| *** jdg has quit IRC | 00:02 | |
| *** jdg has joined #openstack-meeting-cp | 00:06 | |
| *** brault has joined #openstack-meeting-cp | 00:08 | |
| *** jdg has quit IRC | 00:11 | |
| *** jdg has joined #openstack-meeting-cp | 00:12 | |
| *** brault has quit IRC | 00:13 | |
| *** felipemonteiro__ has quit IRC | 00:38 | |
| *** aselius has quit IRC | 00:53 | |
| *** jdg has quit IRC | 00:59 | |
| *** diablo_rojo has quit IRC | 01:42 | |
| *** yamahata has quit IRC | 01:44 | |
| *** iyamahat has quit IRC | 01:44 | |
| *** diablo_rojo has joined #openstack-meeting-cp | 01:46 | |
| *** jhesketh has quit IRC | 01:51 | |
| *** jhesketh has joined #openstack-meeting-cp | 01:51 | |
| *** IgorYozhikov has quit IRC | 01:59 | |
| *** IgorYozhikov has joined #openstack-meeting-cp | 02:04 | |
| *** Rockyg has joined #openstack-meeting-cp | 02:18 | |
| *** diablo_rojo has quit IRC | 02:20 | |
| *** brault has joined #openstack-meeting-cp | 02:24 | |
| *** brault has quit IRC | 02:28 | |
| *** nhelgeson has quit IRC | 02:38 | |
| *** jgriffith_ has joined #openstack-meeting-cp | 02:42 | |
| *** jgriffith_ has quit IRC | 03:08 | |
| *** lbragstad has joined #openstack-meeting-cp | 03:12 | |
| *** markvoelker has quit IRC | 03:36 | |
| *** tonyb has quit IRC | 03:42 | |
| *** lbragstad has quit IRC | 03:50 | |
| *** tonyb has joined #openstack-meeting-cp | 03:56 | |
| *** jgriffith has quit IRC | 04:11 | |
| *** jgriffith has joined #openstack-meeting-cp | 04:11 | |
| *** coolsvap has joined #openstack-meeting-cp | 04:33 | |
| *** Rockyg has quit IRC | 04:35 | |
| *** gouthamr has quit IRC | 05:06 | |
| *** markvoelker has joined #openstack-meeting-cp | 05:37 | |
| *** Rockyg has joined #openstack-meeting-cp | 05:38 | |
| *** dhellmann has quit IRC | 05:56 | |
| *** dhellmann has joined #openstack-meeting-cp | 05:58 | |
| *** markvoelker has quit IRC | 06:11 | |
| *** iyamahat has joined #openstack-meeting-cp | 06:28 | |
| *** brault has joined #openstack-meeting-cp | 06:44 | |
| *** markvoelker has joined #openstack-meeting-cp | 07:08 | |
| *** iyamahat has quit IRC | 07:34 | |
| *** markvoelker has quit IRC | 07:42 | |
| *** Rockyg has quit IRC | 08:11 | |
| *** markvoelker has joined #openstack-meeting-cp | 08:38 | |
| *** markvoelker has quit IRC | 09:12 | |
| *** MarkBaker has joined #openstack-meeting-cp | 09:46 | |
| *** sdague has joined #openstack-meeting-cp | 10:00 | |
| *** iyamahat has joined #openstack-meeting-cp | 10:05 | |
| *** markvoelker has joined #openstack-meeting-cp | 10:09 | |
| *** iyamahat has quit IRC | 10:11 | |
| *** MarkBaker has quit IRC | 10:36 | |
| *** markvoelker has quit IRC | 10:42 | |
| *** markvoelker has joined #openstack-meeting-cp | 11:39 | |
| *** markvoelker has quit IRC | 12:12 | |
| *** edmondsw has joined #openstack-meeting-cp | 12:13 | |
| *** MarkBaker has joined #openstack-meeting-cp | 12:14 | |
| *** edmondsw_ has joined #openstack-meeting-cp | 12:16 | |
| *** edmondsw has quit IRC | 12:17 | |
| *** edmondsw_ has quit IRC | 12:22 | |
| *** markvoelker has joined #openstack-meeting-cp | 12:28 | |
| *** edmondsw has joined #openstack-meeting-cp | 12:29 | |
| *** edmondsw has quit IRC | 12:33 | |
| *** edmondsw has joined #openstack-meeting-cp | 12:36 | |
| *** lbragstad has joined #openstack-meeting-cp | 13:06 | |
| *** gouthamr has joined #openstack-meeting-cp | 13:25 | |
| *** coolsvap has quit IRC | 13:36 | |
| *** nikhil_k has joined #openstack-meeting-cp | 14:22 | |
| *** nikhil_k is now known as nikhil | 14:23 | |
| *** rarcea has joined #openstack-meeting-cp | 14:23 | |
| *** felipemonteiro_ has joined #openstack-meeting-cp | 14:37 | |
| *** felipemonteiro__ has joined #openstack-meeting-cp | 14:38 | |
| *** felipemonteiro_ has quit IRC | 14:42 | |
| *** gouthamr_ has joined #openstack-meeting-cp | 15:03 | |
| *** gouthamr has quit IRC | 15:04 | |
| *** gouthamr_ is now known as gouthamr | 15:04 | |
| *** iyamahat has joined #openstack-meeting-cp | 15:24 | |
| *** iyamahat has quit IRC | 15:31 | |
| *** diablo_rojo has joined #openstack-meeting-cp | 15:32 | |
| *** diablo_rojo has quit IRC | 15:39 | |
| *** diablo_rojo has joined #openstack-meeting-cp | 15:40 | |
| *** iyamahat has joined #openstack-meeting-cp | 15:42 | |
| *** mriedem has joined #openstack-meeting-cp | 15:59 | |
| ildikov | #startmeeting cinder-nova-api-changes | 15:59 |
|---|---|---|
| openstack | Meeting started Thu Oct 5 15:59:51 2017 UTC and is due to finish in 60 minutes. The chair is ildikov. Information about MeetBot at http://wiki.debian.org/MeetBot. | 15:59 |
| openstack | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 15:59 |
| *** openstack changes topic to " (Meeting topic: cinder-nova-api-changes)" | 15:59 | |
| openstack | The meeting name has been set to 'cinder_nova_api_changes' | 15:59 |
| mriedem | o/ | 15:59 |
| ildikov | johnthetubaguy jaypipes e0ne jgriffith hemna mriedem patrickeast smcginnis diablo_rojo xyang1 raj_singh lyarwood jungleboyj stvnoyes | 16:00 |
| * johnthetubaguy has a nasty clash this week, but tries to keep his eye on the channel | 16:00 | |
| smcginnis | Here but at a training event so not really here. | 16:00 |
| ildikov | johnthetubaguy: I knew you're still hiding :) | 16:00 |
| jungleboyj | @! | 16:01 |
| _pewp_ | jungleboyj |。･ω･|ﾉ | 16:01 |
| jgriffith | o/ | 16:01 |
| ildikov | ok, let's start | 16:01 |
| ildikov | I guess we can have a quick one today | 16:01 |
| ildikov | so the gate is getting more friendly, so I have my hopes up again | 16:01 |
| ildikov | mriedem: johnthetubaguy: shared_targets spec: https://review.openstack.org/#/c/507670/ | 16:02 |
| ildikov | mriedem: johnthetubaguy: and corresponding code changes: https://review.openstack.org/#/c/509005/ | 16:02 |
| * jungleboyj crosses my fingers. | 16:02 | |
| ildikov | mriedem: johnthetubaguy: the spec has no multi-attach aspects; would we need one on the Cinder side that does? | 16:03 |
| mriedem | i don't understand the question | 16:04 |
| jgriffith | ildikov IMO we'll need multi-attach spec for multi-attach | 16:04 |
| jgriffith | Not for shared_targets reporting | 16:04 |
| johnthetubaguy | the policy stuff | 16:04 |
| ildikov | mriedem: basically where to cover all the aspects of the shared_targets and if it's ok to just figure it out on the Nova side referencing the spec I linked in above? | 16:04 |
| mriedem | it's something to mention in the nova multiattach spec | 16:05 |
| ildikov | jgriffith: johnthetubaguy: policy on multi-attach on both sides? | 16:05 |
| ildikov | as if my understanding is correct we said we will leave R/O for a later time | 16:05 |
| jgriffith | I'm ignoring anything multi-attach right now so "carry on" :) | 16:05 |
| jgriffith | Unless that's the specific topi right now, which is unrelated to the spec or shared_targets patch that's up IMO | 16:06 |
| ildikov | jgriffith: we should do everything now that needs a version bump | 16:06 |
| ildikov | jgriffith: ideally at least | 16:06 |
| jgriffith | ildikov why? | 16:06 |
| jgriffith | we haven't even merged the new attach API's? | 16:06 |
| jgriffith | shouldn't we do that *first*? | 16:06 |
| ildikov | jgriffith: that's what we agreed on and that's mainly what keeps the new attach patch from merging right now | 16:06 |
| jgriffith | ok | 16:07 |
| jungleboyj | ildikov: ++ | 16:07 |
| ildikov | jgriffith: we didn't want to do multiple version bumps and checks if that can be avoided | 16:07 |
| mriedem | the policy stuff isn't a microversoin | 16:07 |
| mriedem | we said you can add policy rules to cinder today | 16:07 |
| ildikov | mriedem: that's not, but the shared_targets thing is | 16:08 |
| mriedem | basic policy rule is, can you even create multi-attach volumes | 16:08 |
| mriedem | the shared_targets cinder spec is approved and code is up though... | 16:08 |
| mriedem | so what else do we need to hold up the nova side for | 16:08 |
| mriedem | ? | 16:08 |
| * jgriffith really doesn't understand | 16:08 | |
| ildikov | ok, if everyone is good, I'm good :) | 16:08 |
| mriedem | pretty sure this was all discussed in detail ad nauseum last week | 16:09 |
| mriedem | i hope someone was taking notes :) | 16:09 |
| jgriffith | mriedem :). as a matter o'fact | 16:09 |
| ildikov | will update the spec on the Nova side with that bit and hopefully the code will get merged so I can update the new attach code too shortly | 16:09 |
| ildikov | jungleboyj: in case we get the shared_targets patch merged along with the version bump in the client, is there anything that would keep us back from cutting yet another client? | 16:10 |
| jungleboyj | ildikov: Not that I am aware of. Think I could make that happen. | 16:11 |
| jgriffith | ildikov jungleboyj technically you don't need a new client anyway | 16:11 |
| jgriffith | just sayin | 16:11 |
| jungleboyj | jgriffith: ? | 16:11 |
| ildikov | jgriffith: the new attach patch will not pass on the gate without a new client | 16:11 |
| ildikov | jgriffith: once I bump the version for the attach calls in it | 16:12 |
| *** diablo_rojo has quit IRC | 16:12 | |
| jgriffith | alrighty | 16:12 |
| ildikov | :) | 16:13 |
| jungleboyj | jgriffith: I try not to argue with ildikov :-) | 16:13 |
| *** MarkBaker has quit IRC | 16:13 | |
| ildikov | jungleboyj: jgriffith: that's what happened before we cut the latest client the last time | 16:13 |
| ildikov | jungleboyj: jgriffith: the client is not built from source on the gate, which is not necessarily a bad thing and there's a test in Nova that checks on the max version, which is going to fail till we don't have a new release out for a version bump | 16:14 |
| ildikov | so one thing I didn't fully get after re-reading the logs from last week | 16:15 |
| jungleboyj | ildikov: Ok, that is good to know and actually makes more sense to me. | 16:15 |
| ildikov | so I understand we don't want boot from a multiattach R/W volume | 16:15 |
| ildikov | that's ok | 16:15 |
| ildikov | what I didn't get is how and where we want to ensure that doesn't happen | 16:16 |
| ildikov | is it on the Cinder side not to be able to even create a volume that's bootable with 'multiattach'=True at the same time? | 16:16 |
| jgriffith | There are a couple of options, I'd probably enforce some things on the Cinder side using the bootable property of the volume | 16:16 |
| ildikov | or will we block the volume getting used for boot for the second time on one on the sides? | 16:17 |
| ildikov | jgriffith: when the volume is created or when it's used? | 16:17 |
| jungleboyj | ildikov: I think we want to go with that option as it is consistent with the default policy we were talking about. | 16:17 |
| johnthetubaguy | I would go for when the volume is created | 16:18 |
| *** Rockyg has joined #openstack-meeting-cp | 16:18 | |
| jgriffith | ? | 16:18 |
| ildikov | johnthetubaguy: that would be my thought as well if that's feasible | 16:19 |
| johnthetubaguy | I think we confirmed when bootable = false Nova rejects you being able to boot from a volume, so that should cover it | 16:19 |
| jungleboyj | So we would not allow multi-attach able volume to be made bootable? | 16:20 |
| jgriffith | OMG | 16:20 |
| jgriffith | look... a volume is marked bootable when it's created from image | 16:20 |
| jungleboyj | Sorry ... | 16:20 |
| jgriffith | the corner case of marking a volume as bootable after the fact is a corner case | 16:21 |
| jgriffith | if you have an attachment finalized on a volume and it's marked as bootable | 16:21 |
| jgriffith | Then if another request to connect is received we fail and respond with an appropriate error message | 16:21 |
| ildikov | isn't it marked as bootable when it's created? | 16:21 |
| smcginnis | Can be later. | 16:22 |
| jungleboyj | jgriffith: Ok. That makes sense to me. | 16:22 |
| jgriffith | Are my messages not showing up in here? | 16:22 |
| jungleboyj | I didn't think that was the same thing that johnthetubaguy was saying. | 16:22 |
| johnthetubaguy | I was thinking more about the Nova attach side of things | 16:23 |
| ildikov | can we block having bootable=True and multiattach=True being set on a volume at the same time? | 16:23 |
| jgriffith | I think perhaps we're each talking about different problems and trying to solve them all at the same time | 16:23 |
| jgriffith | if bootable: multiattach=Prohibited | 16:24 |
| jgriffith | or whatever you like | 16:24 |
| jungleboyj | jgriffith: I agree with you. | 16:24 |
| ildikov | yep | 16:24 |
| jungleboyj | jgriffith: Meaning it can be attached once and subsequent attempts fail. | 16:24 |
| jgriffith | but it doesn't seem like it's difficult to solve that, at least for the primary cases | 16:25 |
| ildikov | or well, simply false | 16:25 |
| jgriffith | ildikov yes, or false depending on how the multi-attach stuff is implemented and what that looks like | 16:25 |
| jgriffith | but I don't know becuase it doesn't exist yet :) | 16:25 |
| jgriffith | jungleboyj yes, correct | 16:25 |
| ildikov | and if the bootable flag can be set later, then don't allow to set it for a volume that created as multiattach | 16:25 |
| jgriffith | the only draw back is possible race conditions | 16:25 |
| jungleboyj | Excellent. I am good with the plan. | 16:26 |
| jgriffith | ohhhh.... donuts! | 16:26 |
| ildikov | jgriffith: where is the race? | 16:27 |
| jgriffith | down the hallway to get donuts | 16:28 |
| jgriffith | :) | 16:28 |
| ildikov | yeah, I certainly want donuts now, that part is ok :) | 16:28 |
| * jungleboyj is jealous | 16:28 | |
| ildikov | jgriffith: I thought to still elaborate on the draw back you raised :) | 16:29 |
| jgriffith | the race is simultaneous bfv calls, but one will fail so it's ok | 16:29 |
| jgriffith | it's just not completely deterministic but I really don't think it's a big deal | 16:29 |
| smcginnis | Yeah, small and unlikely window. | 16:29 |
| jgriffith | smcginnis and somewhat benign anyway | 16:29 |
| ildikov | if multi-attach is false now for BFV then it shouldn't be a bigger problem than it is already | 16:29 |
| ildikov | or am I missing something? | 16:30 |
| jgriffith | sure | 16:30 |
| ildikov | ok, cool | 16:30 |
| smcginnis | The answer to "am I missing something" is almost always true. :) | 16:30 |
| ildikov | jgriffith: would you mind summarizing once again the overall plan on this? | 16:31 |
| ildikov | just to ensure I got it too :) | 16:31 |
| jungleboyj | smcginnis: The answer to 'does it matter' is not always yes. | 16:31 |
| ildikov | smcginnis: :) | 16:31 |
| jgriffith | merge new attach code | 16:32 |
| smcginnis | jungleboyj: Also true. | 16:32 |
| jgriffith | that's the plan | 16:32 |
| ildikov | jgriffith: fair enough :) | 16:32 |
| mriedem | so while you guys rehashed the multiattach policy again this week, | 16:32 |
| mriedem | i reviewed jgriffith's cinder patch | 16:32 |
| mriedem | please oh please dear sweet baby jesus, | 16:32 |
| mriedem | don't do that data migration offline | 16:33 |
| jgriffith | mriedem which one? | 16:33 |
| * jgriffith goes to gerrit | 16:33 | |
| mriedem | https://review.openstack.org/#/c/509005/ | 16:33 |
| *** gouthamr_ has joined #openstack-meeting-cp | 16:33 | |
| *** gouthamr has quit IRC | 16:35 | |
| jgriffith | mriedem if there's someone with the foo to do that without python I'd love to change it | 16:35 |
| mriedem | jgriffith: i can find an example from nova | 16:36 |
| jgriffith | mriedem let me know, when I looked that's how Nova was doing it too I think | 16:36 |
| mriedem | https://github.com/openstack/nova/commit/3674a4268d177230375fa1b581dbdf6f62755cee | 16:37 |
| mriedem | nova is not doing data migrations in the migration scripts | 16:37 |
| mriedem | we do that stuff in the object | 16:37 |
| mriedem | on read | 16:37 |
| mriedem | and we have the online_data_migrations CLI for batching those as needed | 16:37 |
| jgriffith | mriedem don't get offended or anything | 16:39 |
| mriedem | i posted some links in the change | 16:40 |
| jgriffith | thanks | 16:40 |
| mriedem | np | 16:40 |
| mriedem | i'm not offended, | 16:40 |
| mriedem | i just, | 16:40 |
| mriedem | my ass is still sore from years and years of operators ripping us on doing this | 16:40 |
| mriedem | including our internal people | 16:40 |
| jgriffith | mriedem cool, and I really appreciate you pointing it out!! | 16:41 |
| mriedem | i only just put the donut away 2 months ago | 16:41 |
| jgriffith | haha | 16:41 |
| jgriffith | not a good kind of donut either | 16:41 |
| ildikov | ok, it looks like we're on track with this too now | 16:43 |
| mriedem | well, not good to eat | 16:43 |
| jgriffith | good for sitting, not good for munching | 16:43 |
| ildikov | mriedem: thanks for the pointers, etc. | 16:43 |
| ildikov | :( :) | 16:43 |
| mriedem | unless you've got a fetish, but we digress | 16:44 |
| mriedem | end of meeting?! | 16:44 |
| ildikov | I think so | 16:44 |
| ildikov | unless someone objects | 16:44 |
| jgriffith | ha | 16:45 |
| ildikov | ok, so jgriffith will fix the shared_targets patch | 16:45 |
| ildikov | johnthetubaguy will review the live_migrate patch | 16:45 |
| * ildikov will update the multi-attach spec | 16:45 | |
| * ildikov will also look into start a Cinder side of that thing | 16:46 | |
| ildikov | and we will merge the new attach patch as soon as the new client with a new mv is out | 16:46 |
| ildikov | thanks everyone! | 16:46 |
| ildikov | have a good rest of your day! | 16:47 |
| ildikov | #endmeeting | 16:47 |
| *** openstack changes topic to "OpenStack Meetings || https://wiki.openstack.org/wiki/Meetings" | 16:47 | |
| openstack | Meeting ended Thu Oct 5 16:47:18 2017 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 16:47 |
| openstack | Minutes: http://eavesdrop.openstack.org/meetings/cinder_nova_api_changes/2017/cinder_nova_api_changes.2017-10-05-15.59.html | 16:47 |
| openstack | Minutes (text): http://eavesdrop.openstack.org/meetings/cinder_nova_api_changes/2017/cinder_nova_api_changes.2017-10-05-15.59.txt | 16:47 |
| openstack | Log: http://eavesdrop.openstack.org/meetings/cinder_nova_api_changes/2017/cinder_nova_api_changes.2017-10-05-15.59.log.html | 16:47 |
| jungleboyj | Sounds good! Thanks everyone! | 16:47 |
| *** mriedem has left #openstack-meeting-cp | 16:49 | |
| *** nhelgeson has joined #openstack-meeting-cp | 16:52 | |
| *** iyamahat has quit IRC | 17:08 | |
| *** aselius has joined #openstack-meeting-cp | 17:22 | |
| *** iyamahat has joined #openstack-meeting-cp | 17:25 | |
| *** lbragstad has quit IRC | 17:37 | |
| *** yamahata has joined #openstack-meeting-cp | 17:43 | |
| *** lbragstad has joined #openstack-meeting-cp | 18:10 | |
| *** diablo_rojo has joined #openstack-meeting-cp | 18:12 | |
| *** diablo_rojo has quit IRC | 18:19 | |
| *** nikhil has quit IRC | 18:32 | |
| *** diablo_rojo has joined #openstack-meeting-cp | 19:11 | |
| *** Rockyg has quit IRC | 19:12 | |
| *** markvoelker_ has joined #openstack-meeting-cp | 19:15 | |
| *** markvoelker has quit IRC | 19:15 | |
| *** diablo_rojo has quit IRC | 19:41 | |
| *** rarcea has quit IRC | 19:54 | |
| *** xyang1 has joined #openstack-meeting-cp | 20:00 | |
| *** edmondsw has quit IRC | 20:07 | |
| *** edmondsw has joined #openstack-meeting-cp | 20:08 | |
| *** edmondsw_ has joined #openstack-meeting-cp | 20:09 | |
| *** edmondsw has quit IRC | 20:12 | |
| *** edmondsw_ has quit IRC | 20:14 | |
| *** lbragstad has quit IRC | 20:30 | |
| *** edmondsw has joined #openstack-meeting-cp | 20:40 | |
| *** lbragstad has joined #openstack-meeting-cp | 20:40 | |
| *** MarkBaker has joined #openstack-meeting-cp | 20:43 | |
| *** gouthamr_ has quit IRC | 20:44 | |
| *** edmondsw has quit IRC | 21:41 | |
| *** edmondsw has joined #openstack-meeting-cp | 21:41 | |
| *** MarkBaker has quit IRC | 21:42 | |
| *** gouthamr has joined #openstack-meeting-cp | 21:43 | |
| *** edmondsw has quit IRC | 21:46 | |
| *** felipemonteiro__ has quit IRC | 21:46 | |
| *** lbragstad has quit IRC | 22:38 | |
| *** xyang1 has quit IRC | 22:45 | |
| *** iyamahat_ has joined #openstack-meeting-cp | 23:06 | |
| *** iyamahat has quit IRC | 23:08 | |
| *** sdague has quit IRC | 23:18 | |
| *** edmondsw has joined #openstack-meeting-cp | 23:28 | |
| *** edmondsw has quit IRC | 23:32 | |
| *** markvoelker_ has quit IRC | 23:51 | |
| *** kbyrne has quit IRC | 23:57 | |
| *** kbyrne has joined #openstack-meeting-cp | 23:58 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!