14:00:39 #startmeeting sahara 14:00:40 Meeting started Thu Feb 14 14:00:39 2019 UTC and is due to finish in 60 minutes. The chair is tellesnobrega_. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:41 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:44 The meeting name has been set to 'sahara' 14:00:47 o/ 14:01:00 o/ 14:01:15 full attendance today 14:01:17 hi folks 14:01:22 #topics News/Updates 14:03:00 I mostly went through reviews and approved some, and proposed a lot of backports 14:03:27 (I slowed down in the last days because I'm working (again, and this time for real) on the unified native Zuul v3 grenade job) 14:03:55 I've been working on some NTP issues with CDH, the patch is up for review again 14:03:56 I have few patches to test on my queue though (for plugins, mapr and spark) 14:04:15 I saw tosky's comments and replied to them, hopefully it should be good now 14:04:37 other than that, I'm investigating hbase issue with cdh 5.11 and probably 5.13 as well 14:04:46 it simply fails to start, don't really know why 14:04:52 and started testing python3 14:05:04 trying to understand the pickle issue when connecting to instances 14:05:48 i'm working (slowly, not much time) on adding the missing api features to osc, dashboard 14:09:23 hi 14:09:24 I'm back 14:09:34 internet is breaking a lot today 14:10:53 #chair tosky 14:10:55 tellesnobrega: if you set your nickname to the one that you had when you started the meeting, you can set also other people as chairs 14:11:09 lets see if I can 14:11:31 #chair tosky 14:11:32 Current chairs: tellesnobrega_ tosky 14:11:39 oki :) 14:11:40 just in case 14:11:45 any more updates? 14:12:32 not from me 14:12:39 not from me 14:12:56 oh, there was a promotion on RDO and apparently the split sahara packages are working 14:12:57 #topic Python 3 14:13:05 tosky, nice 14:13:55 so, for python3, I wanted to take the to discuss what I have found with some pickle debugging 14:14:24 ah, yep 14:14:28 the tricky issue 14:14:52 so, I started debugging the issue 14:15:58 and the issue happens when we try to run the _connect command on the subprocess with pickle 14:16:27 the only difference I found between are on the args 14:16:43 and to be really sure, I'm not sure if it is an issue 14:17:17 so you used a python3 host with devstack? (I guess Fedora or Ubuntu Bionic) 14:17:46 yes 14:17:52 my devstack is running on centos 14:18:00 but sahara on python3 is running on fedora29 14:18:07 but what about the instances: do we have any plugin where the base OS supports python3? 14:18:23 hum 14:18:28 my open question so far is whether this happen only when you mix python3 (caller) and python2 (callee) 14:18:52 of course maybe it's not important to answer this question to fix the issue 14:19:24 on the other hand, if the code didn't work even with python3-python3, that would be different 14:20:14 apart from that, you may nag some python 3 expert, maybe this is something already know 14:20:26 it may be worth trying that py3-py3 case, yes -- it might be possible to create bionic images for Vanilla without much effort 14:20:27 (but I'm not particularly optimistic on that :) 14:21:08 so, the only difference I found was on the args 14:21:27 on python 2 the args are unicode and on python3 they are string 14:21:44 not sure if that makes a difference at all for pickle, but that is as far as I got 14:24:01 any ideas? 14:24:38 iirc unicode in python2 is the same as string on python3, but maybe they are not exactly the same (or pickle does not encode them in the same way and some magic is needed) 14:25:43 not really an expert 14:26:33 yeah 14:28:05 I will keep trying and probably ping some python3 experts 14:29:04 anything else on this? 14:29:29 I have a smaller python3 fix for sahara-tests which has already a +2: https://review.openstack.org/#/c/628654/ 14:29:41 nice 14:31:34 I will review and approve 14:32:44 done 14:32:49 lets move on 14:32:56 #topic Open Discussion 14:34:06 anything else for today? 14:34:17 the survey question? 14:34:20 yes 14:34:24 thanks jeremyfreudberg 14:34:34 I was asked if we want to update the survey question for sahara 14:34:46 Sahara: What else can be improved in the Sahara project? 14:34:54 that is our current one 14:35:08 jeremyfreudberg, as future PTL you should make that call, I believe 14:35:31 final decision at least since you will be the most responsible for handling its answers later on 14:36:08 I think this question is a little vague, and we could narrow it down a bit to make the responses matter more 14:36:18 what do you two think? 14:37:05 like adding something ".... (for example, ...)" ? 14:37:27 i think there's definitely a balance-- this question does not really inspire people to answer thoughtfully, but we also want to keep it general enough so that it's applicable to everybody (a couple PTGs ago, I proposed making the question about which services should be added to plugins, but that's too specific -- for some users there will be other more important things to mention) 14:38:28 Maybe giving an example is a good addition 14:39:03 any ideas for what the example could be? 14:39:44 for examples, adding new services, adding new data processing tools, making it more user friendly (how so) 14:39:45 I was thinking more a list of examples of possible answers 14:39:47 something like that? 14:40:32 something like that 14:40:42 jeremyfreudberg? 14:42:22 i don't want it to sound too much like a multiple choice question (because then the answers will be short and undetailed), but i think we should somehow allude to the two main categories of suggestions: related to the sahara service itself, and related to the operation of the clusters which sahara creates 14:42:40 (not sure exactly how to express the distinction between those categories in the most elegant words) 14:43:19 that makes sense, lets do this, we have until tomorrow to modify this 14:43:56 lets try to gather some examples today and vote (by email) and we can update if needed 14:44:11 sure 14:44:21 awesome 14:44:26 I will try to write something 14:46:52 before we close: 14:46:55 there are few sahara-tests patches still open (S3 testing, APIv2 for scenario and tempest), but I'd like to ask for a quick check on https://review.openstack.org/#/c/590055/, which should unlock few other patches for python3 support 14:47:43 ups, wrong review 14:47:57 I meant this one: there are few sahara-tests patches still open (S3 testing, APIv2 for scenario and tempest), but I'd like to ask for a quick check on https://review.openstack.org/#/c/590055/, which should unlock few other patches for python3 support 14:48:03 argh, stupid clipboard 14:48:08 https://review.openstack.org/#/c/608211/ 14:48:34 that one I have reviewed 14:48:40 if you can take a look jeremyfreudberg :) 14:49:07 yes, i'll take a look 14:49:14 oh, also: you may have noticed few stable/pike backports whose tempest jobs are failing 14:49:52 yes 14:49:53 I asked gmann, but further investigation is needed, so leave them there for now 14:50:00 ok 14:50:05 one more thing: sahara does not appear on the list attending PTG. http://lists.openstack.org/pipermail/openstack-discuss/2019-February/002618.html . i wanted to make sure that this is intentional (at least based on the current view) 14:50:26 jeremyfreudberg, unfortunately it is 14:50:32 I won't be able to make to the PTG 14:50:52 not sure about tosky, but I'm assuming chances are small right now 14:51:06 for me it is very up in the air too 14:51:17 even if you make it, not much to discuss by yourself 14:51:26 so I decided not to reserve a room 14:51:36 make sense, even if unfortunate 14:51:41 it is sad 14:51:52 I still don't know, my decision process is different than tellesnobrega_'s 14:52:43 we manage to have enough to talk in the hour 14:52:45 :) 14:52:55 anything else for today? 14:53:07 that's it from me 14:54:32 tosky? 14:54:36 nothing else from me 14:55:36 thanks 14:55:41 see you all next week 14:55:56 o/ 14:55:58 lets try to have some question examples by email soon so we can have a decision 14:56:54 #endmeeting