| From: | Gaetano Mendola <mendola(at)bigfoot(dot)com> |
|---|---|
| To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
| Subject: | Re: Single table vacuum full different when vacuum full the whole |
| Date: | 2004-10-18 00:49:52 |
| Message-ID: | 41731330.9090105@bigfoot.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-admin |
Tom Lane wrote:
> Gaetano Mendola <mendola(at)bigfoot(dot)com> writes:
>
>>it seems that a vacuum full on the whole DB is more aggressive.
>
>
> It is not.
>
> A much more plausible theory is that this is the result of concurrent
> changes to the table. It is clear from the "dead row versions" stats
> that there were concurrent transactions ...
That is the more updated/inserted table, and yes there were some concurrent
transaction but, is it plausible that 82 dead rows were responsible of grab
26000 index row:
INFO: index "ua_user_data_exp_id_user_key" now contains 34438 row versions in 886 pages
DETAIL: 27488 index row versions were removed.
instead of:
INFO: index "ua_user_data_exp_id_user_key" now contains 34519 row versions in 886 pages
DETAIL: 1362 index row versions were removed.
Regards
Gaetano Mendola
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Bruce Momjian | 2004-10-18 02:23:39 | Re: set AUTOCOMMIT OFF in psqlrc not having affect |
| Previous Message | Tom Lane | 2004-10-18 00:39:11 | Re: Single table vacuum full different when vacuum full the whole database |