03:04:35 #startmeeting openstack-cyborg 03:04:36 Meeting started Thu Sep 12 03:04:35 2019 UTC and is due to finish in 60 minutes. The chair is Sundar. Information about MeetBot at http://wiki.debian.org/MeetBot. 03:04:37 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 03:04:40 The meeting name has been set to 'openstack_cyborg' 03:04:46 Hi Sundar. 03:04:55 \0 03:05:24 This is the week of Train-3 milestone. 03:06:01 We'll look at the status, pending patches and future plans. Does anybody have any opens to be addressed? 03:06:33 #topic Status so far 03:06:44 maybe, py3 patch. 03:07:04 #info Yumeng 03:07:14 Good news: We got nearly all the nova-integ patches merged. https://review.opendev.org/#/q/project:openstack/cyborg+branch:master+topic:nova-integ Only the Nova notification patch remains. 03:07:59 Thanks Sundar for your great efforts 03:08:05 IMHO, that is close too. We have some cleanup to do in the way we connect to other services. But we can fix that after this week too. 03:08:29 Yumeng: Thanks. All your contributions helped too. 03:09:15 The next most important things after Nova notifn would be tempest (https://review.opendev.org/#/c/667231/) and Python 3. 03:09:36 DO you al agree? Other patches are important too, but they cna be merged after this week too 03:10:26 how about async bind patch? 03:11:10 That also can come in after this week, IMHO. I don;t think it is ready to merge in next 2 days. 03:12:10 Not so good news: It has been an uphill climb to get reviews from nova developers. The view I am hearing is that Nova patches for Cyborg need more sustained review before they can merge. 03:13:11 So far, the fact that most patches haven;t merged was a reason why reviewes were not getting done. Also, tempest CI was a prereq. We are very close to getting them all done , I think. 03:13:34 Any thoughts or comments? 03:14:21 It seems to me that Python 3 patch from Shogo is also close. Has nybody checked it out in their local env? 03:15:21 I checked py3 patch with latest version(P9 merged), it works well. 03:16:19 s_shogo: Great. I'll also try it just in case. I'd invite others to check too. Do you think we can merge it by Friday? 03:17:52 Sundar: I think so, there seems to be no big issue. 03:18:38 +1 03:19:15 Great. BTW, here are my Cycle highlights for Train. https://review.opendev.org/#/c/681283/ They include Python 3 too. 03:19:40 #topic Tempest 03:20:23 We are supposed to show the tempest CI as a prerequisite, like I said. 03:20:59 Yumeng: thanks a lot for all that you and your team have done. Is https://review.opendev.org/#/c/676665/ critical to merge this week? 03:21:58 Definitely, we need to merge https://review.opendev.org/#/c/667231/ soon. 03:22:10 sundar: i will update today. I think we should merge asap 03:22:47 agreed . also 667231 03:22:55 Great. Glad to see your progress on privsep - but I think we can even do that in the next couple of weeks. 03:23:12 No need to burn yourself out :) 03:23:23 about https://review.opendev.org/#/c/673957/ 03:23:50 Eric and sean suggest we we don't need to have a cyborg.privsep.* sub module at all 03:23:59 just a top level cyborg/privsep.py module that init the context and the privileged function should be in normal modules with privileged appended as a prefix or suffix. 03:24:29 but the current fact is nova and cinder still have cyborg.privsep.* sub module 03:24:59 so do we just follow the current way or change it ? what do you think? 03:24:59 Yumeng: should we merge 676665 before 667231? 03:25:07 Yes. I have heard it be said even nova's use of rootwrap etc. is not a good template to follow. Sometimes we are expected to do better than than the legacy projects because we are new. 03:26:47 Yumeng: IMHO, it is better to take some time and do this right, because it will be tough to revisit this again. So, it may be better to follow the experienced input. 03:26:53 yes, pleas merge 676665 before 667231 03:28:01 Sundar: ok. then I will reconsider it. do as they suggest. 03:28:14 Thanks a lot, Yumeng. 03:28:24 Yumeng: Re. 676665, There is a -1 on that. What does it take to get it resolved? 03:28:56 I will do that 03:29:08 Excellent. Thanks :) 03:29:59 What should we do about https://review.opendev.org/#/c/675990/ ? 03:30:16 (That is Python 3. I jumped to a different topic.) 03:31:06 and to all, please help to review and merge this bug fix https://review.opendev.org/#/c/675059/ 03:31:10 That was included in the py3 patch. https://review.opendev.org/#/c/667524/ 03:31:45 Ye. That was part of s_shogo's patch. 03:31:49 So, it should just be abandoned, I suppose. 03:32:09 Agree 03:32:13 Sundar: That's right. 03:32:19 Cool. 03:32:58 Yumeng, wangzhh: Re. https://review.opendev.org/#/c/675059/, Has GPU been tested with the nova-integ patches? 03:35:15 I have tested with part of the nova-integ patches. 03:35:24 Sorry, I don't have GPU env. But I think this part should be OK. Because there is little change this part. 03:35:35 Yumeng, Great. 03:35:43 Ok. Has anybody launched a VM with real GPU attached? Oherise, how do we know it works? 03:35:47 *Otherwise 03:36:30 wangzhh: yes a tiny fix. pls help to review \0/ 03:36:57 hi all 03:37:13 hi shaohe. 03:37:15 Hi shaohe_feng 03:37:32 To summarize the most important patches for this week: 03:37:41 Nova notification: https://review.opendev.org/674520 03:37:48 Python 3: https://review.opendev.org/#/c/667524/ 03:37:55 Tempest: https://review.opendev.org/#/c/676665/ , https://review.opendev.org/#/c/667231/ 03:37:57 Sundar: emmmm... I tested in a devstack env. not a real GPU env 03:38:17 Do we all agree that these are the important patches to merge this week? 03:38:47 agreed' 03:39:16 +1 03:40:01 Zhurong has some suggestions on Nova notification, but I suggest we merge this patch first, and then have time to modify it. 03:40:30 chenke: Totally agreed. Zhurong's point is important and I am working on it. 03:40:59 chenke: Thanks to you and zhurong for lots of active contribution! It'll be good to merge your patches by next week, IMHO, to help further development. 03:41:28 #topic AoB 03:41:38 Anything else, folks? 03:42:00 https://review.opendev.org/#/c/681005/ 03:42:20 will be test on really evn today 03:42:41 Cool, thanks Shaohe. 03:42:42 still some bug, I think it can works today. 03:43:10 it is some complex. 03:43:22 but it implemented as we discussion. 03:43:30 welcome RFC. 03:44:56 Sure. All, please review and help merge the important patches above. Have a good day! 03:44:57 Do we need to merge this patch before nova integrate? https://review.opendev.org/#/c/678177/ 03:45:55 wangzhh: Nova developers have not asked for that. They tend to take security for granted, so this may yet come up. It is important to merge that too. 03:46:04 we need this patch, but maybe nova integrate does not need it . 03:46:31 FIne, so plz also help to review this patch. 03:46:39 sure, 03:46:40 Sure, will do 03:47:09 when is the dead line of async job? 03:47:48 it it really not easy to test it on really env. 03:48:00 too many steps 03:48:30 shaohe_feng: Given that this week is milestone 3, we have to see how things go from here. I agree with you. It is better to take time and ensure it works well. 03:49:05 It does not have to merge this week, if that makes you feel better :) 03:49:30 OK, I will do my best to make it works well today. 03:49:49 Great. Thanks, shaohe_feng 03:49:51 for tomorrow is our Moon festival 03:50:27 I see. Happy Moon festival to all of you! 03:50:58 Thanks, 03:51:04 All right, see you all next week. Bye :) 03:51:09 #endmeeting