On Fri, Jan 3, 2014 at 11:33 AM, Jan Wieck <JanWieck at yahoo.com> wrote:

> On 01/02/14 16:12, Tory M Blue wrote:
> > We are getting outside of slon and into postgres and i'll try to keep
> > this light, here are my postgresql custom config aspects
>
> Postgres and its architecture very much affect the behavior and
> performance of Slony. I'm afraid that they cannot be viewed completely
> isolate from each other.
>
> > The hardware is being refreshed but right now it's an 8 core, 32GB
> > CentOS system, being replaced with some pretty big hardware
> > 256GB/32cores many SSD's. But tuning this now and knowing where I need
> > to tweak when I add the added capacity would be great.
>
> The slon process uses one single database connection to the data
> provider for the copy operation and another single database connection
> to feed the data into the new replica. A single Postgres session only
> utilizes one single core. It therefore makes zero difference if you are
> upgrading from a dual core to 64 cores or more.
>
> The suggestions made about work mem and maintenance work mem are valid.
> If you are using a dedicated DB user for all your slon database
> connections you can tweak these per server via ALTER USER.
>
>
> Regards,
> Jan



Okay that is interesting.. "If you are using a dedicated DB user for all
your slon database connections, you can tweak these per server via Alter
User", what will that do for me? We do use the same user across the board
and i figured I would just modify my postgresql.conf. So i'm a tad confused
:)

And since it appears that my index creation is the big time hole, I'm going
to be testing work mem, by going from 100M to 2GB+ and i'll reduce my
vacuum processes from 7 to maybe 2 or 3, so I don't blow anything out . But
you have intrigued me :)

Thanks
Tory
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.slony.info/pipermail/slony1-general/attachments/20140103/3b76fdd7/attachment.htm