[OSM-dev] osm2pgsql hangs.

[Richard Ive]

Thanks for the reply!

Yes, I tweaked posgresql's config - I mad add a a little more to the shared
buffer.

It just seems weird to me. I did a comparison yesterday between the old and
new machine. The new is obviously very powerful, with a tuned postgresql,
where as the old machine is on it's last legs and does not have the tweaked
postgresql. By the time the old machine had completely finished the import
(about 6 hours), the new machine hadn't even moved passed 'Reading file'.

I'll have a look in what you've said.

Thanks a lot for you help!

Regards,
Richard

2009/11/25 Martijn van Oosterhout <kleptog at gmail.com>

> On Wed, Nov 25, 2009 at 12:50 PM, Richard Ive <richard at xanox.net> wrote:
> > Hi All,
> >
> > I've had an open street map database set up on a server for a while now.
> > I've just got two new servers that are much, much more powerful (16GB
> Ram.
> > Two quad cores..). What I'm trying to do is have the database on one
> > machine, and mapnik on the other.
>
> Nice machine.
>
> Two things: you did tune postgresql a bit, at least increasing the
> shared buffers to several hundred MB, right?
>
> Secondly, you can get much more information about what's going on by
> logging into the DB and doing:
>
> select * from pg_stat_activity;
>
> Then you can see what it's doing. IIRC it's not unusual for loading a
> whole days diff to take several hours. If you're really worried, use
> strace to attach to the process and verify it's not doing anything (it
> should be reading/writing like crazy).
>
> Hope this helps,
> --
> Martijn van Oosterhout <kleptog at gmail.com> http://svana.org/kleptog/
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/dev/attachments/20091126/9f5cbd0a/attachment.html>