14:00:38 #startmeeting cinder 14:00:41 Meeting started Wed Dec 4 14:00:38 2019 UTC and is due to finish in 60 minutes. The chair is rosmaita. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:41 hi rosmaita 14:00:42 Hi 14:00:42 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:43 hi 14:00:44 Hello) 14:00:45 The meeting name has been set to 'cinder' 14:00:46 o/ 14:00:48 o/ 14:01:11 hi :) 14:01:23 o/ 14:01:31 This is my first meeting 14:01:39 welcome whfnst ! 14:01:41 welcome! 14:01:47 o/ 14:01:55 #link https://etherpad.openstack.org/p/cinder-ussuri-meetings 14:02:28 Nice instructions rosmaita ;) 14:02:36 looks like a good turnout, let's wait a minute in case people are having trouble finding the room 14:02:38 smcginnis: ty 14:02:52 o/ 14:03:23 LiangFang: please take a look at the translation instructions in chinese at the top of the etherpad and see if they make sense 14:03:31 (i used google translate) 14:03:32 o/ 14:03:42 ok 14:04:21 ok, i guess we can get started 14:04:28 #topic announcements 14:04:42 first, welcome to our first meeting in our new time slot! 14:04:51 rosmaita, yay! 14:05:13 :) 14:05:41 the translate of chinese is correct 14:05:47 hi 14:05:52 LiangFang: thanks! 14:06:14 hopefully that will be useful to people and keep the etherpads un-messed-up 14:06:21 Yay Google Translate! 14:06:43 google translate is powerful:) 14:06:51 sure is! 14:07:18 LiangFang: is Chinese like French, where there's a governing body that approves new words? 14:07:28 i wonder if they'll introduce an 'etherpad' character 14:07:45 o/ 14:07:57 After 3 weeks in Shanghai I have new appreciation for Google's power. 14:08:05 :) 14:08:22 ok, the PTG summary wiki page has been updated with virtual PTG info 14:08:33 #link https://wiki.openstack.org/wiki/CinderUssuriPTGSummary 14:08:58 it would be good for people to search for their names and/or irc nicks for Actions that they volunteered for 14:09:13 you have to do both, i was kind of inconsistent using nicks and first names 14:09:29 also, we have some actions that are looking for owners 14:09:36 i put them into a handy etherpad 14:09:46 #link https://etherpad.openstack.org/p/cinder-ussuri-ptg-actions 14:10:33 so please take a look ... if you missed the PTG or Virtual PTG and aren't clear on what the action items are, please ping me in IRC or ask at one of the cinder weekly meetings 14:10:35 rosmaita: no governing body approves new words, lots of new words every month in network 14:11:02 LiangFang: let me know if 'etherpad' shows up as a new word! 14:11:06 :) 14:11:08 rosmaita: Thanks for getting that put together so quickly. 14:11:16 LiangFang: ++ 14:11:34 rosmaita: I will get the videos uploaded today. Sorry I have been slow getting to that. 14:11:42 :) 14:11:45 jungleboyj: no problem, i know you've been busy 14:11:59 jungleboyj: You have the power to get it done. ;) 14:12:03 i'll send an update to the ML when we get the video links on the wiki page 14:12:14 and the photo 14:12:28 He! 14:12:30 LiangFang: noted! 14:12:42 thanks 14:13:00 Yes I do. https://usercontent.irccloud-cdn.com/file/RKVQHqtz/image.png 14:13:10 ok, the only other announcement i have is to save the dates for the Virtual Mid-Cycle 14:13:24 we decided to do it as two 2-hour meetings, like the virtual PTG 14:13:28 except spread out more 14:13:45 first one will be either the 3rd or 4th week in January 14:13:55 the spec freeze is last week of January 14:14:14 i was thinking if we have it week before, we can address questions 14:14:23 but that's awful soon 14:14:33 so maybe last week in January is better 14:14:45 let me know what you think 14:14:55 i will send out a poll so we can determine the day and time 14:15:23 the other session will be the week of 16 March 14:15:41 that's on the Ussuri release schedule as the "Cinder new feature status checkpoint" 14:15:50 so it makes sense to check in and see how things are going 14:16:01 poll is good :) 14:16:18 we'll do a poll for that one too, but closer to march 14:16:33 Sounds like a good plan. 14:16:46 * jungleboyj is sad I don't have an excuse to travel to RTP though. 14:16:46 my point in bringing them up now is in case there's some kind of major event somewhere that would interfere with attendance 14:16:54 * rosmaita is too 14:17:33 ok, that's all for announcements 14:17:44 #topic Outreachy internship 14:17:48 enriquetaso: that's you 14:17:57 ok 14:18:03 Welcome to Anastasiya ( anastzhyr ) ! 14:18:13 yesterday was her first day ! 14:18:13 Hello everyone! 14:18:18 welcome! 14:18:19 Welcome anastzhyr 14:18:19 anastzhyr: Welcome! 14:18:23 welcome 14:18:28 From December to March Anastasiya will help us to improve Cinder Tempest. 14:18:40 welcome 14:18:58 I think we'll use a launchpad tag to keet track of the test scenarios 14:19:07 Test and Tempest scenarios: https://bugs.launchpad.net/cinder/+bugs?field.tag=test-coverage. Feel free to add or tag any test idea. 14:19:40 that's all I have 14:19:57 :) 14:20:06 enriquetaso: Thanks for being a mentor! 14:20:07 thanks! 14:20:08 thanks 14:20:25 #topic question about cinder iSCSI 14:20:36 #link https://review.opendev.org/#/c/685271/ 14:20:49 yangziguan: lvbin01: this is your topic 14:21:07 rosmaita: yes 14:21:33 in os.brick module , iscsi.py, ISCSIConnector.connect_volume will call "_connect_to_iscsi_portal" ( connect single path). if you target portal not with "3260" , will endless loop. 14:22:28 gerrit is loading really slowly for me 14:22:29 in my commit, eric Mention of removing 3260 ,but i test , in _connect_to_iscsi_portal will endless loop 14:23:02 multipath_connect have no problem 14:23:15 that is weird... 14:23:54 eric mention of removing 3260 , want to support ipv6 14:23:56 they use the same code afaik 14:24:15 yangziguan: is this an IPv4 or IPv6 issue? 14:24:40 is os.brick code issue 14:24:47 not ipv6 issue 14:25:05 yangziguan: yes, but are you using IPv4 or IPv6 14:25:29 ipv4 to test 14:25:45 now i also want to support ipv6 14:26:03 which one is failing? IPv4 with single path? 14:26:06 my product support ipv6 to iscsi 14:26:10 ipv4 14:26:34 i first join the meeting 14:26:46 how to send a picture 14:27:20 that is a good question 14:27:30 jungleboyj: any suggestions? 14:27:42 :-) In IRC? 14:27:51 also i can sent to all use email 14:28:03 Hmmm, if you are using IRC cloud you just paste it into the chat. 14:28:14 i send it to eric , but he don't reply to me 14:28:19 i think my client is too primitive 14:28:30 That are upload it somewhere and share the link. 14:28:47 yangziguan: You can open a launchpad bug and attach files there 14:29:02 geguileo: good suggestion! 14:29:27 yangziguan: https://bugs.launchpad.net/os-brick/+filebug 14:29:41 hao de 14:29:43 Apparently works for other backends, so something strange is going on here. 14:29:46 ok 14:29:50 yangziguan: you write a one line summary, then a description, and then you can attach logs, images, etc 14:30:12 we have a light agenda today, so if you want to do it now, that would be good 14:30:18 smcginnis: yes, I testted it both with single path and multipath, and worst case it doesn't connect 14:30:28 I've never seen it loop forever :-( 14:30:37 rosmaita: +1 14:31:07 not with 3260 in portal 14:31:07 yangziguan: while you are filing that bug and uploading the image, i will go to the next topic and then come back to you 14:31:17 ok 14:31:29 #topic EOL some branches follow-up 14:32:00 we decided to EOL the driverfixes/m and /n and stable/o and /p 14:32:11 i sent out a final note to the ML, and this time did get a response 14:32:26 #link http://lists.openstack.org/pipermail/openstack-discuss/2019-December/011349.html 14:32:55 the upshot is that there is interest in keeping ocata and pike in Extended Maintenance mode 14:33:11 i think that is fine, it will also keep cinder in sync with the other openstack projects 14:33:25 tosky has a concern that ocata will start failing badly soon 14:33:37 was there just the single answer? It would be interesting to know how long they plan to support both old branches 14:33:50 but when that happens, we'll see what the openstack community wants to do 14:34:19 tosky: i didn't ask because AFAIK, we would be the first project to EOL o and p 14:34:34 re ocata: I think I mentioned it already; most of the changes related to zuulv3 (including devstack and tempest) are available (or have been backported) from pike onwards, ocata was caught in the transition 14:34:53 it may simply happen that we decide to just drop everything which is not unit test and pep8 14:35:14 that may be an acceptable solution according the definition of Extended Maintenance 14:35:30 i believe that's acceptable based on my reading of the docs 14:35:47 but on that note 14:35:58 even the unit tests and pep8 are failing in driverfixes branches 14:36:10 oh, I meant how long that specific operator who uses them is interested in them 14:36:26 i tried a quick fix (run them on xenial) but it did not work 14:36:58 tosky: yes, he did not say 14:37:46 anyway, i kept a -W on my release patch to EOL m,n,o,p 14:38:18 until monday coming up (9 December) 14:38:39 we can see if anyone wants to step up and maintain driverfixes/newton 14:38:50 but that looks unlikely at this point 14:39:10 so my proposal is that we EOL driverfixes/* only 14:39:11 rosmaita: re your fix, you may need to backport some flake8 constraints 14:39:18 but yeah, let's see if there is interest 14:39:43 tosky: that and it looked like there were some packages missing from the index on xenial 14:40:25 although, i don't think we ever allowed windows line endings 14:40:45 the package missing may have been a mirror issue; but again, let's see 14:41:12 tosky: yes, it may or may not be easy to fix, i didn't want to get too far into it 14:41:27 ok, so to summarize: 14:41:45 i won't ask the release team to EOL anything until Monday 9 December 14:42:08 and we will *not* EOL any stable branches 14:42:11 just the driverfixes 14:42:28 any thoughts? 14:43:13 i guess i should say that implied in that is if someone steps up who wants to maintain driverfixes/newton, we'll discuss at next week's meeting 14:43:19 can you please ask the question on how long they are interested in keeping ocata and pike? 14:43:21 (or I can do it) 14:43:22 rosmaita: That makes sense to me. 14:43:57 tosky: if you don't mind, it will make it look like there's a wide community of interest 14:44:07 oki 14:44:30 although, i was going to reply after this meeting to say what we decided 14:44:42 so i can ask as part of that 14:44:52 and you can follow up if the answer isn't clear 14:45:07 tosky: ^^ 14:45:33 ok, let's get back to the previous discussion 14:45:48 https://bugs.launchpad.net/os-brick/+bug/1855126 14:45:48 Launchpad bug 1855126 in os-brick "target portal not with "3260" , will endless loop" [Undecided,New] 14:45:59 #topic question about cinder iSCSI continued 14:46:08 #link https://bugs.launchpad.net/os-brick/+bug/1855126 14:46:13 thanks, yangziguan 14:47:38 Hmmm. 14:47:54 btw, the code link is https://opendev.org/openstack/os-brick/src/branch/master/os_brick/initiator/connectors/iscsi.py#L1022-L1087 14:47:57 yangziguan: For future reference you can provide reference to a line in git. 14:48:07 yangziguan: Like rosmaita Just did. :-) 14:48:34 so line 1071 is the problem? 14:48:57 #link https://opendev.org/openstack/os-brick/src/branch/master/os_brick/initiator/connectors/iscsi.py#L1071 14:49:01 yes 14:49:50 yangziguan: OK, thanks for the bug report!!! 14:51:36 Is there still time? 14:52:03 iscsiadm -m session will get ip:3260, but i not with 3260 , then will goto endless loop 14:52:55 yangziguan: I understand... 14:54:14 why you don't to encount it 14:54:30 why you don't to encounter it 14:55:14 means: why have not 14:55:53 Right. Guessing no one has used something other than port 3260? 14:56:14 wait, I may be missing something 14:56:35 are you using something different than 3260 like 4444, or are you just not returning the port_ 14:56:36 ? 14:56:51 no 14:56:59 alway use 3260 14:57:51 looks like we may need some discussion on the bug report 14:58:09 yangziguan: do you have logs in DEBUG mode, please? 14:58:16 #topic open discussion 14:58:21 whfnst: did you have something? 14:58:25 yangziguan: Feel free to bring up any of these questions in the #openstack-cinder channel. They don't need to wait for a meeting. 14:58:34 smcginnis: ++ 14:58:40 I have a question about volume host.... 14:58:43 I need to Reappearance 14:58:44 And really probably shouldn't be in the meeting, since not everyone here needs to be debugging this together. 14:58:49 yangziguan: yes, please feel free to ping me to discuss this 14:58:55 smcginnis: +1 14:59:33 geguileo: Thank you! 14:59:44 jungleboyj: np 15:00:05 whfnst: let's take your question into the #openstack-cinder channel 15:00:11 we're out of time here 15:00:17 thanks, everyone! 15:00:22 bye! 15:00:22 #endmeeting