[OSM-dev] osm2pgsql segmentation fault

[NopMap]

Hi!

Ok, here we go again without markup:

The osm2pgsql on my server, which has been working nicely for a long time,
is now crashing with a segmentation fault every time and I would like to aks
you for some help and ideas on how to find the reason.

Operationg system is a 64bit debian. The osm2pgsql was compiled from the
then-current SVN on 17th of october and the binary has worked without
problem for about a month. It identifies itself as

osm2pgsql SVN version 0.81.0 (64bit id space)

The only thing strange was that any attempts to run osm2pgsql with multiple
processes resulted in data loss, with the current build as well as with the
previous one, but calling it with 1 process has always worked so far.

The server environment (database, memory, command line, etc.) has not been
changed for about a year and has always worked well. I have tried osm2pgsl
repeatedly and the crash happens always at about the same point of the
import. I have tried re-downloading the osm data and trying it with the new
data, but the crash still happens at about the same time. It does not look
like a memory problem to me, the memory consumption of the server at the
time of the crash is only even half of the what is used during a normal,
working run.

The crash message is

22547 Broken pipe             gzip -d -c /osm/composer/input/update.osm.gz
     22548 Segmentation fault      | /osm/tools/osm2pgsql/osm2pgsql -c
--slim -d topo -p data -C 4000 --number-processes=1 -S
/osm/mapnik/topo_import.style /dev/stdin &>/osm/import.log


The last output of osm2pgsql is

Processing: Node(627330k 251.3k/s) Way(0k 0.00k/s) Relation(0 0.00/s)
Processing: Node(627340k 251.3k/s) Way(0k 0.00k/s) Relation(0 0.00/s)
Processing: Node(627350k 251.3k/s) Way(0k 0.00k/s) Relation(0 0.00/s)
Processing: Node(627360k 251.3k/s) Way(0k 0.00k/s) Relation(0 0.00/s)


Has anybody observed similar behaviour? Any ideas on what this might be?

Do you have a current debian binary I might try?

As I am not a linux/C developer, I am at a loss on how to proceed with a
further analysis to find the exact code location or the exact data item that
is causing the crashes. Any hints would be very welcome.

thanks
             Nop 



--
View this message in context: http://gis.19327.n5.nabble.com/osm2pgsql-segmentation-fault-tp5736195p5736201.html
Sent from the Developer Discussion mailing list archive at Nabble.com.