After index rebuild on a node, automatically replicate the updated index to other cluster nodes in the background

Summary

In instances where nodes have not had downtime or required a restart, but the index has been rebuilt on one node, automatically replicating the index in the background to the other nodes is useful to administrators who wish to maximize uptime.

If a manual index rebuild is triggered, it should be sufficient to flag that node's index as the most current copy for replication to the other nodes without downtime.

In this scenario, only the node with the already-rebuilt index will have a current index, and thus it would be recommended to load balance users to only that node during said index replication process, to ensure their search results are current and complete.

Environment

Confluence Data Center 6.x

Current State

Currently, the index replication only occurs on startup of a node, thus requiring a node to be restarted in order to either automatically pull a fresh copy of the index. Manually copying the index to other nodes is another option but does require downtime.

Reference Documentation

• Administer Your Data Center Search Index
• How To Rebuild The Content Indexes From Scratch On Confluence Data Center
• Confluence Data Center Technical Overview - Indexes

Related to: CONFSERVER-54760