| *** derekh has quit IRC | 01:09 | |
| *** jcooley_ has quit IRC | 02:25 | |
| *** dkehn_ has quit IRC | 03:50 | |
| *** dkehn_ has joined #openstack-ironic | 03:50 | |
| *** anniec has quit IRC | 04:09 | |
| *** dkehn__ has joined #openstack-ironic | 04:17 | |
| *** dkehn has quit IRC | 04:20 | |
| *** jcooley_ has joined #openstack-ironic | 04:20 | |
| *** dkehn_ has quit IRC | 04:21 | |
| *** dkehn has joined #openstack-ironic | 04:27 | |
| openstackgerrit | Yongli He proposed a change to openstack/ironic: Use six.moves.urllib.parse instead of urlparse https://review.openstack.org/72889 | 04:52 |
|---|---|---|
| *** jcooley_ has quit IRC | 04:54 | |
| *** jcooley_ has joined #openstack-ironic | 04:58 | |
| openstackgerrit | Yongli He proposed a change to openstack/ironic: Remove extraneous vim configuration comments for horizon and tools https://review.openstack.org/73160 | 05:02 |
| *** dkehn__ is now known as dkehn_ | 05:10 | |
| *** jcooley_ has quit IRC | 05:21 | |
| *** jcooley_ has joined #openstack-ironic | 05:23 | |
| *** anniec has joined #openstack-ironic | 05:51 | |
| *** anniec_ has joined #openstack-ironic | 05:54 | |
| *** anniec has quit IRC | 05:56 | |
| *** anniec_ is now known as anniec | 05:56 | |
| openstackgerrit | Jenkins proposed a change to openstack/ironic: Imported Translations from Transifex https://review.openstack.org/71192 | 06:06 |
| *** killer_prince has joined #openstack-ironic | 06:21 | |
| *** jcooley_ has quit IRC | 07:00 | |
| *** jcooley_ has joined #openstack-ironic | 07:01 | |
| *** jcooley_ has quit IRC | 07:06 | |
| *** jcooley_ has joined #openstack-ironic | 07:18 | |
| *** jcooley_ has quit IRC | 07:25 | |
| *** dkehn__ has joined #openstack-ironic | 08:22 | |
| *** mdurnosvistov has joined #openstack-ironic | 08:23 | |
| *** dkehn has quit IRC | 08:25 | |
| *** dkehn_ has quit IRC | 08:25 | |
| *** dkehn has joined #openstack-ironic | 08:27 | |
| *** dkehn_ has joined #openstack-ironic | 10:54 | |
| *** dkehn__ has quit IRC | 10:57 | |
| *** dkehn has quit IRC | 10:57 | |
| *** dkehn has joined #openstack-ironic | 10:59 | |
| *** mdurnosvistov has quit IRC | 15:27 | |
| * NobodyCam makes coffee.. | 16:11 | |
| devananda | g'mornin | 16:17 |
| devananda | fairly sure i have a fix for our gate | 16:18 |
| devananda | I spun up a RH6 instance in HPcloud to test on, and can reproduce the gate failure now | 16:18 |
| devananda | not ever run fails, so I run them in a loop. that loop has not yet made it past 5 runs without failing | 16:19 |
| devananda | until this patch ... | 16:19 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Correct PXEPrivateMethodsTestCase.setUp https://review.openstack.org/73872 | 16:26 |
| devananda | NobodyCam: ^ | 16:26 |
| *** jbjohnso has joined #openstack-ironic | 16:26 | |
| * NobodyCam takes breath | 16:27 | |
| NobodyCam | :) | 16:27 |
| NobodyCam | and looks | 16:27 |
| devananda | lol | 16:27 |
| NobodyCam | morning devananda | 16:27 |
| devananda | g'morning! | 16:27 |
| devananda | gate failures were my fault, introduced in https://review.openstack.org/#/c/70468/ | 16:28 |
| NobodyCam | ahh without inheriting from | 16:28 |
| NobodyCam | DbTestCase | 16:28 |
| devananda | and not mocking the driver load | 16:28 |
| devananda | this is why we must mock stevedore in unit tests .... | 16:29 |
| devananda | interestingly, we initially mocked stevedore because it wouldn't load anything repeatably in a .venv (at least, not for us) | 16:29 |
| devananda | it just fails randomly now | 16:29 |
| NobodyCam | guess we can't really come up with a test for that can we? | 16:30 |
| devananda | i'll file the bug in stevedore and see what dhellmann thinks | 16:32 |
| NobodyCam | :) | 16:32 |
| devananda | for now, we need to not create a TaskManager instance in a unit test without first calling get_mockable_extension_manager | 16:32 |
| NobodyCam | ahh, think its worth an agenda item to recap? | 16:34 |
| devananda | oh, heh | 16:34 |
| devananda | actually https://review.openstack.org/#/c/70481/3/ironic/tests/drivers/test_pxe.py introduced the bug | 16:34 |
| devananda | part of that patch series | 16:34 |
| * devananda updates commit message | 16:34 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Correct PXEPrivateMethodsTestCase.setUp https://review.openstack.org/73872 | 16:35 |
| devananda | NobodyCam: probably better as small email | 16:35 |
| NobodyCam | :) | 16:35 |
| devananda | some folks have been affected who don't attend the meeting | 16:35 |
| NobodyCam | ya | 16:35 |
| devananda | NobodyCam: are you goign to be around tomorrow, even though it's a holiday? | 16:36 |
| NobodyCam | mostly.. :) | 16:37 |
| NobodyCam | may run out with John for a while | 16:38 |
| devananda | i'll be here :) | 16:38 |
| devananda | code jam and team meeting and stuff ... | 16:39 |
| devananda | i'll be largely offline tues/wed though, hopping down to CA for the day | 16:39 |
| NobodyCam | what time we doing review / code jam | 16:39 |
| NobodyCam | ack :) | 16:39 |
| devananda | 8am iirc | 16:39 |
| NobodyCam | :-p evil | 16:40 |
| *** aignatov_ has quit IRC | 16:46 | |
| *** datajerk1 has joined #openstack-ironic | 17:00 | |
| NobodyCam | woo hoo :) | 17:02 |
| *** pquerna_ has joined #openstack-ironic | 17:03 | |
| *** pquerna_ has joined #openstack-ironic | 17:03 | |
| NobodyCam | one +2 | 17:04 |
| *** aignatov_ has joined #openstack-ironic | 17:04 | |
| *** SpamapS_ has joined #openstack-ironic | 17:04 | |
| *** aignatov_ is now known as aignatov | 17:04 | |
| *** jbjohnso has quit IRC | 17:08 | |
| *** datajerk has quit IRC | 17:08 | |
| *** SpamapS has quit IRC | 17:08 | |
| *** pquerna has quit IRC | 17:08 | |
| *** jbjohnso has joined #openstack-ironic | 17:15 | |
| devananda | approved | 17:27 |
| devananda | heading out for brunch in a bit, but have my phone and will be back in ~2hr, if something needs to be fixed further | 17:28 |
| *** saju_m has joined #openstack-ironic | 17:31 | |
| NobodyCam | I'll restart checks | 17:31 |
| *** saju_m has quit IRC | 17:41 | |
| *** saju_m has joined #openstack-ironic | 17:42 | |
| openstackgerrit | A change was merged to openstack/ironic: Correct PXEPrivateMethodsTestCase.setUp https://review.openstack.org/73872 | 17:52 |
| *** saju_m has quit IRC | 17:55 | |
| *** jcooley_ has joined #openstack-ironic | 18:05 | |
| *** saju_m has joined #openstack-ironic | 18:11 | |
| *** jbjohnso has quit IRC | 18:38 | |
| *** jcooley_ has quit IRC | 18:39 | |
| *** pquerna_ is now known as pquerna | 18:50 | |
| *** anniec has quit IRC | 19:09 | |
| *** max_lobur has joined #openstack-ironic | 19:31 | |
| max_lobur | Hi Folks | 19:46 |
| max_lobur | anybody here? | 19:46 |
| max_lobur | NobodyCam: r u around? :) | 19:49 |
| max_lobur | I was looking to the fix for gate | 19:50 |
| max_lobur | wanted to make sure I correctly understand it | 19:50 |
| max_lobur | we had a lot of task_manager (and driver factory) calls even before https://review.openstack.org/#/c/70468/ | 19:52 |
| max_lobur | so why I worked before | 19:52 |
| max_lobur | the only difference is that before we had only one "not shared" task, and others were shared | 19:54 |
| max_lobur | and now we have a lot of "not shared" | 19:54 |
| max_lobur | this only changes the db load, but not the driver factory... | 19:54 |
| max_lobur | devananda: r u around? | 19:59 |
| devananda | max_lobur: hi | 20:14 |
| devananda | max_lobur: this patch changes the driver loading by not trying to load extensions from the system | 20:14 |
| devananda | max_lobur: take a look at tests/conductor/utils | 20:14 |
| max_lobur | devananda: hi there! | 20:14 |
| max_lobur | yep, got it | 20:15 |
| max_lobur | so this was because of real driver loading | 20:15 |
| devananda | yes | 20:15 |
| devananda | at least, i'm pretty darn sure of that | 20:15 |
| max_lobur | so how it worked before - we had a number of tests and everything did go smooth | 20:15 |
| devananda | all the failures happened in the same test class, which was the only test class calling TaskManager and NOT patching it | 20:15 |
| max_lobur | they also used real task_manager | 20:15 |
| devananda | and they stopped failnig when it is patched properly | 20:15 |
| max_lobur | maybe too much driver calls | 20:16 |
| max_lobur | and it started failing | 20:16 |
| max_lobur | like critical number | 20:16 |
| devananda | no -- we didn't have any tests loading an extension without patching the driver load | 20:16 |
| devananda | until we landed the series of _update_neutron patches taht I wrote | 20:16 |
| devananda | your patch for TaskManager wasn't the cause (though we did find and fix a problem with it) | 20:16 |
| max_lobur | yep, that's interesting too | 20:17 |
| max_lobur | I'm wondering - if we assume that there are some errors in pathes that we landed recently | 20:17 |
| devananda | the problem was PXEPrivateMethodsTestCase not calling mgr_utils.get_mocked_node_manager(driver='fake_pxe') | 20:17 |
| max_lobur | how much more we will find | 20:17 |
| devananda | this is why testing is important - we should never assume code is perfect | 20:18 |
| max_lobur | ahh, you're right | 20:18 |
| max_lobur | now I see | 20:18 |
| max_lobur | class PXEDriverTestCase(db_base.DbTestCase): | 20:18 |
| max_lobur | mgr_utils.get_mocked_node_manager(driver='fake_pxe') | 20:19 |
| max_lobur | it mocked node_manager | 20:19 |
| max_lobur | but PXEPrivateMethodsTestCase not | 20:20 |
| devananda | right | 20:20 |
| devananda | and it was OK before we created any TaskManager inside of PXEPrivateMethodsTestCase | 20:20 |
| max_lobur | yep | 20:20 |
| max_lobur | I see | 20:20 |
| devananda | and when I moved some unit tests there, and added more, which created TaskManager, I forgot to patch it | 20:20 |
| max_lobur | so you spinned RHEL vm to reproduce it? | 20:20 |
| devananda | yep | 20:20 |
| devananda | well, CentOS | 20:20 |
| max_lobur | is zuul running thos on CentOS? | 20:21 |
| devananda | $ cat /etc/centos-release | 20:21 |
| devananda | CentOS release 6.3 (Final) | 20:21 |
| devananda | yep | 20:21 |
| max_lobur | hmm, I didn't know | 20:21 |
| max_lobur | I assumed the VM is exactly the same | 20:21 |
| devananda | well, i think so -- zuul uses HPCloud, and I did this in HPCloud's CentOS image | 20:21 |
| devananda | so | 20:21 |
| devananda | n | 20:21 |
| devananda | no | 20:21 |
| devananda | py26 and py27 are not tested in teh same VM | 20:22 |
| max_lobur | I'm running my tests inside ubuntu server 12.04 x64 | 20:22 |
| max_lobur | interesting | 20:22 |
| max_lobur | one more question if you have time | 20:22 |
| devananda | so on centos 6.3, with a 2.6.32 kernel and py26, it fails about 40% of the time before my patch | 20:22 |
| devananda | i use this bash script to loop -- http://paste.openstack.org/show/66169/ | 20:22 |
| devananda | need to run out very soon, have plans at 1 | 20:23 |
| max_lobur | https://review.openstack.org/#/c/73872/2/ironic/tests/drivers/test_pxe.py why we inherited from db_base | 20:24 |
| max_lobur | https://github.com/openstack/ironic/blob/master/ironic/tests/db/base.py | 20:24 |
| max_lobur | it only creates contects | 20:24 |
| max_lobur | btw we creating it in our test too - currently overriding | 20:24 |
| max_lobur | just for consistency with others? | 20:25 |
| max_lobur | I guess we may remove context initialization in PXEPrivateMethodsTestCase | 20:25 |
| devananda | hrm | 20:25 |
| devananda | it used to do setup/cleanup of DB | 20:26 |
| devananda | test env runs multiple threads, needs to do per-thread db setup/cleanup | 20:26 |
| devananda | *per test setup/cleanup | 20:26 |
| devananda | i will look at it later | 20:26 |
| max_lobur | k, thanks | 20:28 |
| max_lobur | glad we figured out it | 20:28 |
| *** jcooley_ has joined #openstack-ironic | 20:29 | |
| openstackgerrit | Sirushti Murugesan proposed a change to openstack/ironic: Adds support for deploying Disk Images https://review.openstack.org/73054 | 20:32 |
| *** jcooley_ has quit IRC | 20:35 | |
| max_lobur | https://review.openstack.org/#/c/66621/ failed 26 | 20:37 |
| max_lobur | not sure how to check whether it includes the fix or not | 20:37 |
| NobodyCam | i was to quick with the rechecks | 20:38 |
| max_lobur | same https://review.openstack.org/#/c/73279/ | 20:38 |
| max_lobur | ahh | 20:38 |
| NobodyCam | ya rechecked befor it actually landed | 20:38 |
| max_lobur | hey hey NobodyCam :) | 20:38 |
| NobodyCam | hey hey max_lobur :) | 20:38 |
| max_lobur | how's your weekend going :) | 20:38 |
| NobodyCam | Great so far | 20:39 |
| *** saju_m has quit IRC | 20:39 | |
| NobodyCam | hows about you | 20:39 |
| max_lobur | not bad :) just need to go sleep earlier today | 20:40 |
| max_lobur | which is almost impossible :) | 20:40 |
| NobodyCam | :) and set up that local test env ...lol..j/k | 20:40 |
| max_lobur | I already did :) | 20:42 |
| max_lobur | yesterday | 20:42 |
| NobodyCam | lol :) | 20:42 |
| openstackgerrit | Sirushti Murugesan proposed a change to openstack/ironic: Adds support for deploying Disk Images https://review.openstack.org/73054 | 20:43 |
| *** SpamapS_ is now known as SpamapS | 20:47 | |
| *** SpamapS has joined #openstack-ironic | 20:47 | |
| *** saju_m has joined #openstack-ironic | 20:54 | |
| max_lobur | k, night Ironic | 20:59 |
| max_lobur | see you tomorrow | 20:59 |
| NobodyCam | night max_lobur | 20:59 |
| *** max_lobur has quit IRC | 20:59 | |
| *** saju_m has quit IRC | 21:16 | |
| openstackgerrit | A change was merged to openstack/ironic: alembic with initial migration and tests https://review.openstack.org/67415 | 21:25 |
| *** jcooley_ has joined #openstack-ironic | 21:25 | |
| *** jcooley_ has quit IRC | 21:30 | |
| *** jcooley_ has joined #openstack-ironic | 22:19 | |
| *** jcooley_ has quit IRC | 22:25 | |
| *** mdurnosvistov has joined #openstack-ironic | 22:42 | |
| *** mdurnosvistov has quit IRC | 23:29 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!