[postgis-users] Server not responding

[Andreas Forø Tollefsen]

Hi Andy,

There is no errors in the log. Also, the cpu load is 100 %. Looking at the
pg_stat_activity show that the query is still running.
The locks are also there.
Maybe I should reinstall everything. As i said, this query has worked
before, and still do. It just stops in the loop at random steps in the
loop.
Another thing. After I terminate the frozen query, I cannot run the script
again. It just stops in the first one or two years. Restarting the server
solves this. Some memory problems maybe?

2011/10/5 Andy Colson <andy at squeakycode.net>

> On 10/5/2011 8:52 AM, Andreas Forř Tollefsen wrote:
>
>> Hi all,
>>
>> I have a query that previously have been running without problems.
>> However, after installing everything from scratch it no longer work.
>> Ubuntu 11.04, Postgresql 8.4.8, PostGIS 2.0 (7932), Geos 3.3.1, GDAL
>> 1.8.1.
>>
>> The script measures distances from centroids of a vector grid to the
>> nearest international border to a neighboring country.
>>
>> It uses psycopg2 module in python to loop through the years. So it
>> inserts the distances for each year into the borddist table.
>>
>> However, the problem after the reinstall is that it usually loops
>> through some years, and then everything stops. The usual processing time
>> is ~20 minutes per year.
>> Could this be a memory issue in postgresql.conf? Or is there a problem i
>> am not aware of? This used to work on the previous installation, but we
>> wanted to update the server to latest.
>>
>>
> Can you check the PG logs and see if there are errors?  When you say
> everything stops, is there a postgres process still running?  Is it blocked?
>
> Have you tried:
>
> select * from pg_locks;
> and
> select * from pg_stat_activity;
>
> Is PG using CPU time?  or does it really crash?
>
> -Andy
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-users/attachments/20111006/e2c605bd/attachment.html>