Spikes in Oldest Replication Slot Lag

[Mohammad Hafizullah]

I am quite new to the Postgres, and need some help with WAL Replication.

I have a Postgres 9.6 instance on RDS, and also configured DMS (Database Migration Service) to move data to S3 bucket. The problem I am facing is, if any data manipulation activity stops on the Postgres DB then the storage used by transaction logs start increasing on a rapid rate; and also Oldest Replication Slot Lag starts increasing too. 

As soon as I add/update a row in one of the table in DB, the Oldest Replication Slot lag comes down to zero and all the storage gets cleared; and storage used by Transaction Logs comes down too.

I am not able to understand the logic behind this behavior, if transaction is not happening on the DB then where does these transaction logs comes from? how come transaction logs grows too much that they can accommodate quite a few GB of storage space in 24 hours.

Please help.

[alex....@parkhub.com]

I am running into this exact issue.  I am running DMS for RDS postgres to another RDS postegres DB.  I am running a full load and also ongoing replication.  OldestReplicationSlotLag and TransactionLogDiskUsage continually increase and I am not sure why.  Have you resolved this or had any additional answers for this?

[Christophe Pettus]

These issues aren't really anything to do with WAL-E; I'd suggest the RDS or DMS forums for the best way of handling these problems.

> --
> You received this message because you are subscribed to the Google Groups "wal-e" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to wal-e+un...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

--
-- Christophe Pettus
x...@thebuild.com