Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2018-10-18
21:13:07 cdent So I started a check experimental on it, but then realized the tempest-pg-full won't run the functional tests that exercise the problem. Is there anything else that will?
21:13:26 mriedem nope
21:13:44 mriedem the only mysql/pg-specific stuff that runs in func/unit tests are the schema migration tests
21:14:18 mriedem the alternative is using DISTINCT in there as efried pointed out
21:15:07 efried I tried ^ locally and it worked, but no idea what that means.
21:15:18 efried I've been wanting to ask jay what the difference is.
21:15:39 efried He +2d the change, so he must've thought GROUP BY was fine, but...
21:16:13 cdent it's perfectly fine as is for mysql, but _might_ be a problem for pg
21:17:35 cdent I can poke at it (/me looks at queue ... ) Monday if nobody else wants/needs to. mriedem do you think this will tickle the problem?
21:18:02 cdent as in: do we need to care about this?
21:18:47 mriedem what will tickle the problem?
21:19:11 mriedem i don't know enough about group by rules for pg
21:19:41 mriedem you could probably recreate it with just placement and a pg db,
21:20:05 mriedem by having a consumer with >1 allocation against a resource provider and there not being any consumers table record for the allocations
21:20:09 mriedem so just hack it up manually
21:22:23 cdent What I meant was: is this the type of group by that postgresql will wail at? It will be easy enough to mess with manually. But for me, I won't have time before Monday
21:23:20 mriedem i don't know the answer to that
21:23:38 mriedem zzzeek_ might know off the top of his head
21:26:33 cdent i've left a note on the review, if nothing happens before next week, I'll dig
22:02:32 openstackgerrit Matt Riedemann proposed openstack/nova master: Remove the CachingScheduler https://review.openstack.org/611723
22:02:57 mriedem johnthetubaguy: mgagne: ^
22:04:20 mgagne +1 for me
22:04:42 mgagne didn't review the technical side of your removal, just the idea
22:07:05 mriedem email sent to the ops list as well
22:30:38 openstackgerrit Merged openstack/os-vif master: Fix random test_unplug_ovs failures https://review.openstack.org/611017
22:30:39 openstackgerrit Merged openstack/os-vif master: Do not call linux_net.delete_net_dev on Windows https://review.openstack.org/610916
22:47:37 openstackgerrit Sundar Nadathur proposed openstack/nova-specs master: Nova Cyborg interaction specification. https://review.openstack.org/603955
23:00:36 openstackgerrit Dan Smith proposed openstack/nova master: Return a minimal construct for nova service-list when a cell is down https://review.openstack.org/584829
23:32:15 openstackgerrit Merged openstack/nova stable/queens: Handle volume API failure in _post_live_migration https://review.openstack.org/611084
23:46:08 openstackgerrit Merged openstack/nova master: Fix typo in libvirt.hw_machine_type help https://review.openstack.org/611422
23:46:15 openstackgerrit Merged openstack/nova master: Fix block_device_mapping_v2 mention in server create API reference https://review.openstack.org/611433
#openstack-nova - 2018-10-19
00:58:59 openstackgerrit Brin Zhang proposed openstack/nova master: Add restrictions on updated_at when getting migrations https://review.openstack.org/607798
01:05:35 openstackgerrit Brin Zhang proposed openstack/nova master: Add restrictions on updated_at when getting instance action records https://review.openstack.org/607801
02:04:14 openstackgerrit Merged openstack/python-novaclient master: Recommend against using --force for evacuate/live migration https://review.openstack.org/611436
03:04:08 openstackgerrit Merged openstack/nova stable/rocky: Ignore uuid if already set in ComputeNode.update_from_virt_driver https://review.openstack.org/611337
03:34:01 openstackgerrit Merged openstack/nova stable/rocky: Use unique consumer_id when doing online data migration https://review.openstack.org/611315
05:35:06 openstackgerrit Takashi NATSUME proposed openstack/nova master: Fix best_match() deprecation warning https://review.openstack.org/611204
05:46:11 openstackgerrit Brin Zhang proposed openstack/nova master: Add restrictions on updated_at when getting migrations https://review.openstack.org/607798
05:48:31 openstackgerrit Brin Zhang proposed openstack/nova master: Add restrictions on updated_at when getting instance action records https://review.openstack.org/607801
06:36:43 jaosorior Could I get a review for this https://review.openstack.org/#/c/609591/ ?
07:09:25 bauzas Good morning Nova
07:17:59 openstackgerrit Zhenyu Zheng proposed openstack/nova-specs master: Detach and attach boot volumes - Stein https://review.openstack.org/600628
08:19:47 openstackgerrit Merged openstack/nova master: Migrate nova v2.0 legacy job to zuulv3 https://review.openstack.org/610403
10:13:56 openstackgerrit Tetsuro Nakamura proposed openstack/nova-specs master: Spec: Support filtering by forbidden aggregate https://review.openstack.org/603352
10:30:56 openstackgerrit huanhongda proposed openstack/nova master: AZ operations: check host has no instances https://review.openstack.org/611833
10:33:52 openstackgerrit huanhongda proposed openstack/nova master: AZ operations: check host has no instances https://review.openstack.org/611833
11:39:33 openstackgerrit Radoslav Gerganov proposed openstack/nova master: Preserve compute stats used by the scheduler https://review.openstack.org/611852
12:39:03 openstackgerrit Matt Riedemann proposed openstack/nova master: Document each libvirt.sysinfo_serial choice https://review.openstack.org/611426
12:43:37 mriedem so uh, do we want to do this stable-only ironic inventory workaround thing in rocky? https://review.openstack.org/#/c/609043/
12:43:50 mriedem and queens and pike
12:44:11 mriedem tl;dr once you've migrated all of your ironic instances to resource classes, you don't want to report vcpu/ram/disk inventory anymore on those nodes
12:48:58 SteelyDan I'd say so
12:53:49 mriedem i wasn't sure if the option should be deprecated immediately? seems kind of weird, but it will just be gone when you get to stein.
12:53:53 mriedem not sure how much it matters
12:56:42 SteelyDan me either
12:56:44 mriedem bauzas: were you witholding a +W on https://review.openstack.org/#/c/610088/ for some reason?
12:56:46 SteelyDan mriedem: does this ring any bells? http://logs.openstack.org/58/591658/12/check/tempest-full-py3/5224550/controller/logs/screen-n-api.txt.gz?level=TRACE#_Oct_18_20_49_20_086138
12:56:47 mriedem even though you were +2?
12:57:10 SteelyDan there's a ton of unrelated red in that log, so ignore the rest, but the FK error there doesn't seem related to the patch
12:57:20 SteelyDan and there are also rabbit connection failures later in the log
12:58:14 mriedem hmm, no, also seems weird that we'd get cell0 FK errors for a resize operation....
12:58:24 mriedem which shouldn't have anything to do with cell0
12:58:48 SteelyDan it's an action,
12:58:50 mriedem unless it's trying to record an action in the api,
12:58:54 SteelyDan but yeah that clearly looks like we're talking to the wrong db
12:58:56 mriedem and defaulting to the cell0 db connectoin in nova.conf,
12:59:06 mriedem but when we look up the instance we should target the context to cell1
12:59:21 SteelyDan oh I bet I know
12:59:26 mriedem so, it looks like it's probably shitting b/c it's trying to create an action in cell0 for an instance in cell1
12:59:39 SteelyDan this changes it so we're not permanently targeting the context.. so the resize goes on to be untargeted
13:00:22 SteelyDan didn't think of that until you made the connection to cell0
13:00:34 SteelyDan that's also probably why we can't connect to fake:/// or whatever rabbit
13:00:44 mriedem uh huh
13:02:51 openstackgerrit Dan Smith proposed openstack/nova master: Return a minimal construct for nova show when a cell is down https://review.openstack.org/591658
13:02:52 openstackgerrit Dan Smith proposed openstack/nova master: Return a minimal construct for nova service-list when a cell is down https://review.openstack.org/584829
13:16:23 bauzas mriedem: just a miss I guess
13:16:24 bauzas fixed
13:18:06 mriedem danke
13:18:19 bauzas bitte
13:30:37 openstackgerrit Daniel Abad proposed openstack/nova master: Fix ironic client ironic_url deprecation warning https://review.openstack.org/611872
14:01:00 melwitt
14:18:29 openstackgerrit Stephen Finucane proposed openstack/osc-placement master: Enforce key-value'ness for 'allocation candidate list --resource' https://review.openstack.org/611883
14:18:30 openstackgerrit Stephen Finucane proposed openstack/osc-placement master: tox: Hide deprecation warnings from stdlib https://review.openstack.org/611884
14:49:43 spatel I got this error when i reboot one of my instance any idea what is this? http://paste.openstack.org/show/732501/
15:27:25 cfriesen spatel: looks like corrupt filesystem
15:28:07 cfriesen or at least corrupt something on disk
15:31:48 mriedem_afk i've triaged https://bugs.launchpad.net/nova/+bug/1798805 but not sure if we would ever do anything about it
15:31:48 openstack Launchpad bug 1798805 in OpenStack Compute (nova) "Nova scheduler schedules VMs on nodes where nova-compute is down" [Wishlist,Triaged]
15:33:07 finucannot bauzas: If I delete a compute nodes RP, it'll be recreated, right?
15:34:04 SteelyDan mriedem_afk: I think that means disable in the "break its needs" sort of sense
15:34:31 SteelyDan I'm strongly in favor of keeping disable purely for scheduler reasons, but I think it's saying if the compute is actually down, we shouldn't take action on instances
15:34:35 SteelyDan which is probably reasonable
15:35:01 SteelyDan or make our super awesome bug-free state sync periodic power the instance on when the compute comes back
15:36:03 imacdonn I kinda sorta wish there was a "don't schedule anything new" status, which is different from "don't try to interact with me at all"
15:36:26 SteelyDan imacdonn: that's what disable means
15:36:29 SteelyDan it's the only thing it means
15:36:43 imacdonn the former, you mean
15:36:59 SteelyDan it means don't schedule anything there
15:37:04 imacdonn right
15:37:31 SteelyDan you're saying you wish there was a "kneecap this compute"

Earlier   Later