| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2018-03-05 | |||
| 15:43:42 | stephenfin | Yup, that all makes sense. It was that bit of info (what didn't work with neutron) that I was after | |
| 15:44:24 | mriedem | i expect i'll have to blacklist a bunch of tests, and/or make the job non-voting before we can merge that patch | |
| 15:46:32 | stephenfin | Makes sense | |
| 15:46:33 | stephenfin | I never caught up with mordred to talk about the zuulv3 native rework of those either :( Hopefully he's free some time next week once I've POCd some of the NUMA-aware vSwitch stuff | |
| 15:46:47 | jgriffith | mriedem: hey.. well that's up to you. I was going to try and finish it up this morning, but I'm certainly always open to help if it's something you want to work on. | |
| 15:46:58 | mriedem | stephenfin: i would put zuulv3 native rework of nova CI jobs way way down on the priority list | |
| 15:47:07 | jgriffith | mriedem: if you're just giving me a vote of "no confidence" well.. then.. uuummm.. | |
| 15:47:10 | jgriffith | ok | |
| 15:47:13 | jgriffith | :) | |
| 15:47:15 | mriedem | stephenfin: since 98% of nova people don't know how to maintain CI jobs, less zuulv3 native | |
| 15:47:32 | mriedem | jgriffith: heh, no, just didn't know how busy you were | |
| 15:47:43 | mriedem | jgriffith: figured this might be pro bono work for you | |
| 15:47:47 | jgriffith | mriedem: crazy busy, but I do plan to work it today | |
| 15:48:01 | jgriffith | mriedem: I could use your help on the policy piece though | |
| 15:48:10 | stephenfin | mriedem: I'd figured they should be easier to write and maintain. Also, it scratches a personal itch (I want to know how this stuff works). Much lower priority than the NUMA vSwitch stuff though, yes | |
| 15:48:21 | stephenfin | *should* :) | |
| 15:48:25 | mriedem | jgriffith: ok i rambled a lot in the comments, but i think you just need to use the volume api to get the volume details and then use the attachments to get the servers to count things | |
| 15:48:35 | mriedem | jgriffith: don't think we need to worry about any policy stuff | |
| 15:48:47 | mriedem | jgriffith: we should be able to figure this out w/o relying on config changes | |
| 15:48:53 | jgriffith | I saw this morning, need to go through your latest notes. | |
| 15:49:13 | mriedem | stephenfin: the legacy d-g job defs are very easy to write and maintain IMO | |
| 15:49:26 | mriedem | especially for people that have known how the CI framework has worked for the last 8 years | |
| 15:49:34 | jgriffith | mriedem: ok, let me have a look at what you added and go from there. If it ends up I can't figure out a solution I'm happy for you to hack on it or just get your help on enlightening me :) | |
| 15:49:48 | mriedem | jgriffith: ack | |
| 15:50:29 | stephenfin | mriedem: Good to note. I was starting from scratch there so it was all new | |
| 15:51:35 | jgriffith | mriedem: oh! You added the user-context addition! That's perfect | |
| 15:52:00 | mriedem | jgriffith: yup so we should be able to hit the volume API from the virt driver | |
| 15:52:01 | jgriffith | Tjat | |
| 15:52:12 | jgriffith | That's what I meant when I said policy changes by the way | |
| 15:52:17 | mriedem | ok i figured | |
| 15:52:24 | jgriffith | perfect.. I'll roll on it and get somethign up before lunch | |
| 15:52:25 | jgriffith | thanks! | |
| 15:52:29 | mriedem | yw | |
| 16:10:15 | openstackgerrit | Merged openstack/nova master: Updated from global requirements https://review.openstack.org/549694 | |
| 16:29:53 | stephenfin | yes | |
| 16:29:59 | stephenfin | whoops | |
| 16:42:21 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: Fix the notification devref location in exception https://review.openstack.org/549816 | |
| 16:46:43 | hrw | hm. patches... | |
| 16:47:19 | hrw | https://review.openstack.org/#/c/545034/ awaits for someone brave to give +w+2 to my pcie ports patch | |
| 16:56:21 | gibi | mriedem, Kevin_Zheng: fyi, I proposed the bp to add request_id to the instance action notification: https://blueprints.launchpad.net/nova/+spec/add-request-id-to-instance-action-notifications | |
| 16:57:25 | mriedem | gibi: cool, thanks - do you plan on getting 'formal' approval in the nova team meeting this week? | |
| 16:57:32 | mriedem | mel is on vacation this week btw | |
| 16:57:49 | gibi | mriedem: sure, I will bring it up on the meeting | |
| 16:57:49 | mriedem | although we already discussed it in person, | |
| 16:57:54 | mriedem | so likely not really needed | |
| 16:58:12 | gibi | it doesn't hurt | |
| 16:59:02 | mriedem | i've approved it | |
| 16:59:43 | gibi | mriedem: thanks | |
| 17:00:08 | gibi | mriedem: if you are interested in a bit heavier read then I also reworked the minimum bandwidth support spec during the weekend based on the PTG discussion https://review.openstack.org/#/c/502306/ | |
| 17:00:35 | mriedem | omg | |
| 17:00:41 | mriedem | i don't think i can handle that kind of weight right now | |
| 17:01:37 | gibi | mriedem: no problem. I agree that this is not good read for the first day after PTG | |
| 17:02:41 | openstackgerrit | Matt Riedemann proposed openstack/nova master: DNM: Test nova-api eventlet deprecation https://review.openstack.org/549770 | |
| 17:03:01 | mriedem | gibi: thanks for taking that on though, i'm glad you're owning it | |
| 17:03:31 | gibi | mriedem: yeah, my employer is pushing me to take it seriously | |
| 17:04:10 | mriedem | my employer also cares about it, but i haven't been physically kicked yet | |
| 17:04:36 | mriedem | it will be easier to kick Kevin_Zheng | |
| 17:04:40 | gibi | mriedem: mlavalle seems interested too and If I know correctly you two have the same employer | |
| 17:04:45 | mriedem | yup | |
| 17:05:44 | mlavalle | gibi: yes, I am interested. Will push it from the Neutron side | |
| 17:05:47 | stephenfin | mriedem: What needs to happen to move this forward now? https://review.openstack.org/#/c/483994 | |
| 17:06:55 | gibi | mlavalle: thanks! I think when you read the nova spec you will find couple of thing that needs to be discussed on the neutron side | |
| 17:06:59 | mriedem | stephenfin: um | |
| 17:07:16 | stephenfin | I'm currently deploying DevStack to see if I can reproduce that issue from Tushar Patil. However, the deprecation doesn't really depend on that | |
| 17:07:24 | mriedem | stephenfin: would be nice to know if tpatil can confirm novnc 1.0.0 works for him | |
| 17:07:29 | mlavalle | gibi: ack. will look for them | |
| 17:08:05 | gibi | mlavalle: I have Bence Romsics (rubasov) who can help with the neutron discussion / implementation work | |
| 17:08:21 | mriedem | stephenfin: ovh was saying they have the novnc fix deployed in their cloud though and it works | |
| 17:09:00 | stephenfin | Right. I can't verify as I can't read Japanese :) | |
| 17:09:03 | mlavalle | gibi: fantastic! | |
| 17:09:07 | mriedem | stephenfin: so i think the thing i was looking for in the deprecation was we need wording in the config option help text that you need novnc >= 1.0.0 | |
| 17:09:15 | mriedem | stephenfin: maybe takashin can help | |
| 17:09:21 | mriedem | he was also pushing a related spec for NTT related to this | |
| 17:09:28 | stephenfin | That's not a bad idea | |
| 17:09:34 | mriedem | it's a *great* idea | |
| 17:09:46 | gibi | mlavalle: I have to run now, talk to you later | |
| 17:26:08 | Spazmotic | whew, third shifters shouldn't be up t his late | |
| 17:59:30 | jgriffith | mriedem: hey.. there's a little bit of a snag in the plan | |
| 17:59:52 | jgriffith | mriedem: it seems the attached_host field wasn't being translated into the viewbuilder from cinder :( | |
| 18:00:15 | jgriffith | Not sure if it was dropped somewhere or if we just missed it altogether | |
| 18:00:29 | jgriffith | mriedem: easy enough to add, but then we get into the whole versioning debate | |
| 18:00:43 | mriedem | jgriffith: we don't want to rely on the attached_host field, | |
| 18:00:53 | jgriffith | mriedem: oh | |
| 18:01:01 | jgriffith | well then I suppose that's good news :) | |
| 18:01:02 | mriedem | because we discussed at the ptg that we need to control that via policy rule, and default that to admin-only | |
| 18:01:37 | jgriffith | mriedem: ok... so on the disconnect, I have a list of attachment-ids | |
| 18:01:50 | jgriffith | and of course can get the list of attachments from said id's | |
| 18:02:06 | mriedem | jgriffith: so i think the flow is: get volume details, from the volume['attachments'], get the list of server uuids attach to that volume, get the list of instances on that host (InstanceList.get_uuids_by_host - something like that), and then cross-reference those 2 lists | |
| 18:02:11 | jgriffith | so are you suggesting then iterate through bdms? | |
| 18:02:19 | mriedem | no i don't think we need bdms | |
| 18:02:33 | jgriffith | mriedem: ahh | |
| 18:02:41 | mriedem | i think we just need to see if there are >1 servers attached to the same volume, and if >1 of those servers are on *this* host | |
| 18:02:42 | jgriffith | I think I see what you're getting at | |
| 18:02:46 | mriedem | if so, don't call disconnect | |
| 18:03:16 | jgriffith | Ok, I think that will work; let me try and put it all together, thanks! | |
| 19:03:50 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Make nova-cells-v1 run with neutron https://review.openstack.org/549789 | |
| 19:43:59 | openstackgerrit | Eric Berglund proposed openstack/nova master: WIP: PowerVM Driver: Localdisk https://review.openstack.org/549300 | |
| 19:58:31 | openstackgerrit | Chris Dent proposed openstack/nova master: Move placement exceptions into the placement package https://review.openstack.org/549862 | |
| 20:06:42 | openstackgerrit | Eric Berglund proposed openstack/nova master: WIP: PowerVM Driver: Localdisk https://review.openstack.org/549300 | |
| 20:19:57 | openstackgerrit | John Griffith proposed openstack/nova master: Check for multiattach before removing connections https://review.openstack.org/548427 | |