Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2018-01-05
16:06:00 openstackgerrit melanie witt proposed openstack/nova stable/pike: libvirt: Re-initialise volumes, encryptors, and vifs on hard reboot https://review.openstack.org/531407
16:06:01 kashyap Currently, I'm in split-brains mode, and trying to finishing something before the 'cache is flushed
16:06:43 mriedem kashyap: i did, and it was just something i should have realized wouldn't apply here earlier on
16:06:47 mriedem since imagebackend != volumes
16:08:03 mriedem johnthetubaguy: oh i think i see,
16:08:20 mriedem the ComputeFilter is running every 60 seconds due to the caching scheduler's periodic task interval,
16:08:20 figleaf fried_rice: sorry, missed your question earlier - network issues.
16:08:31 mriedem which means it's potentially using a stale version of the compute service reference,
16:08:33 figleaf fried_rice: mriedem answered what I would have said, so no worries
16:08:36 johnthetubaguy mriedem: sorry, yeah, that's it
16:08:38 mriedem which means it's last_seen_up would be old
16:08:42 johnthetubaguy mriedem: yeah
16:09:22 johnthetubaguy mriedem: hence the note about it not seeing instances being deleted, and being really bad if you run two schedulers
16:09:53 mriedem heh
16:10:03 leakypipes johnthetubaguy: make checks payable to fried_rice.
16:10:04 mriedem if you're wondering why i'm asking https://review.openstack.org/#/c/531020/
16:10:26 Roamer` johnthetubaguy, did you catch my earlier messages about your procedural -2 on https://review.openstack.org/140733/? or am I being annoying already? :)
16:10:27 mriedem people are still using the cachingscheduler but we don't really test it functionally anywhere, so i'm trying to flush out some squirrels
16:11:13 leakypipes totally off-topic, but why does gerrit allow someone to hit a "quote" button on a review comment that is on a source file but won't provide a "quote" button to respond to a review comment that is on the review (not a source file)...
16:11:50 johnthetubaguy Roamer: ah, sorry, I missed those, was visiting a customer this morning
16:12:31 melwitt leakypipes: there's a button but it doesn't say "quote". it's a curved arrow next to the comment timestamp
16:12:33 mriedem Roamer`: you should probably be talking to me - i haven't been tracking that blueprint for nova in queens at all
16:12:41 johnthetubaguy mriedem: I don't think this once has been re-approved: https://blueprints.launchpad.net/nova/+spec/libvirt-storpool-volume-attach
16:12:51 mriedem Roamer`: where is the patch that adds the storpool driver to cinder?
16:13:03 leakypipes melwitt: well, holy crap.
16:13:08 mriedem johnthetubaguy: yeah, because i didn't know about it until now and we're <2 weeks from FF
16:13:09 mriedem :)
16:13:11 leakypipes melwitt: I thought that was to close the comment...
16:13:26 Roamer` mriedem, yeah, actually things happened 1. a bit suddenly at the start of December, and then 2. a bit slower than we expected with getting our CI system running, so our timing is a bit less than impeccable
16:13:52 Roamer` mriedem, https://review.openstack.org/#/c/220155/ is Cinder and https://review.openstack.org/#/c/192639/ is os-brick
16:13:52 mriedem Roamer`: you also need an os-brick release
16:14:24 leakypipes melwitt: heh, and the more you click the button, the more gerrit repeats the comment in the reply box. fun!
16:14:40 melwitt leakypipes: hah, awesome
16:15:17 Roamer` mriedem, and, yeah, my last comment to 140733 was "oof, yeah, Zuul cannot pass without an os-brick release"
16:15:38 mriedem the unit tests are also all failing
16:15:56 Roamer` yep, since our driver refers to initiator.STORPOOL
16:16:16 Roamer` I could make it pass the "storpool" literal string, then it might fail in a slightly different way somewhat further
16:16:26 Roamer` but I think that since os-brick has the constants, it'd be best to use them
16:17:13 sean-k-mooney leakypipes: did you not tell me once you used gertty mainly for reviews or has the new mac forced you to the web in protest.
16:17:31 openstackgerrit Merged openstack/nova master: conf: hyperv: fix a comment typo https://review.openstack.org/531329
16:17:57 Roamer` mriedem, to make it clear: yes, the unit tests are also failing because the released version of os-brick doesn't know about the StorPool driver... if I patch it manually within the tox venv, the tests pass
16:19:00 mriedem Roamer`: so you need an os-brick release and a bump to the required minimum version of os-brick in global-requirements in the openstack/requirements repo,
16:19:06 mriedem non-client library freeze is jan 18
16:19:13 leakypipes sean-k-mooney: I don't use the Mac for anything other than Verizon Slack and email. I use gertty as well but also use the Gerrit web UI just as much nowadays now that the whole "comment on this and the whole page goes haywire" bug was solved by mriedem.
16:19:13 mriedem nova feature freeze is jan 25
16:19:42 mriedem s/mriedem/smcginnis/
16:19:48 Roamer` mriedem, so you're saying that if I bother the os-brick folks enough, there might be a chance? Thanks!
16:20:18 mriedem Roamer`: i'm not terribly keen on approving a nova blueprint 1 week before feature freeze
16:20:32 sean-k-mooney leakypipes: wait that has been fixed?
16:20:32 mriedem since it's a libvirt volume driver, it might get an exception
16:20:41 mriedem Roamer`: but the ducks need to be in a row first
16:20:54 Roamer` mriedem, okay, thanks a lot! I'll be on my way to bother others now
16:21:37 Roamer` hmm, does this mean that, at least for the next couple of days, it might be a good idea for johnthetubaguy to remove his procedural -2, or shall we think about that only if we do get our ducks in a row first?
16:23:29 leakypipes cdent: answered :)
16:23:40 cdent thanks
16:24:11 openstackgerrit Matt Riedemann proposed openstack/nova master: Remove the deprecated scheduler_driver_task_period option https://review.openstack.org/531413
16:24:12 mriedem johnthetubaguy: see how that wording sounds ^
16:24:18 mriedem Roamer`: it's fine for johnthetubaguy to remove his -2
16:24:25 mriedem we can replace it with mine :)
16:24:55 openstackgerrit Merged openstack/nova master: Remove a duplicate line in a unit test https://review.openstack.org/531317
16:24:57 Roamer` :)
16:25:26 leakypipes sean-k-mooney: yeah. go into your profile -> settings -> diff preferences and set Render to "Slow"
16:25:45 leakypipes mriedem: is that right? ^
16:26:03 mriedem yup
16:28:12 leakypipes cdent: great questions, btw
16:28:48 cdent thanks, I'll try to make some more, by monday's meeting we can have a grand chat
16:29:26 sean-k-mooney leakypipes: mriedem awsome that bug was painful.
16:29:37 leakypipes sean-k-mooney: yeah, it sucks.
16:30:16 leakypipes it's kinda funny. the Gerrit UI seems to deliberately try and emulate late 1990s style.
16:30:53 mriedem what is 2018 style? animal avitars like the iphone x?
16:31:00 leakypipes mriedem: you know it.
16:31:05 mriedem i'll gladly fork out $1000 for that piece of shit
16:31:09 leakypipes not.
16:35:59 openstackgerrit melanie witt proposed openstack/nova stable/ocata: libvirt: Re-initialise volumes, encryptors, and vifs on hard reboot https://review.openstack.org/531422
16:41:04 cdent leakypipes: made a response, in which besides opening up some more worm cans, I note that the good questions still stand
16:45:08 mriedem melwitt: i think we should split the index add out of this https://review.openstack.org/#/c/334614/ and do that in a separate change
16:45:31 melwitt rats. okay
16:47:26 sean-k-mooney leakypipes: well gerrit is written in java useing gwt so its not exactly html5 native. its better then github pull requests though for code review so im not complaining
16:47:27 leakypipes cdent: re: allocations for vCenter-owned (not OpenStack-owned) resources... the allocation records in placement don't care who creates them, of course. so yes, if there were some external agent that claimed resources for some non-OpenStack consumers, that would be fine I suppose. As long as that agent uses the POST /allocations interface to ensure read and write consistency, that doesn't bother me at all.
16:47:59 leakypipes cdent: I just want to make sure that for *OpenStack consumers*, it's the scheduler (or conductor for move operations) that is the thing that claims resources for the system.
16:48:34 leakypipes cdent: for all the reasons that we moved to scheduler doing claims in Pike...
16:49:14 sean-k-mooney leakypipes: speaking of code review. regarding your nrp api patch that merged in early december, is the best way to play around with that just use curl/postman and do it manually or have there been any osc patches created?
16:49:56 leakypipes sean-k-mooney: :) that's actually the patch series we're currently discussing :)
16:50:03 melwitt mriedem: I changed the name of the ConsoleConnection object too. let me know if that wasn't a good idea
16:50:21 leakypipes sean-k-mooney: https://review.openstack.org/#/q/topic:bp/nested-resource-providers+(status:open)
16:50:28 mriedem melwitt: i like it, going through that diff now
16:50:43 leakypipes sean-k-mooney: I'm currently working on integrating n-r-p with the allocation candidates stuff. fried_rice is focusing on the integration with the virt driver piece.
16:51:04 leakypipes sean-k-mooney: so besides playing around with the hierarchy stuff in providers, there's not much you can yet do.
16:52:37 sean-k-mooney leakypipes: yep i was brefily talking to fried_rice about a poc im hoping to do in the next 6-8 to use it to support vhost-vfio interfaces.
16:53:01 leakypipes ++
16:54:06 sean-k-mooney *6-8 weeks... units are important
16:54:45 mriedem melwitt: which service should i see these log statements? cauth? https://review.openstack.org/#/c/320063/25/nova/objects/console_auth_token.py@116
16:54:56 mriedem or will this not actually run yet in ci in this change?
16:56:27 melwitt mriedem: yeah I think you won't see it until the last change in the series, in the novnc service log
16:57:07 mriedem ah yes
16:57:08 mriedem http://logs.openstack.org/12/527812/1/check/tempest-full/26f5e5f/controller/logs/screen-n-novnc-cell1.txt
16:57:15 mriedem ess_path=None,port=5900,token='***',updated_at=None) {{(pid=6917) validate /opt/stack/nova/nova/objects/console_auth_token.py:159}}
16:57:15 mriedem Jan 05 01:44:25.338490 ubuntu-xenial-citycloud-sto2-0001695415 nova-novncproxy[10280]: DEBUG nova.objects.console_auth_token [None req-166854c4-3f01-4db1-890b-212b7ba78e95 None None] Validated token - console connection is ConsoleAuthToken(access_url_base='http://10.0.1.151:6080/vnc_auto.html',console_type='novnc',created_at=2018-01-05T01:44:25Z,host='127.0.0.1',id=1,instance_uuid=4453afe4-e6eb-4479-a073-0f99a76c71f1,internal
16:57:26 mriedem NOICE
16:57:33 melwitt :D

Earlier   Later