Discourse Docker startup ERROR on K8S

AI-generated summary

Divyaanand_Sinha is encountering a Discourse Docker startup error on Kubernetes, with logs showing workers being killed by SIGKILL and occasional 502 Bad Gateway errors. They’re seeking help interpreting the unicorn_stderr.log output, which shows repeated worker crashes. pfaffman notes that Kubernetes deployments are unsupported and suggests the issue could stem from the Docker image, database config, or ingress. Stephen adds that on standard installs, such errors often trace back to broken plugins and advises rebuilding the app after disabling non-essential plugins — but explicitly states that Kubernetes-related issues are outside the scope of support. Divyaanand_Sinha follows up asking if the unicorn log errors are meaningful or just symptomatic of broader configuration issues — which remains unanswered, as the consensus is that Kubernetes deployments aren’t supported and troubleshooting is not feasible here.

I trying to run discourse in a docker on k8s. However after deploying the docker I am facing the below Error when I try to access discourse.

I checked /var/www/discourse/log/unicorn_stderr.log and see the following error.

I, [2020-06-19T06:03:12.458829 #95]  INFO -- : Refreshing Gem list
I, [2020-06-19T06:03:22.012156 #95]  INFO -- : listening on addr=0.0.0.0:3000 fd=10
I, [2020-06-19T06:03:49.744882 #95]  INFO -- : master process ready
E, [2020-06-19T06:04:06.155398 #95] ERROR -- : reaped #<Process::Status: pid 200 SIGKILL (signal 9)> worker=0
E, [2020-06-19T06:04:06.744237 #95] ERROR -- : reaped #<Process::Status: pid 214 SIGKILL (signal 9)> worker=1
E, [2020-06-19T06:04:06.744368 #95] ERROR -- : reaped #<Process::Status: pid 228 SIGKILL (signal 9)> worker=2
E, [2020-06-19T06:04:06.744453 #95] ERROR -- : reaped #<Process::Status: pid 242 SIGKILL (signal 9)> worker=3
E, [2020-06-19T06:04:11.048128 #95] ERROR -- : reaped #<Process::Status: pid 274 SIGKILL (signal 9)> worker=5
E, [2020-06-19T06:04:40.041398 #95] ERROR -- : reaped #<Process::Status: pid 311 SIGKILL (signal 9)> worker=7

Didn’t find much in the other log files. Can someone give me some idea on this?

And sometimes I am also getting 502 Bad Gateway

There are lots of things it could be,from the image docker image you’re using, to the database config, to the ingres, to name a few. This is why only the standard docker install is supported here. Have you made that work before?

Oops errors on a standard install usually arise from a broken plugin. Comment everything apart from docker manager out, the ./launcher rebuild app

We can’t help you with kubernetes.

Thanks, but the error I am seeing in unicorn_stderr.log, can that be an issue? Is there any specific reason for that or again it could be due to many things?