| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2018-08-06 | |||
| 13:16:37 | cdent | mdbooth: sorry, I've got no clues. I've not been much involved in osc-placement's birth | |
| 13:16:39 | efried | mdbooth: --test-path=... | |
| 13:17:17 | mdbooth | efried: Is that passed to testr? | |
| 13:17:46 | mdbooth | https://docs.openstack.org/pbr/latest/user/using.html#testing This suggests it might be tox.ini, although there's a deprecation notice | |
| 13:18:11 | efried | mdbooth: have we not switched osc-placement to stestr yet? | |
| 13:18:56 | mdbooth | efried: pbr is the darkest most arcane magik | |
| 13:19:03 | mdbooth | I have no idea :) | |
| 13:19:05 | cdent | are you building rpms mdbooth, and thus forced through setup.py? | |
| 13:19:11 | mdbooth | cdent: Yes | |
| 13:19:15 | efried | mdbooth: I would run to mtreinish if it were me. | |
| 13:19:26 | mdbooth | Well not *forced*. It's just a script, I can run whatever I want. | |
| 13:19:53 | mdbooth | But everything else runs setup.py test, so I'd prefer to make that work for consistency | |
| 13:19:58 | cdent | you might also try efried's suggestion with a --<space> before the --test-path | |
| 13:20:15 | mdbooth | What is the argument to --test-path? | |
| 13:20:45 | efried | not having the repo in front of me, I would guess ./osc_placement/tests/unit | |
| 13:21:25 | mdbooth | efried: Ah, ok. | |
| 13:24:44 | mdbooth | Eurgh... It doesn't work and I'm way too hot and I just want to lie in an ice bucket and weep until it all goes away. | |
| 13:40:51 | mdbooth | So I have a behaviour difference here between the CI environment and locally. CI runs 'python2 setup.py test' and it does All The Things(tm). Locally it does nothing: http://paste.openstack.org/show/727412/ | |
| 13:41:35 | mdbooth | CI: (search for 'python2 setyp.py test') https://logs.rdoproject.org/81/15181/3/check/legacy-rdoinfo-DLRN-check/83a6ba9/buildset/centos-rpm-master/repos/95/77/9577cd899541b4c2a5b9fa74a59f4c346bd5addf_dev/rpmbuild.log | |
| 13:42:02 | mdbooth | Any idea what environment factors affect setup.py test? | |
| 13:50:36 | efried | n-sch meeting in 10 minutes in #openstack-meeting-alt | |
| 13:51:21 | dansmith | lyarwood: mriedem I left comments on that review.. definitely seems like obligatory pings to cinder for every build is a less-than-awesome change, and would rather see it be something we do on reschedule if we need | |
| 13:51:51 | dansmith | also, queried about whether or not the GET is enough. I guess I would expect there are some attachment states that can't be reversed blindly by the boot process? | |
| 13:52:59 | lyarwood | dansmith: ack thanks | |
| 13:55:45 | openstackgerrit | sahid proposed openstack/nova master: hardware: fix memory check usage for small/large pages https://review.openstack.org/532168 | |
| 13:56:00 | openstackgerrit | Lee Yarwood proposed openstack/nova master: DNM/WIP compute: Reduce likelihood of bdm creation race during attach https://review.openstack.org/589164 | |
| 13:59:47 | mriedem | dansmith: lyarwood: that's where i'm leaning - re-create attachments in conductor build_instances if we're rescheduling | |
| 13:59:56 | dansmith | yeah | |
| 14:00:03 | dansmith | t'would make more sense to me to do that | |
| 14:00:48 | lyarwood | mriedem / dansmith ; kk, tbh I've spent very little time looking around within conductor and just assumed it wouldn't be the right place, I'll try to move things over now and respin. | |
| 14:04:59 | mdbooth | OS_TEST_PATH | |
| 14:07:58 | efried | mdbooth: \o/ nice one | |
| 14:08:29 | mdbooth | Although I can't see why rpmbuild isn't using the default | |
| 14:14:58 | mdbooth | mriedem: Wow, brave. | |
| 14:15:42 | mriedem | already mostly done | |
| 14:15:54 | mdbooth | Any good nuggets? | |
| 14:16:46 | mriedem | mdbooth: i'll post a summary with a link to my spreadsheet in the ML once i'm done | |
| 14:16:55 | mriedem | there are definitely some things we can upstream to nova | |
| 14:17:21 | mriedem | given it was a snapshot based on pike, there are also several bug fixes they have backported so not really forks, just cherry picks we don't have in stable/pike | |
| 14:17:40 | mriedem | and other bug fixes that weren't reported upstream | |
| 14:17:56 | mriedem | e.g. https://review.openstack.org/#/c/588689/ | |
| 14:18:05 | mriedem | and https://review.openstack.org/#/c/588657/ | |
| 14:18:26 | mdbooth | Nice | |
| 14:18:32 | mdbooth | Ah, so you do at least have a git repo | |
| 14:18:43 | mdbooth | You're not just staring at a massive unified diff | |
| 14:18:54 | mriedem | i am | |
| 14:18:55 | mriedem | https://github.com/starlingx-staging/stx-nova/commit/71acfeae0d1c59fdc77704527d763bd85a276f9a | |
| 14:19:07 | mriedem | oh yeah i do have the actual git repo cloned | |
| 14:19:11 | mriedem | so i can search things for context | |
| 14:19:14 | mriedem | but i'm mostly starting with the diff | |
| 14:20:04 | mriedem | i'm also starting to glaze over things toward the end here out of sheaer fatigue | |
| 14:20:06 | mriedem | *sheer | |
| 14:26:56 | gibi | dansmith: hi! I'm looking at some migation revert functional test case and it seems we are still hitting the legacy allocation handling code path when reverting the allocation on the destination | |
| 14:27:08 | gibi | dansmith: I think it is a bug in https://github.com/openstack/nova/blob/8688b25ca7379391cba28fab30b5a628957e673e/nova/compute/manager.py#L3952-L3955 | |
| 14:27:40 | gibi | dansmith: cn_uuid points to the destinantion host based but the migration.uuid only supposed to hold allocation on the source host | |
| 14:27:53 | gibi | dansmith: So the if condition never be true | |
| 14:28:41 | gibi | dansmith: Did I missed something or it is really a bug? | |
| 14:30:56 | dansmith | gibi: there's something tricky about this, let me re-load context | |
| 14:37:47 | openstackgerrit | Lee Yarwood proposed openstack/nova master: conductor: Recreate volume attachments during a reschedule https://review.openstack.org/587071 | |
| 14:40:51 | dansmith | gibi: this is the context I was thinking of, although reading that review comment now doesn't seem to make sense: https://review.openstack.org/#/c/498948/9/nova/compute/manager.py@3648 | |
| 14:41:07 | dansmith | gibi: however, I rarely add code like this without some test fail making me do it, | |
| 14:41:11 | dansmith | so I'm not sure | |
| 14:41:32 | dansmith | maybe raise there and run all the test_servers tests to see if something fails? | |
| 14:43:24 | dansmith | it's possible that it was there for some case which has since been removed, but.. worth a shot | |
| 14:43:51 | dansmith | I was thinking that we actually call that on both the source and destination | |
| 14:43:58 | dansmith | because of finish_resize_revert_on_destination(), | |
| 14:44:01 | dansmith | but i don't see it in there | |
| 14:44:07 | gibi | dansmith: I will check resize-same-host sceneario. a simple migrate revert func test now goes to the legacy path | |
| 14:44:08 | dansmith | so I wonder if it was at some point | |
| 14:44:23 | dansmith | gibi: we have those tests in tree though right? | |
| 14:44:30 | dansmith | to verify the migration-holding allocations? | |
| 14:44:32 | gibi | dansmith: we have functional tests | |
| 14:44:47 | gibi | dansmith: for same host resize and migrate too | |
| 14:45:16 | dansmith | gibi: right, which validate the non-legacy path works yeah? | |
| 14:45:51 | gibi | dansmith: which validates that allocations handled properly in non-legacy case. But I think those tests are still hitting the legacy path | |
| 14:46:10 | gibi | let me reproduce both with migrate and same host resize | |
| 14:46:29 | dansmith | hmm | |
| 14:46:41 | dansmith | so the tests aren't noticing that we're leaking an allocation or something? | |
| 14:47:20 | gibi | dansmith: I think we are eventually not leaking as the legacy codepath also handles the alloction properly | |
| 14:48:01 | dansmith | okay I'm not sure how that could be, if it's not handling the migration uuid, but I'll wait for your analysis :) | |
| 14:48:02 | gibi | dansmith: https://github.com/openstack/nova/blob/8688b25ca7379391cba28fab30b5a628957e673e/nova/compute/manager.py#L3982 this will remove the allocation on the destination | |
| 14:48:25 | dansmith | on the dest, yeah, but the source? | |
| 14:48:55 | openstackgerrit | Merged openstack/nova master: Reload oslo_context after calling monkey_patch() https://review.openstack.org/587772 | |
| 14:49:15 | gibi | dansmith: yeah, I'm confused now. I will do the reproduction | |
| 14:50:24 | mriedem | the allocation on the source should get moved by conductor right? or is this for the "old computes" scenario? | |
| 14:51:11 | dansmith | mriedem: he thinks we're taking the old doubled patch during some cleanups | |
| 14:51:25 | dansmith | but I'm not sure how we could, without failing those tests (or having big holes in them | |
| 14:51:59 | dansmith | well, and his point is there is a clause in the delete_after_move function we can't explain | |
| 14:52:04 | dansmith | which he thinks might be related | |
| 14:56:09 | dansmith | gibi: I added an exception to that clause and we _do_ hit it in the functional tests | |
| 14:56:16 | dansmith | so that _is_ true somewhere :) | |
| 14:56:27 | dansmith | during revert resize | |
| 14:57:02 | dansmith | in three tests | |
| 14:57:40 | dansmith | test_migrate_revert, test_resize_revert, and test_resize_revert_reverse in ServerMovingTests | |
| 15:05:54 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Define irrelevant-files for tempest-full-py3 job https://review.openstack.org/589039 | |
| 15:06:31 | gibi | dansmith: did you put the raise in the legacy path? | |
| 15:06:47 | dansmith | gibi: no, in the clause you think can't ever be true | |
| 15:07:05 | dansmith | gibi: this: https://pastebin.com/GC17gjG4 | |
| 15:08:46 | gibi | dansmith: I put it in front of the legacy code and hit the same tests you hit | |
| 15:09:01 | gibi | dansmith: it doesn't make any sense | |