Earlier  
Posted Nick Remark
#openstack-nova - 2018-10-01
12:47:57 mdbooth Which would be eventlet-fu
12:48:18 mdbooth But I wonder if its execution could be blocked by, eg, a long-running periodic task
12:49:51 sean-k-mooney mdbooth: we had a downstream bugs(3) related to osp7 upgrades that saw similar behavior. compute filter elimated host becase host was still down after upgrade. in the downstream case it was causeing tempset test to fail as the heartbeat had not been recived yet and the host was down. this seams similar.
12:50:28 mdbooth sean-k-mooney: The immediate cause would certainly be the same. I wonder if the underlying cause is, though.
12:50:33 mdbooth Did you get to the bottom of it?
12:51:40 sean-k-mooney mdbooth: no in the down stream case we saw that subsequent schduling attempts suceeded as the heatbeat had been processed so we closed as not a but as the cloud had not stablised after upgrade before running the tests
12:52:05 sean-k-mooney e.g. it corrected it self a few seconds after teh tempest test was started
12:57:49 bauzas mdbooth: honestly, I don't see the problem with the service group .join() call
12:58:44 bauzas mdbooth: is this only a transient issue ?
12:59:10 bauzas mdbooth: if that's more than that, try to have a DNM patch adding some log like herehttps://github.com/openstack/nova/blob/master/nova/servicegroup/drivers/db.py#L91
13:11:01 gryf is anyone know what is happening with this blueprint: https://blueprints.launchpad.net/nova/+spec/numa-aware-live-migration ?
13:12:16 mnaser a little idle, simple clean backport if someone has a second or two: https://review.openstack.org/#/q/I811e84af46d678c3fdbf94ee400eabe659fc3d4e
13:14:19 openstackgerrit Takashi NATSUME proposed openstack/python-novaclient master: WIP: Fix test_instance_action functional test failure https://review.openstack.org/606947
13:17:30 mdbooth Folks, anybody else hitting the bug in parallel evacuate in CI I have a fix up for it here: https://review.openstack.org/#/c/605436/
13:18:08 mdbooth It's not a one line fix, though, so you'll need to block out some time to review it. However, probably not as much time as we're losing to this test failing continuously.
13:29:30 mdbooth Hmm, I just had to recheck all 3 commits from this morning for that bug :/
13:53:22 efried n-sch/placement meeting in 7 minutes in #openstack-meeting-alt
14:45:33 openstackgerrit Takashi NATSUME proposed openstack/python-novaclient master: Fix test_instance_action functional test failure https://review.openstack.org/606947
14:57:57 openstackgerrit Chris Dent proposed openstack/nova master: Do not encode server create location reponse header https://review.openstack.org/606967
15:25:04 mriedem bauzas: let's not run the linuxbridge job on all nova changes https://review.openstack.org/#/c/606616/
15:25:08 mriedem ^ fixes a regression
15:25:23 bauzas mriedem: /me looks
15:26:13 bauzas mriedem: oh shit, reviewing the py3-first changes was horribly difficult, my bad if I made a regression :(
15:26:31 mriedem it's ok, one line within hundreds, hard to spot
15:27:42 bauzas mriedem: https://review.openstack.org/#/c/606616/1/.zuul.yaml@186 why this line couldn't be only needed ?
15:28:17 bauzas ^(?!nova/network/.*)(?!nova/virt/libvirt/vif.py).*$ means 'all but nova/network or nova/virt/libvirt/vif.py AFAIU
15:28:38 bauzas but anyway, a nit
15:29:19 bauzas +Wd
15:30:21 mriedem yeah could do that in a follow up
15:39:06 openstackgerrit Matt Riedemann proposed openstack/nova master: Drop nova-multiattach job https://review.openstack.org/606981
15:41:38 openstackgerrit Chris Dent proposed openstack/nova master: Do not encode server create location response header https://review.openstack.org/606967
15:47:15 openstackgerrit Matt Riedemann proposed openstack/nova master: Drop nova-multiattach job https://review.openstack.org/606981
15:52:07 openstackgerrit Matt Riedemann proposed openstack/nova master: Drop nova-multiattach job https://review.openstack.org/606981
15:55:10 openstackgerrit Matt Riedemann proposed openstack/nova master: Remove redundant irrelevant-files from neutron-tempest-linuxbridge https://review.openstack.org/606989
15:55:30 openstackgerrit Balazs Gibizer proposed openstack/nova master: Consider nested allocations during allocation cleanup https://review.openstack.org/606050
15:55:30 openstackgerrit Balazs Gibizer proposed openstack/nova master: Ignore forcing of live migration for nested instance https://review.openstack.org/605785
15:55:31 openstackgerrit Balazs Gibizer proposed openstack/nova master: Run negative server moving tests with nested RPs https://review.openstack.org/604125
15:55:31 openstackgerrit Balazs Gibizer proposed openstack/nova master: Ignore forcing of evacuation for nested instance https://review.openstack.org/606111
16:07:15 openstackgerrit Stephen Finucane proposed openstack/nova master: conf: Allow 'nova-xvpvncproxy' to be called with CLI args https://review.openstack.org/606929
16:07:15 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Rewrite the console doc https://review.openstack.org/606148
16:07:16 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Add minimal documentation for MKS consoles https://review.openstack.org/606993
16:07:16 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Add minimal documentation for RDP consoles https://review.openstack.org/606992
16:10:34 openstackgerrit Surya Seetharaman proposed openstack/nova master: Update --max-rows parameter description for archive_deleted_rows https://review.openstack.org/606995
16:21:33 stephenfin This console documentation stuff is a damn rabbit hole :/
16:39:28 tssurya dansmith: thanks for the review on nova list patch; would you happen to have some ideas for https://review.openstack.org/#/c/591658/7/nova/compute/api.py@2327 ? melwitt and I briefly discussed the possibility to add the "type of exception" instead of the sentinel being returned..
16:39:32 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Add minimal documentation for RDP consoles https://review.openstack.org/606992
16:39:32 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Rewrite the console doc https://review.openstack.org/606148
16:39:33 openstackgerrit Stephen Finucane proposed openstack/nova master: conf: Allow 'nova-xvpvncproxy' to be called with CLI args https://review.openstack.org/606929
16:39:33 openstackgerrit Stephen Finucane proposed openstack/nova master: doc: Add minimal documentation for MKS consoles https://review.openstack.org/606993
16:44:13 tssurya dansmith: sorry for the wrong timing of the ping, I have to leave office now, but I will update the nova list patch asap and see if I can just catch the InstanceNotFound exception seperately for the above and the then mabe we can disuss a generic solution.
16:44:30 tssurya maybe*
18:24:53 cdent this very minor bug fix is just about ready to pass all its tests, if anyone is looking for something easy to review: https://review.openstack.org/#/c/606967/
18:28:23 efried cdent: Your comments is garbledes
18:29:15 cdent again?
18:30:07 efried is there be a tests somewhere that demonstrate the bug?
18:30:07 cdent ah, misplaced )
18:31:04 cdent the tests I changed?
18:31:27 efried no, I mean where were you seeing the concatenation errors you mention in the bug report?
18:31:48 efried the changed tests just prove that you changed the code in the way that you changed the code :)
18:32:23 mriedem cdent: question inlline
18:32:25 mriedem *inline
18:33:44 cdent efried: exactly, it changed from being bytes to being not bytes. the code violated the spec, I referenced the spec
18:34:49 efried yeah, cool, but the bug report says you actually encountered a real problem as a result of the code being wrong. Something about concatenating the str url with the unicode header under wsgi intercept?
18:36:12 cdent efried: in one of the iterations of https://review.openstack.org/#/c/601614/
18:36:23 cdent http://logs.openstack.org/14/601614/25/check/placement-tempest-gabbi/d9a37a4/controller/logs/screen-n-api.txt.gz#_Sep_29_09_57_29_463012
18:37:15 mriedem dansmith: want to this the queens version of that add host to aggregate case revert change (and the one after it): https://review.openstack.org/#/c/605266/1
18:37:17 cdent efried: but mriedem raises a good question. It appears like things may be more broken than I hoped (I guess I never learn)
18:37:24 mriedem s/this/hit/
18:37:44 dansmith mriedem: omg I so want
18:41:16 dansmith heh
18:41:39 dansmith mriedem: lee was +2 then removed.. I went to +W but then figured I must be missing something?
18:41:52 mriedem i think melwitt asked the same thing last week
18:42:33 mriedem i can only assume b/c the rocky change hadn't merged yet
18:42:58 mriedem you might as well re-apply that +W b/c it's in the gate
18:43:05 dansmith gah okay
18:43:09 mriedem let's use the rule of stable core backport == +2
18:43:54 mriedem dansmith: in case you haven't seen my email yet, i'm trying to flush the open ocata backports so we can EM https://etherpad.openstack.org/p/nova-ocata-em
18:44:25 dansmith okay
18:44:33 mriedem if you want to do some mindless stable branch reviews on your first day back
18:44:50 cdent efried: sigh. you know what you were saying about "senile" last week? Those gabbits are not using wsgi intercept at all that's a live devstack. I guess I will spin one up
18:45:49 efried glad I could help in not at all a way I intended or even understand.
18:48:14 cdent efried: what it means in this case it is that it is likely the problem is present in "real" situations
18:48:20 cdent so now I'm going to check
18:48:24 efried cool
18:52:13 cfriesen mriedem: once you're done the ocata stuff, any chance you could look at https://review.openstack.org/#/c/605164/ ? since you triaged the bug :)
18:53:44 mriedem that was the old me
18:54:12 mriedem i also tried fixing that and couldn't recreate it in the functional test...
18:57:58 cfriesen so, need to reproduce for real first on master?
19:04:48 mriedem not necessarily
19:05:06 mriedem the fix is the same thing i was doing,
19:05:13 mriedem i just couldn't get a recreate in the functional test,
19:05:24 mriedem but was also doing it while at one of my kids activity things
19:05:30 mriedem like a champion dad
19:07:37 cfriesen mriedem: you were actually doing the delete/purge API calls, right? (as opposed to this patch which is doing a direct DB operation)
19:09:38 mriedem i deleted the marker record from the instances table but probably forgot to archive it
19:20:51 mriedem cfriesen: i just need to wrap up what i'm doing locally first
19:21:01 cfriesen mriedem: no worries
19:39:41 openstackgerrit Jack Ding proposed openstack/nova master: Add HPET timer support for x86 guests https://review.openstack.org/605902
19:42:43 openstack Launchpad bug 1795425 in OpenStack Compute (nova) "create server api sends location header as bytestring in py3" [Undecided,In progress] - Assigned to Chris Dent (cdent)
19:42:43 cdent mriedem, efried : the location header issue is weirder/worse than I feared: https://bugs.launchpad.net/nova/+bug/1795425

Earlier   Later