This can also happen if you have a long-running transaction - Postgres can't clean up the xlogs until every transaction from before them has finished.
On 7 October 2014 01:50, Daniel Farina <[email protected]> wrote: > On Thu, Sep 4, 2014 at 7:27 AM, Kevin Harriss <[email protected]> wrote: >> On our postgres slave we have it configured to build initial with wal-e and >> then once caught up use a connection with the primary to stay updated. I >> noticed today that our pg_xlog directory is over 60GBs and no segments have >> been removed since June. I just confirmed that the sync with the primary is >> working correctly it just doesn't seem to clean up the xlogs. My >> recovery.conf is below. I also noticed these lines in the postgres log but >> couldn't find out what the return code meant, archive_cleanup_command >> "pg_archivecleanup /var/lib/postgresql/9.3/archive/ %r": return code 32512. >> Any ideas on what could be causing these segment files to stick around and >> how to fix it. >> >> >> >> standby_mode = 'on' # enables stand-by (readonly) mode >> >> # Connect to the master postgres server using the replicator user we >> created. >> primary_conninfo = 'host=<primary_connection> port=<primary_port> >> user=<primary_user> password=<primary_password>' >> >> # Specifies a trigger file whose presence should cause streaming replication >> to >> # end (i.e., failover). >> trigger_file = '/tmp/pg_failover_trigger' >> >> # Shell command to execute an archived segment of WAL file series. >> # Required for archive recovery if streaming replication falls behind too >> far. >> restore_command = '/usr/bin/envdir /etc/wal-e.d/env /usr/local/bin/wal-e >> wal-fetch "%f" "%p"' >> archive_cleanup_command = 'pg_archivecleanup >> /var/lib/postgresql/9.3/archive/ %r' > > You shouldn't need archive cleanup since that's, as I understand it, > to remove stuff from the archive, e.g. like "wal-e delete". I > personally do not use archive_cleanup, instead preferring to run > "delete" commands via cron or a similar system. > > pg_xlog in particular is not typically cleaned up by > archive_cleanup_command. I'm not sure why your pg_xlog is growing so > large, perhaps you have wal_keep_segments at some ginormous number? > > -- > 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 [email protected]. > For more options, visit https://groups.google.com/d/optout. -- Russ Garrett [email protected] -- 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 [email protected]. For more options, visit https://groups.google.com/d/optout.
