16:00:18 #startmeeting Cinder 16:00:19 Meeting started Wed Nov 21 16:00:18 2018 UTC and is due to finish in 60 minutes. The chair is jungleboyj. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:20 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:23 The meeting name has been set to 'cinder' 16:00:27 Hi 16:00:29 hello 16:00:38 o/ 16:00:41 Courtesy ping: jungleboyj diablo_rojo, diablo_rojo_phon, rajinir tbarron xyang xyang1 e0ne gouthamr thingee erlon tpsilva ganso patrickeast tommylikehu eharney geguileo smcginnis lhx_ lhx__ aspiers jgriffith moshele hwalsh felipemonteiro lpetrut lseki _alastor_ whoami-rajat yikun rosmaita 16:00:51 hi 16:00:53 hi 16:01:18 @! 16:01:19 <_pewp_> jungleboyj (◍˃̶ᗜ˂̶◍)ノ” 16:01:32 hi 16:01:43 o/ 16:01:54 o/ 16:02:15 geguileo: Make your flight last week? 16:02:30 jungleboyj: yup, I did :-) 16:02:38 :-) Good. 16:03:05 Ok. smcginnis Is on a plane somewhere. So, we have everyone else I expect today. 16:03:09 So, lets get started 16:03:18 #topic Announcements 16:03:33 I have the Cinder on-boarding presentation posted: 16:03:44 #link https://www.slideshare.net/JayBryant2/cinder-onboarding-room-berlin-11132018 16:04:00 Thank you to rosmaita who reviewed it and found some mistakes. 16:04:11 :) 16:04:11 I have fixed those so they will be right when I start with that for Denver. :-) 16:04:43 rosmaita: That was after I fixed other issues found during the presentation. Ooops. 16:04:43 Might have phoned that one in. 16:04:54 There was good turnout for it and good questions so I think it was worth the time to do. 16:05:34 So, that is all I have there. 16:05:58 I am working on getting the audio from the forum sessions posted. I would like to say I will have that done today, but we will see. 16:06:09 I will work to get it posted soon. 16:06:34 just out of curiosity, have we gotten any contributors in the past from the onboarding sessions? 16:07:46 rosmaita: Yes, a couple. 16:07:53 great! 16:08:11 I think that Brian Pawlik started with an on-boarding session. 16:08:24 Same for the DRBD driver developers. 16:08:29 that's about 200% better than glance 16:08:50 So, not a lot, but something. 16:09:16 Also the group working on the encryption changes sat in on the session in Berlin so I think that was helpful for them. 16:09:32 We had a good discussion around their needs and actually may have solved their issues. 16:09:57 :-) So ... it has been useful time. 16:10:36 Ok. Moving on: 16:10:39 jungleboyj: yeah, they did 16:10:50 jungleboyj: because they talked with the Oslo team 16:10:57 geguileo: Awesome. 16:11:00 which is what we recommended they should do 16:11:12 And then they spent some time talking to eharney as well. 16:11:14 they'll be proposing a new olso project 16:11:25 geguileo: Good news. 16:11:37 that's good news 16:12:15 That makes more sense than the SDK idea. 16:12:30 jungleboyj: ++ 16:13:01 Anyway ... 16:13:16 #topic Paying attention to Bugs 16:13:40 Ok, so this was smcginnis but I think what he is trying to say is pretty obvious. 16:14:02 If you look at the link 3Par has a lot of bugs that haven't been addressed. 16:15:00 The contribution wiki clearly says that bugs need to be addressed. 16:15:19 Is there anyone here that works with 3Par? 16:15:21 _hemna: ^^ 16:15:27 hemna^^ 16:15:50 jungleboyj: to be addressed is to be confirmed? 16:16:17 enriquetaso: Confirmed, responded to and a plan to mitigate the issue. 16:16:53 i see some of them assigned long ago but no progress since a year. 16:17:13 whoami-rajat: Right. 16:17:49 jungleboyj: thanks 16:18:07 Some of them have had no response. 16:18:16 https://bugs.launchpad.net/cinder/+bug/1645136 this can be closed i think 16:18:17 Launchpad bug 1645136 in Cinder "create_cloned_volume of HPE 3PAR Cinder Driver has a performance bottleneck when source/target volume size are different" [Undecided,New] 16:18:19 I am sure that 3Par is not the only example here. 16:19:25 whoami-rajat: True. I will close that one out. 16:20:26 Anyway, we don't need to go through all of these right now. 16:20:37 jungleboyj: how can we proceed for these driver specific bugs? 16:21:20 Point is, if you are a driver maintainer, please take some time to review the list of bugs and respond to any bugs that may be filed against your driver. 16:21:40 jungleboyj: whoami-rajat the policy is that if there's not an active maintainer of the driver then it gets deprecated 16:22:00 actively maintaining means dealign with bugs in LP related to said driver 16:22:04 I am going to follow up with a note to the mailing list and indicate that this is required. 16:22:29 jgriffith: ok. Thanks for the info. 16:22:36 jgriffith: I am thinking I should indicate that drivers that are not dealing with their bugs will be marked unsupported just like not having CI running. 16:22:50 In other words how you deal with the driver related bugs is by maintaining your driver. The Cinde repo isn't a dumping ground for stuff 16:23:05 jgriffith: ++ 16:23:12 jungleboyj: yeah, we've done that before actually 16:24:08 Ok. So... 16:24:30 #action jungleboyj to e-mail ml about keeping up with bugs and that if you don't you will be marked unsupported. 16:24:47 #action jungleboyj and smcginnis to review open bugs to figure out who the offenders are. 16:25:38 Everyone ok with the plan above? 16:25:59 +1 16:26:02 +1 16:26:04 +1 16:26:08 ++ 16:26:25 +1 16:26:28 Excellent. 16:26:36 #topic User Survey Feedback Blog Post 16:26:48 #link https://docs.google.com/document/d/1d4_nUfuZacGABG2hqqCz3hC238uoVqdloi_qVGRNoRY/edit?usp=sharing 16:27:09 I have created a blog post to respond to our user survey feedback based on discussion we had at the Forum. 16:27:27 I would appreciate the rest of the team taking a look at it and making comments/updates. 16:27:44 jungleboyj: I read it really fast, could miss something. but it looks great! I'll read it more carefully after later today 16:27:56 I would like to get it sent off to SuperUser soon so that it is still fresh. 16:28:01 e0ne: Awesome. Thank yoU! 16:28:26 jgriffith and geguileo I would appreciate your input as well. 16:28:27 jungleboyj: thanks for writting it ! 16:28:46 geguileo: No problem. I have wanted to start doing that kind of thing. So, it was a good time to do it. 16:30:15 Any comments or questions to address on the document now? 16:31:07 ++ on keeping backups scheduling out of cinder 16:31:26 rosmaita: :-) Yeah, so many people asked for that and it is just not what we do. 16:32:02 Anyway, please review and leave your comments. I will get it updated and have one last call on it next week before submission. 16:32:40 That was all I had on the agenda for topics. Anything else to discuss or should we talk quickly about bugs? 16:32:53 I have something for "Ceph driver should respect the `--incremental` option for backups", but I'm not sure if this is the place to ask for more opinions about the bug. 16:33:03 #link https://review.openstack.org/#/c/612503/ 16:33:38 enriquetaso: I think we could talk about that. 16:34:01 #topic Ceph driver should respect the `--incremental` option for backups 16:34:12 jungleboyj: Looks good. i will read it in detail.i found some minor spelling mistakes, can we correct them ourselves to not to bother you with it? 16:34:26 whoami-rajat: Go ahead and fix them. 16:34:34 :) 16:34:59 jungleboyj: i've requested edit access. 16:35:37 Granted. 16:35:50 I'll appreciate any opinions about changing the default behavior of the backups and reviews of course :) 16:35:52 enriquetaso: Go ahead. 16:35:53 jungleboyj: Thanks! 16:37:52 enriquetaso: you have to explain on what you want us to give you feedback here 16:38:03 geguileo: Thank you. :-) 16:38:49 enriquetaso: Are you just looking for follow-up reviews? 16:39:51 oh, i'm looking for reviews u.u geguileo and jungleboyj 16:40:09 Ok. Can take a look and review. 16:40:21 enriquetaso: only for reviews? or do you want to hear about keeping backward compatibility? 16:40:27 Just wanted to make sure you weren't looking for something more specific. 16:40:50 jungleboyj: there is an issue with keeping it backward compatible 16:41:01 or making it behave like the other drivers that support incremental backup 16:41:03 Doh. Ok. What is the issue? 16:41:30 ceph backups cannot create full backups (except the first one) 16:41:37 since they are doing incremental by default 16:41:40 the patch fixes this 16:41:50 but if it behaves like other drivers 16:41:54 Right. Remember that discussion. 16:42:10 only doing incremental when requested, then it's not backward compatible 16:42:21 and if it's backward compatible, then it's not consistent with other drivers 16:42:39 :( 16:42:41 Hmmm. 16:42:52 :-( Indeed. 16:43:06 so it's a matter of preference/choosing our favourite evil 16:43:11 I'm wondered I we didn't catch anything of this on our gates 16:43:33 thanks geguileo, yup. So my patch make the driver consistent with other drivers 16:43:59 What are other drivers doing? 16:44:09 they create full by default 16:44:21 if you specify incremental, then they do incremental using the latest full as the base 16:44:22 geguileo: Ok. Makes sense. 16:44:30 sorry, latest full, or latest incremental 16:44:50 And I forget, why can't we make Ceph do the same? 16:44:51 but ceph just tries to do incremental if it can, and does full if it can't 16:45:00 I'm +1 for consistency even if we have to change our default behaviour 16:45:02 jungleboyj: we can do it, but then we change the default behavior 16:45:49 maybe it's enough to add an upgrade note 16:45:59 e0ne: Yeah. 16:46:02 to make sure that any automation is changed 16:46:05 I think consistency is best. 16:46:54 enriquetaso: looks like what you have is OK, maybe update the release note, and change the "features" section for an upgrade one 16:47:26 great! 16:47:31 geguileo: ++ 16:47:31 thanks!! 16:47:38 geguileo++ 16:47:55 Honestly, it seems like a bug that it didn't work the same as other drivers in the first place. 16:48:12 It's reported as a bug 16:48:13 https://bugs.launchpad.net/cinder?field.searchtext=ceph+incremental&search=Search&field.status%3Alist=NEW&field.status%3Alist=INCOMPLETE_WITH_RESPONSE&field.status%3Alist=INCOMPLETE_WITHOUT_RESPONSE&field.status%3Alist=CONFIRMED&field.status%3Alist=TRIAGED&field.status%3Alist=INPROGRESS&field.status%3Alist=FIXCOMMITTED&field.assignee=&field.bug_reporter=&field.omit_dupes=on&field.has_patch=&field.has_no_package 16:48:13 = 16:48:13 looks like many people want the incremental functionality with ceph. 16:48:30 full backup* 16:48:44 jungleboyj: it's not a real bug 16:48:54 jungleboyj: Ceph was the first backup driver that did incremental 16:49:05 Oh ... ok. :-) 16:49:08 jungleboyj: then we introduced incremental, but the driver was not updated 16:49:16 jungleboyj: so it was inconsistent with the others 16:49:33 geguileo: Gotcha. Ok. So, I think it is good to get it caught up. 16:49:34 in any case, this will fix it, and our drivers will be more consistent :-) 16:49:44 \o/ 16:50:05 Moving on 16:50:11 #topic Bug Triage 16:50:21 So, I wanted to bring up the following bug: 16:50:31 #link https://bugs.launchpad.net/openstack-ansible/+bug/1802070 16:50:32 Launchpad bug 1802070 in openstack-ansible "Add resource_filter field in cinder.conf for non-admin user can retrieve volume & snapshot volume" [Undecided,In progress] - Assigned to KimMinsik (for-beatitudo) 16:50:48 OpenStack Ansible needed to make a change to fix an issue they were seeing. 16:51:01 It seems to me, however, they are working around a Cinder bug. 16:51:32 I don't remember who was working on the resource_filter stuff. I am thinking it might have been tommylike ? 16:52:39 jungleboyj: I agree that we need to fix it in cinder too 16:53:07 e0ne: That was my thought as well. 16:53:59 It seems like there should be some default behavior if the file doesn't exit. 16:54:10 jungleboyj: +1 16:54:20 A 400 isn't right. 16:54:58 e0ne: Are you familiar enough with that code path to fix it up? 16:55:12 Since this was causing issues for Horizon it would be up your alley to fix/test. 16:55:21 jungleboyj: at lease I think so:) 16:55:36 I'll take it as my priority for tomorrow morning 16:55:38 :-) 16:56:05 e0ne: Ok. Cool. I can open a bug based on the one in OpenStack ansible to track it against Cinder. 16:56:09 Make sense? 16:56:24 I just add Cinder project to the existing bug 16:56:39 *added 16:56:45 Oh ... yeah. That probably makes more sense. 16:57:15 e0ne: Cool. Thanks for taking that one. 16:57:35 jungleboyj: np. I'm happy to be useful for the team :) 16:58:28 geguileo: Did you open a follow up bug to 16:58:29 #link https://bugs.launchpad.net/cinder/+bug/1782714 16:58:30 Launchpad bug 1782714 in Cinder "Properties of an attached volume are lost after live migration" [Medium,New] 16:58:30 e0ne: You are always useful to the team. :-) 16:58:41 jungleboyj: thanks, Jay! 16:59:05 jungleboyj: yes, and I proposed a fix 16:59:28 Ok. Cool. 16:59:28 * jungleboyj is behind on reviews. :-( 16:59:41 jungleboyj: it's in the last comment of the bug for reference 16:59:49 it's a 4 lines patch 16:59:50 lol 17:00:01 Nice! Ok. I will follow up on that. 17:00:05 thanks 17:00:11 jungleboyj: Sorry for interrupting, there is a frequent gate failure that yikun and I have been looking at, https://bugs.launchpad.net/cinder/+bug/1803648 . can the cinder team take a look at the patches for it? Thanks. 17:00:12 Launchpad bug 1803648 in Cinder "gate failure : testtools.matchers._impl.MismatchError: 'volume.retype' != 'backup.createprogress'" [Undecided,In progress] - Assigned to Rajat Dhasmana (whoami-rajat) 17:00:14 And on that note. We are out of time. 17:00:39 whoami-rajat: Sure. Can look at that as well. 17:00:39 Ok. Lets wrap up. 17:00:42 Thanks team for a good meeting. 17:00:53 Hope those of you in the US have a Happy Thanksgiving! 17:01:00 Talk to you all next week. 17:01:06 jungleboyj: Thanks! :) 17:01:25 Happy thanksgiving! 17:01:27 #endmeeting