14:00:04 <tellesnobrega> #startmeeting sahara
14:00:05 <openstack> Meeting started Thu Jan  4 14:00:04 2018 UTC and is due to finish in 60 minutes.  The chair is tellesnobrega. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:07 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:09 <openstack> The meeting name has been set to 'sahara'
14:00:19 <jeremyfreudberg> o/
14:01:51 <tellesnobrega> waiting for tosky
14:03:12 <tosky> i/
14:03:14 <tosky> o/
14:03:21 <tellesnobrega> there we go
14:03:31 <tellesnobrega> #topic News/Updates
14:04:32 <tellesnobrega> I'm working on image gen for MapR and dashboard part of decommission of specific node (having trouble but hopefully making progress)
14:05:28 <tosky> I'm helping tellesnobrega testing the MapR generation (we are starting to see the light); also testing ambari 2.4 generation with less sucess
14:05:44 <jeremyfreudberg> i finished up S3 JB in openstackclient and sahara-dashboard.  finished up [keystone_authtoken] patch. now hopefully i will be able to focus on force delete (not really that hard) and APIv2 in saharaclient (harder than i want it to be)
14:06:06 <tosky> and trying to clean some issues in my queue: I just opened https://bugs.launchpad.net/sahara/+bug/1741233 with a related patch
14:06:08 <openstack> Launchpad bug 1741233 in Sahara "Reduce the log messages from amqp and oslo.messaging" [Undecided,New]
14:06:18 <tosky> and I'm going back to the CLI tests not working with v3
14:06:22 <tosky> keystonev3
14:07:24 <tellesnobrega> jeremyfreudberg, about APIv2, maybe we should send another email on the thread so Monty can actually see, the last one probably went undetected by him
14:07:40 <tosky> or ping him/nag him on IRC
14:07:46 <jeremyfreudberg> yeah, i need to talk to mordred for sure
14:07:57 <jeremyfreudberg> the holiday season did not make it easy for either of us, i'm sure
14:08:17 <tellesnobrega> definitely
14:08:20 <tosky> yep
14:08:39 <jeremyfreudberg> we need to get that moving soon because we have to have a final release of client library at Q-3
14:08:44 <tosky> same for zuulv3 updates, which I need to look at again
14:08:46 <tellesnobrega> yes
14:09:02 <tellesnobrega> deadline is Jan 26
14:09:16 <tellesnobrega> we get about 3 weeks to get this done
14:09:33 <jeremyfreudberg> it is definitely do-able, I just need an injection of Monty power to get it done
14:10:06 <tellesnobrega> sounds good
14:11:00 <tellesnobrega> I'm looking for help on dashboard, I will ping horizon people to see if I can get things sped up
14:11:17 <jeremyfreudberg> tellesnobrega, what's the exact problem you're facing?
14:12:30 <tellesnobrega> the biggest one is I don't really understand dashboard stuff
14:13:25 <tellesnobrega> what I want to do is basically the same flow as create cluster, where you first select the plugin and version and later it goes to create cluster stuff
14:13:31 <tellesnobrega> I want to do the same for scale
14:13:57 <tellesnobrega> you select the count of node groups than it goes to the new tab with the instances to select
14:14:26 <jeremyfreudberg> i don't even remember anything about the current scaling stuff in UI
14:14:27 <tellesnobrega> don't really know how to connect the two workflows, not sure if there should be two workflows or just steps in the same workflows
14:14:47 <tellesnobrega> its just the node groups, and the count of instances for each
14:14:57 <tellesnobrega> so you can increase or decrease
14:15:08 <jeremyfreudberg> right
14:16:50 * mordred waves to jeremyfreudberg
14:17:12 * jeremyfreudberg waves back to mordred
14:17:31 <mordred> ah - the api thing ... crap, I probably did miss the most recent message
14:17:46 <jeremyfreudberg> mordred, no worries
14:18:02 <mordred> there it is!
14:18:03 <tellesnobrega> mordred, definitely not your fault, too close to Christmas
14:18:20 <tellesnobrega> I'm just going to change topic so the talk is on the right place
14:18:33 <tellesnobrega> #topic APIv2
14:19:35 <mordred> jeremyfreudberg, tellesnobrega: I can definitely get you some links to some places that are using the discovery code from keystoneauth 'correctly'
14:20:40 <jeremyfreudberg> mordred: that would be great
14:20:48 <mordred> there's also another option to consider that may be a *bit* crazy, but maybe isn't? ...
14:21:02 <jeremyfreudberg> crazy can be good
14:21:25 <mordred> which is to not add v2 support to python-saharaclient at all and to just add support to openstacksdk and point people to it
14:22:15 <tellesnobrega> I don't think I understand how that would work
14:22:38 * tosky smells a bit of conflict of interests from mordred
14:22:40 <tosky> :D
14:22:54 <mordred> I'm not 100% sure I do either - it just occured to  me as I was thinking about what I'd link you to for discovery examples...
14:22:58 <mordred> tosky: :)
14:23:18 <jeremyfreudberg> right, i would assume if any project were to be doing version discovery right, it would be openstacksdk
14:23:26 <jeremyfreudberg> does it do the endpoint manipulation stuff too?
14:23:37 <jeremyfreudberg> with the project id and such, i mean
14:23:40 <tosky> the only problem is: at this point, porting to saharaclient/v2 should be fast enough (most methods will almost the same), how different is going to be the API exposed by openstacksdk?
14:23:41 <mordred> the code in the master branch does
14:23:59 <mordred> tosky: oh, good point
14:24:05 <tosky> I mean, we can implement it for sure and slowly move there like we did for openstackclient, I'm just not sure about the initial implementation
14:24:10 <tosky> the timing for it
14:25:05 <mordred> maybe for now I'll just give you some links to some things - and then when I'm done with this pagination patch I'm working on right now, I'll make a POC patch to see if just consuming sdk in python-saharaclient would make anything easier/better
14:25:32 <tellesnobrega> that sounds great mordred
14:25:33 <tellesnobrega> thanks
14:26:04 <mordred> sure nuff! seven years in to openstack you'd think we'd have a solid answer for this ... maybe in another seven we will
14:26:24 <jeremyfreudberg> mordred, great
14:26:26 <jeremyfreudberg> thanks!
14:26:35 <tellesnobrega> I don't think there is a solid answer for much stuff in OpenStack
14:28:31 <tellesnobrega> I guess we can move on
14:28:43 <jeremyfreudberg> yep
14:29:20 <tellesnobrega> #topic Open Discussion
14:29:59 <tellesnobrega> Did you get a chance to read shu yingya's email?
14:30:20 <tosky> about https://bugs.launchpad.net/sahara/+bug/1741233 and https://review.openstack.org/#/c/531137/, do you think that I can go on and (if passes) backport to pike? I have an hard time debugging anything right now without
14:30:22 <openstack> Launchpad bug 1741233 in Sahara "Reduce the log messages from amqp and oslo.messaging" [Undecided,New]
14:30:36 <jeremyfreudberg> cycle-with-intermediary¶
14:30:37 <jeremyfreudberg> oops
14:30:43 <jeremyfreudberg> java.lang.NoClassDefFoundError: org/apache/commons/httpclient/HttpMethod
14:31:23 <jeremyfreudberg> tosky, i think it's fine, but i'm not the expert on that kind of thing
14:31:25 <tellesnobrega> tosky, I think so
14:33:04 <tosky> jeremyfreudberg: as discussed on https://review.openstack.org/#/c/490971/, I kept stevedore as INFO, overriding the default
14:33:11 <tosky> so let's see if the gates agree
14:33:58 <jeremyfreudberg> tosky, great
14:34:09 <tosky> aaand also I will work again on zuulv3 jobs; the job for sahara-tests is ready, with the exception of the multinode job, but that can be fixed later
14:34:34 <tosky> I will shamelessly use this space and having mordred around to check if there are some news about the new publishing method for artifacts, or if I should code using the current way
14:34:51 <tosky> so to fix sahara-extras, publish the oozie tarball and remove the last bits of sahara-files
14:37:43 <tellesnobrega> looks like mordred went away
14:37:56 <jeremyfreudberg> i think it is rather early where he is
14:38:08 <tellesnobrega> probably
14:39:21 <tosky> yeah - and I nagged him enough already about this
14:39:28 <tellesnobrega> lol
14:40:54 <jeremyfreudberg> we can discuss the shu yingya hadoop-swiftfs thing for a bit
14:40:59 <jeremyfreudberg> as telles suggested
14:41:03 <tellesnobrega> yes
14:41:20 <tellesnobrega> does anyone understands what is going on there?
14:41:30 <tosky> don't steal my questions!
14:41:45 <tosky> :)
14:41:58 <jeremyfreudberg> lol, i don't really know either, it's something with the dependencies
14:42:30 <jeremyfreudberg> building hadoop-swiftfs against the hadoop 2.8 libraries and things go wonky i guess
14:42:31 <tosky> he asked about building openstack-hadoop.jar, but the script in sahara-extra should do it
14:42:57 <tellesnobrega> yes, maybe we need a little more context from him
14:43:00 <jeremyfreudberg> tosky, i think he has it building now, it just builds something that doesn't totally work
14:43:58 <tellesnobrega> hum, makes sense. Maybe the new oozie version he is using removed some dependencies
14:44:15 <tosky> could it be that the current openstack-hadoop.jar is not working with hadoop 2.8 and he can't build it?
14:44:39 <tosky> also, openstack-hadoop is not about oozie
14:44:45 * tosky a bit more confused than before
14:44:59 <jeremyfreudberg> in the email the job was launched by oozie but in fact the traceback was within hadoop-swiftfs
14:45:09 <tellesnobrega> oh, that makes more sense
14:45:16 <mordred> tosky, tellesnobrega: sorry, went away for a second
14:45:53 <mordred> however, nothing new to report there yet, so use the current way for now
14:46:08 <tosky> ack, thanks!
14:46:13 <jeremyfreudberg> https://github.com/apache/hadoop/commit/62579b69a0a294ba1ea14cf76c650b640f89f331 << I wonder if shuyingya has seen that
14:47:47 <tosky> but that looks like 3.0.0
14:48:15 <jeremyfreudberg> tosky, that is true
14:48:34 <jeremyfreudberg> (but still, removing the dependency causing the error is still nice)
14:49:55 <jeremyfreudberg> actually a handful of these commits https://github.com/apache/hadoop/commits/trunk/hadoop-tools/hadoop-openstack/pom.xml all talk about httpclient
14:50:03 <tellesnobrega> next time we see shu yingya we should ask him for more details so we can pin point the problem
14:50:11 <jeremyfreudberg> that seems to be one of the few things apache folks have been doing to hadoop-swiftfs, is managing dependencies
14:51:34 <jeremyfreudberg> but regardless, we do need to more info from shu yingya, to really know what the right fix is
14:51:49 <tellesnobrega> agreed
14:53:15 <jeremyfreudberg> anything else?
14:53:34 <jeremyfreudberg> ptg stuff?
14:54:33 <tosky> oh, right, PTG
14:54:39 <tellesnobrega> PTG
14:54:48 <tellesnobrega> I will be there, tosky as well
14:55:00 <tellesnobrega> are you going to be there? jeremyfreudberg
14:55:19 <tellesnobrega> I will start an etherpad this week so we can start putting topics there
14:55:39 <tellesnobrega> I asked for 2 days, just like last time, should be enough time
14:55:48 <jeremyfreudberg> most likely i will be there. of course it takes place during classes so it's always a challenge
14:56:03 <jeremyfreudberg> on that note, i might want to switch our two days to Thursday and Friday
14:56:12 <jeremyfreudberg> instead of the usual Wednesday and Thursday
14:56:14 <jeremyfreudberg> but i'm not sure yet
14:56:50 <tellesnobrega> that should be possible, I just need to check my flight time on friday
14:56:57 <tellesnobrega> but I think it is late enough
14:57:26 <jeremyfreudberg> it might not be necessary
14:57:28 <jeremyfreudberg> just giving heads up
14:57:32 <jeremyfreudberg> hopefully i can say by next week
14:57:35 <tellesnobrega> thanks
14:57:48 <jeremyfreudberg> since i'm sure the ptg people want to know if we are changing days too
14:57:58 <tellesnobrega> definitely
14:58:31 <tellesnobrega> I will send out an email with the etherpad so we can start putting it together
14:58:42 <jeremyfreudberg> cool
14:59:41 <tosky> what is the deadline for requesting the days? tellesnobrega, did you already request the basic two days as in Denver?
14:59:57 <tellesnobrega> I did already, but I can try to change it
15:00:04 <tellesnobrega> not sure there is an specific deadline
15:00:19 <tellesnobrega> we are at the top of the hour. Thanks all for joining and lets hurry up our tasks, and start getting ready for PTG.
15:00:37 <tosky> thanks all o/
15:00:42 <tellesnobrega> #endmeeting