search

berngp / docker-zabbix

Docker Container running a Zabbix Server and Zabbix Web UI.
Other
160 stars 66 forks source link

After a while container will not start anymore #12

Open Franselbaer opened 9 years ago

Franselbaer commented 9 years ago

After I did some stops and starts it'll exit immediately after starting the conainer. Had this multiple times on multiple plattforms.

I use this container often for testing purposes, or onetime tests but if i stop/start too often i can only throw away...

Best regards

Franselbaer commented 9 years ago

Seems to be an ubuntu networkmananger issue

Franselbaer commented 9 years ago

Nope. I started a new instance on a fresh ubuntu and let it run for 21 hours. The next day i stopped it and couldn't start it anymore.

Franselbaer commented 9 years ago

Reproduced: started a new instance, set up a cron to stop/start it every 5 minutes and after a half an hour it doesn't start anymore

berngp commented 9 years ago

Thanks @Franselbaer I am currently running it on a CentOS. Could you please share with me which version of Ubuntu and Docker you are using?

Franselbaer commented 9 years ago

Hey @berngp. Thanks for the response. I tested with ubuntu 14.10 and debian 7 with docker taken from repository (I think 1.2). Unfortunately you cannot get inside if it doesn't start anymore therefore I couldn't investigate more yet. If I find the time the next days I'll try to let the container write logs to an outside location.

berngp commented 9 years ago

Thank you @Franselbaer, I do my due diligence to replicate the issue. We can have mounts where the logs would be written to, I was planning on doing this anyway for MySQL too.

berngp commented 9 years ago

@Franselbaer the monit & AppArmor issue referenced in #13 was causing the syslogs to grow very quickly and might be related. Could you please run the container with the following flags.

--cap-add SYS_PTRACE  --security-opt apparmor:unconfined
berngp commented 9 years ago

@Franselbaer let me know if the changes above helped you out.

meltonpineda commented 9 years ago

Hi @berngp , I'm having the same problem. I already add the flags that you suggested but the problem still occurs. I'm running it on Centos 6.6.

berngp commented 9 years ago

@meltonpineda we need to move out of monit, is already causing several stability issues. As reference #22

berngp commented 9 years ago

@azurewraith I am migrating to supervisord but haven't yet finished. I'll push a labeled image that we can test out and see if supervisord helps out on stabilizing the container.

azurewraith commented 9 years ago

@berngp Thanks, that would be helpful. I'm assuming there is a supervisord branch as well?

berngp commented 9 years ago

@azurewraith sorry for the late reply. I have the feature/22-supervisord to address the migration from monit to supervisor. It still not working but will keep track to such effort through that branch.