*** felipemonteiro has joined #openstack-meeting-cp | 00:18 | |
*** felipemonteiro_ has joined #openstack-meeting-cp | 00:19 | |
*** felipemonteiro has quit IRC | 00:23 | |
*** felipemonteiro_ has quit IRC | 00:26 | |
*** gouthamr has joined #openstack-meeting-cp | 00:42 | |
*** harlowja has quit IRC | 00:45 | |
*** nhelgeson has quit IRC | 02:22 | |
*** aselius has quit IRC | 02:29 | |
*** amrith has quit IRC | 02:32 | |
*** amrith has joined #openstack-meeting-cp | 02:36 | |
*** benj_ has quit IRC | 02:40 | |
*** benj_ has joined #openstack-meeting-cp | 02:41 | |
*** gouthamr has quit IRC | 03:17 | |
*** markvoelker_ has quit IRC | 03:24 | |
*** lbragstad has quit IRC | 03:44 | |
*** lamt has quit IRC | 05:05 | |
*** lamt has joined #openstack-meeting-cp | 05:05 | |
*** lamt is now known as Guest62992 | 05:05 | |
*** markvoelker has joined #openstack-meeting-cp | 05:25 | |
*** markvoelker has quit IRC | 05:58 | |
*** MarkBaker has joined #openstack-meeting-cp | 06:36 | |
*** MarkBaker has quit IRC | 06:41 | |
*** markvoelker has joined #openstack-meeting-cp | 06:55 | |
*** markvoelker has quit IRC | 07:30 | |
*** felipemonteiro_ has joined #openstack-meeting-cp | 07:31 | |
*** felipemonteiro__ has joined #openstack-meeting-cp | 08:03 | |
*** felipemonteiro_ has quit IRC | 08:06 | |
*** markvoelker has joined #openstack-meeting-cp | 08:26 | |
*** markvoelker has quit IRC | 09:00 | |
*** MarkBaker has joined #openstack-meeting-cp | 09:08 | |
*** felipemonteiro__ has quit IRC | 09:16 | |
*** MarkBaker has quit IRC | 09:20 | |
*** markvoelker has joined #openstack-meeting-cp | 09:57 | |
*** sdague has joined #openstack-meeting-cp | 09:59 | |
*** markvoelker has quit IRC | 10:29 | |
*** MarkBaker has joined #openstack-meeting-cp | 10:33 | |
*** MarkBaker has quit IRC | 10:46 | |
*** _pewp_ has quit IRC | 11:08 | |
*** _pewp_ has joined #openstack-meeting-cp | 11:09 | |
*** markvoelker has joined #openstack-meeting-cp | 11:26 | |
*** markvoelker has quit IRC | 12:00 | |
*** markvoelker has joined #openstack-meeting-cp | 12:57 | |
*** markvoelker has quit IRC | 13:28 | |
*** markvoelker has joined #openstack-meeting-cp | 13:28 | |
*** edmondsw has joined #openstack-meeting-cp | 13:37 | |
*** MarkBaker has joined #openstack-meeting-cp | 13:38 | |
*** gouthamr has joined #openstack-meeting-cp | 14:25 | |
*** lbragstad has joined #openstack-meeting-cp | 14:50 | |
*** xyang1 has joined #openstack-meeting-cp | 14:50 | |
*** felipemonteiro__ has joined #openstack-meeting-cp | 15:13 | |
*** felipemonteiro__ has quit IRC | 15:17 | |
*** felipemonteiro__ has joined #openstack-meeting-cp | 15:25 | |
*** felipemonteiro_ has joined #openstack-meeting-cp | 15:26 | |
*** felipemonteiro__ has quit IRC | 15:30 | |
*** MarkBaker has quit IRC | 15:58 | |
*** mriedem has joined #openstack-meeting-cp | 15:58 | |
ildikov | #startmeeting cinder-nova-api-changes | 16:00 |
---|---|---|
openstack | Meeting started Thu Dec 14 16:00:12 2017 UTC and is due to finish in 60 minutes. The chair is ildikov. Information about MeetBot at http://wiki.debian.org/MeetBot. | 16:00 |
openstack | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 16:00 |
*** openstack changes topic to " (Meeting topic: cinder-nova-api-changes)" | 16:00 | |
openstack | The meeting name has been set to 'cinder_nova_api_changes' | 16:00 |
ildikov | johnthetubaguy jaypipes e0ne jgriffith hemna mriedem patrickeast smcginnis diablo_rojo xyang1 raj_singh lyarwood jungleboyj stvnoyes | 16:00 |
jungleboyj | @! | 16:01 |
_pewp_ | jungleboyj (*´・д・)ノ | 16:01 |
mriedem | o/ | 16:01 |
ildikov | good morning gentlemen :) | 16:01 |
jungleboyj | Good morning. | 16:01 |
ildikov | we can quickly go through things I think | 16:02 |
ildikov | we have a bug fixing round that has bits and pieces in both Nova and Cinder for the new attach code | 16:02 |
ildikov | #link https://review.openstack.org/#/q/topic:bug/1737779+(status:open+OR+status:merged) | 16:02 |
mriedem | there is a cinder one now too | 16:03 |
ildikov | #link https://review.openstack.org/#/c/527852/ | 16:03 |
mriedem | https://review.openstack.org/#/c/527852/ | 16:03 |
mriedem | ahead of me | 16:03 |
ildikov | mriedem: yep, just didn't have the link handy :) | 16:03 |
mriedem | i'll update ^ once i'm out of meetings | 16:03 |
mriedem | and ML threads | 16:03 |
mriedem | and tc channels | 16:03 |
mriedem | oh my | 16:03 |
jungleboyj | mriedem: Great and then I will review. | 16:03 |
ildikov | mriedem: thanks for stepping up on that one, I wasn't sure whether or not we would want to drop attachment_specs | 16:03 |
ildikov | mriedem: sorry about the meeting :( :) | 16:04 |
ildikov | mriedem: as for the Nova side, do we have anything else beyond testing on those patches? | 16:05 |
mriedem | just needs reviews | 16:05 |
mriedem | i already asked for that in the nova meeting earlier today | 16:05 |
mriedem | i'll bug melwitt later | 16:05 |
ildikov | ah ok, missed that part | 16:06 |
ildikov | thanks! | 16:06 |
ildikov | as for reviews, we have the Cinder side spec for multi-attach still open: https://review.openstack.org/#/c/523608/ | 16:06 |
ildikov | with a recent comment to move it Rocky from tommylikehu | 16:07 |
ildikov | jungleboyj: what's your take on this? | 16:07 |
jungleboyj | ildikov: In multiple meetings. | 16:08 |
ildikov | jungleboyj: ok, no worries, we can get back to it, when you have a minute | 16:09 |
ildikov | mriedem: I started to update the libvirt patch | 16:09 |
jungleboyj | ildikov: Why was that put in there? | 16:09 |
ildikov | jungleboyj: I'm not sure I understand your question | 16:10 |
ildikov | jungleboyj: the spec is not approved yet | 16:10 |
mriedem | i haven't read the cinder spec yet | 16:10 |
mriedem | i wasn't sure why there was a cinder spec honestly | 16:10 |
smcginnis | It's up to jungleboyj, but I think we will still want that spec even though it's past spec freeze. | 16:10 |
ildikov | jungleboyj: and it seems we're over some deadline now that would suggest to move it Rocky now? | 16:10 |
mriedem | i thought nova just needed the shared_targets field on the volume to disconnect safely | 16:10 |
smcginnis | mriedem: Someone told jgriffith to write one. | 16:10 |
* smcginnis looks at ildikov | 16:10 | |
jgriffith | I was just following orders | 16:11 |
mriedem | is there anything in the cinder spec that nova actually needs to add multiattach support? | 16:11 |
ildikov | mriedem: we talked about that spec two weeks ago or three, johnthetubaguy had comments that got fixed | 16:11 |
ildikov | mriedem: it's mainly about the policies and how to specify 'multiattach' at volume creation and after | 16:11 |
jungleboyj | I don't think the spec was that far off and I think that we were close to having the pieces in place to integrate for Rocky. | 16:12 |
mriedem | ok | 16:12 |
mriedem | if it's mostly policy knobs, sure | 16:12 |
smcginnis | I belive the idea was to make sure to get written down some of our assumptions and decisions so we didn't have to rehash it 100 times. | 16:12 |
ildikov | mriedem: not a long read, if you have a minute to glance through if you see anything odd, I think it should be ok b now | 16:12 |
smcginnis | Just 50. | 16:12 |
jungleboyj | Long story short, I think we need to get the spec in for Queens. | 16:12 |
jgriffith | smcginnis: 75 | 16:12 |
mriedem | you need the policy rules for sure | 16:12 |
mriedem | because some nova backends don't support multiattach so deployments will need to turn it off | 16:12 |
smcginnis | jgriffith: 60, that's my final offer. | 16:12 |
ildikov | smcginnis: jgriffith: my bad, I like having some notes on what we agreed to do :) | 16:13 |
jgriffith | smcginnis: sold | 16:13 |
jgriffith | ildikov: you weren't the only one that asked for it so don't worry about it | 16:13 |
smcginnis | ildikov: No, it does make sense. We just need someone to approve it. | 16:13 |
ildikov | 60 is a nice number :) | 16:13 |
jgriffith | and it is good to have it written down so somebody can work it out | 16:13 |
*** aselius has joined #openstack-meeting-cp | 16:13 | |
ildikov | jgriffith: yeah, I guess I wanted notes I can follow... :) | 16:14 |
ildikov | mriedem: the only argument we had was handling 'multiattach' as volume type at creation time and then blocking the ability to modify it once the volume is attached | 16:15 |
jungleboyj | I will go approve it if that means we can move on. | 16:15 |
jungleboyj | jgriffith: ildikov if I approve it do we think we can still get it in, in queens? | 16:15 |
mriedem | i think definitely on the latter | 16:15 |
mriedem | changing a multiattach volume that is in-use likely screws up stuff on the nova side | 16:16 |
ildikov | jungleboyj: that would be the idea | 16:16 |
mriedem | 409 IMO | 16:16 |
ildikov | jungleboyj: mriedem has been missing some sleep for a few days/weeks now to make it happen for Queens | 16:16 |
jungleboyj | ildikov: Ok, I will review and we should target approving today. | 16:16 |
jungleboyj | mriedem is the best! | 16:16 |
ildikov | mriedem: yeah, we cannot play with the domain xml, or at least I don't want to | 16:16 |
ildikov | mriedem: so no modifying that value once the volume is attached, if it's back to available then I don't care who does what with it until the next attach | 16:17 |
ildikov | jungleboyj: cool, thanks | 16:17 |
mriedem | ildikov: same | 16:17 |
ildikov | and yes, mriedem is the best!!! | 16:17 |
mriedem | i'll have to read the spec about the volume type thing | 16:17 |
mriedem | i thought we said last time that the existing boolean flag on the volume was enough | 16:18 |
mriedem | even though the volume type might not support it | 16:18 |
mriedem | maybe fail the volume create in that case? | 16:18 |
mriedem | or fail the PUT /volumes change | 16:18 |
ildikov | mriedem: it is still a flag on the volume from the volume object perspective | 16:18 |
jungleboyj | Can we retrype a volume when it is attached? | 16:18 |
jungleboyj | jgriffith: ^^^ | 16:18 |
mriedem | jungleboyj: yes | 16:18 |
mriedem | that's the swap volume API in nova | 16:18 |
ildikov | mriedem: jgriffith clarified it in the spec as I got confused as well | 16:18 |
jungleboyj | Holy shit. Didn't know that. | 16:19 |
mriedem | retype in cinder posts to the swap volume api in nova | 16:19 |
mriedem | to attach the new block device in the guest and remove the old one | 16:19 |
jungleboyj | Ok, so we don't want to allow that on multi-attached volumes? | 16:19 |
jgriffith | jungleboyj: we don't want to allow that for multi-attach type | 16:20 |
jgriffith | jungleboyj: read the spec :) | 16:20 |
jungleboyj | jgriffith: Going. That is fine with me. | 16:20 |
smcginnis | mriedem: That's just retype with migration, right? | 16:21 |
jgriffith | smcginnis: That should be the case yes | 16:22 |
jgriffith | smcginnis: which you may recall :) | 16:22 |
jgriffith | just saying | 16:22 |
* jungleboyj has the spec open and is going to approve. | 16:22 | |
jungleboyj | after reading again. | 16:22 |
jungleboyj | mriedem: Any concerns with that? | 16:23 |
mriedem | sorry, in tc | 16:24 |
mriedem | smcginnis: it's also migration w/o retype | 16:24 |
mriedem | swap volume is called from cinder if you (1) retype a volume or (2) don't retype but migrate to another backend | 16:25 |
mriedem | as far as i understand | 16:25 |
smcginnis | But just plain old retype is fine and would make it very simple to change the flag. | 16:25 |
mriedem | i think you' can't do 1 without 2 | 16:25 |
mriedem | but you can do 2 without 1 yes? | 16:25 |
jgriffith | mriedem: migrate is a form of retype, but yes, we're saying the same thing | 16:25 |
smcginnis | There's 1) retype, 2) retype with --migration-policy on-demand, and there's 3) migration. | 16:25 |
ildikov | smcginnis: +1 to easily change the flag when the conditions are met | 16:28 |
mriedem | i'll try to read the spec this afternoon | 16:29 |
jungleboyj | Ok, the spec says flat out that you can't retype. I think that is the safest thing. | 16:29 |
mriedem | if jungleboyj can hold off on the approval | 16:29 |
jgriffith | smcginnis: I left the migrate command out because it shouldn't include any changes to the multiattach type setting | 16:29 |
jungleboyj | mriedem: Yep, will do. | 16:29 |
mriedem | i'm way overcommitted atm | 16:29 |
mriedem | and need to start actually doing some stuff | 16:29 |
jgriffith | smcginnis: there's no retype involved, so it should just keep the same setting etc, but maybe not | 16:29 |
ildikov | mriedem: fair enough | 16:30 |
smcginnis | jgriffith: Yeah, I think the same settings in that case. | 16:30 |
ildikov | mriedem: I think the libvirt patch is mainly ready except moving up the support-matrix.ini changes | 16:30 |
smcginnis | So #1 should be no big deal, #2 would have to raise SorryCantDoThat(), and migrate isn't involved. | 16:30 |
ildikov | mriedem: there was a concern on how to puch down the 'multiattach' info to the libvirt driver, which I sorted out by putting it into the connection_info dict | 16:31 |
jgriffith | smcginnis: yeah, that's what I came up with at least | 16:31 |
smcginnis | jgriffith: ++ | 16:31 |
ildikov | mriedem: that might be the only questionable thing in it | 16:31 |
ildikov | smcginnis: jgriffith: +1 | 16:32 |
mriedem | puch down? | 16:32 |
mriedem | is that hungarian? | 16:32 |
jungleboyj | smcginnis: ++ | 16:32 |
ildikov | doesn't seem to be :) | 16:32 |
mriedem | pass down? | 16:32 |
mriedem | ildikov: will have to look | 16:32 |
mriedem | ildikov: maybe leave a comment on the patch where you have a question? | 16:33 |
mriedem | to highlightit | 16:33 |
jungleboyj | I am good on the Spec unless we want to clarify the retype considerations but I think being conservative there is fine. Don't need to respin. | 16:33 |
smcginnis | jungleboyj: We can always do an update later. | 16:34 |
jungleboyj | ++ | 16:34 |
ildikov | mriedem: I will add a comment to be sure | 16:34 |
ildikov | mriedem: I just would like to be sure we don't have concerns with that part anymore | 16:35 |
jgriffith | jungleboyj: line #100 - 110 | 16:35 |
jgriffith | actually just 107 | 16:35 |
mriedem | tab opened to https://review.openstack.org/#/c/523608/ for later | 16:36 |
jungleboyj | jgriffith: I saw that but I thought that was only a problem if it is retype with migration ? | 16:36 |
jungleboyj | jgriffith: But I am fine with leaving it as is. | 16:36 |
ildikov | mriedem: thanks! | 16:37 |
jgriffith | jungleboyj: ahh, I see; yeah we can add some clarification in there; but in-use should cover that case IMO | 16:38 |
jungleboyj | jgriffith: Ok. I think staying on the safe side is best. | 16:38 |
jgriffith | retyping an in-use volume is the problem, not necessarily migrating, migrating an in-use volume however is no bueno, but that's a result of retype. Geesh... so simple | 16:38 |
jungleboyj | Yep! | 16:40 |
*** yamahata has joined #openstack-meeting-cp | 16:44 | |
ildikov | ok, it seems that we're on the same page with this | 16:45 |
ildikov | or are there any further concerns? | 16:45 |
jungleboyj | I don't think so. | 16:45 |
jungleboyj | Just need to get the spec through and keep working things. | 16:46 |
ildikov | ok, cool | 16:46 |
ildikov | jgriffith: I guess you're out of bandwidth regarding anything written in the Cinder spec | 16:46 |
ildikov | jgriffith: is that a valid assumption? | 16:46 |
jgriffith | ildikov: I'm just out of bandwidth to spend a year polishing the turd | 16:46 |
ildikov | jgriffith: fair enough I guess | 16:47 |
jgriffith | so yes, that's valid; I won't be picking up that work any time in the next few weeks | 16:47 |
jgriffith | I started it last week, but I won't get back to it any time soon I don't think | 16:48 |
ildikov | jgriffith: if there are any bits and pieces I can start from plz point me to it | 16:48 |
ildikov | jgriffith: otherwise I'll figure it out | 16:48 |
jgriffith | ildikov: yeah, I'll shoot you some info in Cinder channel | 16:49 |
ildikov | jgriffith: sounds good | 16:49 |
ildikov | jgriffith: thank you | 16:49 |
jungleboyj | ildikov: Thank you. | 16:49 |
ildikov | ok, I don't have anything more for today | 16:50 |
ildikov | anything from anyone else? | 16:50 |
jungleboyj | ildikov: I don't think so. Thank you for all you do! | 16:50 |
jungleboyj | jgriffith: Thank you for continuing to help. I know this has been a big turd to polish. | 16:51 |
ildikov | jungleboyj: just trying to keep it together :) | 16:51 |
jungleboyj | :-) | 16:51 |
ildikov | ok, let's close this for today and do some work | 16:51 |
ildikov | thanks everyone! | 16:51 |
jungleboyj | ildikov: ++ | 16:51 |
ildikov | #endmeeting | 16:51 |
*** openstack changes topic to "OpenStack Meetings || https://wiki.openstack.org/wiki/Meetings" | 16:51 | |
openstack | Meeting ended Thu Dec 14 16:51:57 2017 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 16:51 |
openstack | Minutes: http://eavesdrop.openstack.org/meetings/cinder_nova_api_changes/2017/cinder_nova_api_changes.2017-12-14-16.00.html | 16:52 |
openstack | Minutes (text): http://eavesdrop.openstack.org/meetings/cinder_nova_api_changes/2017/cinder_nova_api_changes.2017-12-14-16.00.txt | 16:52 |
openstack | Log: http://eavesdrop.openstack.org/meetings/cinder_nova_api_changes/2017/cinder_nova_api_changes.2017-12-14-16.00.log.html | 16:52 |
*** mriedem has left #openstack-meeting-cp | 17:00 | |
*** xyang1 has quit IRC | 17:01 | |
*** edmondsw has quit IRC | 17:06 | |
*** nhelgeson has joined #openstack-meeting-cp | 17:06 | |
*** edmondsw has joined #openstack-meeting-cp | 17:06 | |
*** edmondsw_ has joined #openstack-meeting-cp | 17:10 | |
*** edmondsw has quit IRC | 17:11 | |
*** edmondsw_ has quit IRC | 17:14 | |
*** edmondsw has joined #openstack-meeting-cp | 17:22 | |
*** edmondsw has quit IRC | 17:25 | |
*** edmondsw has joined #openstack-meeting-cp | 17:40 | |
*** felipemonteiro_ has quit IRC | 17:43 | |
*** edmondsw has quit IRC | 17:53 | |
*** edmondsw has joined #openstack-meeting-cp | 17:54 | |
*** xyang1 has joined #openstack-meeting-cp | 17:55 | |
*** xyang1 has quit IRC | 17:59 | |
*** yamahata has quit IRC | 18:06 | |
*** edmondsw has quit IRC | 18:14 | |
*** edmondsw has joined #openstack-meeting-cp | 18:21 | |
*** gouthamr has quit IRC | 18:24 | |
*** edmondsw has quit IRC | 18:25 | |
*** edmondsw has joined #openstack-meeting-cp | 18:27 | |
*** harlowja has joined #openstack-meeting-cp | 18:29 | |
*** edmondsw has quit IRC | 18:32 | |
*** stvnoyes has joined #openstack-meeting-cp | 18:32 | |
*** edmondsw has joined #openstack-meeting-cp | 18:33 | |
*** gouthamr has joined #openstack-meeting-cp | 18:37 | |
*** edmondsw has quit IRC | 18:37 | |
*** edmondsw has joined #openstack-meeting-cp | 18:39 | |
*** edmondsw has quit IRC | 18:43 | |
*** edmondsw has joined #openstack-meeting-cp | 18:45 | |
*** felipemonteiro_ has joined #openstack-meeting-cp | 18:56 | |
*** felipemonteiro__ has joined #openstack-meeting-cp | 18:58 | |
*** felipemonteiro_ has quit IRC | 19:02 | |
*** edmondsw has quit IRC | 19:12 | |
*** edmondsw has joined #openstack-meeting-cp | 19:13 | |
*** openstack has joined #openstack-meeting-cp | 20:31 | |
*** ChanServ sets mode: +o openstack | 20:31 | |
*** edmondsw has joined #openstack-meeting-cp | 20:31 | |
*** edmondsw has quit IRC | 20:36 | |
*** gouthamr has quit IRC | 20:47 | |
*** yamahata has joined #openstack-meeting-cp | 21:09 | |
*** openstack has joined #openstack-meeting-cp | 21:11 | |
*** ChanServ sets mode: +o openstack | 21:11 | |
*** gouthamr has joined #openstack-meeting-cp | 21:44 | |
*** gouthamr_ has joined #openstack-meeting-cp | 21:45 | |
*** gouthamr has quit IRC | 21:48 | |
*** gouthamr_ is now known as gouthamr | 22:04 | |
*** diablo_rojo has quit IRC | 22:04 | |
*** diablo_rojo has joined #openstack-meeting-cp | 22:07 | |
*** gouthamr has quit IRC | 22:28 | |
*** felipemonteiro_ has joined #openstack-meeting-cp | 22:50 | |
*** felipemonteiro__ has quit IRC | 22:53 | |
*** felipemonteiro_ has quit IRC | 22:56 | |
*** felipemonteiro_ has joined #openstack-meeting-cp | 23:14 | |
*** harlowja has quit IRC | 23:32 | |
*** felipemonteiro_ has quit IRC | 23:35 | |
*** harlowja has joined #openstack-meeting-cp | 23:39 | |
*** harlowja has quit IRC | 23:40 | |
*** harlowja has joined #openstack-meeting-cp | 23:42 | |
*** harlowja has quit IRC | 23:50 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!