Jira DataCenter in docker (multiple nodes)

[Luka Čadež]

Hello,

I have a problem running Data Center in docker (with two nodes). I used the files from

https://github.com/codeclou/docker-atlassian-jira-data-center/tree/master/versions/8.13.4

and instead of their load balancer I used nginx and created my own LB/proxy.

When I go and run one node it works OK...I complete the setup enter the trial key and all is good. Except on the DB container i get errors in logs for example:

ERROR:  duplicate key value violates unique constraint "cluster_lock_name_idx"
DETAIL:  Key (lock_name)=(com.atlassian.jira.cluster.zdu.DefaultClusterUpgradeStateManager.clusterUpg
radeState) already exists.

But even with those errors I can use jira (from what I've tested ofcourse).

Then when I run the second node it starts up normaly but then when I want to log in I enter the credentials and it just refreshes the screen (does not log me in).

Does anyone have any experience with this kind of setup? Any help would be appreciated.

Luka

[jesse_j_borden_ctr · Accepted solution]

Came here for a similar database error, but not what I was looking for. However, I had the issue you are talking about before. I had to enable the sticky sessions on the web server (in my case K8s ingress). For Nginx it is the affinity and affinity-mode setting. https://kubernetes.github.io/ingress-nginx/examples/affinity/cookie/

[Luka Čadež]

Big thanks for that session persistence setting - now it works. Just have to sort out the db errors

[Gonchik Tsymzhitov]

+1 I am experiencing the same thing

[Martin Cleaver]

I get the same, with the same version (8.13.x),  but seemingly not in a load balancer context. 

Any hints?

[Tommy van Extel]

Same here, also see https://jira.atlassian.com/browse/JRASERVER-69899