Hello,
I've just read about WAL and tried to set these 2 commands
for my test database (which is doing nothing 99% of time):
archive_command = 'cp -v %p /var/lib/pgsql/data/archive/%f'
archive_timeout = 300 # force a logfile segment switch after this
# many seconds; 0 is off
And after few days it is already crowded in the archive dir:
# ll /var/lib/pgsql/data/archive/|wc -l
1098
# du -hs /var/lib/pgsql/data/archive/
18G /var/lib/pgsql/data/archive/
Is there some archive_XXXX command for rotating WAL files
available (can't find it in the docs) or is it my responsibility
to get rid of the old files (and how do I identify them then?)
Or should I maybe just set archive_timeout to 0?
(the doc isn't clear enough for me what happens then)
My target is to have backups for the any point in the last 4 weeks.
Thank you
Alex
PS: I'm using NetApp filers with snapshots and:
# rpm -qa|grep postg
postgresql-libs-8.2.12-1PGDG.rhel5
compat-postgresql-libs-4-1PGDG.rhel5
postgresql-8.2.12-1PGDG.rhel5
postgresql-server-8.2.12-1PGDG.rhel5
# cat /etc/*release
CentOS release 5.2 (Final) <-- compatible to RHEL 5.2