18:00:09 <tellesnobrega> #startmeeting sahara
18:00:10 <openstack> Meeting started Thu Mar 30 18:00:09 2017 UTC and is due to finish in 60 minutes.  The chair is tellesnobrega. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:00:12 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
18:00:15 <openstack> The meeting name has been set to 'sahara'
18:00:26 <tosky> o/
18:00:36 <raissa> o/
18:00:57 <shuyingya> o/
18:01:03 <esikachev> hi
18:01:11 <vgridnev> hello
18:02:25 <tellesnobrega> #topic News/Updates
18:02:45 <mariannelm> o/
18:02:54 <tellesnobrega> I'm working on cdh image generation, now I'm testing it.
18:03:05 <tellesnobrega> probably going to start working on updating spark and storm version
18:03:24 <esikachev> i am working on improvements for sahara-ci stability
18:03:31 <tosky> so, after last week's outages, both sahara-image-elements (broken due to changes in dib-lint from dib) and sahara (broken due to the new eventlet) gates are now fixed
18:04:00 <vgridnev> Working on jobs in upstream Jenkins
18:04:26 <tosky> also testing internally (on RHEL) and finding issues from time to time
18:04:52 <shuyingya> nothing new from my side.
18:05:43 <tellesnobrega> shuyingya, is your colleague working on cdh version upgrade to 5.10?
18:06:10 <raissa> busy testing sahara internally
18:06:55 <shuyingya> not yet. The intern has to go back to school to finish his paper.
18:07:05 <tellesnobrega> ok
18:07:10 <mariannelm> I'm working with APIv2 (with shuyingya), and planning to work on adding some documentation for the data source / job binary abstraction interface
18:07:28 <mariannelm> * working on APIv2
18:08:10 <tellesnobrega> I still haven't sent the email about microversions, but I'll try to send it tomorrow first thing
18:09:38 <tellesnobrega> anyone else has news?
18:09:59 <shuyingya> tellesnobrega, I have dealt with the internal affairs of company yesterday. So I can work on CDH refactoring and upgrading
18:10:32 <tellesnobrega> shuyingya, great to hear, if anyone is interested in working on storm or spark let me know
18:11:29 <shuyingya> as you know, mariannelm has lots of interest on APIv2, so I can take some time to work on plugin upgrading first
18:11:46 <tellesnobrega> yes, that sounds great
18:12:33 <tellesnobrega> about boston summit, apparently there will be only me and jeremy there and hopefully Elise, I requested onboarding room for sahara so we can try introduce new people to our project and get them working
18:12:45 <mariannelm> yes sure, tellesnobrega feel free to email me the details about it
18:13:59 <tellesnobrega> I forgot, mariannelm will be there as well
18:14:08 <tellesnobrega> she will probably help us out there
18:15:05 <tellesnobrega> I will also present a projects update talk there, I may ask for your input on a topic or other, so please help me out :)
18:16:45 <tellesnobrega> anyone else has news?
18:18:33 <tellesnobrega> #topic Open Discussion
18:18:50 <esikachev> I need some help with debugging Sahara-ci from developers side because now I am working on another project and have a little time for working on Sahara-ci
18:19:49 <vgridnev> What kind of help you need?
18:19:49 <tellesnobrega> esikachev, I see, I haven't done any work on sahara-ci itself but I can try to help out with that
18:19:50 <esikachev> take a look on sahara-ci results https://review.openstack.org/#/c/445954/
18:20:23 <esikachev> only small plugins can working correct
18:20:58 <esikachev> also, core-team take a look on https://review.openstack.org/#/c/443554/ :)
18:21:10 <tosky> could it be due to resource constraints?
18:21:43 <esikachev> i think, no
18:21:59 <esikachev> because we have a lot of resources
18:23:58 <vgridnev> I see a lot of failures related to eep jobs, can that be caused by datasources pluggability changes?
18:24:28 <esikachev> let's try to create revert-patch
18:24:29 <mariannelm> don't know vridnev, do you have the logs?
18:24:39 <tosky> esikachev: or trigger against ocata and/or newton
18:24:48 <esikachev> ok
18:24:50 <esikachev> )
18:25:04 <vgridnev> Let's revert that
18:25:58 <tellesnobrega> vgridnev, you mean revert the data pluggability changes and see if it passes?
18:26:08 <vgridnev> Yes
18:26:36 <vgridnev> If that is the general reason of failures, that has to be repaired
18:27:52 <tellesnobrega> makes sense, but do we know that is the source of the problem? did it pass the tests when it was submitted?
18:29:53 <mariannelm> hey everyone, I got to go now, but if possible please discuss the status of job binary internal's deprecation, I'll see the logs, thanks \o/
18:30:26 <vgridnev> I don't know for sure, but that might be possible
18:31:24 <tellesnobrega> vgridnev, I'm ok with reverting the changes and see if it fixes the issue
18:31:47 <tosky> just sending the revert patch should say something
18:32:33 <tosky> I shamelessy request some eyes (and votes) for two RHEL-related fixes for sahara-image-elements: https://review.openstack.org/#/c/448741/ and https://review.openstack.org/#/c/448644/
18:32:33 <tellesnobrega> there were three patches, should we revert all of them?
18:32:44 <tosky> (tested internally, planning to backport to Ocata too)
18:32:57 <vgridnev> yes, let's revert all of them
18:33:04 <tellesnobrega> vgridnev, ok
18:33:56 <tellesnobrega> I will take care of that
18:34:22 <tellesnobrega> and we can see if it helps the CI, if it does, we have to discuss how to bring those in again without breaking it
18:34:30 <tosky> you can chain the 3 revert patches
18:34:49 <tosky> the results from the last one should be enough
18:35:06 <tosky> I mean, we don't need to merge the revert to see if it fixes things, right?
18:35:34 <tellesnobrega> probably not
18:35:42 <tellesnobrega> here are the patches
18:35:48 <tellesnobrega> https://review.openstack.org/#/c/427447/
18:35:53 <tellesnobrega> https://review.openstack.org/#/c/427359/
18:36:03 <tellesnobrega> https://review.openstack.org/#/c/427480/
18:36:35 <tosky> one is the spec
18:36:40 <tosky> not required for the test
18:36:45 <tosky> the other two can be chained
18:38:16 <tellesnobrega> tosky, none of these is a spec file
18:38:45 <tosky> uh, where did I click?
18:38:55 <tellesnobrega> tosky, not really sure
18:39:23 <tosky> ok, ok
18:39:29 * tosky shuts up
18:39:52 <tellesnobrega> I will take a look on how to chain those 3 figure out the order that it should happen and see how it goes
18:40:05 <shuyingya> Hi all, what do you think about adding HBase support in vanilla?
18:41:03 <tellesnobrega> shuyingya, do you have use case?
18:42:17 <shuyingya> not yet. Just m product manager asks me why not support HBase in vanilla like CDH
18:43:15 <tellesnobrega> I don't see why not. You could write a spec file and we would discuss that there
18:43:54 <shuyingya> Uhm, sound great. thx
18:45:28 <shuyingya> tellesnobrega what's you plan about microversion?
18:45:46 <tellesnobrega> about microversions, this is something that was proposed in the original spec on APIv2. I started working on it, but would like to know you input on it, if this is something we still want
18:47:08 <tellesnobrega> I mean, the goal to avoid having to bump versions on small api changes, and give support to different set of behaviors depending on which api version is being requested
18:49:01 <tellesnobrega> vgridnev, do you have thoughts on that? I know tosky has some concerns
18:50:05 <shuyingya> uhm, I am not sure yet. waiting for you email.
18:50:09 <vgridnev> Do whatever you want, we actually don't have a lot changes on api
18:51:24 <tellesnobrega> yes, we don't change much our api, we might not even need this, wanted to know your input, but I will write the email and we can have a more detailed discussion
18:51:37 <tellesnobrega> how does that sound?
18:53:05 <tosky> yep, let's start from the email
18:53:27 <shuyingya> great. make sense.
18:53:29 <vgridnev> Basically, I don't have strict opinion on this, just because we don't do something that totally will break everything
18:53:51 <tellesnobrega> vgridnev, I agree with you on this
18:54:42 <shuyingya> https://review.openstack.org/#/c/448113/ about this patch, do you have any idea about storing the job bianries?
18:55:33 <tellesnobrega> The idea is to remove internal in favor of swift or other storage
18:55:44 <shuyingya> I think use db to store blob is not good idea, do we need use swift or manila?
18:56:57 <tellesnobrega> if I remember correctly we want to not use internal db for that and use swift or manila or whatever comes next
18:57:44 <tellesnobrega> for APIv2 we shouldn't have the internal job binary option anymore
18:58:46 <shuyingya> uhm, that's true. But the job binaries operation use the resource of job binary internal.
18:59:30 <tellesnobrega> hmm, I have to take a better look into it. We can discuss it later, since we are almost out of time. Is that ok?
18:59:40 <shuyingya> ok.
18:59:46 <shuyingya> I need to go to sleep
18:59:48 <shuyingya> haha
18:59:53 <tellesnobrega> shuyingya, awesome
18:59:57 <shuyingya> goodbye everyone
19:00:08 <tellesnobrega> thanks guys, I will follow up an email on microversions and job binary internal
19:00:15 <tellesnobrega> thank all
19:00:22 <tellesnobrega> #endmeeting