| From: | Heikki Linnakangas <heikki(at)enterprisedb(dot)com> |
|---|---|
| To: | Rafael Martinez <r(dot)m(dot)guerrero(at)usit(dot)uio(dot)no> |
| Cc: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, pgsql-performance(at)postgresql(dot)org |
| Subject: | Re: Need to run CLUSTER to keep performance |
| Date: | 2007-11-12 16:31:41 |
| Message-ID: | 47387FED.8090408@enterprisedb.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-performance |
Rafael Martinez wrote:
> DETAIL: 83623 dead row versions cannot be removed yet.
Looks like you have a long-running transaction in the background, so
VACUUM can't remove all dead tuples. I didn't see that in the vacuum
verbose outputs you sent earlier. Is there any backends in "Idle in
transaction" state, if you run ps?
In 8.1, CLUSTER will remove those tuples anyway, but it's actually not
correct. If the long-running transaction decides to do a select on
hosts-table later on, it will see an empty table because of that. That's
been fixed in 8.3, but it also means that CLUSTER might no longer help
you on 8.3. VACUUM FULL is safe in that sense in 8.1 as well.
--
Heikki Linnakangas
EnterpriseDB http://www.enterprisedb.com
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Andrew Sullivan | 2007-11-12 16:32:23 | Re: Curious about dead rows. |
| Previous Message | Rafael Martinez | 2007-11-12 16:11:59 | Re: Need to run CLUSTER to keep performance |