Re: Bad perfomance of pl/pgsql-function on new server

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Wil Peters <info(at)itaudit(dot)demon(dot)nl>
Cc: pgsql-performance(at)postgresql(dot)org
Subject: Re: Bad perfomance of pl/pgsql-function on new server
Date: 2003-03-29 23:25:43
Message-ID: 1001.1048980343@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

Wil Peters <info(at)itaudit(dot)demon(dot)nl> writes:
> On the new server - although this server has far better specs! - the
> function does NOT perfom as well as on the old server. Searches take
> several minutes, where on the old server a few SECONDS where needed.

Is the new installation really equivalent to the old? I'd wonder about
differences in multibyte compilation option, database locale and
encoding, etc. Any of these could result in a huge hit in text-pushing
performance.

Another traditional post-upgrade problem is forgetting to VACUUM
ANALYZE; but that probably shouldn't affect this function, since it's
not issuing any database queries.

(Personally I'd have written this sort of function in plperl or pltcl,
either of which are far more appropriate for text-string-mashing than
plpgsql. But that's not really answering your question.)

regards, tom lane

In response to

Browse pgsql-performance by date

  From Date Subject
Next Message Marco Colombo 2003-03-31 12:06:50 WAL monitoring and optimizing
Previous Message Tom Lane 2003-03-29 23:13:28 Re: Index not used, performance problem