Worker processes timing out since upgrade to 3.1 (#2089) · Issues · gpt / large_projects / gitlabhq1

Worker processes timing out since upgrade to 3.1

Created by: AngeloStavrow

I upgraded (per the instructions) from 3.0 to 3.1 but I'm having issues getting Gitlab to respond.

Salient details: Gitlab is running on Ubuntu 12.04 server on a Dell Optiplex GX240 w/ 512MB RAM, on an intranet (not accessed over the Internet). App status passes with flying colours.

From unicorn.stderror.log, I keep seeing:

E, [2012-11-27T10:53:06.356483 #3935] ERROR -- : worker=0 PID:3945 timeout (61s > 60s), killing
E, [2012-11-27T10:53:06.401877 #3935] ERROR -- : reaped #<Process::Status: pid 3945 SIGKILL (signal 9)> worker=0
I, [2012-11-27T10:53:07.012398 #4738]  INFO -- : worker=0 ready

This is resulting in 502 Bad Gateway errors when I try to access the web app.

I have timeout 60 (the default) in unicorn.rb.

Any thoughts?

UPDATE: Another thing I've never seen before -- when I run $ sudo service gitlab start I get a message: nohup: redirecting stderr to stdout. No idea why.