On Fri, Jan 31, 2020 at 11:20 AM <soumik.bhattacharjee@xxxxxxx> wrote:
Dear Experts,
Post the setup of streaming replication –MASTER and one SLAVE in PostgreSQL 12 , the pg_wal in both MASTER/SLAVE are - filling up.
Do you have any replication slots on the replica? When you created the replica, you were supposed to exclude the directory pg_replslot. If you didn't the replica might have "inherited" slots it should not have.
My MASTER is in archive log mode and SLAVE in no archive log mode.
MASTER
postgres=# show archive_command;
archive_command
------------------------------------------------------------------------------------
test ! -f /data/tablespaces/archivedir/%f && cp %p /data/tablespaces/archivedir/%f
Is this command succeeding? Do you see messages in the log file about it failing? Are new files showing up in '/data/tablespaces/archivedir/'?
Does the master have any old replication slots which have been abandoned or are falling behind? Check the view pg_replication_slots.
Is it advisable to change default wal location (pg_wal)? Or any other options apart from using pg_archivecleanup?
pg_archivecleanup is for cleaning up the archive (which is apparently '/data/tablespaces/archivedir/'). It is *not* for cleaning up pg_wal.
Cheers,
Jeff
