Update takes ~20 minutes and progress update breaks (temporarily) at step multisite:migrate

I suspect this is because we didn’t drop the index concurrently.

A normal DROP INDEX acquires an ACCESS EXCLUSIVE lock on the table, blocking other accesses until the index drop can be completed.

This should be fixed in PERF: Update migration to drop index concurrently. by tgxworld · Pull Request #28955 · discourse/discourse · GitHub.

@schneeland Thank you for taking the time to report the issue :+1:

2 Likes