Skip to content

Draft: datastore: Clean completed & dead replication jobs

Toon Claes requested to merge tc-replication-cleanup-queue-again into master

In [1] changes were made so 'completed' and 'dead' jobs are no longer kept in the database, and are cleaned up. In a follow-up commit [2] a migration was added to clean up all 'completed' and 'dead' jobs from the replication queue. But this method did not take the order of deployment into account. So the migration did run before the code was deployed, which resulted some new 'completed' and 'dead' rows to be added for a little while.

This change repeats the DELETE query from the migration added in [2] to clean up 'completed' and 'dead' rows once and for all.

  1. 8f8ae302 (Replication job acknowledge removes 'completed' and 'dead' events., 2020-08-05)
  2. 2d3fc806 (Replication job acknowledge removes 'completed' and 'dead' events., 2020-09-03)

Changelog: performance
Issue: #3665 (closed)

Edited by Toon Claes

Merge request reports