| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-03-06 | |||
| 18:23:26 | mriedem | and i doubt anyone there will be new enough to have cells v2 | |
| 18:23:47 | johnthetubaguy | we have previously had Nova folks at most of the ops summits, but a little bit by chance, I did the manchester one, for example | |
| 18:24:01 | mriedem | and sdague was in NY | |
| 18:24:06 | johnthetubaguy | mriedem: yeah I assumed that too, not sure why | |
| 18:24:08 | mriedem | and melwitt i think | |
| 18:24:20 | mriedem | well this is kind of last minute | |
| 18:24:25 | johnthetubaguy | yeah | |
| 18:24:40 | mriedem | cells v2 is going to be presented at the summit, | |
| 18:24:55 | mriedem | dansmith has a talk for it specifically, and i will be bringing it up in my talk about the new and notable in ocata | |
| 18:24:57 | johnthetubaguy | yeah, I think thats the right way to get that info | |
| 18:25:14 | johnthetubaguy | I duno what they want from the cells v2 session | |
| 18:25:21 | mriedem | i'm ignoring the fact that some operators are probably not going to be going to the summit b/c they traded for the ops meetup or PTG | |
| 18:25:42 | johnthetubaguy | the videos help some | |
| 18:26:46 | mriedem | sure, i just read an article from the register last night that linked to alaski's talk on cells v2 in austin | |
| 18:28:29 | mriedem | https://www.theregister.co.uk/2017/03/01/under_octata_deep_dive/ | |
| 18:28:31 | mriedem | oh octata | |
| 18:28:39 | mriedem | the release that was | |
| 18:29:07 | mriedem | also, side note, but i'm a bit worried about the amount of unicorn promises made in ^ about what you can do with placement in octata | |
| 18:29:13 | mriedem | jaypipes: you might find ^ interesting | |
| 18:29:46 | mriedem | they jumped the gun on qualitative traits-based scheduling by at least a release | |
| 18:30:02 | cdent | youch | |
| 18:30:47 | cdent | let's play startup: any press is better than no press | |
| 18:31:00 | mriedem | plus it's press for alaski | |
| 18:31:11 | mriedem | i also noticed "The Cola lifecycle management environment" | |
| 18:31:15 | mriedem | i wonder if that was meant to be kolla | |
| 18:31:39 | jaypipes | mriedem: that's what happens when you interview non-ATCs about technical stuff. | |
| 18:31:55 | cdent | "lifecycle management environment" | |
| 18:31:57 | cdent | that's a lovely phrase | |
| 18:32:17 | mriedem | you don't have to pick between slicing or dicing anymore | |
| 18:32:24 | cdent | \o/ | |
| 18:32:50 | alaski | oh man. I don't want people watching me speak | |
| 18:33:28 | cdent | hope you had a clean shirt | |
| 18:33:49 | mriedem | he looked great | |
| 18:34:36 | cdent | of course he did | |
| 18:37:55 | dims_ | go alaski ! | |
| 18:41:29 | openstackgerrit | Lee Yarwood proposed openstack/nova master: compute: Detach volumes on _rebuild_default_impl failure https://review.openstack.org/442105 | |
| 18:42:12 | mriedem | jgriffith: you should take a look at this https://review.openstack.org/#/c/304715/13/specs/pike/approved/scaleio-ephemeral-storage-backend.rst@73 | |
| 18:42:27 | openstackgerrit | Gábor Antal proposed openstack/nova master: Remove mox in unit/cells/test_cells_manager.py https://review.openstack.org/399036 | |
| 18:42:29 | jgriffith | mriedem looking | |
| 18:42:34 | mriedem | jgriffith: that's saying it's going to take 2-3 releases to add a cinder imagebackend to nova | |
| 18:42:34 | jgriffith | oh dear | |
| 18:42:43 | mriedem | and because of that, we shouldn't wait on adding the scaleio imagebackend | |
| 18:42:56 | sdague | johnthetubaguy: we've tried to get geographic coverage | |
| 18:43:27 | mriedem | jgriffith: also fyi we've had 2 blueprints for the cinder ephemeral backend in the past https://blueprints.launchpad.net/nova/+spec/nova-ephemeral-cinder https://blueprints.launchpad.net/nova/+spec/nova-ephemeral-storage-with-cinder | |
| 18:43:47 | sdague | however, it is the danger of doing an ops summit kind of far away from where dev teams are, we're just not going to get that coverage | |
| 18:44:10 | jgriffith | mriedem so what you're saying is I need to get my *&^% together and get a spec with some comittemnts to timelines :) | |
| 18:44:51 | mriedem | jgriffith: well, the assertions made in the scaleio spec are kind of scary and i assume you would disagree with at least part of them | |
| 18:45:15 | mriedem | jgriffith: so just bringing that to your attention, but yeah at some point sooner than later we'll want to start looking at a spec | |
| 18:45:17 | jgriffith | mriedem yeah, I'm going through it now | |
| 18:45:23 | jgriffith | mriedem thanks for raising it | |
| 18:45:26 | mriedem | https://review.openstack.org/#/c/91321/ was also an old spec, | |
| 18:45:30 | mriedem | but apparently lacking a lot of details | |
| 18:45:50 | jgriffith | mriedem best kind of spec is the spec with no details | |
| 18:45:51 | jgriffith | then you can't be held accountable :) | |
| 18:45:51 | mriedem | probably not even worth getting distracted by it | |
| 18:50:10 | mriedem | that reminds me, i need to hedge in my searchlight spec about the fact you won't be able to list deleted instances if nova is using searchlight | |
| 18:50:17 | mriedem | dansmith: ^ did you read my email on that? | |
| 18:50:43 | mriedem | i probably need that in the cells meeting agenda | |
| 18:51:11 | edmondsw | mriedem, tx for the ping. It's a good question. I will talk to some folks here to figure out how important log translations are. My current understanding is that yes, we need those | |
| 18:51:45 | mriedem | edmondsw: you'll have to pony up some translators probably :) | |
| 18:52:02 | edmondsw | mriedem I've actually been trying to make that case... | |
| 19:21:35 | openstackgerrit | Logan V proposed openstack/nova master: Add LVM thin pool support https://review.openstack.org/442126 | |
| 19:32:25 | mriedem | cfriesen: is that from your team? ^ | |
| 19:33:30 | dansmith | mriedem: I didn't, but I saw it and thought "I should read that" | |
| 19:33:48 | mriedem | dansmith: ok, i'm just about to push the updated spec mentioning the issue, along with some other issues | |
| 19:33:52 | dansmith | cool | |
| 19:37:20 | openstackgerrit | Matt Riedemann proposed openstack/nova-specs master: List instances using Searchlight (spec) https://review.openstack.org/441692 | |
| 19:37:28 | mriedem | cdent: gibi: ^ comments addressed, thanks for the review on that | |
| 19:40:27 | logan- | mriedem: (re: 442126) not from cfriesen's team, just me. hopefully I didn't miss any steps in the docs before pushing that. first patch in nova. | |
| 19:41:04 | mriedem | logan-: ok, cfriesen and the windriver team have something similar i'm told, and we talked about it at the ptg so figured it was his team pushing that up | |
| 19:41:11 | mriedem | since there is general interest | |
| 19:42:02 | mriedem | logan-: this is small enough that we probably don't need a spec, but you should probably mention the blueprint during the weekly nova meeting durin the 'open discussion' section | |
| 19:42:16 | mriedem | http://eavesdrop.openstack.org/#Nova_Team_Meeting | |
| 19:42:26 | logan- | sounds good. I wasn't sure if it was specless or not, so I figured i'd just push it and see. will make sure to be at the meeting this week. | |
| 19:42:46 | logan- | thanks | |
| 19:42:52 | mriedem | yw | |
| 19:49:37 | openstackgerrit | Sarafraj Singh proposed openstack/nova master: conf: remove console_driver opt https://review.openstack.org/441485 | |
| 20:01:09 | openstackgerrit | Pushkar Umaranikar proposed openstack/nova-specs master: Stop using Neutron admin token https://review.openstack.org/439890 | |
| 21:02:15 | openstackgerrit | Jay Pipes proposed openstack/nova master: virt: implement get_inventory() for Ironic https://review.openstack.org/441544 | |
| 21:02:15 | openstackgerrit | Jay Pipes proposed openstack/nova master: virt: add get_inventory() virt driver API method https://review.openstack.org/441543 | |
| 21:02:17 | jaypipes | mriedem: ^^ addressed your concerns | |
| 21:03:28 | mriedem | jgriffith: do you have time to talk with alexandrelevine about the scaleio ephemeral driver for nova? | |
| 21:04:13 | alexandrelevine | jgriffith: hi | |
| 21:04:33 | jgriffith | alexandrelevine hi | |
| 21:04:44 | jgriffith | mriedem alexandrelevine can you give me about a 1/2 hour? | |
| 21:04:52 | jgriffith | maybe less? | |
| 21:05:05 | alexandrelevine | jgriffith: Ok. | |
| 21:06:11 | jgriffith | alexandrelevine never mind, I'm available (you forget to include a dial in for a meeting invite you're out of luck) | |
| 21:06:24 | alexandrelevine | :) | |
| 21:06:40 | jgriffith | alexandrelevine :) | |
| 21:06:49 | jgriffith | soon they won't invite me at all :) | |
| 21:06:54 | alexandrelevine | So there is a problem with the ScaleIO ephemeral driver. And it's mostly historical and resource-wise. I'll explain. | |
| 21:07:00 | jgriffith | alexandrelevine sure | |
| 21:07:29 | alexandrelevine | We started doing it 1.5 years ago when we worked for EMC. And we were postponed because at the time a refactoring started in the same area - imagebackend.py | |
| 21:08:03 | alexandrelevine | We were told that it'll go fast and well but as often it didn't. So we were postponed once. And then twice because of communicational and organizational reasons. | |
| 21:08:10 | alexandrelevine | And now seems to be the third time. | |
| 21:08:29 | alexandrelevine | All this time we had to support patches for new versions of nova and ScaleIO aside from the main code. | |
| 21:08:49 | jgriffith | alexandrelevine so to be clear I have the luxury of being able to claim "not being involved" in any past efforts :) | |
| 21:08:51 | alexandrelevine | Now we've left EMC and we don't have enough resources to keep doing this cumbersome tasks. | |
| 21:09:05 | jgriffith | so while I feel for ya, I don't necessarily have anything to say about that | |