Re: VACUUM ANALYZE blocking both reads and writes to a table - Mailing list pgsql-performance

From Peter Schuller
Subject Re: VACUUM ANALYZE blocking both reads and writes to a table
Date
Msg-id 20080630154318.GB15756@hyperion.scode.org
Whole thread Raw
In response to Re: VACUUM ANALYZE blocking both reads and writes to a table  (Alvaro Herrera <alvherre@commandprompt.com>)
Responses Re: VACUUM ANALYZE blocking both reads and writes to a table
List pgsql-performance
Actually, while on the topic:

>     date: 2007-09-10 13:58:50 -0400;  author: alvherre;  state: Exp;  lines: +6 -2;
>     Remove the vacuum_delay_point call in count_nondeletable_pages, because we hold
>     an exclusive lock on the table at this point, which we want to release as soon
>     as possible.  This is called in the phase of lazy vacuum where we truncate the
>     empty pages at the end of the table.

Even with the fix the lock is held. Is the operation expected to be
"fast" (for some definition of "fast") and in-memory, or is this
something that causes significant disk I/O and/or scales badly with
table size or similar?

I.e., is this enough that, even without the .4 bug, one should not
really consider VACUUM ANALYZE non-blocking with respect to other
transactions?

(I realize various exclusive locks are taken for short periods of time
even for things that are officially declared non-blocking; the
question is whether this falls into this category.)

--
/ Peter Schuller

PGP userID: 0xE9758B7D or 'Peter Schuller <peter.schuller@infidyne.com>'
Key retrieval: Send an E-Mail to getpgpkey@scode.org
E-Mail: peter.schuller@infidyne.com Web: http://www.scode.org


Attachment

pgsql-performance by date:

Previous
From: Peter Schuller
Date:
Subject: Re: VACUUM ANALYZE blocking both reads and writes to a table
Next
From: Moritz Onken
Date:
Subject: Re: Planner should use index on a LIKE 'foo%' query