**** BEGIN LOGGING AT Sat Feb 09 02:59:58 2013
Feb 09 03:07:06 <fcooper> has anyone had bitbake get stuck in a loop and spawn 30+ bitbake children ?
Feb 09 03:07:36 <fcooper> This happens even if I run "bitbake" from the build directory
Feb 09 03:08:34 <nerdboy> um, no...
Feb 09 03:08:59 <nerdboy> maybe your setenv/local.conf is goofy?
Feb 09 03:10:35 <nerdboy> check the BB_ENV_EXTRAWHITE export and make sure it includes SCRATCH
Feb 09 03:13:15 <fcooper> no it does not
Feb 09 03:14:58 <nerdboy> i don't think it's strictly required, but if you want oebase and scratch/downloads in separate trees, you need to add scratch to that export
Feb 09 03:15:30 <nerdboy> otherwise it should still work with everything under oebase
Feb 09 03:15:47 <nerdboy> as long as you have write privileges there...
Feb 09 03:15:58 <fcooper> Having them together is fine with me
Feb 09 03:16:48 <nerdboy> well, if your local.conf isn't fubar'd then it should work with pretty much as-is
Feb 09 03:16:55 <fcooper> I placed some echos within oe-core/scripts/bitbake and I see that it is looping non stop
Feb 09 03:16:57 <nerdboy> s/with//
Feb 09 03:17:59 <nerdboy> can you pastebin your setenv and local.conf?
Feb 09 03:18:22 <nerdboy> and i assume you have the toolchain in the front of you path?
Feb 09 03:18:56 <fcooper> http://www.hastebin.com/qasegitobu.bash   http://www.hastebin.com/riheyomuve.vala
Feb 09 03:19:12 <fcooper> Yep toolchain is in the front of my path
Feb 09 03:24:20 <nerdboy> are you trying to build as a jenkins task?
Feb 09 03:25:04 <nerdboy> also, have you tried reducing bbthreads and -jN to something a little smaller?
Feb 09 03:25:47 <fcooper> Yep. I have a automated build that has been working for 8+ months. Now I've logged in and try to run the jenkins commands manually
Feb 09 03:26:22 <nerdboy> i build in jenkins every night, but i have some fairly elaborate jenkins job and build scripts that do a full git clone of all the recipe trees and build scripts
Feb 09 03:26:37 <nerdboy> ah
Feb 09 03:26:38 <fcooper> I think i found the problem...
Feb 09 03:26:42 <nerdboy> so what changed?
Feb 09 03:27:06 <fcooper> Well where the problem is located. in oe-core/scripts/bitbake this line: 
Feb 09 03:27:06 <fcooper> PSEUDOBINDIR=`bitbake -e | grep STAGING_BINDIR_NATIVE=\" | cut -d '=' -f2 | cut -d '"' -f2`
Feb 09 03:27:40 <fcooper> The script gets stuck there. What change was migrating from denzil to danny
Feb 09 03:36:49 <fcooper> Hmm its actually this line "PSEUDO_BINDIR=$PSEUDOBINDIR PSEUDO_LIBDIR=$PSEUDOBINDIR/../lib/pseudo/lib PSEUDO_PREFIX=$PSEUDOBINDIR/../../ PSEUDO_DISABLED=1 $PSEUDOBINDIR/pseudo $BITBAKE $@"  At the very end of the oe-core/scripts/bitbake.
Feb 09 03:37:38 <fcooper> I added a echo right before and after that statement and run bitbake and the console keeps on spitting out the the first echo statement and never the second one
Feb 09 03:44:00 <nerdboy> weird...
Feb 09 03:44:36 * nerdboy 's wife is being distracting on purpose
Feb 09 03:49:00 <nerdboy> better send her out for pizza...
Feb 09 03:49:31 <fcooper> Kill two birds with one stone
Feb 09 03:51:24 <nerdboy> so do you have a cluster for that setup?
Feb 09 03:51:46 <fcooper> Its a 24 core machine
Feb 09 03:52:14 <nerdboy> my build server has 8 cores and 16 GB of memory and it gets saturated with a lot less than that...
Feb 09 03:53:30 <nerdboy> parallel make doesn't actually help much except on a few big jobs
Feb 09 03:55:05 <nerdboy> but -j7 and 4 bbthreads is somewhat less than fully loaded and takes about 2:45 hrs to build our release image
Feb 09 03:55:54 <fcooper> I just enjoy not having to wait 5 hours to build qt anymore
Feb 09 03:57:12 <nerdboy> qt/webkit/firefox/oo are all painful
Feb 09 03:57:32 <nerdboy> especially on older/low-end boxes
Feb 09 03:58:12 <nerdboy> <- gentoo geek
Feb 09 03:58:56 <nerdboy> seems like i need to spec a new build server...
Feb 09 03:59:19 <fcooper> Can never have too many or too fast build servers
Feb 09 03:59:28 <fcooper> *cant
Feb 09 03:59:38 <nerdboy> see any random package failures?
Feb 09 04:00:18 <nerdboy> i was getting a consistent python failure with different -jN/thread values
Feb 09 04:00:32 <nerdboy> restart and it builds fine
Feb 09 04:01:39 <fcooper> There was one recipe that had a failure and it was due but we implemented a fix and submitted it to oe-core awhile ago. Other than that one we never had a problem
Feb 09 04:02:46 <nerdboy> i haven't seen that failure since i changed it
Feb 09 04:03:56 <nerdboy> *changed parallel make and bbthreads
Feb 09 04:06:53 <fcooper> Hmm that is weird
Feb 09 04:07:37 <fcooper> I duplicated my denzil build in one directory and then my danny build in another directory. setenv, local.conf and bblayers.conf are exactly the same
Feb 09 04:07:53 <fcooper> Denzil build doesn't have that bitbake issue while the danny build does
Feb 09 04:09:22 <nerdboy> okay, you win
Feb 09 04:09:32 <nerdboy> that;s weirder
Feb 09 04:12:17 <fcooper> Even weirder i switches the repos in sources to the danny branches and it worked 
Feb 09 04:12:22 <fcooper> *switched
Feb 09 04:13:59 <nerdboy> fire up the espresso machine and diff the branches
Feb 09 04:26:59 <nerdboy> damn, i need to upgrade the toaster motherboard
Feb 09 04:28:04 <nerdboy> it gets weirdly cpu-bound during when some machines write tarballs over nfs
Feb 09 04:28:19 <nerdboy> s/during//
Feb 09 04:29:04 <nerdboy> keeps spiking to 100%
Feb 09 13:56:42 <Shinda> hello
Feb 09 16:55:42 <nerdboy> moin
Feb 09 17:05:21 <fcooper> nerdboy: I finally found the problem...
Feb 09 17:05:38 <nerdboy> and?
Feb 09 17:06:30 <fcooper> 1 sec. I am posting a bug and fix to bugzilla. I'll send it to you
Feb 09 17:07:54 * nerdboy is still making a tablet-abuse-arm-build-host
Feb 09 17:14:10 <fcooper> nerdboy: https://bugzilla.yoctoproject.org/show_bug.cgi?id=3872
**** ENDING LOGGING AT Sun Feb 10 02:59:58 2013