Hi, There is a problem for autobuilder on OpenSuse 13.1 while building world image. Bitbake is crushing during a task (it doesn't matter which one) with error: ERROR: Fakeroot process (28226) exited unexpectedly (-9), shutting down... I looked in dmesg and saw the following error: [ 3854.585762] Out of memory: Kill process 2921 (python) score 45 or sacrifice child [ 3854.585765] Killed process 28226 (python) total-vm:291460kB, anon-rss:38856kB, file-rss:512kB Once again, this is happening only on OpenSuse 13.1 Autobuilder commit used is: b6f4f74a7475e7180a8e8d1d51b6a4dfb563706d Cheers, Ionut C
Ross will chase this down with IonutC
Hello, For the same OS and same world image the second main error why it crashes is this: NOTE: recipe sudo-1.8.10p3-r0: task do_install: Started Traceback (most recent call last): File "/home/autobuilder/yocto-autobuilder/yocto-worker/nightly-world/build/bitbake/bin/bitbake-worker", line 361, in <module> worker.serve() File "/home/autobuilder/yocto-autobuilder/yocto-worker/nightly-world/build/bitbake/bin/bitbake-worker", line 269, in serve self.build_pipes[pipe].read() File "/home/autobuilder/yocto-autobuilder/yocto-worker/nightly-world/build/bitbake/bin/bitbake-worker", line 218, in read worker_fire_prepickled(self.queue[:index+8]) File "/home/autobuilder/yocto-autobuilder/yocto-worker/nightly-world/build/bitbake/bin/bitbake-worker", line 54, in worker_fire_prepickled worker_flush() File "/home/autobuilder/yocto-autobuilder/yocto-worker/nightly-world/build/bitbake/bin/bitbake-worker", line 63, in worker_flush written = os.write(worker_pipe, worker_queue) OSError: [Errno 32] Broken pipe [Errno 32] Broken pipeTraceback (most recent call last): File "/home/autobuilder/yocto-autobuilder/yocto-worker/nightly-world/build/bitbake/bin/bitbake-worker", line 368, in <module> worker_flush() File "/home/autobuilder/yocto-autobuilder/yocto-worker/nightly-world/build/bitbake/bin/bitbake-worker", line 63, in worker_flush written = os.write(worker_pipe, worker_queue) OSError: [Errno 32] Broken pipe command timed out: 100000 seconds without output, attempting to kill Costin
I wrote a small script to monitor memory usage to see if we can work out what is eating all the memory.
Hello, I just finished a distro testing session and for this time with commits: poky 52b788c6df9201764130ab744bf67b770b24b896 meta-qt3 3016129d90b7ac8517a5227d819f10ad417b5b45 no issues encountered. So for now on comment#3 I have nothing to add. I will wait for a new weekly testing and will add results then (if bug is reproducing), otherwise I will mark this bug as resolved. Best regards, Costin
Comment #4 shows this is in review.
Nobody has been able to reproduce this for a while, so closing.