Has anyone been successful getting the Confluence containers to run on kubernetes?

[ryan gnatt]

I'm trying to migrate an old version of Confluence to a kubernetes deployment and having a surprising amount of difficulty getting past the install phase. Every time I complete the database configuration web-form the container crashes. (I have tried mysql, mariadb, postgresql and even the flat local file database, and the container still crashes.)

Things I have verified:

• localhost is resolvable in /etc/hosts and in the container

• write permissions to the confluence data directory are fine

• selinux is in permissive mode on the container host. (I know, bad policy, but I'm currently testing)

• access to the database server is fine

• memory limits are golden (16G memory and 2 cpus dedicated)

This almost certainly has to be related to kubernetes in some way, because the container runs fine when run on a host via docker. Has anyone ran into these issues and successfully deployed Confluence? I have even tried creating an un-official container image and installing Confluence there, but this issue persists.

[Praveen]

Hey Ryan, Hope you got this to work? I am studying on setting up auto scaling for Jira and Confluence, any inputs/best practices from you would be much helpful.

Thanks.

[MB]

Clustering and setup is difficult. In short, you would need: 

* Persistent Volume Templates for each node to persist the confluence local home dir
* leverage hazelcast clustering, sneeky one here: https://github.com/Bonn93/atlassian-kubernetes/blob/connie-testing/confluence/confluence_sts.yml
* Use the official images and set the XMX/XMS to atleast 2GB, it's funky below that and seems to crash