500 Internal Server Errors

[Roland Martin]

Hi Luis,

Thanks for your thoughts - unfortunately it didn't make any difference
(anyway logically it shouldn't anyway because they're all in EPSG 27700 and
use OIDs as the unique column).

Interestingly it's very intermittent - I couldn't get the error to occur at
all yesterday. I've now created the layers afresh in PostGIS from an
entirely different procedure, and it's happening again.

It could also be the nature of the layers - I've now filtered out a number
of zero length line features, and there may be other topographical errors.

Cheers,
Roland.



On 13/07/06, Luis Treviño <ltrevinoh at gmail.com> wrote:
>
> Roland,
>
> Try this on your DATA stmt:
>
> DATA "geometry from background as foo using unique <gid> using srid=<your
> srid>"
> just replace the <gid> name with your oid column and put the number of the
> srid you're using.
> Hope this helps.
>
> Regards,
> Luis
>
>
>
> 2006/7/13, Roland Martin <roland.w.martin at gmail.com>:
>
> > Hi Luis,
> >
> > Thanks for your reply. I don't think that's the problem, but here are
> > some excerpts nonetheless:
> >
> >
> >
> >   LAYER                     # this one works
> >     NAME "Contours"
> >     GROUP "Height"
> >     STATUS OFF
> >     CONNECTIONTYPE postgis
> >     CONNECTION "dbname=lba user=postgres password=postgres"
> >     DATA "geom from height_data"
> >     TYPE LINE
> >     LABELITEM "height"
> >     CLASS
> >       EXPRESSION ('[layer_name]' = 'C900LX000')
> >       COLOR 204 51 0
> >       LABEL
> >         FONT sans
> >         SIZE small
> >         COLOR 100 100 100
> >         OUTLINECOLOR 255 255 255
> >         POSITION auto
> >         ANTIALIAS false
> >       END
> >     END
> >   END
> >
> >   LAYER                           # this one doesn't
> >     NAME "Background Drawing"
> >     GROUP "CAD Drawings"
> >     STATUS OFF
> >     CONNECTIONTYPE postgis
> >     CONNECTION "dbname=lba user=postgres password=postgres"
> >     DATA "geometry from background"
> >     TYPE LINE
> >     CLASS
> >       COLOR 51 102 0
> >     END
> >   END
> >
> > The geometry column names are different because I generated the tables
> > from 2 different sources - but that shouldn't make any difference...
> >
> > Cheers,
> >
> > Roland.
> >
> >
> >
> >
> > On 13/07/06, Luis Treviño <ltrevinoh at gmail.com > wrote:
> > >
> > >  Roland:
> > >
> > > Could be the "DATA" sql statement on your map file. It would be easier
> > > if you can show part of your map file that involves the postgis layer.
> > >
> > > Regards,
> > > Luis
> > >
> > >
> > > 2006/7/12, Roland Martin <roland.w.martin at gmail.com>:
> > >
> > > > OK, this is probably yet another stupid query to which I should know
> > > > the answer... however, I'm on a roll here.
> > > >
> > > > I've got an application which is intermittently (i.e. most of the
> > > > time, but not always) throwing a 500 Internal Server Error. The application
> > > > is almost a direct clone of another which does not exhibit the same error.
> > > >
> > > > I've spent most of the afternoon trying to figure out why this might
> > > > be, and as far as I can tell, it's something to do with the PostGIS database
> > > > which holds most of the data. If I remove all the PostGIS layers, it seems
> > > > to work fine (although interestingly doesn't work if they're all in the map
> > > > file but turned off - does MapScript try to load the layers regardless?)
> > > >
> > > > The database contains converted CAD (DWG) data with typically stupid
> > > > contents (meaningless fields, escape characters, etc). I wondered if it
> > > > might be to do with the sizes of the tables (>30 000 records in each), but
> > > > dissolving them doesn't seem to have made much difference. At the moment I'm
> > > > leaning towards thinking it may be because of the aforementioned escape
> > > > characters (looking down the tables, a number just contain the entry " which
> > > > might upset things; others start with an asterisk). However, some of the
> > > > tables causing the issue don't contain anything nearly as nasty.
> > > >
> > > > The error is being thrown on loading the map; approximately 80-90%
> > > > of the time.
> > > >
> > > > Also of interest may be the following gunk from the error logs:
> > > > [Wed Jul 12 16:29:11 2006] [error] [client x] Premature end of
> > > > script headers: php-cgi.exe, referer: http://x/lbia_dev/
> > > > [Wed Jul 12 16:51:08 2006] [error] [client x] Premature end of
> > > > script headers: php-cgi.exe, referer: http://x/lbia_dev/
> > > > (etc)
> > > >
> > > > Anyone have any thoughts on this?
> > > >
> > > > Ta,
> > > > Roland.
> > > >
> > > >
> > > > MS4W 2.0
> > > > MapServer 4.8.3
> > > > Chameleon for MS4W 2.4
> > > > PgSQL 8.1
> > > > etc 1.0
> > > >
> > > >
> > > >
> > > > --
> > > > Please note my new email address. Use other addresses at your peril.
> > > >
> > > >
> > >
> > >
> > >
> >
> >
> >
> > --
> > Please note my new email address. Use other addresses at your peril.
> >
>
>


-- 
Please note my new email address. Use other addresses at your peril.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/mapserver-users/attachments/20060714/3b34bc2d/attachment.htm>