Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2019-09-05
13:50:54 efried oh, the "unlikely" one perhaps
13:51:16 openstackgerrit Balazs Gibizer proposed openstack/nova master: Allow migrating server with port resource request https://review.opendev.org/671497
13:51:17 mriedem bauzas: it's not going to make train
13:51:18 bauzas efried: well, nevermind my question then, you know what, we'll see what's left
13:51:54 cdent efried: i put some unit tests on that host manager change of jays, above. it was as painful as I feared it would be, but I think it got it. at least cover says so
13:51:58 dansmith mriedem: so you think you can take a trip through the numa lm set this morning?
13:52:08 mriedem dansmith: it's on the todo list
13:52:16 dansmith mriedem: thanks
13:52:50 efried dansmith: can you run through the vpmem obj patches again please?
13:53:10 mriedem so,
13:53:22 mriedem efried: dansmith: besides numa lm,
13:53:36 dansmith efried: I just dropped a -1.9 on one
13:53:38 mriedem it seems the two other series that dansmith was reviewing was the forbidden aggregates stuff and sort of the vpmem stuff, right?
13:53:54 mriedem wouldn't it be a better use of dan's time to focus on forbidden aggregates if he's going to review one of those?
13:54:05 dansmith um..
13:54:11 openstackgerrit Balazs Gibizer proposed openstack/nova master: Do not query allocations twice in finish_revert_resize https://review.opendev.org/678827
13:54:13 mriedem not that i'm picking what dan should work on in the next week,
13:54:19 dansmith apparently you are :)
13:54:22 mriedem but forbidden aggregates isn't merged yet right?
13:54:34 mriedem it was all +2ed at some point
13:55:46 dansmith I haven't looked at that set in a while, but I can circle back
13:55:48 efried ugh, forbidden aggregates slid down my list and out of sight at some point.
13:55:51 mriedem my point is if you're going to ask dan to review something he's reviewed before, maybe make it the forbidden aggs one
13:55:52 efried yeah, same
13:56:27 dansmith mriedem: well, I had gone back and forth on the vpmems bottom patches already recently
13:56:28 mriedem the vpmems series seeems way too risky to me at this point
13:56:41 efried mriedem: for me, vpmem is higher priority, and also more reliant on dansmith's expertise. If we needed to finish up forbidden aggs without him, it would be possible. vpmems (clearly) not so much.
13:56:42 mriedem sure, but do we expect to land that whole series in the next week?
13:56:44 mriedem w/o regression?
13:57:04 dansmith efried: the vpmem series has a -1 on the fourth patch and is really large above that, is that really going to land?
13:57:09 openstackgerrit Balazs Gibizer proposed openstack/nova master: Allow resizing server with port resource request https://review.opendev.org/679019
13:57:13 dansmith efried: because if not, we shouldn't merge the db and object stuff, IMHO
13:57:17 efried dansmith: it has been pretty close for a while
13:57:34 dansmith mriedem: yeah, I dunno about that, I was just looking to see what was above the two they were asking me to look at and ... it looks like a lot
13:57:37 efried mriedem: there should be no regression on that, it's code that isn't hit unless you activate it. I don't think it's risky from that perspective.
13:57:41 mriedem if dan hasn't been in the 2nd half of the series i wouldn't call it close
13:57:57 efried mriedem: we don't need Dan for the second half. We have three cores on it already
13:58:07 efried specialized review for ovo stuff
13:58:11 efried is where we need dan.
13:58:33 dansmith just skimming the top,
13:58:41 dansmith it seems like there's plenty of room for regression in this stuff
13:58:51 dansmith despite being turned on only if you use it, it touches a lot of places
13:59:33 efried well, it's not like this is brand new, it's been through many rounds of review
13:59:45 dansmith like, changes the order and timing of talking to placement from the compute manager
13:59:51 dansmith that's not gated on this feature, that's fun for everyone
14:01:03 mriedem anything that deals with the RT flow is non-trivial
14:01:10 mriedem and prone to regression
14:01:16 mriedem i'm still fixing RT regressions since ocata
14:01:37 efried meeting now
14:02:06 efried If you like, we can discuss there why we're trying to freeze a feature a week before feature freeze
14:03:06 dansmith narrowing focus to get some things through the door instead of spreading focus among everything and getting nothing through the door?
14:06:44 mriedem gibi: comments in https://review.opendev.org/#/c/656422/21
14:06:50 mriedem let me know if we should just defer to a follow up
14:09:29 gibi mriedem: ack. I think I will defer to follow up there to ease the rebase pain
14:11:32 mriedem gibi: ok +2
14:11:38 gibi mriedem: thx
14:22:09 openstackgerrit Brin Zhang proposed openstack/nova master: Add user_id and project_id colume to Migration https://review.opendev.org/673990
14:22:09 openstackgerrit Brin Zhang proposed openstack/nova master: Add operator user_id/project_id to the migrations https://review.opendev.org/679413
14:22:10 openstackgerrit Brin Zhang proposed openstack/nova master: Filter migrations by user_id/project_id https://review.opendev.org/674243
14:28:31 brinzhang mriedem: https://review.opendev.org/#/c/674243/21 has bean resolved the merge conflict.
14:51:55 sean-k-mooney donnyd: it looks like its not completing and going to node_failure fairly consitently.
14:52:10 donnyd hrm
14:52:23 donnyd I will try to give that flavor a spin and make sure it actually works
14:57:38 donnyd Well I didn't have any issues getting it to spin up, so I guess we dig a little deeper
14:57:54 donnyd sean-k-mooney: do you have a log file handy?
14:58:52 sean-k-mooney donnyd: node_failure means that nodepool was not able to provision a node for the zuul job
14:59:04 sean-k-mooney so it failed before starting the job
14:59:22 sean-k-mooney there would be a log entry in nodepool but not in what is uploaded
14:59:43 donnyd http://grafana.openstack.org/d/3Bwpi5SZk/nodepool-fortnebula?orgId=1
14:59:57 donnyd I only see 3 failures from the last 6 hours
15:00:47 sean-k-mooney if you search for 679656 in https://zuul.opendev.org/t/openstack/status you wil see the latest one
15:01:16 sean-k-mooney this can happen if zuul is jsut at capastity for the provider
15:01:27 sean-k-mooney and it timed out waiting
15:01:47 sean-k-mooney i was seeing some node faiures before you update the lables too
15:02:08 sean-k-mooney so maybe its just worse due to the shared pool
15:02:58 sean-k-mooney it was about 1 in 2 or 1 in 3 that failed previously
15:03:03 donnyd sean-k-mooney: I am not sure where I am supposed to be looking
15:03:22 donnyd that is a pretty high failure rate no matter what
15:03:27 donnyd shouldn't be failing that often
15:04:06 donnyd maybe someone in infra can give us some insights. I just provisioned 10 without a single failure
15:04:07 sean-k-mooney well as i said it was a node failutre meaning noodpool could not provide a node to zuul to run the job on. that can happen due to hitting quota
15:04:20 efried kashyap: can we talk about https://blueprints.launchpad.net/nova/+spec/allow-secure-boot-for-qemu-kvm-guests ?
15:04:32 kashyap efried: Saddled with calls; but do ask away
15:04:41 kashyap efried: Actively working on it...
15:04:55 sean-k-mooney donnyd: FN is proably fine this could be on infras side.
15:04:57 donnyd sean-k-mooney: let me up the quota and see if that is helpful
15:04:58 efried kashyap: do you hope to land it in the next week or should I defer to ussuri?
15:05:27 sean-k-mooney i need to jump on a call but for now ill wait and check the jobs later when the load drops a bit
15:05:32 donnyd ok
15:05:54 openstackgerrit Kobi Samoray proposed openstack/nova master: Avoid fetching metadata when no subnets found https://review.opendev.org/679247
15:06:16 sean-k-mooney we intend to run this job nightly at 6am with ocational manual triggering so it should not generally be a proablem
15:10:16 donnyd sean-k-mooney: infra said they were hitting quota issues, so i just oversubscribed it for instances by 30% and pulled the rest of the quota entirely
15:10:58 donnyd openstack.exceptions.HttpException: HttpException: 403: Client Error for url: https://openstack.fortnebula.com:13774/v2.1/e8fd161dc34c421a979a9e6421f823e9/servers, Quota exceeded for instances: Requested 1, but already used 70 of 70 instances
15:11:11 donnyd sean-k-mooney: that issue should be solved
15:11:37 openstackgerrit Balazs Gibizer proposed openstack/nova master: migrate: Add bw min service level check of source compute https://review.opendev.org/680395
15:11:38 openstackgerrit Balazs Gibizer proposed openstack/nova master: resize: Add bw min service level check of source compute https://review.opendev.org/680396
15:13:25 openstackgerrit Stephen Finucane proposed openstack/nova master: Validate CPU config options against running instances https://review.opendev.org/680107
15:13:25 openstackgerrit Stephen Finucane proposed openstack/nova master: trivial: Use sane indent https://review.opendev.org/680229
15:13:26 openstackgerrit Stephen Finucane proposed openstack/nova master: objects: Add 'NUMACell.pcpuset' field https://review.opendev.org/680108
15:13:26 openstackgerrit Stephen Finucane proposed openstack/nova master: hardware: Differentiate between shared and dedicated CPUs https://review.opendev.org/671800

Earlier   Later