Discourse run error in k8s

Product Support Self-hosting

AI-generated summary

Yaoqi_Yan encountered a runtime error while deploying Discourse in Kubernetes, citing issues with Redis and PostgreSQL containers failing to stay up. The error included messages like “chgrp: invalid group: ‘syslog’” and “kill: (114) - No such process”, suggesting misconfiguration or environment incompatibility. pfaffman clarified that running Discourse in Kubernetes is unsupported in the official community, as the standard install doesn’t account for Kubernetes-specific environments. They advised that unless scaling is critical for variable traffic, K8s is unnecessary. pfaffman further noted that the error likely stems from trying to run internal Redis and PostgreSQL containers inside the Discourse image — which isn’t compatible with K8s without external services. Yaoqi_Yan confirmed the issue was resolved by switching to external Redis and PostgreSQL services, and pfaffman celebrated the successful deployment. The key takeaway: Discourse in K8s requires external database and cache services and is not officially supported — use with caution or consider alternatives like Bitnami for managed support.

1

I would like to install discourse in K8s. The Docker works well in local environment but report bellow error after upload to k8s. Looks like related to postures and redis. Can anybody help?

run-parts: executing /etc/runit/1.d/00-ensure-links
run-parts: executing /etc/runit/1.d/00-fix-var-logs
run-parts: executing /etc/runit/1.d/01-cleanup-web-pids
run-parts: executing /etc/runit/1.d/anacron
run-parts: executing /etc/runit/1.d/cleanup-pids
Cleaning stale PID files
run-parts: executing /etc/runit/1.d/copy-env
Started runsvdir, PID is 47
ok: run: redis: (pid 61) 0s
ok: run: postgres: (pid 62) 0s
chgrp: invalid group: ‘syslog’
supervisor pid: 57 unicorn pid: 114
config/unicorn_launcher: line 71: kill: (114) - No such process
config/unicorn_launcher: line 15: kill: (114) - No such process
(57) exiting
timeout: down: redis: 0s, normally up, want up
ok: run: redis: (pid 169) 0s
timeout: down: postgres: 1s, normally up, want up
timeout: down: redis: 0s, normally up, want up
timeout: down: redis: 0s, normally up, want up
timeout: down: redis: 1s, normally up, want up

2

That’s not supported here. There are too many things that might be wrong to be able to support anything other than the standard install. Unless you’re sure to need scaling for a community that has periods of much higher activity than others you really don’t need k8s.

Did you build an image with launcher or are you using bitnami? If it’s bitnami, you can get support from them.

3

build with launcher ~~~

4

You’d need to use an external postgres and redis. It looks like you’re trying I use a container that includes redis and postgres?

5

It works with external rids and Postgres. Thanks Jay!

6

Great! Glad you’re up and running!