| *** FL1SK has joined #openstack-mistral | 00:10 | |
| *** dprince has quit IRC | 00:22 | |
| *** rbrady has quit IRC | 00:43 | |
| *** jamielennox is now known as jamielennox|away | 00:50 | |
| *** jamielennox|away is now known as jamielennox | 01:21 | |
| kong | ping rakhmerov | 01:28 |
|---|---|---|
| kong | rakhmerov: a question for named lock implementation. according to https://github.com/openstack/mistral/blob/master/mistral/db/v2/sqlalchemy/api.py#L1538, what if two threads are trying to create lock using the same name? | 01:30 |
| *** hparekh has joined #openstack-mistral | 04:25 | |
| *** ist has joined #openstack-mistral | 06:06 | |
| rakhmerov | kong: I'm here | 06:59 |
| rakhmerov | the second thread will block until the lock is released (1st TX is finished) | 07:00 |
| rakhmerov | and please what you mean by threads | 07:00 |
| rakhmerov | please clarify.. | 07:00 |
| ist | Do you have any idea why tripleo buildbot fail in following review: https://review.openstack.org/#/c/417314 | 07:46 |
| ist | it seems some host is unreachable, I just do not know what could be the root cause... | 07:47 |
| d0ugal | ist: looking | 08:17 |
| ist | thanks! | 08:17 |
| openstackgerrit | Dougal Matthews proposed openstack/mistral: Insecure flag added to openstack context https://review.openstack.org/417314 | 08:18 |
| d0ugal | I just rechecked it once more and I'll track this build, I see it failed a few times in a row | 08:18 |
| *** jpich has joined #openstack-mistral | 08:31 | |
| *** jpich has quit IRC | 08:37 | |
| *** jpich has joined #openstack-mistral | 08:38 | |
| *** ist has quit IRC | 08:58 | |
| *** ist has joined #openstack-mistral | 09:08 | |
| d0ugal | It failed again :( | 09:10 |
| d0ugal | I am trying to get help in #tripleo | 09:11 |
| openstackgerrit | Renat Akhmerov proposed openstack/python-mistralclient: Add user and project domain name parameters for target cloud https://review.openstack.org/419332 | 09:16 |
| ddeja | d0ugal, I've just read your discussion with thrash | 09:18 |
| d0ugal | ddeja: Morning :) | 09:18 |
| d0ugal | Yeah, we got a bit confused. | 09:18 |
| ddeja | some bad assumptions | 09:18 |
| ddeja | I was thinking that 'fake' is only used for testing purpose | 09:19 |
| d0ugal | Yeah | 09:19 |
| ddeja | and therefore all of the flags are in mistral.conf | 09:19 |
| ddeja | so I don't have to treat the 'fake' driver any special way | 09:19 |
| ddeja | since the rpc_utils are only about preparing the config options for rpc layer | 09:20 |
| rakhmerov | ist: hi | 09:20 |
| ddeja | d0ugal: I'll send a patch in a minute | 09:20 |
| d0ugal | ddeja: thanks! | 09:20 |
| ddeja | d0ugal: is there a bug for that? | 09:20 |
| d0ugal | ddeja: no, I don't think so | 09:20 |
| rakhmerov | ist: on https://review.openstack.org/#/c/417314/ not only tripleO gate fails, do you think the other failure is related with your changes? | 09:20 |
| d0ugal | ddeja: I think we were unsure if the problem was with us or the code :) | 09:21 |
| ddeja | d0ugal: OK, so I'll submit one | 09:21 |
| d0ugal | ddeja: thanks | 09:21 |
| d0ugal | rakhmerov, ist: I can confirm the TripleO failure is unrelated. I am still looking into it (or still trying to find somebody that can help me!) | 09:22 |
| rakhmerov | d0ugal, ddeja: hi guys, sorry being late for your discussion | 09:22 |
| rakhmerov | what's wrong with 'fake'? | 09:22 |
| rakhmerov | if you can tell with a couple of words.. | 09:22 |
| d0ugal | rakhmerov: I think the issue is that it still attempts to connect to rabbit | 09:22 |
| d0ugal | but my understanding could be wrong. | 09:23 |
| rakhmerov | you mean even if we have 'fake' in config? | 09:23 |
| ddeja | d0ugal, rakhmerov, it's not trying to connect to rabbit | 09:23 |
| d0ugal | rakhmerov: yes | 09:23 |
| d0ugal | oh :) | 09:24 |
| ddeja | but there is an assumption that some flags in config would be provided | 09:24 |
| ddeja | like 'rabbit_host', 'rabbit_port' and so on | 09:24 |
| rakhmerov | aah, yeah | 09:24 |
| *** shardy has joined #openstack-mistral | 09:25 | |
| rakhmerov | I think I even made a comment when doing some refactoring in RPC a couple of months ago | 09:25 |
| rakhmerov | true | 09:25 |
| ddeja | I'll fix it in a minute | 09:25 |
| rakhmerov | because in any case we call that method get_oslo_config or something.. | 09:25 |
| rakhmerov | or get_rabbit_config | 09:25 |
| ddeja | hm, it may not be this easy to fix it | 09:26 |
| rakhmerov | :)) | 09:26 |
| ddeja | it may brake some unit tests | 09:26 |
| ddeja | I have to check | 09:26 |
| rakhmerov | ddeja: this is important, please try to fix it. As a matter of fact, now what we have in terms of flexible RPC layer is not 100% flexible | 09:27 |
| rakhmerov | in this aspect | 09:27 |
| rakhmerov | we always expect rabbit | 09:27 |
| rakhmerov | ok, thanks for clarification.. | 09:28 |
| ddeja | rakhmerov: sure | 09:28 |
| ddeja | rakhmerov: BTW, I've looked onto kombu failure in ist's patch | 09:28 |
| rakhmerov | ok | 09:28 |
| ddeja | but I have never seen this test fail befora | 09:28 |
| ddeja | so it *may* be related to the patch itself | 09:28 |
| rakhmerov | ok | 09:30 |
| rakhmerov | ist: here? Can you have a look at this? | 09:31 |
| ist | I just looking into it and it seems following testcase fails: | 09:32 |
| ist | mistral_tempest_tests.tests.api.v2.test_executions.ExecutionTestsV2.test_update_nonexistent_execution | 09:32 |
| ist | i think so it is not related to my patch | 09:33 |
| kong | rakhmerov: i was trying to understand the named lock. suppose there are 2 threads(or green threads, whatever) are both creating named lock record using same name, they will both succeed? in sequence? | 09:37 |
| rakhmerov | ist: ok | 09:37 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/mistral: Updated from global requirements https://review.openstack.org/419344 | 09:37 |
| rakhmerov | kong: nope, not both | 09:37 |
| kong | the 2nd will fail? | 09:38 |
| rakhmerov | sec, I'll find an article | 09:38 |
| rakhmerov | no, it will hang | 09:38 |
| kong | rakhmerov: appriciate | 09:38 |
| ddeja | d0ugal, thrash: https://bugs.launchpad.net/mistral/+bug/1655895 | 09:38 |
| openstack | Launchpad bug 1655895 in Mistral "Rabbit related flags need to be set when using fake driver" [Undecided,New] | 09:38 |
| rakhmerov | kong: http://blog.udby.com/archives/14 | 09:40 |
| * kong is reading | 09:40 | |
| kong | rakhmerov: thanks! | 09:40 |
| rakhmerov | please read it, after that it'll be clear to you how it works | 09:40 |
| rakhmerov | just one thing in addition: with green threads there are problems, they are not even related to these locks, I can tell in details too | 09:41 |
| kong | rakhmerov: will get back to you if i have further questions after reading | 09:42 |
| kong | thanks again | 09:42 |
| rakhmerov | so that's why we handle RPC messages in engine w/o using green threads | 09:42 |
| rakhmerov | ok | 09:42 |
| rakhmerov | kong: these locks are fully tested, they work | 09:42 |
| kong | cool | 09:42 |
| openstackgerrit | Jeremy Liu proposed openstack/python-mistralclient: Update .gitignore https://review.openstack.org/419375 | 09:55 |
| openstackgerrit | Dawid Deja proposed openstack/mistral: Stop excpecting to have rabbit flags set when using fake driver https://review.openstack.org/419388 | 10:10 |
| ddeja | thrash: Hi, could you please check if this fixes your problem with fake driver? https://review.openstack.org/#/c/419388 | 10:12 |
| ddeja | @CC d0ugal ^^ ;) | 10:12 |
| d0ugal | ddeja: yay, thanks! | 10:13 |
| d0ugal | fwiw, I don't think thrash will be around for a few hours yet :) | 10:13 |
| d0ugal | but I'll try it in a bit, just in the middle of something else. | 10:13 |
| ddeja | d0ugal: OK, hope he'll check it once he's around | 10:13 |
| d0ugal | yup, sure he will - I've added him as a reviewer. | 10:14 |
| ddeja | d0ugal: also timeout or exchange in this: https://github.com/openstack/mistral/blob/master/mistral/utils/rpc_utils.py#L107-L110 | 10:14 |
| ddeja | may caouse a problem when using fake | 10:15 |
| ddeja | cause* | 10:15 |
| ddeja | so if the patch itself won't work, they can be safely deleted as long as fake is used | 10:16 |
| ddeja | and I'll make adjustements in my patch | 10:16 |
| *** AlexeyAbashkin has joined #openstack-mistral | 10:17 | |
| *** shardy has quit IRC | 10:51 | |
| *** shardy has joined #openstack-mistral | 11:14 | |
| rakhmerov | kong: sorry for making https://review.openstack.org/#/c/414508 a long discussion, I just find it very important and want to make sure we discuss it within a bigger crowd | 11:21 |
| rakhmerov | ddeja, d0ugal: btw, please also take a look at https://review.openstack.org/#/c/414508 | 11:21 |
| rakhmerov | d0ugal: it's also a little bit related to our actions refactoring | 11:22 |
| d0ugal | Sure | 11:23 |
| d0ugal | Lots of -1 | 11:23 |
| rakhmerov | yeah, -1s are not so important | 11:23 |
| rakhmerov | I guess we're just trying to find the best design | 11:23 |
| rakhmerov | ddeja: can you please explain your changes in https://review.openstack.org/#/c/419388/1/mistral/utils/rpc_utils.py ? | 11:24 |
| rakhmerov | why do we still need to try to call _get_rabbit_info_from_oslo in case of "fake"? | 11:24 |
| rakhmerov | seems weird to me | 11:24 |
| openstackgerrit | Renat Akhmerov proposed openstack/python-mistralclient: Add user and project domain name parameters for target cloud https://review.openstack.org/419332 | 11:36 |
| *** shardy has quit IRC | 11:55 | |
| *** shardy has joined #openstack-mistral | 11:56 | |
| rakhmerov | d0ugal, hparekh, kong: guys, can you please review and approve https://review.openstack.org/#/c/414533/ ? | 12:05 |
| rakhmerov | it fixes the gate | 12:05 |
| rakhmerov | sharatss: seems like it doesn't make to recheck patches for now (you just did for https://review.openstack.org/#/c/419388/ ), I guess we need help from tripleO to fix the gate | 12:14 |
| rakhmerov | d0ugal: do you know if that's being investigated? | 12:14 |
| sharatss | rakhmerov: yes got it | 12:15 |
| sharatss | rakhmerov: just found debugging-tripleo-gate-failures-in-mistral doc written by d0ugal | 12:15 |
| d0ugal | sharatss: :-D | 12:15 |
| d0ugal | rakhmerov, sharatss: there are people working on it | 12:16 |
| d0ugal | https://bugs.launchpad.net/tripleo/+bug/1655919 | 12:16 |
| openstack | Launchpad bug 1655919 in tripleo "CI: Openvswitch agent fails because can't import tinyrpc.server module" [Critical,Triaged] | 12:16 |
| d0ugal | That is the bug FWIW | 12:16 |
| sharatss | d0ugal: thanks :) | 12:17 |
| thrash | ddeja: sure... I'll check it out right away. :) | 12:22 |
| rakhmerov | d0ugal: ok, thanks :) | 12:24 |
| *** dprince has joined #openstack-mistral | 12:35 | |
| ddeja | rakhmerov: I was out for 1.5hour, I'll lok in 15 minutes | 12:43 |
| rakhmerov | ok | 12:43 |
| ddeja | rakhmerov: about still calling _get_rabbit_info_from_oslo | 12:43 |
| thrash | ddeja: wfm | 12:44 |
| ddeja | thrash: great, thanks :) | 12:44 |
| thrash | now... when I run 'run_functional_test.sh' it's complaining about credentials. :( | 12:44 |
| ddeja | rakhmerov: If I remember correctly, some tests rely on info gathered inside this function | 12:45 |
| ddeja | but I'm not sure, I can check | 12:45 |
| ddeja | thrash: after applaying the patch? | 12:45 |
| thrash | ddeja: yes | 12:45 |
| thrash | to be fair, I was getting that before. :D | 12:45 |
| ddeja | ufff | 12:46 |
| ddeja | I was terrified that some functional tests rely on rpc_utils | 12:46 |
| thrash | haha | 12:46 |
| thrash | no, sorry to get your heart racing. | 12:46 |
| ddeja | thrash: I remember that I have trouble with running functional tests on anything but devstack | 12:47 |
| rakhmerov | ddeja: ok, I see | 12:47 |
| rakhmerov | do you remember what tests? | 12:47 |
| ddeja | thrash: and I've run them using tempest, not calling directly run_functional_tests | 12:47 |
| ddeja | rakhmerov: that was some tests regarding the kombu driver | 12:48 |
| ddeja | but I may mixed something | 12:48 |
| ddeja | and I wanted to have this patch ASAP, so I went the 'safe' way | 12:48 |
| ddeja | I'll check in a minute if we can delete it completetly in case of fake | 12:48 |
| thrash | ddeja: interesting... | 12:49 |
| thrash | ddeja: that seems to be working... Calling tempest directly. | 12:49 |
| thrash | ddeja: hold on that... it's running unit tests atm | 12:49 |
| *** gongysh has joined #openstack-mistral | 12:52 | |
| ddeja | thrash: I'm glad I could help, a little | 12:53 |
| thrash | ddeja: definitely a start. :) | 12:54 |
| thrash | ddeja: this: nosetests -v mistral_tempest_tests/tests/api/v2/ | 12:57 |
| thrash | gets me this: InvalidCredentials: Invalid Credentials | 12:57 |
| thrash | Details: Credentials are: {'username': None, 'project_name': None, 'user_id': None, 'tenant_name': None, 'tenant_id': None, 'project_id': None} Password is not defined. | 12:57 |
| ddeja | thrash: I have one idea | 13:00 |
| ddeja | why you got this, give mi a second to confirm | 13:00 |
| ddeja | thrash: well, it works on my devstack | 13:03 |
| ddeja | thrash: do you run it from the tempest directory?\ | 13:03 |
| thrash | ddeja: no... Just running it from mistral dir with tempest installed in venv | 13:06 |
| *** gongysh has quit IRC | 13:07 | |
| *** rbrady has joined #openstack-mistral | 13:07 | |
| *** rbrady has quit IRC | 13:07 | |
| *** rbrady has joined #openstack-mistral | 13:07 | |
| ddeja | thrash: I remember that running tempest tests driectly from tempest dir was importnat | 13:10 |
| thrash | ddeja: let me try that. | 13:10 |
| thrash | ddeja: let me just ask this... is it possible to run the functional tests outside of an openstack environment? | 13:12 |
| ddeja | thrash: I got this when I run tests outside of tempest dir Details: Credentials are: {'username': None, 'project_name': None, 'user_id': None, 'tenant_name': None, 'tenant_id': None, 'project_id': None} Password is not defined. | 13:12 |
| ddeja | thrash: I don't know | 13:12 |
| thrash | ddeja: :D | 13:12 |
| ddeja | I just know how to run them when I need to | 13:13 |
| ddeja | :D | 13:13 |
| thrash | ddeja: fair enough | 13:13 |
| thrash | ddeja: I don't think I'll pursue this further. I'll just stick to devstack. | 13:13 |
| ddeja | OK | 13:14 |
| openstackgerrit | Dawid Deja proposed openstack/mistral: Stop excpecting to have rabbit flags set when using fake driver https://review.openstack.org/419388 | 13:24 |
| ddeja | rakhmerov: It looks like I was wrong | 13:24 |
| ddeja | all tests passes if i simply return empty dict for 'fake' driver | 13:25 |
| *** gongysh has joined #openstack-mistral | 13:36 | |
| *** hparekh has quit IRC | 13:39 | |
| *** AlexeyAbashkin has quit IRC | 13:59 | |
| *** AlexeyAbashkin has joined #openstack-mistral | 14:02 | |
| *** shardy has quit IRC | 14:05 | |
| *** shardy has joined #openstack-mistral | 14:06 | |
| ddeja | folks, we still don't know why 3-O gate is failing, do we? | 14:25 |
| *** bobh has joined #openstack-mistral | 14:30 | |
| *** AlexeyAbashkin has quit IRC | 14:45 | |
| *** gongysh has quit IRC | 14:48 | |
| *** ist has quit IRC | 15:01 | |
| *** chlong has joined #openstack-mistral | 15:03 | |
| *** jaosorior has joined #openstack-mistral | 15:22 | |
| *** AlexeyAbashkin has joined #openstack-mistral | 15:28 | |
| *** bobh has quit IRC | 15:50 | |
| *** bobh has joined #openstack-mistral | 15:53 | |
| *** AlexeyAbashkin has quit IRC | 16:27 | |
| *** chlong has quit IRC | 16:52 | |
| *** thrash is now known as thrash|biab | 17:09 | |
| *** chlong has joined #openstack-mistral | 17:26 | |
| *** jpich has quit IRC | 17:49 | |
| *** thrash|biab is now known as thrash | 17:55 | |
| *** AlexeyAbashkin has joined #openstack-mistral | 17:59 | |
| *** weshay is now known as weshay_bbiab | 18:13 | |
| *** chlong has quit IRC | 18:19 | |
| *** chlong has joined #openstack-mistral | 18:32 | |
| *** openstackgerrit has quit IRC | 18:33 | |
| *** AlexeyAbashkin has quit IRC | 18:55 | |
| *** AlexeyAbashkin has joined #openstack-mistral | 18:59 | |
| *** AlexeyAbashkin has quit IRC | 19:01 | |
| *** shardy has quit IRC | 19:18 | |
| *** weshay_bbiab is now known as weshay | 19:25 | |
| *** chlong has quit IRC | 19:41 | |
| *** chlong has joined #openstack-mistral | 19:56 | |
| *** jrist has quit IRC | 20:33 | |
| *** chlong has quit IRC | 20:44 | |
| *** AlexeyAbashkin has joined #openstack-mistral | 20:44 | |
| *** jrist has joined #openstack-mistral | 20:46 | |
| *** d0ugal has quit IRC | 20:52 | |
| *** d0ugal has joined #openstack-mistral | 20:54 | |
| *** d0ugal has joined #openstack-mistral | 20:54 | |
| kong | ping rakhmerov | 20:56 |
| kong | just try my luck to see if you are here | 20:56 |
| *** chlong has joined #openstack-mistral | 20:58 | |
| *** AlexeyAbashkin has quit IRC | 20:59 | |
| kong | after reading the blog you sent to me and did some testing, i found this: given there are 3 threads are inserting a record to names lock table with the same name in 3 different transactions. the 1st one will insert successfully and the other 2 are both waiting. | 20:59 |
| kong | rakhmerov: after 1st thread delete the record and do commit, there will be only one thread between 2nd and 3rd threads can insert record, the other one will fail with exception. | 21:02 |
| kong | rakhmerov: btw, i'm using mysql with 'repeatable read' isolation. | 21:03 |
| *** d0ugal has quit IRC | 21:03 | |
| kong | is that the expected behavior? | 21:03 |
| *** chlong has quit IRC | 21:04 | |
| *** chlong has joined #openstack-mistral | 21:18 | |
| *** rbrady is now known as rbrady-afk | 21:19 | |
| *** dturner has joined #openstack-mistral | 21:26 | |
| *** bobh has quit IRC | 21:31 | |
| *** sharatss has quit IRC | 22:06 | |
| *** sharatss has joined #openstack-mistral | 22:06 | |
| *** d0ugal has joined #openstack-mistral | 22:07 | |
| *** chlong has quit IRC | 22:29 | |
| *** dprince has quit IRC | 22:34 | |
| *** chlong has joined #openstack-mistral | 22:37 | |
| *** openstackgerrit has joined #openstack-mistral | 22:42 | |
| openstackgerrit | Brad P. Crochet proposed openstack/mistral: Put useFixture in the proper place https://review.openstack.org/419680 | 22:42 |
| openstackgerrit | Brad P. Crochet proposed openstack/mistral: Use utcnow() instead of now() https://review.openstack.org/419681 | 22:48 |
| *** jaosorior has quit IRC | 23:22 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!