Insight Index integrity check on Data Center with multiples nodes

Description

By default, "Restore Insight index from file" and "Store cache on shutdown" is enabled.

Reference: Configuring global Jira settings

On Jira shutdown, Insight will persist the index on disk for faster startup and on the Jira start up, Insight will load the objects into memory from the file instead of doing a full Insight Re-index.

On a DC cluster with more than one instance, “Store cache on shutdown” would never work on a big production system, because most likely there will be changes in Insight made by the other nodes within the cluster.

If one node is stopped for maintenance or any other reason, on Jira start up,
Insight will always perform an Integrity check of the files against the database (source of truth) and possibly there will be an Integrity check failed. Some changes may have occurred while the node was down and a full Insight Re-Index will be kicked out.

In large instances, it could cause that Insight Reindex might take a while to complete affecting Insight operations on the node.

Suggested Solution

Something like Jira Data Center cache replication could be implemented in Insight.