On Fri, 2007-10-26 at 18:06 -0500, Kevin Grittner wrote:
> >>> On Fri, Oct 26, 2007 at 5:47 PM, in message <695.1193438855@sss.pgh.pa.us>,
> Tom Lane <tgl@sss.pgh.pa.us> wrote:
>
> > And after
> > each archive_timeout, we test to see if we need to flush the current WAL
> > segment out to the archive; which is determined by whether the write
> > pointer is currently exactly at the start of a segment or not.
>
> Hmmm... We would actually prefer to get the WAL file at the
> specified interval. We have software to ensure that the warm
> standby instances are not getting stale, and that's pretty simple
> with the current behavior. We don't have a bandwidth or storage
> space issue because we zero out the unused portion of the WAL file
> and gzip it -- an empty file's about 16 KB. Checking that the whole
> system is healthy gets a lot more complicated if we stop sending
> empty WAL files.
>
> Could this at least be a configurable option?
>
A good point.
Keep in mind that even in the current system, your configuration is
variable based on the checkpoint_timeout setting.
Regards,
Jeff Davis