Massive insert created 975 WAL segments.. what?
Massive insert created 975 WAL segments.. what?
От:
Wells Oliver <wellsoliver@gmail.com>
Дата:
We were doing a test and inserted 1.5 million rows. In doing so, postgres created 974 WAL segments of 16MB apiece. The relevant configuration from my master's postgresql.conf:
--
Wells Oliver
wellsoliver@gmail.com
wal_level = hot_standby
archive_mode = on
archive_command = 'test ! -f /mnt/postgresql-logs/%f && cp %p /mnt/postgresql-logs/%f'
max_wal_senders = 3
wal_keep_segments = 16
So, now I have 16GB of WAL segments, 974 of them. Will postgresql clean this up? Will it remove these files? Will it create more with big inserts?
What is the best way to manage this?
Wells Oliver
wellsoliver@gmail.com
Re: Massive insert created 975 WAL segments.. what?
От:
Ray Stell <stellr@vt.edu>
Дата:
On Sep 11, 2012, at 8:10 PM, Wells Oliver wrote: > So, now I have 16GB of WAL segments, 974 of them. Will postgresql clean this up? Will it remove these files? Will it create more with big inserts? > > What is the best way to manage this? http://www.postgresql.org/docs/9.2/static/pgarchivecleanup.html you can use this in conjunction with your backup facility/scipts. Don't want to clean them up until you have a backup, right?
Re: Massive insert created 975 WAL segments.. what?
От:
Simon Riggs <simon@2ndQuadrant.com>
Дата:
On 12 September 2012 01:10, Wells Oliver wrote: > We were doing a test and inserted 1.5 million rows. In doing so, postgres > created 974 WAL segments of 16MB apiece. The relevant configuration from my > master's postgresql.conf: > > wal_level = hot_standby > archive_mode = on > archive_command = 'test ! -f /mnt/postgresql-logs/%f && cp %p > /mnt/postgresql-logs/%f' > max_wal_senders = 3 > wal_keep_segments = 16 > > So, now I have 16GB of WAL segments, 974 of them. Will postgresql clean this > up? Will it remove these files? Will it create more with big inserts? > > What is the best way to manage this? These are transaction log files, created by database writes. They protect you if you crash and also allow replication/backup. These will be rotated every checkpoint, which by default is 5 minutes. So they'll be long gone by now. -- Simon Riggs http://www.2ndQuadrant.com/ PostgreSQL Development, 24x7 Support, Training & Services