Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2018-03-13
14:04:55 edleafe cdent: ok, gimme a few
14:05:48 tblakes gibi: No this isn't about preemptible instances. I'm just wondering about the basic case where a VM deploy fails with NoValidHost. I haven't been able to find where that failure notification should get sent out.
14:09:24 gibi tblakes: searching the code...
14:14:15 edleafe cdent: ok, flase alarm. A missing quote mark was the culprit. The output made me think that gabbi didn't substitute the value
14:14:22 edleafe false alarm, even
14:14:41 gibi tblakes: I think this is the place https://github.com/openstack/nova/blob/107ad403d46c9d6961c5b15c68df0fa24014a832/nova/conductor/manager.py#L588
14:15:02 gibi tblakes: this emits the legacy compute_task.build_instances notification
14:15:14 gibi tblakes: example of the payload is here http://paste.openstack.org/show/688874/
14:15:46 cdent edleafe: if you see the original string in the error output it can only be two thing: the sub couldn't happen because it wasn't able to be parsed as a sub (your case?) or it was parseable but the info couldn't be retrieved (the environ is not set)
14:16:04 gibi tblakes: transformation of that legacy notification is still ongoing in https://review.openstack.org/#/c/482629/
14:16:12 cdent in the former case we can't error that the sub didn't happen because gabbi doesn't know it was supposed to be a sub
14:16:38 gibi tblakes: I don't know about any change around these code path in Queens
14:17:58 gibi tblakes: at least not for the notification
14:20:06 edleafe cdent: makes sense
14:20:22 edleafe cdent: now I'm down to debugging sqla failures. Joy!
14:21:41 tblakes gibi: Thanks for pointing me to the code in build_instances!
14:21:46 gibi stephenfin: about the crypto removal patches, is this a followup of the bp remove-nova-cert
14:22:05 gibi stephenfin: or how we ended up with so many dead code?
14:22:27 gibi tblakes: no problem, I'm happy to help
14:26:19 kashyap johnthetubaguy: Thanks for review here: https://review.openstack.org/#/c/534384/. Just reading through
14:27:39 kashyap johnthetubaguy: A quick note on the name: `cpu_model_extra_flags`. I used the term 'flags' there because, all these extra features are enumerated under 'flags' when you run `cat /proc/cpuinfo`
14:28:04 kashyap But anyway; that's relatively minor :-). I'll get to the rest of the core changes first
14:28:12 johnthetubaguy kashyap: ah, got you, I could go either way
14:28:16 johnthetubaguy cool
14:29:33 kashyap Thanks
14:32:05 openstackgerrit Eric Fried proposed openstack/nova-specs master: Allow for merging traits and aggregates https://review.openstack.org/552122
14:32:53 SamYaple i cant seem to find this information online, but assuming i have multiple cellv2 cells, can an instance migrate (or preferably live-migration) between the cells?
14:36:26 stephenfin gibi: I'm not really sure. I spotted it by chance and realized I could go delete it all
14:36:48 gibi stephenfin: meanwhile I checked and I think nova-cert was the last user
14:37:01 gibi stephenfin: anyhow seems safe to remove
14:37:13 stephenfin gibi: Good to hear :)
14:43:19 gibi stephenfin: I've sent all 3 patches to the gate
14:43:32 stephenfin gibi: Excellent. Thank you :)
14:43:48 gibi stephenfin: going back to reading your mypy example :)
14:49:56 openstackgerrit Nguyen Hai proposed openstack/nova-specs master: Enhance nova-specs repo and webpage https://review.openstack.org/551802
14:51:28 jaypipes stephenfin: will be 4 mins late to nmeetubg
14:51:31 jaypipes meeting
14:54:37 openstackgerrit Ed Leafe proposed openstack/nova master: Add 'member_of' param to GET /allocation_candidates https://review.openstack.org/552098
14:55:20 dansmith edleafe: oh that reminds me you pinged me yesterday during a meeting and I forgot
14:55:56 edleafe dansmith: if your sqla-fu is strong, I need to find out why the added join isn't working
14:56:02 edleafe jaypipes ^^
14:56:28 dansmith it's weak, but let me look over it
14:56:36 edleafe Speaking of meetings... off I go into that black hole..
14:56:43 dansmith edleafe: I don't think you need to make this patch depend on any of mine
14:56:49 dansmith it can go in straight away
14:58:38 edleafe dansmith: sure, I can switch that
14:58:54 tssurya dansmith : mriedem gave his blessings on https://review.openstack.org/#/c/546684/ ; whenever you have the time could you please circle back for your re-approval ?
14:59:18 dansmith tssurya: I saw, it's in my queue
14:59:38 tssurya dansmith: thank you
15:00:15 dansmith edleafe: yeah that join is over my head for sure
15:03:08 edleafe dansmith: I was trying to follow the "join chain" logic in the rest of the code, but sqla doesn't like the "in" clause, it seems
15:03:40 dansmith edleafe: and member_of is a list of strings yeah?
15:04:11 edleafe dansmith: yep
15:04:40 dansmith edleafe: we should try query.summon(guru=jaypipes).execute()
15:05:08 edleafe dansmith: that reads like we're going to execute Jay
15:05:21 dansmith heh
15:10:45 Spaz-Home Man didn't deny it
15:16:18 cdent edleafe: you getting an error or bad results on that join chain? have you debugged to look at the generated table? is the join between the rp_tbl and rp_agg_tbl happening somewhere else already or do you need to hork that in there somewhere too? looks like that may be missing
15:17:14 edleafe cdent: getting an error from sqla on the join clause itself.
15:17:49 cdent it looks like it is an orphan
15:19:27 edleafe cdent: lemme paste the error
15:19:38 cdent ✔
15:21:34 edleafe cd http://paste.openstack.org/show/699991/
15:22:22 edleafe cdent: ^^
15:25:29 cdent edleafe: the third argument to join is the onclause, and you've got a table
15:26:06 cdent so it trying to write a query but a table has no boolean characteristic
15:26:09 cdent http://docs.sqlalchemy.org/en/rel_1_1/core/selectable.html?highlight=sa%20join#sqlalchemy.sql.expression.join
15:28:30 edleafe cdent: ah, ok. I was trying to copy Jay's existing join chain logic
15:28:56 edleafe let me try breaking that up into two steps
15:29:44 edleafe ...after meetings are done, of course!
15:30:21 cdent edleafe: I think basically you need more than one join
15:31:48 edmondsw melwitt should I be doing anything to get eyes on the PowerVM virt driver spec for Rocky besides where I added to https://etherpad.openstack.org/p/rocky-nova-priorities-tracking?
15:34:37 mriedem sahid: let me know if you have any questions about my comments in https://review.openstack.org/#/c/485522/
15:40:48 jaypipes edleafe: on a call with stephenfin and gibi. gimme about 15 mins?
15:43:28 sahid mriedem: ack, will have a look and let you know
15:44:47 edleafe jaypipes: cdent may have pointed me in the right direction. Wait until my next revision
15:47:26 sahid mriedem: in that meantime if i can ask you to have a look at this: https://review.openstack.org/#/c/539605/ i think we are all agree
15:48:46 mriedem hmm
15:50:27 sahid does that "hum" means you are not agree... :)
15:53:19 jaypipes stephenfin, gibi: https://etherpad.openstack.org/p/nfDwesfaMs
16:07:46 openstackgerrit Merged openstack/python-novaclient master: Updated from global requirements https://review.openstack.org/552384
16:10:34 openstackgerrit Merged openstack/nova master: Transform live_migration.post.dest notifications https://review.openstack.org/469784
16:11:19 mriedem sahid: comments inline
16:13:53 mriedem sahid: i won't claim to know much about virtio-net multiqueue
16:14:05 mriedem so my comments and questions are mostly about the big picture and for understanding
16:15:33 dansmith mriedem: do you care to scrutinize the os-vif and networky bits of this? https://review.openstack.org/#/c/489451/30
16:16:13 mriedem it's not using os-vif is it?
16:16:26 mriedem also https://blueprints.launchpad.net/nova/+spec/live-migration-in-xapi-pool isn't approved
16:17:10 mriedem is this the thing to drop the aggregate up-call for xen live migration?
16:19:11 Spaz-Home ?
16:19:11 Spaz-Home Oh there's a blueprint for that
16:19:13 Spaz-Home that makes me happy if so
16:19:32 dansmith mriedem: yeah, we need to approve it specless I think, we discussed that at PTG
16:20:07 Spaz-Home I think i'd like to talk to Niachuan about that..
16:20:59 dansmith mriedem: what I meant about os-vif is using your os-vif awareness to review all the networky bits of that
16:21:40 mriedem dansmith: comments inline
16:21:46 mriedem that patch doesn't remove the aggregate up-call at all
16:21:58 mriedem also, i don't have os-vif awareness :)
16:21:59 dansmith I know, see my comment
16:22:02 mriedem except to say i'm aware it exists

Earlier   Later