RE: pg_xlog piling

We had a similar issue with 9.2. Managed to track it enough to figure out it was happening when streaming replication falls behind.

We log ship to a dir, log_archive, and have an archive_command.sh file in that dir.

Added a line to the archive_command.sh to call pg_archivecleanup for the xlog partition in addition to the existing pg_archivecleanup for log_archive dir. For some reason (have not had time to research) we have to fully path the location of the pg_archivecleanup.

If this is helpful, I can show the lines of the archive_command.sh file.

Paula Price
Senior Database Administrator

From: Venkata B Nagothi [mailto:nag1010(at)gmail(dot)com]
Sent: Tuesday, January 30, 2018 3:25 PM
To: chris <chrisk(at)pgsqlrocket(dot)com>
Cc: David G. Johnston <david(dot)g(dot)johnston(at)gmail(dot)com>; Jorge Torralba <jorge(dot)torralba(at)gmail(dot)com>; pgsql-admin(at)lists(dot)postgresql(dot)org
Subject: Re: pg_xlog piling

On Wed, 31 Jan 2018 at 8:32 am, chris <chrisk(at)pgsqlrocket(dot)com<mailto:chrisk(at)pgsqlrocket(dot)com>> wrote:

postgres=# select * from pg_stat_replication;

pid | usesysid | usename | application_name | client_addr |

client_hostname | client_port | backend_start | state |

sent_location | write_location | flush_location | replay_locatio

n | sync_priority | sync_state

------+----------+----------+------------------+----------------+-----------------+-------------+-------------------------------+-----------+---------------+----------------+----------------+---------------

--+---------------+------------

5303 | 10 | postgres | walreceiver | 198.161.184.74 |

| 57049 | 2017-12-14 11:13:16.339948-07 | streaming | 2FD3/41A38140 |

2FD3/41A38140 | 2FD3/41A38140 | 2F47/7015D490

| 0 | async

(1 row)

postgres=#

# - Archiving -

archive_mode = on # allows archiving to be done

archive_command = 'cp %p /ComplianceDB/pitr/walstage/%f; mv /ComplianceDB/pitr/walstage/%f /ComplianceDB/pitr/wal/%f' # (change requires restart)
#archive_command = '' # command to use to archive a logfile segment
# placeholders: %p = path of file to archive
# %f = file name only
# e.g. 'test ! -f /mnt/server/archivedir/%f && cp %p /mnt/server/archivedir/%f'
#archive_timeout = 0 # force a logfile segment switch after this
# number of seconds; 0 disables
Thanks

Interesting. Do you see any files called *.ready in pg_xlog/archive_status directory ? Normally pg_xlog fills up when archive_command is not successful.

Do you have parameter wal_keep_segments configured to higher value ?

Regards,
Venkata B
--

Regards,

Venkata B N
Database Consultant