Hi there

We are currently looking into upgrading to Jira Data Center version. Therefor we tried to write our own Helm charts for it, so we could run it on our internal Kubernetes infrastructure.

When starting a Jira container in clustered mode, it tries to save it's node entity into the postgres database. At this point the error occurs:

Caused by: org.ofbiz.core.entity.GenericEntityException: while inserting: [GenericEntity:ClusterNode][nodeState,ACTIVE][ip,test-jira-software-0.test-jira-software.jira-namespace.svc.cluster.local][cacheListenerPort,40001][nodeVersion,8.5.6][nodeId,test][nodeBuildNumber,805006][timestamp,1596797586988] (SQL Exception while executing the following:INSERT INTO clusternode (NODE_ID, NODE_STATE, TIMESTAMP, IP, CACHE_LISTENER_PORT, NODE_BUILD_NUMBER, NODE_VERSION) VALUES (?, ?, ?, ?, ?, ?, ?) (ERROR: value too long for type character varying(60)))

The error is happening because the value `test-jira-software-0.test-jira-software.jira-namespace.svc.cluster.local` has more than 60 characters.

Since this is a reverse lookup the entire fqdn is taken. In the case of kubernetes you end up with this long string, because that's the cluster internal fqdn to the container/pod.

There are multiple ways to solve this issue:

• Expand the field clusternode to a higher varying value (like 200)
• Instead of doing a reverse lookup, use the actual IP address.
• Add a configuration field in the cluster.properties to overwrite this field.

This would be very helpful not just for kubernetes use-cases, but for everyone who uses long fqdn names.

Let me know if you need more detail on this case.