On Wed, Nov 2, 2011 at 8:21 PM, Greg Smith <greg@2ndquadrant.com> wrote:
> In cases where there are little or no writes to the WAL, checkpoints will be
> skipped even if checkpoint_timeout has passed. At least one new WAL segment
> must have been created before an automatic checkpoint occurs. The time
> between checkpoints and when new WAL segments are created are not related in
> any other way. If file-based WAL shipping is being used and you want to
> bound how often files are sent to standby server, to reduce potential data
> loss you should adjust archive_timeout parameter rather than the checkpoint
> ones.
I think this is good, although "where there are little or no writes to
the WAL" seems a bit awkward to me - how about "where little or no WAL
has been written"?
I would probably delete "to reduce potential data loss" from the last
sentence, since I think that sentence has a few too many clauses to be
easily parseable.
Should we also put a similar sentence into the documentation for
checkpoint_timeout?
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company