Skip to content

Zero downtime reindexing should target indexes that exist

What does this MR do and why?

A support ticket (internal link) had mentioned having errors about missing the Commit index. They had recently upgraded from 15.4 to 15.6 and done a zero downtime reindex. The gitlab:elastic:info rake task showed that the current migration that was running was DeleteOrphanedCommits which is before the commits index is created in MigrateCommitsToSeparateIndex. I believe this can lead to an incompatible state.

This MR changes the reindexing process to only target classes where the separate indexes have been created by checking the appropriate migrations.

Screenshots or screen recordings

N/A

How to set up and validate locally

Numbered steps to set up and validate the change are strongly suggested.

MR acceptance checklist

This checklist encourages us to confirm any changes have been analyzed to reduce risks in quality, performance, reliability, security, and maintainability.

Edited by Terri Chu

Merge request reports