Re: [Openstack] Lossage in nova test suite?
Ok, I take that back. I do see one issue. It looks like each test run is leaving 2 stuck runner.py's for me… even if the tests complete successfully. On Jun 4, 2012, at 4:57 PM, Chris Behrens wrote: The only thing I notice is about a 50% increase in the unit test run time very recently… I don't know when that started.. maybe today. Not seeing the 100s of python processes. Hm! - Chris On Jun 4, 2012, at 4:00 PM, Kevin L. Mitchell wrote: Today I've noticed some significant problems with nova's test suite leaving literally hundreds of python processes out. I'm guessing that this has to do with the unit tests for the multiprocess patch, which was just approved. This could be causing problems with jenkins, too… Anybody have any other insights? -- Kevin L. Mitchell kevin.mitch...@rackspace.com ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] Lossage in nova test suite?
Sorry, replying too soon before fully investigating. They *are* forking over and over again repeatedly as fast as possible… making them hard to kill. They fork a child and the parent exits immediately.. and that repeats endlessly. My VM is too slow right now that even a tight pkill loop doesn't hit them before they've forked again. http://paste.openstack.org/show/18338/ Only recourse may be to reboot my VM. On Jun 4, 2012, at 4:57 PM, Chris Behrens wrote: The only thing I notice is about a 50% increase in the unit test run time very recently… I don't know when that started.. maybe today. Not seeing the 100s of python processes. Hm! - Chris On Jun 4, 2012, at 4:00 PM, Kevin L. Mitchell wrote: Today I've noticed some significant problems with nova's test suite leaving literally hundreds of python processes out. I'm guessing that this has to do with the unit tests for the multiprocess patch, which was just approved. This could be causing problems with jenkins, too… Anybody have any other insights? -- Kevin L. Mitchell kevin.mitch...@rackspace.com ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] Lossage in nova test suite?
Should we revert this change till we get it cleared up? On 6/4/12 8:29 PM, James E. Blair cor...@inaugust.com wrote: On 06/04/2012 04:00 PM, Kevin L. Mitchell wrote: Today I've noticed some significant problems with nova's test suite leaving literally hundreds of python processes out. I'm guessing that this has to do with the unit tests for the multiprocess patch, which was just approved. This could be causing problems with jenkins, tooŠ Anybody have any other insights? Yes, several Jenkins slaves have been taken out by running nova unit tests. The one that I am able to log into seems to be continuously respawing python processes. Other slaves are inaccessible due to having exhausted their RAM. I note that all of the tests run after that change merged carry this warning notice from Jenkins: Process leaked file descriptors. See http://wiki.jenkins-ci.org/display/JENKINS/Spawning+processes+from+build for more information So I think it's fair to say that Jenkins corroborates your suspicion that change introduced a problem with leaking processes. This is affecting any Jenkins slave that the nova unit tests job runs on, which in turn affects jobs for unrelated projects that happen to later run on that slave. In addition to correcting this problem, I believe we should add a build step to Jenkins to ensure that all of the test processes have terminated correctly. -Jim ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] Lossage in nova test suite?
On Mon, Jun 4, 2012 at 5:47 PM, Gabe Westmaas gabe.westm...@rackspace.comwrote: Should we revert this change till we get it cleared up? +1 On 6/4/12 8:29 PM, James E. Blair cor...@inaugust.com wrote: On 06/04/2012 04:00 PM, Kevin L. Mitchell wrote: Today I've noticed some significant problems with nova's test suite leaving literally hundreds of python processes out. I'm guessing that this has to do with the unit tests for the multiprocess patch, which was just approved. This could be causing problems with jenkins, tooŠ Anybody have any other insights? Yes, several Jenkins slaves have been taken out by running nova unit tests. The one that I am able to log into seems to be continuously respawing python processes. Other slaves are inaccessible due to having exhausted their RAM. I note that all of the tests run after that change merged carry this warning notice from Jenkins: Process leaked file descriptors. See http://wiki.jenkins-ci.org/display/JENKINS/Spawning+processes+from+build for more information So I think it's fair to say that Jenkins corroborates your suspicion that change introduced a problem with leaking processes. This is affecting any Jenkins slave that the nova unit tests job runs on, which in turn affects jobs for unrelated projects that happen to later run on that slave. In addition to correcting this problem, I believe we should add a build step to Jenkins to ensure that all of the test processes have terminated correctly. -Jim ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] Lossage in nova test suite?
On 06/04/2012 05:47 PM, Gabe Westmaas wrote: Should we revert this change till we get it cleared up? Here's a proposal to do that: https://review.openstack.org/#/c/8166/ -Jim ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] Lossage in nova test suite?
On Mon, Jun 04, 2012, James E. Blair cor...@inaugust.com wrote: On 06/04/2012 05:47 PM, Gabe Westmaas wrote: Should we revert this change till we get it cleared up? Here's a proposal to do that: https://review.openstack.org/#/c/8166/ I approved it because of how bad Jenkins is right now, but it probably won't merge because the gate-nova-python27 job fails immediately right now. Jenkins needs some TLC it seems to get things back to where stuff can get merged again. JE ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp