| From: | Rod Taylor <pg(at)rbt(dot)ca> |
|---|---|
| To: | Kevin Grittner <Kevin(dot)Grittner(at)wicourts(dot)gov> |
| Cc: | pgsql-hackers(at)postgresql(dot)org, Neil Conway <neilc(at)samurai(dot)com>, Peter Brant <Peter(dot)Brant(at)wicourts(dot)gov> |
| Subject: | Re: fsutil ideas |
| Date: | 2006-02-24 17:24:50 |
| Message-ID: | 1140801890.5092.148.camel@home |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
> > I'm curious as to how you monitor for total transaction time length
> to
> > ensure that vacuum is able to do its thing, particularly when the
> > transaction is active (not IDLE).
>
> We run a database vacuum nightly and review it the next day.
Ahh.. different issues again I guess.
I have a few structures that need to be vacuumed every 10 minutes or
less. If a transaction runs longer than 20 minutes (thus effectively
disabling vacuum for 20 minutes) I run into pretty big problems.
I watch for table bloat but I haven't figured out a nice way of tracking
down the postgresql process with the oldest transaction running short of
patching PostgreSQL to report the XID for a connection in
pg_stat_activity.
--
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Tom Lane | 2006-02-24 17:48:34 | Re: fsutil ideas |
| Previous Message | Tom Lane | 2006-02-24 17:22:57 | Re: fsutil ideas |