Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2018-06-11
15:09:11 efried dansmith: I foudn anothre tyop.
15:09:22 efried dansmith: But also content-wise, left some questions.
15:09:59 tssurya mriedem: thanks (#CERN totally appreciates) for the initiative regarding the handling the cell down spec on the mailing list and the discussion during the nova-team meeting, I will update the spec soon (just got back from Greece this morning).
15:11:46 mriedem tssurya: yw
15:12:13 mriedem unfortunately no user/ops feedback in the ML or spec
15:12:25 openstackgerrit Dan Smith proposed openstack/nova master: Document the internal online_migrations function behaviors https://review.openstack.org/574268
15:15:14 dansmith mriedem: I think you should just go +W that patch of jay's.. my only remaining comment was about functional tests which I'm sure you're on top of
15:16:40 mriedem smcginnis: dims: should we be also bumping lower-constraints in a project when requirements.txt is updated?
15:16:44 mriedem https://review.openstack.org/#/c/486475/ for example
15:17:22 dansmith oh did I forget to do that?
15:17:26 smcginnis mriedem: Yeah, I believe that should be done in lower-constraints.
15:17:31 mriedem dansmith: you did, but that's not your patch
15:17:35 dansmith oh
15:17:40 dansmith I was going to say
15:18:05 dansmith surely the requirements job should check that?
15:18:25 mriedem there is a lower-constraints job
15:18:39 mriedem but i didn't see it run on https://review.openstack.org/#/c/566696/
15:19:05 smcginnis I'm honestly not sure what the job checks now with different versions set between requirements and lower-constraints.
15:19:35 smcginnis It may just check that requirements isn't less than lower-constraints, but seems like it should also check the other way as well.
15:19:45 smcginnis Or versions should just be removed from requirements.
15:19:46 mriedem lower-constraints just runs unit tests with the lower constraints installed, right? whereas py27 runs with upper-constraints
15:19:55 smcginnis Yeah
15:20:00 mriedem we need more requirements files
15:20:04 mriedem that's all i know
15:20:06 openstackgerrit Curt Moore proposed openstack/nova master: Add ability to download Glance images into the libvirt image cache via RBD https://review.openstack.org/574301
15:20:12 smcginnis Yeah, half a dozen more and we should be solid.
15:22:30 mriedem dansmith: nvm you have that in htere
15:22:35 mriedem the job didn't run on your patch though for some reason
15:22:48 dansmith it trusts me?
15:22:55 mriedem no, shit failed hard
15:23:05 mriedem openstack-tox-lower-constraints finger://ze02.openstack.org/d6e860451b074239a28aae41f39381c8 : POST_FAILURE in 9m 56s
15:23:38 mriedem the xenapi patch last ran jobs in may before we had the lower-constraints job so that's probably why it wasn't caught there
15:23:46 dansmith ah and doesn't get reported up top I guess?
15:23:57 mriedem i would have expected those POST_FAILURES to get reported up top
15:24:02 mriedem they used to with lots of red
15:24:09 mriedem maybe new gerrit / zuul doesn't show that now?
15:29:22 dansmith idk, seems confusing
15:29:32 dansmith because zuul -1s and reports few jobs, all passing
15:30:02 dansmith efried: left comments on your spec from skimming
15:31:24 efried dansmith: Thanks. Are you going to do a patch set adding words for things you said you were going to compose?
15:31:53 dansmith efried: I commented on that one place, which I think is enough, but I think you have things you need to tweak anyway right?
15:32:23 efried dansmith: I do. Okay, I'll add that in, if you think that's enough.
15:32:36 efried terse guy that you are.
15:32:45 dansmith efried: you covered what it'll do pretty well I think
15:32:55 efried okay, good
15:34:47 mriedem naichuans: is it possible to get a test run for the xenserver ci with image_handler=vdi_remote_stream to make sure this code passes tempest?
15:36:44 efried dansmith: If (found, done) == (!0, 0), it means there are non-migratable records. What am I supposed to do if I get that?
15:37:05 dansmith efried: who is "I" in that case?
15:37:18 dansmith efried: if "I" is the CLI user, I think it's handled in the docs for that command
15:37:23 efried dansmith: I don't know. What does the code do? Does it raise?
15:37:56 efried dansmith: I'm more asking for my edification, not suggesting the docs are inaccurate/inadequate.
15:38:33 dansmith efried: these functions migrate stuff, what they do and why they might be unable to migrate some records depends on what they're migrating
15:38:48 dansmith efried: if I'm migrating instances, I might refuse to touch instances with task_state!=None for example
15:39:15 dansmith efried: so I might say "there are 10 more instances to do, but I could only do 8" and then next I might say "there are 2 more to do but I could only do 0"
15:40:32 mriedem once done is 0 across the board, the CLI returns 0 and yo'ure done
15:40:58 efried dansmith: I guess I'm trying to understand whether there are cases where that means "the migration failed, push the red button and blow yourself up" or "go do XYZ manually" or "meh, don't worry about it, these things happen".
15:41:06 mriedem efried: if this helps https://review.openstack.org/#/c/574268/1/nova/cmd/manage.py@387
15:41:57 mriedem return ran and 1 or 0
15:41:59 efried mriedem: Right, that's the sentence that's prompting my question. "We've done all the ones we can do." So... what about the rest of them?
15:42:02 mriedem ran is based on done
15:42:12 dansmith mriedem: not really.. if done is zero it just means you're not able to do more work, but it doesn't mean everything is migrated
15:42:41 dansmith efried: it depends
15:43:00 dansmith efried: in my instance example, it means "try later once those instances are no longer transitioning state"
15:43:02 mriedem dansmith: then that's not clear from the command help https://docs.openstack.org/nova/latest/cli/nova-manage.html#nova-database
15:43:08 mriedem "Perform data migration to update all live data. Return exit code 0 if migrations were successful or exit code 1 for partial updates. This command should be called after upgrading database schema and nova services on all controller nodes. If the command exits with partial updates (exit code 1) the command will need to be called again."
15:43:35 dansmith mriedem: it's a minor loop vs. major loop type of thing
15:43:45 mriedem so,
15:43:47 efried ah, that indicates that (!0, 0) means "keep doing this until (0, 0)"
15:43:50 dansmith mriedem: nonzero means ansible should call again immediately.. zero means there's no more you can do right now
15:43:59 dansmith I can update those words too
15:44:02 mriedem it sounds like what you'd really need, if you were automating this, is rc=2 == total != 0, but done == 0
15:44:30 efried no, it sounds like rc=1 means that already.
15:44:41 mriedem but rc=1 is only based on "done != 0"
15:44:54 efried oh, okay. So yeah.
15:44:58 efried what you said.
15:45:20 dansmith mriedem: we could, but anyone currently just checking for nonzero would get stuck, because it'd be changing what zero really means
15:45:23 mriedem otherwise if there is no other return code for the total != 0, done == 0 case, i can only know that by manually running things
15:45:41 dansmith early on in this, we didn't have migrations that couldn't complete fully, until we did, so..
15:45:45 mriedem dansmith: sure, osa uses [0, 4] for things like that
15:46:00 mriedem i realize it would be backward incompatible, and upgrade release note, etc
15:46:14 dansmith mriedem: I don't know what that means
15:46:26 mriedem which part?
15:46:33 dansmith oh you mean [0,4] for either return value, sure,
15:46:36 dansmith but it has to change
15:46:40 mriedem right
15:46:56 mriedem if you're just saying != 0 in your tooling, that's kind of on you
15:47:06 mriedem anyway, i don't mean to open a can of worms
15:47:13 dansmith well, you did
15:47:16 dansmith worms. everywhere.
15:47:50 dansmith although efried is like the child that just starts throwing cans on the floor of the supermarket aisle
15:47:59 dansmith denting cans left and right until one opens
15:48:10 efried lol
15:48:20 efried like a child in my ignorance, at least.
15:48:51 efried My vote is that we merge this as is and (maybe) figure out what to do about that other thing later.
15:49:14 mriedem sure, throw a TODO in there about the rc=2 idea i guess
15:50:29 openstackgerrit Dan Smith proposed openstack/nova master: Document the internal online_migrations function behaviors https://review.openstack.org/574268
15:50:34 mriedem efried: can you double check https://review.openstack.org/#/c/486475/36/doc/source/admin/configuration/hypervisor-xen-api.rst and if i'm right, i'll update the patch for my comments and respin so we can flush it
15:50:53 efried ...
15:51:04 mriedem flush from the runways spot i mean
15:53:46 efried mriedem: You're right. History there was that they were going to switch the default, but I said nay. They wound it back, but missed a couple spots. Good eye.

Earlier   Later