Experimental IRC log happs-2009-06-10

Available formats: content-negotiated html turtle (see SIOC for the vocabulary)

Back to channel and daily index: content-negotiated html turtle

These logs are provided as an experiment in indexing discussions using IRCHub.py, Irc2RDF.hs, and SIOC.

00:44:33<dsfox>I guess I'll stick to debianizing everything I use
02:56:20<mightybyte>Yeah, linode.
14:11:08<kowey>is this something that affects all happstack users? http://hackage.haskell.org/trac/ghc/ticket/698
14:20:24<dcoutts>kowey: you mean, do people find it a serious problem in their specific applications?
14:20:57<kowey>right
14:21:18<dcoutts>obviously it particularly affects apps that inflate their mem use for individual requests
14:21:30<dcoutts>and not all web apps have that characteristic, indeed most do not
14:22:22<dcoutts>kowey: so eg the hackage-server doesn't have that problem
14:23:03<dcoutts>because while the state uses quite a bit of memory, individual requests do not need that much temporary storage
14:24:35<kowey>whereas gitit is affected when (guessing wildly) it has to parse a huge XML file from darcs changes, and for some reason wants to grab lots of memory to do it?
14:24:40<kowey>(this is idle speculation, of course)
14:33:49<kowey>so what I'm scratching my head over (without really being equipped for it) is what's causing these gitit OOMs on darcs.net
14:34:55<kowey>gwern mentioned this bug as one of the things he noticed... but where I can see how that would prevent our memory usage from ever dropping
14:36:08<kowey>if our individual requests don't get progressively larger, and gitit/filestore itself doesn't have some sort of leak, then no real problem, right?
15:10:11<dcoutts>kowey: right, this issue that ghc does not release memory does not affect OOM, because that's about peak usage
15:11:54<dcoutts>so if you're getting OOM you need to look for a different cause, ie why serving one request takes so much memory
15:13:05<kowey>thanks! (and sorry for the confused rambling)
15:13:21<kowey>I also just noticed I haven't been reading that bug carefully... just saw this from gwern:
15:13:29<kowey>"There are a few pages like 'Recent changes' which get the entire revision history; each time this happens, the memory usage goes up a bit..."

Back to channel and daily index: content-negotiated html turtle