Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-03-10
17:44:01 andrewbogott Is 'Host xxx has more disk space than database expected' ever anything important? That message shows up in several bug reports but best I can tell the message is a red herring in all cases.
17:44:16 andrewbogott (I have 4 out of 13 compute nodes that throw that warning regularly.)
17:47:19 openstackgerrit Eric Fried proposed openstack/nova master: PowerVM Driver: SSP emphemeral disk support https://review.openstack.org/443189
17:47:59 efried thorst_afk adreznec esberglu_ ^^ live tested, works, ready for subteam review.
17:59:47 openstackgerrit Dan Smith proposed openstack/nova master: Make os-availability-zones know about cells https://review.openstack.org/442163
17:59:47 openstackgerrit Dan Smith proposed openstack/nova master: Get instance availability_zone without hitting the api db https://review.openstack.org/439754
17:59:49 openstackgerrit Dan Smith proposed openstack/nova master: Make scheduler target cells to get compute node instance info https://review.openstack.org/439891
17:59:50 openstackgerrit Dan Smith proposed openstack/nova master: Add identity helper property to CellMapping https://review.openstack.org/444038
17:59:50 openstackgerrit Dan Smith proposed openstack/nova master: Make conductor ask scheduler to limit migrates to same cell https://review.openstack.org/438025
19:04:29 cfriesen someone pointed out to me that InstanceActionEvent.pack_action_event_finish() uses a "message" field, but the InstanceActionEvent model in the DB uses "details". Is this a bug?
19:07:53 mriedem_afk andrewbogott: this is the comment by that warning
19:07:54 mriedem_afk # can occur when an instance in database is not on host
19:08:16 mriedem andrewbogott: which is similar to the _sync_power_states periodic task in the compute manager, but that might not be a warning
19:08:18 mriedem would have to check
19:08:35 andrewbogott mriedem: thanks. I did a side-by-side comparison on one of the hosts already but maybe I missed a discrepancy.
19:08:58 andrewbogott At least… you'd expect 'nova list —host <hostname>' to show the db contents and not the disk contents, right?
19:10:10 openstackgerrit John Griffith proposed openstack/nova master: Add Cinder API version detection https://review.openstack.org/444465
19:10:30 openstackgerrit OpenStack Proposal Bot proposed openstack/nova master: Updated from global requirements https://review.openstack.org/443357
19:11:49 jgriffith ildikov mriedem ^^ don't know if that's more along the lines of what you'd like to see or not
19:12:23 openstackgerrit OpenStack Proposal Bot proposed openstack/os-vif master: Updated from global requirements https://review.openstack.org/395347
19:18:24 openstackgerrit Jay Pipes proposed openstack/nova master: virt: implement get_inventory() for Ironic https://review.openstack.org/441544
19:19:42 leakypipes cdent: hi. I'm getting a single unit test failure for nova.tests.unit.test_fixtures.TestPlacementFixture.test_responds_to_version locally on ^^ patch. it looks like this: http://paste.openstack.org/show/602305/
19:20:09 leakypipes cdent: nothing in the change to that patch should affect the placement API as far as I can tell. have you seen that error before?
19:20:28 cdent i've not, one sec
19:23:24 cdent leakypipes: are the functional tests running okay?
19:23:32 leakypipes cdent: yes.
19:26:12 cdent imma need to try this stuff myself leakypipes
19:26:31 leakypipes cdent: np
19:36:44 leakypipes JayF: thx for the reviews on the ironic patch. responded to your query on the latest revision.
19:37:24 JayF leakypipes: it's fine, "valid -1 on a jaypipes nova patch" will make a fine line on my resume ;)
19:37:34 leakypipes JayF: lmao
19:39:27 JayF leakypipes: re: your comment, that's exactly what I suspected, but wanted to be sure. Also get_inventory() will only ever be called on nodes that pass get_available_resources() which guarantees all our logic guarding nodes in maintenance/wrong states/etc will be intact
19:39:55 leakypipes JayF: correct. node_is_available() gates any call to get_inventory()
19:44:12 cdent leakypipes: does it fail consistently? does it only fail when you run all the tests?
19:44:36 leakypipes cdent: fails consistently when I run all unit tests. lemme try running it singularly.
19:44:38 cdent in other words, works for me when I test just that file and also a) make sure I don't have an /etc/nova/nova.conf and -r my tox
19:45:26 cdent I resisted this fixture. I figured we should just keep on doing the wsgi-intercept thing everywhere.
19:50:08 leakypipes cdent: yeah, fails when I just run tox -epy27 -- test_fixtures
19:50:23 cdent interesting, wfm
19:50:29 cdent i'm running all of them now
19:51:00 openstackgerrit Dan Smith proposed openstack/nova master: Teach HostAPI about cells https://review.openstack.org/442162
19:51:00 openstackgerrit Dan Smith proposed openstack/nova master: Sort CellMappingList.get_all() for safety https://review.openstack.org/443174
19:51:01 openstackgerrit Dan Smith proposed openstack/nova master: Make server_groups determine deleted-ness from InstanceMappingList https://review.openstack.org/443293
19:51:01 openstackgerrit Dan Smith proposed openstack/nova master: Add get_by_instance_uuids() to InstanceMappingList https://review.openstack.org/443292
19:51:02 openstackgerrit Dan Smith proposed openstack/nova master: Add workaround to disable group policy check upcall https://review.openstack.org/442736
19:51:02 openstackgerrit Dan Smith proposed openstack/nova master: Remove Mitaka-era service version check https://review.openstack.org/442861
19:51:03 openstackgerrit Dan Smith proposed openstack/nova master: Clean up ClientRouter debt https://review.openstack.org/444487
20:01:50 cdent GET /resource_providers?resources=VCPU:24
20:07:55 cdent leakypipes: works when running all too
20:08:17 cdent do you have any evidence from the fixture of the internal failure (the exception within the 500)
20:19:42 leakypipes cdent: none :(
20:20:03 cdent you done a fresh tox?
20:21:09 cdent :( I dunno.
20:31:17 cdent leakypipes: I gotta go, if/when you figure it out I'm curious to know. I think it's just you though: unit tests passed in the gate for that commit
20:31:31 leakypipes yeah...
20:31:39 leakypipes cdent: I'll let ya know what I discover.
20:34:06 tonyb sfinucan: np
20:37:06 openstackgerrit Rafael Folco proposed openstack/nova master: Raising http codes on old microversion https://review.openstack.org/444497
20:37:06 openstackgerrit Rafael Folco proposed openstack/nova master: DELETE all inventory for a resource provider https://review.openstack.org/416669
20:51:32 openstackgerrit John Griffith proposed openstack/nova master: Add Cinder API version detection https://review.openstack.org/444465
21:19:59 openstackgerrit Eric Fried proposed openstack/nova master: PowerVM Driver: SSP emphemeral disk support https://review.openstack.org/443189
21:20:40 efried thorst_afk ^^ addressed comments. Hoping for that +1 before you disappear into babyland.
21:20:50 thorst_afk efried: k
21:33:35 efried harlowja Another TaskFlow-havin' change set, if you're interested ^^
21:33:41 harlowja cools
21:34:01 efried harlowja Be sure to read thorst's commentary ;-)
21:34:47 harlowja kk
21:35:43 harlowja woah, lots of commentary
21:35:44 harlowja lol
21:36:16 harlowja hopefully taskflow has helped some :-P
21:36:33 efried harlowja As you can see, we definitely dig it.
21:36:49 harlowja ya, i mean i don't like having all the LinearFlow and such objects
21:37:02 harlowja i was going to try to make it more natural, but haven't quite figured that out
21:37:02 harlowja ha
21:37:32 harlowja it'd be interesting to do something like
21:37:48 efried harlowja I was telling thorst yesterday - I've had a TODO for a long time to rework our TaskFlow usage into a much more complicated unordered parallel flow with appropriate interdependencies.
21:37:59 efried He basically said, "You had my veto at 'complicated'"
21:38:13 harlowja lol
21:38:51 harlowja ya, if i could i'd somehow plug what taskflow in at more of the language level
21:39:11 efried harlowja FYI, this is the flow I was talking about: https://github.com/openstack/nova-powervm/blob/master/nova_powervm/virt/powervm/driver.py#L384
21:39:41 harlowja so that somehow u could give say a set of interconnected functions to some kind of object and it would try to extrapolate the flow(s)
21:40:16 efried harlowja You basically have that, with provides/requires. And I'm sure it works perfectly with no bugs.
21:40:17 harlowja ya, it'd be interesting to try to make such kind of functions more elegant
21:40:25 harlowja well i was thinking of something like
21:40:48 efried But I've fiddled with it a little bit and remember finding it a bit complicated to figure out how to set up the right engine/flow so it would actually be useful and run stuff in parallel.
21:41:26 harlowja b()
21:41:26 harlowja `def a():
21:41:27 harlowja now given a, is there a way to 'decompile' that into a corresponding set of linear flows (for example)
21:41:27 harlowja c()`
21:41:27 harlowja def b():
21:41:47 efried harlowja You're talking about static analysis of the code, though?
21:41:51 harlowja sort of
21:42:12 efried My knee-jerk is that you'd be passing the 80/20 point there ;-)
21:42:18 harlowja fair enough
21:42:19 harlowja ha
21:43:14 harlowja afaik the parallelism, basiclaly if u have things that can run in parallel (determined by ordering relationships); then changing the engine will run them in parallel (if there is no paralleism possible, then changing the engine won't do much)
21:44:43 efried harlowja Well, we did do some of that, down in pypowervm. You may actually get a kick out of this: https://github.com/powervm/pypowervm/blob/develop/pypowervm/utils/transaction.py#L498
21:45:04 efried (thorst_afk we were just talkin about ya)
21:45:15 harlowja interesting
21:45:52 harlowja anything in taskflow that may make such stuff easier/more pretty/other?

Earlier   Later