apache memory gradually increasing over time with python/django/tastypie -
i'm getting gradual increase of memory on time in apache running django/tastypie. traffic increase, expect see gradual decrease of memory on time.
<ifmodule mpm_worker_module> startservers 6 minsparethreads 12 maxsparethreads 24 threadlimit 48 threadsperchild 24 maxclients 24 maxrequestsperchild 0 maxmemfree 1024 </ifmodule> daemon process:
wsgidaemonprocess www.domain.com processes=24 threads=24 reports new relic


is thinking flawed? if memory stables , goes linear, don't care if memory frees on time long never goes past threshold. have 2 load balanced machines , both same setup , doing same thing.
everything running great otherwise , super fast.
thanks, mark
one day of runtime low number of requests makes hard tell if you're leaking memory. me graphs reasonable, there many variables it's hard tell certain.
if you're concerned leaks can try using maximum-requests wsgidaemonprocess option restart processes after serve many requests.
also, need 24 wsgi processes? each 1 have own python interpreter, , on time consume 24 * (1 python/django/tastypie stack). i'd suggest load testing , reducing number of processes accordingly. check out this post graham dumpleton.
Comments
Post a Comment