Rainer J. H. Brandt wrote:
A while ago we had a client running PostgreSQL on a Quad v440, and the
performance was horrific. There are some known issues with PostgreSQL
and solaris. AFAIK there was a kernel patch to alleviate some of the
issues, but i'm unsure of the state of this as its been sometime since
i've had any experience with PostgreSQL and solaris.
Can anyone remember specific information?
It been too long to remember exact specifics, but it was something very
similar to this problem.
http://sunportal.sunmanagers.org/pipermail/summaries/2003-March/003254.html
You might also want to have a look at this, although most of these
issues are very old and have probably been resolved by now.
http://www.phil.uni-erlangen.de/tree/postgresql/majordomo/pgsql-novice-digest/v01.n098
The only detail I remember about performance under Solaris is a discussion
about slow qsort:
http://archives.postgresql.org/pgsql-hackers/2002-04/msg00103.php
A quick search didn't turn up information about whether the build
process now replaces Sun's qsort, or whether Sun has improved qsort
since the problem was noticed by postgres users in 2002.
It'd be interesting to invest some time studying the sources at
opensolaris.org.
Rainer
----------------------------------------------------------------------
Rainer J. H. Brandt email: rjhb@xxxxxxx
Brandt & Brandt Computer GmbH web: www.bb-c.de
Kamberg 111 phone: +49 2448 919126
D 53940 Hellenthal mobile: +49 172 9593205
---------------------------(end of broadcast)---------------------------
TIP 3: Have you checked our extensive FAQ?
http://www.postgresql.org/docs/faq
--
Marc McIntyre
MySource Matrix Lead Developer
