The URL that is suggested for use in JIRA Data Center behind a load balancer, as mentioned in https://confluence.atlassian.com/display/ENTERPRISE/JIRA+Data+Center+Load+Balancer+examples states to use /status and it will return "state":"RUNNING" for a healthy node that should be in the LB pool. In my testing, I found that when a node is performing a foreground re-index that locks JIRA, it is actually unavailable for processing user requests. Accessing a node that is performing a locking re-index redirects to an error page and gives an HTTP 503. When that is occurring, I would not want that node to have user requests routed to it.
Having multiple nodes for JIRA Data Center, we like foreground re-indexing as it is much faster with the amount of data we have. So when a node is locked for re-indexing we'd like it to take itself out of the pool. This issue is to suggest either:
- the /status page be updated to provide a different status during indexing, possibly "state":"INDEXING". This would allow a LB health check to take the node out of the pool during indexing
- update the Data Center LB documentation (above) to instruct users to use /secure/errors.jsp as the health check URL as it more accurately conveys whether a node is capable of servicing requests