Jira Data Center cluster requires constant administrator authentication after switching to multi-nod

[Matt Parks]

I am running Jira DC 8.20.1 on a single RHEL7 machine. We are in the process of migrating to RHEL8, so I wanted to set up two nodes and run it as a cluster.

I have opened up the EhCache ports and configured our F5 to redirect traffic to both nodes. I have also created a shared homedir, which was copied from the original homedir on the RHEL7 machine.

I am able to log into Jira and everything seems to be working well from a user perspective but, when I access any area where I have to put in my administrator credentials, I literally have to put them in every single time. Even if I navigate to a new area less than five seconds after putting in my credentials.

Is there a configuration piece that I'm missing? I didn't see this behavior before setting up the cluster.

[Evgenii · Accepted solution]

When setting up Jira Data Center as a cluster, there are a few additional configurations you should consider to ensure a seamless user experience. In your case, the issue you're facing with having to repeatedly provide administrator credentials is likely related to session affinity or sticky sessions.

In a clustered environment, multiple requests from a single user may be distributed across different nodes. To maintain session state and avoid repeated authentication, you need to configure session affinity on your load balancer to ensure that subsequent requests from the same user are directed to the same node where the session was initially established.

Once session affinity is properly configured, subsequent requests from the same user should consistently land on the node where the initial session was established. This will eliminate the need for repeated administrator credentials input.

I think, it's the main reason.

[Matt Parks]

I can confirm that our security team needed to configure sticky sessions on the F5.

Thanks for the quick response!