Problem Definition

Currently, starting up a node in Jira data center is building the EagerOfBizUserCache and EagerOfBizGroupCache by pulling existing cache from other nodes.
This can cause problems when we have corrupted cache and the only way to rebuild the cache is by doing a cold restart (stopping all nodes) causing downtime.

Suggested Solution

It would be great if the cache was build from scratch (DB) during the node startup so we don't need to have any cluster downtime to properly rebuild node's cache.
Rolling restart will heal the cluster.

Workaround

• Rebuild the cache by doing a cold restart (stopping all nodes). This will cause downtime.