- Newest
- Most votes
- Most comments
So I will answer myself.
I had 5 DMS tasks with 5 logical replication slots in PG (one for each task, according to the Cloudwatch logs). But DMS said that "all slots are in use" which was wrong.
As 2 of my tasks didn't need to be set as CDC, I deleted these and keep three tasks. Then I deleted the two unnecessary logical slots in PG (with a select pg_drop_replication_slot(XXX)). Finally, I was able to resume my two failed tasks and they are running now.
So my guess is that unlike the doc said (https://docs.aws.amazon.com/dms/latest/userguide/CHAP_Source.PostgreSQL.html#CHAP_Source.PostgreSQL.Prerequisites), you don't need 5 replication slots for 5 tasks but more. I think I perhaps could have changed the max_replication_slots of the source server, but as I can't reboot it again, it couldn't be a solution.
Relevant content
- Accepted Answer
- asked 3 months ago
- AWS OFFICIALUpdated 2 years ago
- AWS OFFICIALUpdated 2 years ago