2009/2/20 simon <[email protected]>: > > On Feb 19, 9:25 am, Graham Dumpleton <[email protected]> > wrote: >> BTW, are you using a VPS. Wander is related issue to one I know about, >> resulting from how VPS handles virtual memory allocations. > > It's a dedicated server. > >> > BTW, are you attempting to use WSGIImportScript to preload your WSGI >> > script and therefore Django code modules? > > No, I haven't tried that. But I assume traffic causes all the modules > to load up pretty quick?
Preloading will eliminate initial requests handled by a process being stalled while loading and initialisation occurs. Real busy now, but will explain later some more. We do though need to work out a way to force Django to do all its own initialisation at the preload time. By default is still does it initialisation in a lazy fashion on first request. I sort of remember working this out before for different issue and details in the list archive somewhere. Graham > The funny thing with the ThreadStackSize finding was that even > beforehand, the server seemed fine. Only after a day or so and a run > of burst traffic would it start to buckle. Every CPU would just spin > apache2 on 100%, with load average up over 30. I first tried reducing > processes and threads back to the bare minimum we needed to keep > plenty of memory left spare, but it'd still spin out and eat up stacks > of memory and grind the CPUs to a halt. Maybe it's the app, but it's > been up for nearly a week now with ThreadStackSize set lower and it's > doing fine. --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "modwsgi" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/modwsgi?hl=en -~----------~----~----~----~------~----~------~--~---
