Le lundi 04 février 2008 à 16:36 -0600, David Morton a
écrit :
> "Heap Only Tuple (HOT) dramatically reduces the
database maintenance
> issues associated with frequently updated data,
reducing the need to
> vacuum and giving substantial throughput improvements
for some
> applications."
>
> It may be a good candidate for the bayes database -
anyone interested
> in testing?
I'm watching this feature too, but at this time, not so much
docs are
out for public reading.
In mid 2007, on Pg lists, there was reports about 100% speed
improvement
with HOT ! But was internals tests only and pgbench based
(not real
world).
I'll keep an eye on this, and maybe start a test server.
> Another thing at the bottom is CyberCluster - I wonder
if we are close
> to having a real database cluster option?
More informations on this on the developper's wiki :
http://developer.postgresql.org/index.php/Replication,_
Clustering,_and_Connection_Pooling
CyberCluster is supported by one man, and provide a nice
solution on the
top of PGCluster :
http://pg
foundry.org/projects/cybercluster/
and :
http://www.cybertec.at/english/pr_cybercluster_e.html
Also, there is a good mature project :
http://pgpool.
projects.postgresql.org/
I've only played with Slony-I few years ago, and this
introduce a very
high load and latency on systems, so we havent not pushed it
in
production.
Regards
--
Alexandre
_______________________________________________
Maia-users mailing list
Maia-users renaissoft.com
http://www.renaissoft.com/mailman/listinfo/maia-users
|
