| *** nosnos has joined #openstack-qa | 00:36 | |
| *** sthaha has joined #openstack-qa | 00:43 | |
| *** sthaha has joined #openstack-qa | 00:43 | |
| openstackgerrit | Eric Windisch proposed a change to openstack/tempest: Fix colon in create volume logging output https://review.openstack.org/45149 | 00:52 |
|---|---|---|
| *** ewindisch- has quit IRC | 01:22 | |
| *** cyeoh_ is now known as cyeoh | 01:37 | |
| *** svarnau_ has quit IRC | 01:43 | |
| *** jhenner has joined #openstack-qa | 01:49 | |
| *** jhenner has quit IRC | 01:49 | |
| *** ericw has joined #openstack-qa | 02:34 | |
| ericw | uh yeah, about that patch.. | 02:34 |
| ericw | you can merge it, but it was just a demo for the AWS hackathon | 02:35 |
| *** adalbas has quit IRC | 02:40 | |
| *** coolsvap has joined #openstack-qa | 02:48 | |
| coolsvap | can somebody approve https://review.openstack.org/44743, I have a cinder review dependant on this | 02:49 |
| *** jhenner has joined #openstack-qa | 03:02 | |
| openstackgerrit | Zhi Kun Liu proposed a change to openstack/tempest: add negative volume tests https://review.openstack.org/44610 | 03:06 |
| *** ericw has quit IRC | 03:12 | |
| *** ericw has joined #openstack-qa | 03:18 | |
| *** torandu has quit IRC | 03:18 | |
| *** torandu has joined #openstack-qa | 03:20 | |
| *** torandu has quit IRC | 03:26 | |
| *** jhenner has quit IRC | 03:41 | |
| *** prekarat1 has joined #openstack-qa | 04:16 | |
| *** ericw has quit IRC | 04:23 | |
| openstackgerrit | A change was merged to openstack/tempest: add flavor creation tests https://review.openstack.org/44746 | 04:47 |
| openstackgerrit | A change was merged to openstack/tempest: Add tests for Swift's StaticWeb middelware https://review.openstack.org/44826 | 04:50 |
| openstackgerrit | Nachi Ueno proposed a change to openstack-dev/devstack: Support OpenSwan in Neturon VPNaaS https://review.openstack.org/42265 | 05:05 |
| openstackgerrit | Nachi Ueno proposed a change to openstack-dev/devstack: Support OpenSwan in Neturon VPNaaS https://review.openstack.org/42265 | 05:21 |
| Anju | cyeoh : https://review.openstack.org/#/c/44610/3/tempest/api/volume/test_volumes_actions.py | 05:22 |
| Anju | is this a crct scenario | 05:22 |
| openstackgerrit | A change was merged to openstack/tempest: fix test_flavors_extra_specs failure https://review.openstack.org/44797 | 05:23 |
| cyeoh | so test_attach_volumes_with_invalid_volume_id seems pretty reasonable to me | 05:23 |
| cyeoh | if passing an invalid volume id a reasonable response would be 400 or 404 perhaps | 05:24 |
| cyeoh | Anju: what is it about the test that concerns you? | 05:24 |
| Anju | yeah scenario seems good now | 05:25 |
| Anju | hmmm cyeoh : now code is look ing good | 05:28 |
| Anju | :) | 05:28 |
| *** ken1ohmichi has quit IRC | 05:29 | |
| Anju | but cyeoh : according to me it should be in the neagtive tests file | 05:29 |
| Anju | api/volume/test_volumes_negative.py | 05:30 |
| *** SergeyLukjanov has joined #openstack-qa | 05:40 | |
| coolsvap | Anju: who can approve https://review.openstack.org/44743 ? | 05:44 |
| Anju | coolsvap : add sdague in reviewers list | 05:45 |
| Anju | coolsvap: when reviewers get time they will go through it | 05:47 |
| cyeoh | Anju: well when I looked at the negative tests file it looked like it was only testing the volumes api, not the servers api | 05:53 |
| coolsvap | Anju: Thanks added to sdague to reviewers | 05:54 |
| cyeoh | and generally speaking I don't think we really need a separate file for negative tests, after all we have a per test tag now | 05:54 |
| cyeoh | so I don't consider it a big issue... | 05:54 |
| openstackgerrit | A change was merged to openstack/tempest: refactor - _is_timed_out using instance timeout https://review.openstack.org/43434 | 05:54 |
| Anju | yeah :) | 05:55 |
| Anju | its not a big issue | 05:55 |
| cyeoh | coolsvap: btw you probably want to use something like partial-bug: rather than closes-bug: in the future for bugs where you need multiple patches to fix them completely | 06:01 |
| openstackgerrit | A change was merged to openstack/tempest: Updating HACKING.rst https://review.openstack.org/42296 | 06:01 |
| coolsvap | cyeoh: sure I will take care for that in future | 06:02 |
| Anju | cyeoh : please check mine also | 06:02 |
| Anju | :) | 06:02 |
| Anju | cyeoh : https://review.openstack.org/#/c/44330/ | 06:03 |
| cyeoh | Anju: o | 06:06 |
| cyeoh | Anju: I'll have a look | 06:06 |
| Anju | :( | 06:06 |
| Anju | ok ok | 06:06 |
| Anju | I thought 0 marks | 06:07 |
| *** dkranz has quit IRC | 06:11 | |
| *** sthaha has quit IRC | 06:13 | |
| *** dkranz has joined #openstack-qa | 06:20 | |
| mkoderer | cyeoh: is xmlns=XMLNS_11 optional for root elements? | 06:36 |
| mkoderer | cyeoh: regarding https://review.openstack.org/#/c/44330/ | 06:36 |
| mkoderer | cyeoh: so I will correct my -1 then | 06:36 |
| cyeoh | mkoderer: at least for Nova apis (thought this is cinder) it depends on how the api extension is written | 06:37 |
| cyeoh | some do, some don't. But I checked the tempest run and the xml test is passing, so at least for this specific one its not required | 06:37 |
| cyeoh | its something we've been fixing up with the Nova v3 api to be consistent | 06:37 |
| cyeoh | (to always require it) | 06:38 |
| mkoderer | cyeoh: ok I see, I already thought that its just nit | 06:38 |
| Anju | mkoderer : i can show u the reponse , request body in xml | 06:39 |
| mkoderer | it's ok I will change my vote :) | 06:39 |
| Anju | I need to wrote something in the comment | 06:40 |
| cyeoh | mkoderer: short of looking at the server code, the simplest way to check is just to see if the test passes or not. If the extension requires the namespacing you'll end up with a 400 | 06:40 |
| mkoderer | cyeoh: ok thanks for clarification :) | 06:40 |
| Anju | thanks : cyeoh , mkoderer | 06:41 |
| *** SergeyLukjanov has quit IRC | 06:45 | |
| openstackgerrit | Ryan Hsu proposed a change to openstack-dev/devstack: Add support for uploading thin provisioned vmdks to glance https://review.openstack.org/45181 | 06:54 |
| openstackgerrit | nayna-patel proposed a change to openstack/tempest: Added test to check list/show extensions-neutron https://review.openstack.org/45030 | 06:59 |
| *** prekarat1 has quit IRC | 07:02 | |
| *** prekarat has joined #openstack-qa | 07:02 | |
| *** anteaya has quit IRC | 07:04 | |
| *** jhenner has joined #openstack-qa | 07:36 | |
| *** mkollaro has joined #openstack-qa | 07:49 | |
| *** jhenner has quit IRC | 07:51 | |
| *** prekarat has quit IRC | 08:04 | |
| *** prekarat has joined #openstack-qa | 08:04 | |
| *** yassine has joined #openstack-qa | 08:06 | |
| openstackgerrit | Julien Danjou proposed a change to openstack-dev/devstack: Use 10gen repository to get MongoDB 2.4 https://review.openstack.org/42642 | 08:19 |
| *** jhenner has joined #openstack-qa | 08:46 | |
| *** dosaboy_ is now known as dosaboy | 09:16 | |
| openstackgerrit | Cyril Roelandt proposed a change to openstack-dev/devstack: Swift: configure Ceilometer when it is enabled https://review.openstack.org/45201 | 09:21 |
| openstackgerrit | SlickNik proposed a change to openstack-dev/devstack: Added Trove (Database as a Service). https://review.openstack.org/38169 | 09:23 |
| openstackgerrit | nayna-patel proposed a change to openstack/tempest: Added test to check list/show extensions-neutron https://review.openstack.org/45030 | 09:51 |
| openstackgerrit | A change was merged to openstack/tempest: Restrict Volume type deletion with volumes assoc https://review.openstack.org/44743 | 10:00 |
| openstackgerrit | SlickNik proposed a change to openstack-dev/devstack: Added Trove (Database as a Service). https://review.openstack.org/38169 | 10:14 |
| openstackgerrit | Yuuichi Fujioka proposed a change to openstack-dev/devstack: Rename ceilometer alarm service name https://review.openstack.org/45214 | 10:43 |
| *** sthaha has joined #openstack-qa | 11:13 | |
| openstackgerrit | nayna-patel proposed a change to openstack/tempest: Test cases for V3 Project Actions https://review.openstack.org/41345 | 11:14 |
| *** nosnos has quit IRC | 11:24 | |
| *** prekarat has quit IRC | 11:52 | |
| openstackgerrit | A change was merged to openstack-dev/devstack: xenapi: enable block device access for stack user https://review.openstack.org/44268 | 11:52 |
| *** coolsvap has quit IRC | 12:00 | |
| openstackgerrit | nayna-patel proposed a change to openstack/tempest: Test cases for V3 Project Actions https://review.openstack.org/41345 | 12:35 |
| Anju | sdague : ping | 12:40 |
| sdague | pong | 12:40 |
| mkoderer | sdague: I am currently writing on a blog post about the new way of stress testing | 12:44 |
| sdague | mkoderer: nice | 12:44 |
| mkoderer | sdague: maybe you could help me to push is a bit when it's released | 12:44 |
| sdague | sure, will do | 12:44 |
| mkoderer | sdague: I think we could close the blueprint, there isn't really anything left | 12:45 |
| mkoderer | and everything else can be done with simple bugs | 12:46 |
| sdague | mkoderer: I'm happy with that, mark it done | 12:46 |
| mkoderer | dkranz: would you agree in closing the stress test bp? | 12:47 |
| *** weshay has joined #openstack-qa | 12:53 | |
| sdague | mkoderer: he's actually out today | 13:00 |
| Anju | sdague : are u agree in deleting the file server basic_ops.py | 13:06 |
| *** sthaha has quit IRC | 13:06 | |
| Anju | sdagur ,mkoderer : https://github.com/openstack/tempest/blob/master/tempest/scenario/test_server_basic_ops.py | 13:06 |
| sdague | Anju: you need to give me more context | 13:06 |
| *** afazekas has joined #openstack-qa | 13:06 | |
| Anju | actually these tests are in server actions.py | 13:07 |
| Anju | sdague : these tests arer in server actions.py | 13:10 |
| sdague | Anju: that's ok, some duplication between scenario and api is allowed, as they use different clients | 13:11 |
| Anju | so sdague acn we add an top_strat action also here | 13:12 |
| Anju | sdague can we add an stop_start action also here before terminating an instance | 13:13 |
| sdague | sure, if that's not covered in other scenario tests | 13:13 |
| openstackgerrit | A change was merged to openstack/tempest: Added missing xml tests of volume https://review.openstack.org/44330 | 13:13 |
| Anju | sdague : afazeaks was not agree | 13:14 |
| Anju | not mkoderer | 13:14 |
| sdague | well, it's not clear that start stop is a huge value add there, so I'd focus on other places | 13:14 |
| Anju | but sdague : will this test not create a redundancy | 13:15 |
| sdague | Anju: so the focus should be on adding tests that have impact. The basic scenario test is a good test to run on it's own to show that something is at least working. | 13:16 |
| mkoderer | sdague Anju I am in a meeting | 13:16 |
| mkoderer | I will respond later | 13:17 |
| sdague | but it doesn't need every nova api call in it | 13:17 |
| *** adalbas has joined #openstack-qa | 13:17 | |
| sdague | mkoderer: cool, no worries | 13:17 |
| Anju | hmm : sdague : point is valid | 13:20 |
| *** raies has quit IRC | 13:35 | |
| *** burt has joined #openstack-qa | 13:42 | |
| *** mriedem has joined #openstack-qa | 13:51 | |
| *** adalbas has quit IRC | 13:53 | |
| openstackgerrit | Julien Danjou proposed a change to openstack-dev/devstack: Switch Ceilometer default backend to MySQL https://review.openstack.org/43851 | 13:53 |
| *** adalbas has joined #openstack-qa | 13:59 | |
| *** krtaylor has joined #openstack-qa | 14:06 | |
| *** jasondotstar has joined #openstack-qa | 14:07 | |
| *** jhenner has quit IRC | 14:08 | |
| mkoderer | sdague: regarding scenario tests - in the readme file is stated that a scenario test should invoke multiple API backends as main goal | 14:31 |
| sdague | mkoderer: right, fair | 14:34 |
| sdague | and start stop definitley doesn't fit there | 14:34 |
| mkoderer | sdague: would it be more clearer to say it should be a real-life scenario | 14:34 |
| mkoderer | sdague: IMHO having a test on just one API would be ok if it follows a real-life use case | 14:36 |
| sdague | mkoderer: sure | 14:37 |
| mkoderer | sdague: ok so I will add this point to the readme :) | 14:38 |
| mkoderer | sdague: and because of this discussion we had the idea to simply delete test_server_basic_ops.py | 14:39 |
| mkoderer | since it doesn't follow any real life use case | 14:39 |
| *** anteaya has joined #openstack-qa | 14:41 | |
| *** weshay has quit IRC | 14:43 | |
| *** weshay has joined #openstack-qa | 14:43 | |
| mkoderer | sdague: btw I won't make it to the meeting today | 14:45 |
| *** mriedem has quit IRC | 14:50 | |
| *** ericw has joined #openstack-qa | 14:50 | |
| openstackgerrit | Doug Hellmann proposed a change to openstack-dev/devstack: Add flag to turn of chown of egg-info dirs https://review.openstack.org/45248 | 15:06 |
| *** mlavalle has joined #openstack-qa | 15:08 | |
| *** SergeyLu_ has joined #openstack-qa | 15:23 | |
| *** SergeyLu_ has quit IRC | 15:23 | |
| *** SergeyLu_ has joined #openstack-qa | 15:24 | |
| *** SergeyLu_ has quit IRC | 15:29 | |
| *** SergeyLukjanov has joined #openstack-qa | 15:30 | |
| openstackgerrit | Cyril Roelandt proposed a change to openstack-dev/devstack: Swift: configure Ceilometer when it is enabled https://review.openstack.org/45201 | 15:40 |
| openstackgerrit | A change was merged to openstack/tempest: raise assertion error if output is falsy https://review.openstack.org/43155 | 15:59 |
| openstackgerrit | A change was merged to openstack/tempest: Added test case to check floating IP API operations https://review.openstack.org/41038 | 16:01 |
| openstackgerrit | A change was merged to openstack-dev/devstack: Switch Ceilometer default backend to MySQL https://review.openstack.org/43851 | 16:02 |
| *** SergeyLukjanov has quit IRC | 16:13 | |
| openstackgerrit | Eric Windisch proposed a change to openstack/tempest: Re-enable test_stamp_pattern scenario https://review.openstack.org/45250 | 16:34 |
| *** jaypipes has joined #openstack-qa | 16:38 | |
| *** ChanServ sets mode: +o jaypipes | 16:38 | |
| *** yassine has quit IRC | 16:52 | |
| mlavalle | Hi, no qa meeting today? | 17:05 |
| adalbas | mlavalle, mtreinish and dkranz are out today. Not sure if sdague will run the meeting. | 17:07 |
| timello | mlavalle: sdague sent an email one hour ago to the openstack-dev. It seems it has been cancelled. | 17:07 |
| mlavalle | adalbas, timello: thanks guys ;-) | 17:07 |
| *** coolsvap has joined #openstack-qa | 17:09 | |
| *** SergeyLukjanov has joined #openstack-qa | 17:10 | |
| *** mkollaro has quit IRC | 17:16 | |
| afazekas | mlavalle: hi | 17:19 |
| *** mlavalle has quit IRC | 17:23 | |
| *** afazekas has quit IRC | 17:27 | |
| *** mkollaro has joined #openstack-qa | 17:33 | |
| *** coolsvap has quit IRC | 17:56 | |
| *** jhenner has joined #openstack-qa | 18:00 | |
| *** mlavalle has joined #openstack-qa | 18:02 | |
| mlavalle | afazekas: hi | 18:02 |
| *** VinodGupta has joined #openstack-qa | 18:33 | |
| openstackgerrit | Monty Taylor proposed a change to openstack/tempest: Update tempest to latest tox and requirements https://review.openstack.org/45303 | 18:48 |
| *** VinodGupta has quit IRC | 18:48 | |
| *** SergeyLukjanov has quit IRC | 18:54 | |
| *** krtaylor has quit IRC | 19:12 | |
| *** mkollaro has quit IRC | 19:15 | |
| *** krtaylor has joined #openstack-qa | 19:17 | |
| sdague | yeh, sorry for the last minute cancelation | 19:30 |
| sdague | with mtreinish, attilla, dkranz, and mkoderer all not able to make it, it seemed simpler to cancel | 19:31 |
| *** jhenner has quit IRC | 20:03 | |
| *** jhenner has joined #openstack-qa | 20:04 | |
| *** mriedem has joined #openstack-qa | 20:04 | |
| openstackgerrit | Joe Gordon proposed a change to openstack-dev/devstack: Enable multi-threaded nova API servers https://review.openstack.org/45314 | 20:08 |
| *** anteaya has quit IRC | 20:11 | |
| openstackgerrit | Joe Gordon proposed a change to openstack-dev/devstack: Enable multi-threaded nova API servers https://review.openstack.org/45314 | 20:25 |
| openstackgerrit | Joe Gordon proposed a change to openstack-dev/devstack: Enable multi-threaded nova API servers https://review.openstack.org/45314 | 20:28 |
| *** mkollaro has joined #openstack-qa | 20:46 | |
| adalbas | sdague: could you help me understand how to retrieve the value for a extension key, as task_state, in tempest? | 20:49 |
| adalbas | for instance, in json, the returned body from an api call is 'EXT-OS-STS:task_state' and i simply used it as the key for the dict | 20:51 |
| sdague | adalbas: do you have some example code that's not working? usually starting from code is easier for me. | 20:51 |
| adalbas | sdague, the same dict key for xml would be something like '{http://docs.openstack.org/compute/ext/extended_status/api/v1.1}task_state' | 20:53 |
| adalbas | sdague, let me upload the patch and it will be more simple to explain then | 20:53 |
| sdague | yeh, that will make it simpler | 20:53 |
| openstackgerrit | adalbas proposed a change to openstack/tempest: Wait for image to upload before create new image https://review.openstack.org/45326 | 21:09 |
| adalbas | sdague, ^that should be it | 21:11 |
| jog0 | anyone see this recently http://logs.openstack.org/14/45314/3/check/gate-tempest-devstack-vm-postgres-full/c27aa41/testr_results.html.gz | 21:13 |
| jog0 | if not its related to the patch I am working on (multi theaded n-api in devstack) | 21:14 |
| adalbas | jog0, This one could be related, but it has been fixed : https://code.launchpad.net/bugs/1217163 | 21:17 |
| jog0 | adalbas: hmm thanks. | 21:17 |
| jog0 | I will dig into it further | 21:17 |
| jog0 | using multiple n-api servers makes races easier | 21:18 |
| jog0 | which is good for making sure we find bugs | 21:18 |
| adalbas | jog0, i m also currently looking at that in tempest, specially the ones with too many rechecks | 21:19 |
| adalbas | i mean, the races | 21:19 |
| jog0 | heisenbugs FTW | 21:19 |
| sdague | jog0: my guess is you exposed another aggregates race | 21:21 |
| sdague | jog0: it's really a test race, just probably needs more locking in tempest | 21:22 |
| jog0 | sdague: :( yeah | 21:22 |
| sdague | so we have lockutils in tempest, my guess is that mtreinish just didn't get to that case because we didn't trip it | 21:22 |
| jog0 | I see we are logging the req-ids now \o/ | 21:23 |
| jog0 | this will make searching easier | 21:23 |
| jog0 | now if only we didn't have a gazillion lines of n-api logs to download over slow wifi | 21:25 |
| sdague | yeh, well | 21:25 |
| *** jasondotstar has quit IRC | 21:29 | |
| jog0 | sdague: so the stacktrace I see has lockutils in it | 21:32 |
| sdague | jog0: sure, but my guess is that the lock isn't broad enough | 21:33 |
| sdague | i.e. with the extra workers, there are other tests that need to be in the same lock now | 21:33 |
| jog0 | sdague: right | 21:34 |
| jog0 | dumb question do the locks keep holding with fixutres | 21:40 |
| jog0 | when fixtures are doing cleanup | 21:40 |
| openstackgerrit | Joe Gordon proposed a change to openstack-infra/devstack-gate: Add support for large_ops tempest test https://review.openstack.org/43779 | 21:45 |
| sdague | not a dumb question, and I don't know | 21:49 |
| jog0 | sdague: I was hoping it would be a dumb question | 21:51 |
| sdague | maybe an osloer would know, honestly, I haven't worked with the code much, mtreinish did most of that | 21:51 |
| jog0 | lifeless: ping | 21:53 |
| jog0 | he isn't an oslo'er but he knows fixtures | 21:53 |
| *** ericw has quit IRC | 21:57 | |
| lifeless | pong | 21:57 |
| *** ravikumar_hp has joined #openstack-qa | 21:59 | |
| jog0 | when using self.addCleanup in a test, and using a locking decorator | 21:59 |
| jog0 | will the cleanup be done inside the decorated function or outside? | 21:59 |
| jog0 | lifeless: I am testing it and looks like outside | 22:00 |
| lifeless | cleanup is run by the framework | 22:01 |
| ravikumar_hp | I would like to restore - https://review.openstack.org/#/c/37194/ which was abandoned. I could not see restore change in web interface . can some one restore this? | 22:01 |
| lifeless | after all other user code has run | 22:01 |
| lifeless | you need to free the lock in a cleanup | 22:01 |
| lifeless | self.addCleanup(mylock.cleanUp) | 22:01 |
| lifeless | self.addCleanup(mything_which_needs_lock.cleanUp) | 22:01 |
| lifeless | LIFO order | 22:01 |
| sdague | lifeless: so I actually think with the lock utils decorator, the opposite might be happening | 22:02 |
| lifeless | applies to useFixture too, of course, as that builds on this | 22:02 |
| lifeless | sdague: if you're decorating the test function, definitely. | 22:02 |
| sdague | the lock is cleaned up automatically, then cleanup is running without it | 22:02 |
| lifeless | you can only assume the lock is present for the body of the function | 22:02 |
| sdague | yeh, this is the oslo lockutils | 22:02 |
| lifeless | not for tearDown, setUp, or cleanups. | 22:02 |
| *** ericw has joined #openstack-qa | 22:02 | |
| lifeless | if the lock is a context manager | 22:04 |
| lifeless | then it's very easy to adapt that to Fixture, and use 'useFixture' in setUp to acquire the lock and hold it for the entire test code - setup/method/teardown/cleanups | 22:05 |
| lifeless | I would poke at this now for you but I have a flight and a paper for pycon to finish, so EEEEESTRESSED> | 22:05 |
| sdague | lifeless: no worries, I expect we've got enough clues to sort it | 22:05 |
| sdague | thanks | 22:06 |
| *** fungi has quit IRC | 22:08 | |
| jog0 | sdague lifeless: here is an example of fail | 22:10 |
| jog0 | http://paste.openstack.org/show/45843/ | 22:10 |
| lifeless | jog0: can't look now, sorry. | 22:10 |
| jog0 | run that and you see the cleanup is done after the lock | 22:10 |
| jog0 | lifeless: no problem thanks for your help | 22:10 |
| *** fungi has joined #openstack-qa | 22:11 | |
| sdague | jog0: yeh, so we probably have to do manual cleanup instead | 22:11 |
| jog0 | sdague: maybe a lock fixutre? | 22:11 |
| sdague | maybe? | 22:12 |
| sdague | I don't know if you can use lockutils that way | 22:12 |
| sdague | jog0: you got a github link to the problematic code? | 22:12 |
| jog0 | sdague: sdague no i Just wrote a sample | 22:12 |
| jog0 | but I can point to an example of where it fails | 22:13 |
| jog0 | the idea is with mulitple n-apis the lock is released and then cleanup is done | 22:13 |
| jog0 | but the next test runs before cleanup | 22:13 |
| jog0 | sdague: https://github.com/openstack/tempest/blob/master/tempest/api/compute/admin/test_aggregates.py#L178 | 22:15 |
| jog0 | thats an example | 22:15 |
| jog0 | we delete the node outside the lock | 22:15 |
| *** burt has quit IRC | 22:15 | |
| sdague | jog0: right, so we can just call the cleanup function manually at the end (not using addCleanup) as long as we put a try around the asserts and stick it in the finally block | 22:16 |
| sdague | that's probably the near term fix | 22:16 |
| jog0 | sdague: yeah its a little ugly though but I agree | 22:16 |
| sdague | you want to spin the patch, or should I? | 22:20 |
| jog0 | you can do it | 22:20 |
| jog0 | I was just trying to get n-api working with multiple threads | 22:20 |
| jog0 | I am seeing if there as more elagent way to fix this | 22:20 |
| *** weshay has quit IRC | 22:21 | |
| jog0 | so lock in oslo is a context | 22:26 |
| jog0 | meh sounds like there is a better fix but not sure what it is, lifeless knows but he is indisposed | 22:35 |
| lifeless | adapt it to Fixture | 22:35 |
| lifeless | Fixture is also a context, context is not a Fixture | 22:36 |
| lifeless | jog0: I have a presentation to give tomorrow at pycon, and haven't gotten the thing I'm presenting working yet, let alone the presentation about it | 22:36 |
| jog0 | lifeless: no worries, go prepare | 22:36 |
| sdague | jog0: you got a bug registered for the race? | 22:38 |
| sdague | just so I can ref something | 22:38 |
| jog0 | sdague: filing it right now | 22:38 |
| jog0 | sdague: https://bugs.launchpad.net/tempest/+bug/1221469 | 22:40 |
| sdague | jog0: are you able to reproduce locally with your devstack patch? | 22:49 |
| sdague | i.e. so you can test this fix | 22:50 |
| jog0 | sdague: I didn't try locally yet | 22:51 |
| jog0 | but i am fairly sure I can | 22:51 |
| jog0 | since I got two jenkins job to fail identically | 22:51 |
| sdague | cool | 22:51 |
| *** ericw has quit IRC | 22:51 | |
| *** ericw has joined #openstack-qa | 22:53 | |
| openstackgerrit | Sean Dague proposed a change to openstack/tempest: fix race condition between addCleanup and lockutils https://review.openstack.org/45342 | 22:55 |
| sdague | ok, and with that I think I'm on the hook to make dinner, so I'm off for the evening | 22:56 |
| jog0 | sdague: did you test that locally? | 22:56 |
| jog0 | if not i can give it a shot in a few | 22:56 |
| sdague | no, I didn't | 22:57 |
| jog0 | sdague: bye | 22:57 |
| sdague | just made sure it's pep8 clean | 22:57 |
| jog0 | sdague: testing it now | 23:04 |
| jog0 | reproduced bug locally on first try | 23:04 |
| *** mlavalle has left #openstack-qa | 23:06 | |
| jog0 | sdague: fix barfed | 23:07 |
| jog0 | but that is for another day | 23:07 |
| *** mkollaro has quit IRC | 23:13 | |
| openstackgerrit | Joe Gordon proposed a change to openstack/tempest: fix race condition between addCleanup and lockutils https://review.openstack.org/45342 | 23:19 |
| jog0 | I fixed it up for you ^ | 23:20 |
| *** ravikumar_hp has quit IRC | 23:31 | |
| lifeless | jog0: arguably it's not a race. It *never* works. | 23:32 |
| lifeless | jog0: races are for things that work sometimes :) | 23:32 |
| jog0 | lifeless: I just got it working :/ | 23:33 |
| *** jasondotstar has joined #openstack-qa | 23:33 | |
| jog0 | I reproduced for a while then I couldn't | 23:33 |
| jog0 | lifeless: or maybe it was just a fixed race :) | 23:34 |
| lifeless | jog0: my point is that it's deterministic | 23:35 |
| lifeless | jog0: the lock was always released before the cleanup ran | 23:35 |
| jog0 | lifeless: oh right, but that didn't always trigger an error | 23:35 |
| lifeless | jog0: right, but thats due to the state of the thing being cleanedup and thats probably due to the test failure-or-not and cloud-under-test-status | 23:36 |
| lifeless | jog0: was it a cross-process lock ? | 23:36 |
| openstackgerrit | Joe Gordon proposed a change to openstack/tempest: Don't use addCleanup with lockutils https://review.openstack.org/45342 | 23:37 |
| jog0 | lifeless: yes cross process lock | 23:37 |
| lifeless | in which case, yes, there is a race, I didn't get that it was cross process | 23:38 |
| jog0 | lifeless: besides the extra space the new title is still correct I hope | 23:39 |
| jog0 | anyway, I too am out | 23:41 |
| *** jasondotstar has quit IRC | 23:42 | |
| openstackgerrit | Ryan Hsu proposed a change to openstack-dev/devstack: Provide a means of setting vmdk custom properties via image filename https://review.openstack.org/45181 | 23:48 |
| openstackgerrit | Ryan Hsu proposed a change to openstack-dev/devstack: Provide a means of setting vmdk custom properties via image filename https://review.openstack.org/45181 | 23:49 |
| *** ericw has quit IRC | 23:49 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!