Hi all,
I'm wondering if there's an accepted way to monitor a warm standby
machine's lag in 8.4. The wiki[1] has a link[2] to a script which
parses the output of pg_controldata, looking for a line like this:
Time of latest checkpoint: Thu 09 Dec 2010 01:35:46 PM EST
But I'm not sure whether this timestamp is to be trusted as an
indicator of how far behind the standby is in its recovery -- this
timestamp just tells us when the standby last performed a checkpoint,
regardless of how far behind in the WAL stream it is, right?
I haven't come across any other monitoring suggestions for warm
standby on 8.4. I've seen suggestions for hot standby slaves to use:
SELECT pg_last_xlog_receive_location();
but this won't work on an 8.4 warm standby of course. I've searched
around and haven't found[3] any other tips on how to monitor my
standby.
The manual mentions[4] using pg_xlogfile_name_offset() in the context
of implementing record-based log shipping. Would this be useful for
monitoring standby lag? Any other ideas?
Thanks,
Josh
--
[1] http://wiki.postgresql.org/wiki/Warm_Standby
[2] http://www.kennygorman.com/wordpress/?p=249
[3] I was hoping this page would have some relevant info:
http://www.scottrmead.com/blogs/scott/warm-standby-monitoring , but
it's down now :(
[4] http://www.postgresql.org/docs/8.4/static/warm-standby.html#WARM-STANDBY-RECORD
