14:00:17 #startmeeting sahara 14:00:22 Meeting started Thu Oct 25 14:00:17 2018 UTC and is due to finish in 60 minutes. The chair is tellesnobrega. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:23 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:25 The meeting name has been set to 'sahara' 14:00:32 o/ 14:00:39 o/ 14:00:42 hey jeremyfreudberg long time no talk 14:01:00 welcome back tellesnobrega 14:01:04 thanks :) 14:02:04 lets start with updates and we can move to tc talk once someone joins 14:02:09 #topic News/Updates 14:02:58 I'm still testing plugins split, I've successfully created spark and storm clusters 14:03:02 now I'm testing vanilla 14:03:14 ambari is failing for a still unknown and unrelated to the split issu 14:03:32 cdh image is huge and it is not uploading on my env at this moment 14:03:37 mapr I haven't got to it yet 14:03:51 I'm still working on S3 testing; my patch is almost done (I just need to figure out how to cleanly override the job settings for a specific cluster in scenario templates) 14:03:52 once I get all clusters to start I will start working on running scenarios testing 14:04:21 I managed to test S3 (but with RHEL) on Cloudera and MapR; I have some troubles with Ambari, where jobs are killed 14:04:37 and I realized that the oozie console is not installed on Ambari 14:05:32 is (was) oozie not on the template? 14:05:45 also, I've been working on health repair, albeit very, very slowly and sporadically 14:06:07 uhm, I will recheck, but I thought it was - maybe not 14:06:10 thanks for the hint 14:06:36 two tc members! 14:06:37 hi! 14:07:07 o/ 14:07:14 lets move on to the cloud vision discussion 14:07:22 #topic Cloud Vision Discussion 14:07:27 hi zaneb 14:07:41 hello friends 14:07:51 first of, zaneb, do you have anything you want to start with? 14:08:30 jeremyfreudberg: thanks for your feedback on the review 14:09:16 zaneb: happy to help 14:09:17 I think that's a good suggestion, to try to suggest where the complexity level is that having it as-a-Service in OpenStack is worth it 14:10:22 I guess I don't have a lot to say beyond what's in the review and the mailing list thread, I'm mostly interested in what y'all think 14:10:42 is that the way that you see Sahara as adding value to OpenStack, or are there different ways as well? 14:10:52 jeremyfreudberg, tosky? 14:11:21 does the rest of it kind of fit with the pattern of what you imagine users will do with OpenStack, or were you imagining something different? 14:11:29 those kinds of questions :) 14:11:59 cdent: anything to add to that? 14:12:11 zaneb, I have to be honest, I haven't read the whole thing yet, but regarding the sahara part, I agree with you and also with what jeremyfreudberg has said 14:12:51 specially the part regarding the complexity of the deploying such plataforms 14:13:37 this is something that should be touched because it is one of the main motivations of using sahara 14:13:39 i really like what you've written, and i think it covers sahara well, and fits with my brain-pictures... one thing to point out "abstracting the management" can be broadened slightly in the sahara use case, as it's currently worded it doesn't quite cover the abstraction of submitting data processing jobs 14:13:44 and of course trove 14:13:51 yes, it's one layer of abstraction over other layers of abstractions, but it captures the need of a specific category of users/admins 14:13:52 My main concern is that the vision helps people make decisions about what is in scope for their day to day work. 14:13:56 do we think the fact that it involves clustering is what makes stuff complex enough to need a service? or just complexity in general? 14:15:00 jeremyfreudberg: ah, that's interesting 14:15:20 IMHO the good thing for Sahara is that glues together various OpenStack services to help people deploying their cluster 14:15:28 zaneb, there are two levels of complexity, the clustering itself, and the configuration of each data processing tools 14:15:45 which is something that, surely, you could do manually, but you would end up reimplementing the glue that sahara is... in a less general way 14:16:30 it's good to clarify :) 14:18:08 so complexity that may include but is not limited to managing clusters :) 14:19:38 yes, i'll agree, the tools of any of these "abstract specialized ops" may be complex, and the deployment of these tools into a clustered env may be complex, and that entire picture gives a sense of whether it's "complex or not" 14:19:43 I would say that it's managing cluster ->with resources provided by other openstack services<- 14:20:14 tosky: ++ 14:21:58 you're seeing the first draft of that particular section, so this is all great feedback that we can definitely roll in 14:23:16 great 14:23:22 anything else on this topic? 14:24:05 thanks for inviting us over! 14:24:29 thanks for starting that work, it will be a great source of information 14:24:46 that's it from me for now, thanks zaneb and others, i'll keep an eye on the review 14:25:19 * tosky agrees too 14:25:30 thanks everybody :) 14:25:49 thanks zaneb cdent and evrardjp 14:25:57 anytime 14:26:01 I did so much ;) 14:26:21 moving on 14:26:26 I'll stay around if you don't mind, just listening, in case you need anything else :) 14:26:30 tosky, jeremyfreudberg specific topics for today? 14:26:37 evrardjp, sure, no problem at all 14:27:23 i have to leave at 40ish past the hour, not really any specific topics from me 14:28:07 i could probably fill open discussion for a few minutes, so let's go to that unless tosky or tellesnobrega has a topic 14:28:11 just a quick reminder for few patches: https://review.openstack.org/#/c/611276/ and the two backports for queens 14:28:25 nothing else right now 14:28:36 #topic Open Discussion 14:28:40 today is M1 14:28:56 yes, and m1 is now optional 14:29:03 from what I understood from recent email it is optional 14:29:26 but there is a rule for libraries, should we cut a release for python-saharaclient at this point? 14:29:30 do we want m1? 14:29:58 do we have relevant changes in python-saharaclient? 14:30:07 (I think not really) 14:30:31 https://github.com/openstack/python-saharaclient/compare/2.0.0...master 14:30:37 no, nothing to release 14:30:43 no code changes 14:30:49 cool 14:30:59 we will skip m1 for everything them 14:31:00 so no new version to depend on 14:31:11 yes 14:31:19 we will need an M2 though, after the split 14:31:41 yup, m2 is jan 10 14:31:44 (my birthday, actually) 14:31:48 ah! 14:32:02 as a gift you will do the release :D 14:32:10 \o/ 14:32:12 ha 14:32:21 jk 14:32:33 but yes, we will definitely release m2, and we have time to get lots of good things (split, etc.) in there 14:32:40 yes 14:32:51 we need apiv2 there 14:33:05 agree 14:33:23 talking about the split, I started drafting the email for openstack-dev@ on http://etherpad.openstack.org/p/sahara-plugins-split 14:33:51 yes, ongoing infra problems on etherpad, so we'll see if i can load it 14:33:55 that pad was originally created by tellesnobrega to collect the links to the work-in-progress split repositories, but it seemed the right place 14:34:21 ah, loaded, finally 14:34:27 tosky, it is right enough 14:34:43 I need to fill few details about the current packaging status 14:35:07 and I will probably raise the question of how to handle the split on RDO first, only because I'm technically an RDO packager too 14:35:45 email looks good so far 14:35:53 i may think of things as we start to get closer 14:37:30 thanks for starting that draft tosky 14:38:18 one quick topic from me 14:38:28 i think you've all seem gaasmann's story here: https://storyboard.openstack.org/#!/story/2004111 14:38:36 about some process problems 14:38:50 he's already done a good bit of investigation 14:38:51 my question is 14:38:51 nc process hanging? 14:39:15 my question is, have we ever seen this in container tripleo? 14:39:59 I haven't 14:40:26 I don't remember 14:40:39 but it's not a scenario that I tested too much in depth (sadly) 14:41:11 alright, that's fine if you don't have the complete information about it 14:41:17 but it would be interesting to look into 14:41:46 whether or not it's kolla specific or all container specific, or not at all container related 14:42:02 but it sounds like gaasmann is working hard at it, so no big worry yet 14:42:08 yes 14:42:47 i gotta go guys, but great meeting 14:42:57 thanks jeremyfreudberg 14:42:58 thanks jeremyfreudberg! 14:43:03 bye 14:43:23 anything else for today tosky? 14:45:03 if not I guess we can close for today 14:46:52 nothing else from my side 14:47:00 thanks everyone 14:47:04 great meeting 14:47:11 thanks tc folks for joining as well 14:47:28 #endmeeting