this assumption.
Also, from the application, is the LIMIT 25 passed as a constant or is that
also a parameter?
--
Andrew, Supernews
http://www.supernews.com - individual and corporate NNTP services
---(end of broadcast)---
TIP 9: In versions below
On 2007-04-15, Tom Lane <[EMAIL PROTECTED]> wrote:
> Andrew - Supernews <[EMAIL PROTECTED]> writes:
>> Looking at current CVS code the RI check seems to be skipped on update of
>> the _referred to_ table if the old and new values match, but not on update
>> of th
erred to_ table if the old and new values match, but not on update
of the _referring_ table.
--
Andrew, Supernews
http://www.supernews.com - individual and corporate NNTP services
---(end of broadcast)---
TIP 2: Don't 'kill -9' the postmaster
using a
shared_buffers setting of 5), and increase work_mem to 16364, and
see if there are any noticable changes in behaviour.
--
Andrew, Supernews
http://www.supernews.com - individual and corporate NNTP services
---(end of broadcast)---
TIP
with the
index in cache, is on the order of 30ms (where the data is cached in
shared_buffers) to 60ms (where the data is cached by the OS). That's on
a 2.8GHz xeon.
--
Andrew, Supernews
http://www.supernews.com - individual and corporate NNTP services
---(end of broadcast)---
TIP 6: explain analyze is your friend
rder,
thus enabling it to take advantage of prefetch and other sequential-read
optimizations (in the underlying OS and disk subsystem, rather than in
pg itself).
--
Andrew, Supernews
http://www.supernews.com - individual and corporate NNTP services
---(end of broadcast)---
TIP 2: Don't 'kill -9' the postmaster
herefore you have to
patch the sources.
The symptom to look for is: largescale filesystem deadlocks with many
processes (especially syncer) blocked in "nbufkv" state.
--
Andrew, Supernews
http://www.supernews.com - individual and corporate NNTP services
---(end of bro
