[Tilecache] Problem with tilecache_seed.py

[Christopher Schmidt]

On Tue, Nov 11, 2008 at 11:20:16AM -0500, Alexandre Dube wrote:
> In fact, it does get beyond.  Before, using MapServerLayer type, the 
> script crashed after around 1 minute of running.  So, I was starting it 
> again and it quickly returned where it crashed, continued for around 30 
> seconds then crashed again.

Yes, I know that. Before you were suffering memory leaks. This is not a
memory leak, but it may still be a leak of a different kind. The
question is, is this also a fixed number of 'new' tiles generated before
the crash? is it more or less than under the mapserver configuration? 

It's possible that for some reason, something goofy got temporarily
screwed up, but I suspect resource allocation problems of some sort.

> So there may be a problem with my file system on my server, you think ?  
> It could be that, because I'm currently running the script on my own pc 
> and it hasn't crashed yet ( using the same environment, libraries, same 
> MapServer version, using MapServer type, jpg output but on shapfiles 
> though...  Guess I'll have to try those ECW also... )  So far, it's 
> looking good locally...

I would guess that ECW is the reason you're experiencing this. Anything
which isn't ECW is likely to be better tested, since (as I understand
it) ECW is not actually open source, and is therefore a less commonly 
used path. (So, success without ECW is a useless metric.)

Regards,
-- 
Christopher Schmidt
MetaCarta