[mapserver-users] FastCGI issues with MapServer 5.2, Ubuntu 8.04

Andrea Aime aaime at opengeo.org
Fri Aug 22 04:21:03 EDT 2008

Andrea Aime ha scritto:
> Andrea Aime ha scritto:
>> Hi,
>> I'm struggling a trying to configure MapServer 5.2
>> for the performance comparison presentation that we'll
>> hold at FOSS4G 2008
>> (http://conference.osgeo.org/index.php/foss4g/2008/paper/view/256)
>> In particular, no matter what I do, a single, long running
>> mapserv process seems to be handling all of the requests
>> coming down the pipe, regardless of how many client threads
>> I'm using to hit the server. Some details on what I've
>> done so far follow.
> To answer myself, that was happening I added the
> directive into a shapefile layer definition. It
> seems it's not exactly irrelevant, it's actually
> harmful (whilst it's beneficial, and works too,
> for postgis layers).
> When I removed it the fastcgi processes started to get
> spawn as configured in the fastcgi.conf file.

This morning I started benchmarking again and ended up with a single
fastcgi process again, so the above is probably not the cause.
What made apache understand I want 20 processes instead of just one
remains a mystery, it has magically disappeared as it magically
appeared yesterday.

Anyone has a clue? Is fastcgi supposed to be usable in production
under Ubuntu servers, are there known issues/instabilities? (this may
explain why out of the box mapserver is not built with fastcgi
support enabled in Ubuntu).

In the meantime I'm going to try out the FastCgiServer directive
described here: 
hoping it proves more reliable.


More information about the mapserver-users mailing list