Am Montag, 15. Oktober 2007 03:06:10 schrieb Alex Cruz:
> On Sunday 14 October 2007, Joerg Ostertag wrote:
> > ...
> >
> > > Writing way(66k)mapnik-osm-updater.sh: line 318: 677 Segmentation
> > > fault sudo -u "$osm_username" osm2pgsql --database "$database_name"
> > > $planet_file
> > >
> > > Any thoughts as to why this is happening?
...
> -svn up updated with no errors
> -compiled osm2pgsql with no errors
> -I have 2GB of RAM
> -downloaded the planet file (several times)
>
> What I have noticed is that regardless if the planet file is compressed or
> uncompressed (I didn't realize it was 12GB uncompressed),
Yes, it't growing really fast ...
> the process
> always dies when it hits the "Writing way(66k)".
So this points to an consistent error.
> Also, regardless if it's
> the only process running, it consumes all available RAM and starts to swap
> a little, too.
With your 2GB this shouldn't happen. Last time I remember it should consume
less than 1GB of RAM.
> The -s option didn't work either.
I havn't tried any of these options yet.
> Since I have never used gdb to debug, what parameters should I use to
> capture the information? Thanks.
I hope that a simple backtraced (bt) would help.
Start gdb with
gdb --args osm2pgsql planetfile.bz2
run the programm with
r
and as soon as the SIGSEG is seen type in
bt
then cut 'n paste the result and let's see
One more question: Which architecture has you CPU? x86 or different?
-
Joerg
|
