--- Log opened Tue Sep 29 00:00:33 2009
04:24 < Dunearhp_> does anyone know what might cause a happstack instance to max out the cpu?
04:24 < Dunearhp_> i.e. any common culprits
06:07 < Dunearhp_> gah. compiling profiling libraries is painful
06:07 < mightybyte> It really depends on the details of what your app is doing.
06:08 < Dunearhp_> almost nothing, static files + hstringtemplate
06:08 < Dunearhp_> it just started chewing up cpu after runninge for a few hours while I played with my templates
06:09 < mightybyte> And you're using fileServe for the static files?
06:09 < Dunearhp_> yes
06:10 < mightybyte> Hmmm, there have been some issues with fileServe, but some work has been done on that recently.
06:10 < mightybyte> I haven't kept track of it as much since I've never had any problems.
06:11 < mightybyte> You may want to try to distill the problem down to a minimal test case and file a bug report.
06:11 < Dunearhp_> I might be a few weeks behind, when I have profiling in place I'll be able to find the culprit if it comes up again
06:11 < mightybyte> ...or if you don't want to put in a bug report, ask the question on the mailing list.
06:13 < Dunearhp_> at the moment I don't have any specifics beyond that it just started misbehaving while I was editing template files and reloading
06:13 < mightybyte> Ok.
06:13 < Dunearhp_> I didn't even notice the exact time that it started
06:16 < mightybyte> I personally don't make heavy use of static files, and I also don't use hstringtemplate, so without a concrete test case, I can't help you much.
06:16 < Dunearhp_> do you use templating, or generate from haskell?
06:16 < mightybyte> This IRC channel has low enough traffic that it's not always the best place to get your problems solved.  But the mailing list generally gets pretty timely responses.
06:16 < mightybyte> Generate
06:17 < Dunearhp_> I really should get on the list, I've been playing with this for a couple of weeks now
06:17 < mightybyte> You'll only get help here if the right people are around when you ask your question.
06:17 < Dunearhp_> it's fine, bugs can be an opportunity to learn
06:43 < Dunearhp_> I take it the mailing list is the HAppS google group
08:07 < mightybyte> Dunearhp_: Yes.
08:12 < burp> it's also available via gmane
08:12 < burp> gmane.comp.lang.haskell.happs
10:07 < stepcut> Dunearhp_: is the app compiled be being run via runhaskell?
10:08 < stepcut> Dunearhp_: possibly, http://hackage.haskell.org/trac/ghc/ticket/3408
10:10 < Dunearhp_> It was precompiled
10:11 < Dunearhp_> but that may be something for me to investigate
10:13 < stepcut> I have seen happstack apps suck up CPU like that on more than one occasion. But I have not seen anything in the code that would cause it aside from making the GC run too often
10:13 < stepcut> http://hackage.haskell.org/trac/ghc/ticket/2891
10:18 < Dunearhp_> I want to run this on a virtual server, so I might have to monitor it with cpulimit until I can avoid the problem
10:19 < Dunearhp_> it is possible that I triggered a garbage collection issue with the frequency with which I was making queries
10:28 < stepcut> Dunearhp_: dunno. Try the -I1 or -I0 option and see if helps though
10:36 < Dunearhp_> will do
--- Log closed Wed Sep 30 00:00:21 2009