*** rcernin has quit IRC | 00:07 | |
*** rcernin has joined #openstack-sahara | 00:09 | |
*** tosky has quit IRC | 00:09 | |
*** tellesnobrega has quit IRC | 04:24 | |
*** chason has quit IRC | 05:02 | |
*** pcaruana has joined #openstack-sahara | 07:06 | |
*** rcernin has quit IRC | 07:26 | |
*** tosky has joined #openstack-sahara | 08:37 | |
*** tellesnobrega has joined #openstack-sahara | 11:12 | |
tosky | tellesnobrega: hi! About the openstack/release change, this is the context: http://eavesdrop.openstack.org/irclogs/%23openstack-release/%23openstack-release.2019-01-09.log.html#t2019-01-09T10:11:00 | 12:06 |
---|---|---|
tosky | (even if you already voted on it, it's still useful to know) | 12:06 |
tellesnobrega | tosky, thanks | 12:06 |
tellesnobrega | tosky, quick unrelated question | 12:12 |
tellesnobrega | when running unit tests what config file does sahara load? | 12:12 |
tosky | I don't know; don't the base classes of unit tests load some basic parts of the sahara engine? | 12:31 |
tellesnobrega | it does, apparently it doesn't load a file, it just sets the db part, and in some tests it overrides the plugins info | 12:33 |
tellesnobrega | tosky, regarding the oozie patch, for that to work, we have to change sahara-tests to add oozie on the fake plugin | 12:51 |
tellesnobrega | I think | 12:51 |
tosky | oh, let me try - or do you want to try? | 13:09 |
tellesnobrega | if you can try I appreciate | 13:11 |
tosky | tellesnobrega: patch sent, you can add a Depends-On: on your patch | 13:11 |
tellesnobrega | I'm on the program call and updating bug scrub doc | 13:11 |
tellesnobrega | thanks | 13:11 |
tosky | or I guess I can add the Depends-On | 13:11 |
tosky | tellesnobrega: adding the depends-on | 13:12 |
tellesnobrega | tosky, thanks | 13:13 |
*** vkmc has left #openstack-sahara | 13:29 | |
tellesnobrega | tosky, did you add the depends -on? | 13:47 |
*** tmckay has joined #openstack-sahara | 14:11 | |
tellesnobrega | tosky, looks like it worked | 14:50 |
tellesnobrega | just need to fix grenade now | 14:51 |
*** chason_ has joined #openstack-sahara | 15:09 | |
*** pcaruana has quit IRC | 16:03 | |
tosky | oh! | 16:08 |
tosky | tellesnobrega: we may need to invert the order of patches though | 16:09 |
tosky | more complicated: we need to backport the sahara patch on all supported branches | 16:21 |
tellesnobrega | the oozie one? | 16:35 |
tosky | yes, I changed the order | 16:36 |
tosky | the oozie patch in fails in sahara-tests because oozie is not available right now in the fake plugin | 16:36 |
tosky | and if we change it to master only, it will fail on older branches, because the same test is executed on all the branches | 16:37 |
tosky | unless we add more code to differentiate between the fake job run on master (with oozie) and without (no oozie) | 16:37 |
tellesnobrega | I see | 16:37 |
tosky | but it's probably easier to add use to the fake plugin for all branches | 16:37 |
tosky | or disable that test :) | 16:38 |
tellesnobrega | disabling the test is the easier one for sure | 16:38 |
tellesnobrega | simple, and since it is fake, no harm is done | 16:38 |
tellesnobrega | at least from what I can see | 16:38 |
tosky | oki, fine with that | 16:39 |
tosky | we can reeavaluate it later | 16:39 |
tosky | it was an interesting test though | 16:39 |
tellesnobrega | of course | 16:39 |
*** elmiko has quit IRC | 16:39 | |
tellesnobrega | yes | 16:39 |
tellesnobrega | from your perspective, what do you suggest? | 16:39 |
* tosky looks at the calendar | 16:40 | |
tosky | I think that oozie was enabled to exercise some additional code | 16:42 |
tosky | but now that I think of it, and you may have it explained already but I forgot, why did it break? The scenario test used fake even before the split | 16:42 |
tellesnobrega | the problem was with unit tests | 16:44 |
tellesnobrega | some tests used vanilla plugin and I had to change to fake | 16:44 |
tellesnobrega | so I had to add some code to the validate_job_execution on fake_plugin | 16:45 |
tellesnobrega | and I got that chunk of code from vanilla | 16:45 |
tellesnobrega | and there is check for oozie | 16:45 |
tosky | oh, ok, so it is checked also on scenario test | 16:45 |
tosky | but shouldn't the fake plugin ignore oozie if oozie is not enabled? | 16:46 |
tellesnobrega | I think the check breaks that, forcing it to have at least 1 oozie enabled | 16:48 |
tellesnobrega | like a datanode on spark | 16:48 |
tellesnobrega | tosky, https://review.openstack.org/#/c/629028/11/sahara/plugins/fake/edp_engine.py | 16:49 |
tellesnobrega | the issue is in line 45 | 16:49 |
tellesnobrega | but, I can remove that if we remove the test as well | 16:50 |
tosky | tellesnobrega: did you basically copy the validate_job_execution from vanilla? | 16:51 |
tellesnobrega | yes | 16:51 |
tellesnobrega | tosky, so, what do you think is the best way here? keep the test and code, or drop it | 16:57 |
tosky | I'm trying to understand how much testing would be lost | 16:57 |
tellesnobrega | just one test | 16:57 |
tosky | ok, let's remove it for now | 16:58 |
tosky | maybe it can be *later* moved to the unit tests of the vanilla plugin | 16:58 |
tellesnobrega | https://review.openstack.org/#/c/629028/13/sahara/tests/unit/service/validation/edp/test_job_executor.py | 16:58 |
tellesnobrega | tosky, yes | 16:59 |
tellesnobrega | lin 153 | 16:59 |
tellesnobrega | :wqa | 16:59 |
tosky | $ | 16:59 |
tellesnobrega | closing on vim | 16:59 |
tosky | and returning to the prompt | 17:00 |
*** elmiko has joined #openstack-sahara | 18:59 | |
tellesnobrega | tosky, did you see mike's comment? | 19:10 |
elmiko | i can repeat here | 19:10 |
elmiko | i don't _think_ there a problem releasing microversions during a cycle, but the important thing it to ensure that backward compat is not broken | 19:11 |
elmiko | it shouldn't be too bad because iirc the default behavior for not specifying a microversion to give the minimum version | 19:11 |
elmiko | the main key is to not break anyone and ensure backward compat behavior | 19:12 |
elmiko | outside of that, i would think you can add the microversions | 19:12 |
tellesnobrega | good thing is apiv2 is being released just now, probably no one will be using before we release microversioning | 19:13 |
tosky | it's about microversion support | 19:13 |
elmiko | tellesnobrega: yeah, but still have to take into account the situations where someone might have initial v2 and needs to continue using without breaking on microversions | 19:14 |
tellesnobrega | tosky's comment is important | 19:14 |
tellesnobrega | what I meant was adding micronversion support | 19:14 |
elmiko | i still don't think it would be an issue adding support for microversions during the cycle, as long as you take care of the backward compat | 19:15 |
elmiko | but, that might be a sticking point | 19:15 |
tellesnobrega | we are just trying to be safe here, we might even be able to do it now, so hoping to finish in time | 19:16 |
elmiko | for maximum safety, probably best to release the support around a cycle break | 19:17 |
elmiko | if you guys would like, i can bring this up in our api-sig office hours tomorrow? | 19:17 |
tosky | yes, please | 19:17 |
elmiko | ++ | 19:17 |
tellesnobrega | thanks elmiko | 19:17 |
elmiko | so, just to be clear. you would like to release v2 at the start of the cycle and then add support for microversions during the cycle. is that accurate? | 19:18 |
tosky | it may happen that we may need to make v2 STABLE now (M2) and add microversioning support before M3 | 19:19 |
tosky | would that be a problem? | 19:19 |
elmiko | got it, thanks! | 19:19 |
elmiko | and just to make sure i haven't missed anything in the last year(XD), v1 does _not_ have microversion support? | 19:20 |
tellesnobrega | no | 19:25 |
elmiko | got it | 19:27 |
elmiko | i know i'm gonna bring this up and they are going to ask some crazy question, so i'm just trying to be prepared ;) | 19:27 |
tosky | elmiko: is it this one? http://eavesdrop.openstack.org/#API_Special_Interest_Group_office_hours | 19:28 |
tosky | I should be around at 16:00 UTC | 19:28 |
elmiko | yes | 19:32 |
tosky | oki | 19:35 |
tosky | and I hang around on that channel | 19:35 |
tosky | (in that channel?) | 19:35 |
*** openstackgerrit has joined #openstack-sahara | 19:39 | |
openstackgerrit | Telles Mota Vidal Nóbrega proposed openstack/python-saharaclient master: Preparing OSC for APIv2 https://review.openstack.org/584110 | 19:39 |
tellesnobrega | tosky, can you check if this changes fixes the issue? | 19:39 |
tosky | tellesnobrega: sure | 19:39 |
tellesnobrega | thanks :) | 19:40 |
elmiko | tosky: yes! | 19:41 |
tellesnobrega | tosky, also looks like this one is ready https://review.openstack.org/#/c/623300/12 | 20:04 |
tosky | tellesnobrega: right, I was waiting for the +v | 20:28 |
tosky | nothing happens without it anyway | 20:28 |
tellesnobrega | yeah | 20:28 |
tellesnobrega | looks like I have to update some tests on python-saharaclient patch as wel | 20:38 |
tellesnobrega | well | 20:38 |
tellesnobrega | trying to figure out how to fix it | 20:38 |
tosky | when you are done with that, do you have any idea about the error highlighted by Jeremy's comment in revision 9 of https://review.openstack.org/#/c/629337/ ? | 20:41 |
tosky | also, this is good to go: https://review.openstack.org/#/c/628056/ | 20:41 |
tellesnobrega | tosky, not sure how it was introduced by the other patch | 20:46 |
tellesnobrega | there are a lot of errors on apiv2 part | 20:48 |
tellesnobrega | apiv1 has only 1 | 20:48 |
tosky | ignore the hadoop_version errors; I think that the comment is about the Internal Server errors on data sources | 20:49 |
tellesnobrega | ok | 20:51 |
tellesnobrega | I will take a look | 20:51 |
tellesnobrega | I'm still trying to figure out the osc issue yet | 20:51 |
tellesnobrega | you might know it better | 20:51 |
tellesnobrega | as you may have seen on the patch, I changed the jobs to job_template... on the tests I'm doing the same but it keeps saying that get_configs() is not being called | 20:52 |
tosky | I don't know that part too much, but let me send an update for the tempest patch | 20:59 |
tosky | and see if I can reduce the failures | 20:59 |
tellesnobrega | sure | 21:00 |
tellesnobrega | sounds good | 21:00 |
elmiko | tosky man, really burning the midnight oil! | 21:10 |
tosky | now or never | 21:13 |
tosky | well, not never, but... not another cycle o.O | 21:14 |
tellesnobrega | elmiko, he is an inspiration to all of us | 21:14 |
tosky | I just pretend I'm typing something meaningful | 21:17 |
elmiko | haha, tosky ++ | 21:25 |
elmiko | tellesnobrega: definitely =) | 21:26 |
tellesnobrega | tosky, any luck with tempest? | 22:18 |
tosky | tellesnobrega: not yet, I misnamed a variable thus invalidating the results, I sent another patch | 22:26 |
tosky | I also deployed a devstack machine (I should have done it earlier) and I will probably do some local testing there | 22:27 |
tellesnobrega | makes sense | 22:29 |
tellesnobrega | but please | 22:29 |
tellesnobrega | go ahead a get some sleep | 22:29 |
tellesnobrega | we can get to it tomorrow | 22:30 |
tellesnobrega | and if not we can try and push to friday as last resort | 22:31 |
tosky | sure :) | 22:35 |
tellesnobrega | I think I got it for the osc patch | 22:38 |
tellesnobrega | yeah | 22:38 |
tellesnobrega | unit tests passed | 22:38 |
openstackgerrit | Telles Mota Vidal Nóbrega proposed openstack/python-saharaclient master: Preparing OSC for APIv2 https://review.openstack.org/584110 | 22:38 |
*** rcernin has joined #openstack-sahara | 22:52 | |
*** tellesnobrega has quit IRC | 22:58 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!