not seem to work - jobs stay in queue - any idea?

By Administrator on 2013-01-09T08:09:39 (imported from GitLab project)

By Administrator on 2013-01-09T08:09:39 (imported from GitLab)

Created by: netdata-be

Actually I have figured this out yesterday. It seems that our Redis database version is to old, (which is the default redis installed from the repo on debian squeeze 6.0 ). It seems there is a catched error which does not get handled and swallowed.

After putting A LOT of debug statements I could figure this out.

So try to upgrade your redis instance

By Administrator on 2013-01-09T08:28:44 (imported from GitLab project)

By Administrator on 2013-01-09T08:28:44 (imported from GitLab)

Created by: hjanuschka

ok thx for the hint got it fixed

as root - on stable debian

/etc/init.d/redis-server stop
cd /tmp
wget http://redis.googlecode.com/files/redis-2.2.4.tar.gz
cd redis-2.2.4
make PREFIX=/usr
make install PREFIX=/usr
/etc/init.d/redis-server start

edit: /etc/redis/redis.conf and remove the last 2 lines shareobjects and shareobjectspoolsize

edit the init.d script and remove PIDFILE= from resque command after the start command of resque add the following

ps auxww|grep QUEUE=post|head -n1|awk '{print $2}' > $RESQUE_PID

so the pidfile is created - this fixes the "stop" error

stop/start gitlab - push and your fine

By Administrator on 2013-01-09T09:10:20 (imported from GitLab project)

By Administrator on 2013-01-09T09:10:20 (imported from GitLab)

Created by: netdata-be

@randx I suggest this must be documented somewhere, do you have a good place for this then i can create a pull request to update the documentation

By Administrator on 2013-01-09T19:43:57 (imported from GitLab project)

By Administrator on 2013-01-09T19:43:57 (imported from GitLab)

Created by: ccoenen

[edited] Sorry, found the backport repo mentioned in https://github.com/gitlabhq/gitlab-public-wiki/wiki/Trouble-Shooting-Guide (sidekiq section). Upgrading People with Sidekiq-Problems: look at that troubleshooting-guide.

By Administrator on 2013-01-24T16:40:14 (imported from GitLab project)

By Administrator on 2013-01-24T16:40:14 (imported from GitLab)

Created by: riyad

Closing this as "can't reproduce" since 4.1 has replaced Resque with Sidekiq.

By Administrator on 2013-01-24T18:16:41 (imported from GitLab project)

By Administrator on 2013-01-24T18:16:41 (imported from GitLab)