I'm running PostgreSQL 7.1.2 on Solaris 8. The box is an Ultra 60 with
1GB of RAM. The database is on it's own 11GB partition which is
currently 93% full. I've since moved the pg_xlog directory to a
seperate 11GB partition.
We have one table which accounts for 90% of the database size. In
trying to vacuum this table, the pg_xlog directory reaches 100%
(starting from less than 1%).
In reading posts to this news group I've come across one suggestion
which talks about increasing the CHECKPOINT_TIMEOUT and
CHECKPOINT_SEGMENTS size. I will try that shortly.
Does PostgreSQL 7.2 address this problem? Any other suggestions?
Obviously, throwing more disk space at this would work but I currently
don't have that option.
Thanks,
Dave