From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | "news(dot)apexsc(dot)com" <lance(at)companywatch(dot)co(dot)uk> |
Cc: | pgsql-general(at)postgresql(dot)org |
Subject: | Re: VACUUM query |
Date: | 2002-03-07 17:11:44 |
Message-ID: | 10737.1015521104@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
"news.apexsc.com" <lance(at)companywatch(dot)co(dot)uk> writes:
> System: Linux 6.2; Postgresql 7.1.3
> can anyone tell me how to:
> 1. safely abort a VACUUM once it has started?
> 2. safely abort a scheduled vacuumdb once it has started?
In general, you cancel a VACUUM the same way as any other query:
control-C if interactively working in psql, or PQrequestCancel()
in a custom application, or send SIGINT to the individual backend
if you are an admin watching something go nuts.
In 7.1.* there is a known bug that aborting VACUUM partway through
vacuuming a TOAST table may leave things in a bad state. (Actually
there is nothing wrong with the table, but the TOAST code will get
confused because it doesn't check tuple status correctly. Re-vacuuming
will fix it, if you chance to get bit.) So I don't really recommend
doing this on a routine basis, until you update to 7.2.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Eric Webber | 2002-03-07 17:15:52 | Re: [ODBC] ODBC from win2k to postgres on Linux HELL |
Previous Message | Tom Lane | 2002-03-07 17:04:36 | Re: What are functional indices good for? |