Hi,

I am using VM for gitlab 12.1. This morning Users started facing issue communicating with gitlab VM. I had a look at the resource consumption which showed memory was fully consumed. I switched off the VM and increased memory to 4GB from 512MB. After the restart i noticed gitlab and associated services are not able to start. I have tried manually >service gitlab start and can see processes starting but within few seconds all processes get killed. How can i start gitlab again? 

I do have a backup on Amazon and have successfully restored that into another VM. Gitlab worked for sometime and then again 4GB memory got consumed. I tried restarting 2nd VM too but end result was same. 

Any pointers would be helpful. 

Replies (5)

Probably worth looking at the logs

Reply 2

You'll find all the logs in /var/log and see what you can see. IIRC GitLab is a (Ruby-on-)Rails app which I'd expect would have it's own log (although not 100% sure what it'd be called and don't currently have an instance handy). Plus all the usual logs like daemon.log etc. Also sometimes when starting the service fails there will be a message that may give a clue (or at least something to google...)

Fixed:

Reply 3

A space was missing between parameter and value in gitlab.yml

Since then its stable

 

Well spotted...thanks for the update :-)

Reply 4

Curious as to how that would affect performance erratically...you'd figure it would be all or nothing.

You may consider setting your host to your primary domain name. Although a lot of implementations consider www.x.com and x.com to be synonymous they are not. The former is a 2nd level root domain subdomain and the latter is a 1st level root domain subdomain (though some people may consider 'www' to be a special case). Food for thought.

Cool - thanks for sharing

Reply 5

For the record...which was the gitlab.yml entry that was the troublemaker (for future reference)?