On Jun 28, 2010, at 9:18 AM, yvecai wrote:

> 
>>> The problem I have is that some tiles are properly rendered, but other are 
>>> not, like in this example: 
>>> http://yvecai.olympe-network.com/divers/Capture.png
>>>     
>> That is because each request is likely generating a child thread that is 
>> trying to share the same mapnik map, which is not good behavior, unless your 
>> python code is deployed as multiprocess (rather than multithreaded). If you 
>> are interested in multiprocess deployment see mod_wsgi, run in daemon mode, 
>> with threads=1 and processes=N CPU
>>   
> Thank you for your support Diane,

Dane actually. 

> I suspected something like that, but I'm far from expert in this field. 
> Actually, I just wanted a local real-time rendering method to test rendering 
> and I choose Sylvain's method because it was the most documented. I found the 
> readme.txt from mod_tile is a little scary (any chance that I can make it 
> work??).

Yes, I agree that the mod_tile readme doc needs improvement, but its not 
impossible. Just give it a try and let us know where exactly you get stuck.

> 
> At the end, the setup is working fine from yesterday evening.

huh? that is confusing.

> I tried to set up mod_tile with no luck, then I switch back to mod_python and 
> I had just run generate_xml.py again to change some rendering rules.

Running generate_xml.py will have no effect on this problem AFAIK.

> That is not thread or processes related!

I'm pretty sure it is threading related.

> When I switch back to mod_python method, I probably tidied up the apache 
> config files or something like that.

Well, in that case, maybe you could try to figure out what you changed and 
report back?

> 
> Yves


_______________________________________________
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev

Reply via email to