14:00:17 <egafford> #startmeeting sahara 14:00:18 <openstack> Meeting started Thu May 19 14:00:17 2016 UTC and is due to finish in 60 minutes. The chair is egafford. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:19 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:22 <openstack> The meeting name has been set to 'sahara' 14:00:25 <egafford> o/ 14:00:30 <egafford> Hi all! 14:00:32 <Poornima> Hi everyone 14:00:51 <crobertsrh> hello/ 14:01:10 <egafford> #topic News / updates 14:01:12 <Poornima> oh chairman is egafford :) 14:01:48 <egafford> Poornima: true this time around. 14:02:37 * egafford has been writing up a spec for the next stages of the Great Image Generation Push. Hopefully I'll have those posted by next week. Other than that, bunch of reviews. 14:02:47 <egafford> What's everyone else up to? 14:02:51 <crobertsrh> just reviews for me lately 14:03:16 <huichun> Reviewing patches this week 14:03:50 <Poornima> reviewing patches for HDP2.4 and api v2 14:03:57 <tmckay> I have been consumed with other things, but hopefully soon I'll have more time for reviews 14:05:10 <mionkin> i finished research about launching kerberos with cloudera 14:05:11 <elmiko> o/ 14:05:17 <egafford> Cool; sounds like we should be able to crush our review queue handily at least. :) 14:05:22 <egafford> mionkin: Nice! 14:05:28 <Poornima> elmiko, o/ 14:05:44 <elmiko> i've been looking at the oozie hardcoded password bug, sadly it's going to require some heavy lifting to get it fixed 14:06:09 <elmiko> we need to rewrite how we inject the sql file to create the passwords, so... it's going to take some time 14:06:25 <egafford> elmiko: Yeah, looks like it's much deeper than most from the comments. 14:06:26 <elmiko> hey Poornima =) 14:06:41 <elmiko> egafford: yea, i'm bummed about it 14:07:10 <egafford> No action items from last time, so last chance for updates going once... 14:07:16 <egafford> Twice... 14:07:28 <egafford> #topic Review priorities 14:07:37 <egafford> #link https://etherpad.openstack.org/p/sahara-review-priorities 14:08:05 <egafford> We're starting to use this again to track higher-prio reviews, to try to speed along key features. 14:08:44 <elmiko> ack, will take a look 14:08:46 <egafford> Does anyone have patches they believe should be on the sheet for expedited review? 14:09:43 <egafford> For the minutes, the patches up for expedited review at the moment are: 14:09:55 <egafford> #link https://review.openstack.org/#/c/304679/ 14:10:10 <egafford> MapR failure at liberty. 14:10:17 <egafford> #link https://review.openstack.org/#/c/315024/ 14:10:31 <egafford> Ambari on Centos7 (Red Hatters take note :) ) 14:10:45 <egafford> #link https://review.openstack.org/#/c/316010/ 14:10:57 <egafford> Fix to remove config groups in Ambari on cluster scaledown. 14:11:08 <egafford> #link https://review.openstack.org/#/c/201617/ 14:11:17 <egafford> And finally Ambari-Hive integration. 14:12:27 <egafford> So if we do have time for reviews, those are good to hit (also any specs at this stage of the cycle, as that'll help larger features get in more reasonably.) 14:12:54 <egafford> #topic API v2 progress 14:13:07 <egafford> elmiko: How's it going on the v2 front? 14:13:13 <elmiko> slowly, but forward! 14:13:32 <egafford> That's half of what we like to hear! 14:13:44 <elmiko> Poornima and pgadiya have been making some progress on tackling the endpoint changes 14:13:55 <elmiko> and iirc Poornima is looking into the job cancel fix next 14:14:11 <elmiko> i still need to write some specs for the larger work 14:14:26 <elmiko> so, overall, slow but steady progress 14:14:33 <Poornima> since Rename job execution and templates endpoints is merged looking into Re factor cancel job process 14:14:56 <egafford> elmiko: Nice. Do you think the specs will likely start review in the N1 timeframe? 14:15:00 <egafford> Poornima: :) 14:15:36 <elmiko> egafford: probably n2 is more likely 14:16:50 <egafford> elmiko: Seems reasonable. N1 is incredibly soon, btw (May 31- June 2). 14:17:12 <elmiko> yea 14:17:19 <egafford> #topic Open discussion 14:17:20 <elmiko> and i have been swamped with internal stuff 14:17:44 <egafford> elmiko: That seems to be the way of most of us who aren't vgridnev. 14:18:00 <elmiko> hehe 14:18:04 <egafford> Anything else to talk about? 14:18:25 <Poornima> I have been facing issues while creating clusters any help / pointers to speed that will be appreciated 14:18:36 <elmiko> i would like to talk about the oozie password fix, but maybe that would be better for when we have more mirantis folks around? 14:18:54 <elmiko> Poornima: what issues? 14:18:54 <egafford> elmiko: Yeah, possible. We can talk about it now, though. 14:19:48 <Poornima> elmiko, the creature creation fails via dashboard 14:19:58 <Poornima> cluster* 14:20:38 <tmckay> Poornima, do you have specific errors from the dashboard or the sahara log? 14:20:43 <egafford> Poornima: What's the error message? (Is it "No valid host was found: not enough hosts available?" It is about 3/4 of the time...) 14:20:53 <elmiko> heh, you guys beat me to it ;P 14:20:54 <tmckay> if debug logging is on for horizon/sahara, there should be something useful there 14:21:13 <tmckay> and is this a setup where a local dev instance of sahara/horizon is pointed at an openstack somewhere else? 14:21:38 <egafford> Poornima: Yeah, sadly the messaging in Horizon itself is way too brief to get a decent diagnosis most of the time. You always have to go to logs. 14:22:00 <Poornima> tmckay, egafford i have using devstack+sahara swpan with stack user 14:22:37 <NikitaKonovalov> Poornima: what is the hardware spec for the machie you are running it on? 14:23:48 <Poornima> NikitaKonovalov, its fedora23 14:23:51 <Poornima> Name m1.xlarge 14:23:51 <Poornima> VCPUs 8 14:23:52 <Poornima> Root Disk 160 GB 14:23:52 <Poornima> Ephemeral Disk 0 GB 14:23:52 <Poornima> Total Disk 160 GB 14:23:52 <Poornima> RAM 16,384 MB 14:24:19 <NikitaKonovalov> ok it looks like enough power to run 3-4 VMs 14:24:40 <egafford> Poornima: Ah, are you trying to run devstack nested in a VM (from the name)? 14:25:21 <Poornima> egafford, yes 14:26:00 <Poornima> Error: name: u'Cluster template' is not a 'valid_name_hostname' Error ID: 27d20577-9a53-4096-9c16-33321f16dc53 14:27:17 <egafford> Poornima: Hm; that certainly seems more like an input validation error than a nested virt issue. 14:28:28 <Poornima> this is the recent error which i encountered may be i need to dig deep 14:28:38 <egafford> Poornima: We definitely want to help, but does anyone have any other topics for the meeting? (Troubleshooting in #openstack-sahara is always an option. :) ) 14:29:03 <Poornima> egafford, sure will come up with more insides :) 14:29:14 <egafford> Poornima: We can take a look at your inputs and see if we can understand. 14:29:45 <Poornima> thanks for the pointers 14:29:58 <elmiko> i'm curious about how we might solve the oozie password issue 14:30:10 <tmckay> oh, you can't use a space 14:30:38 <tmckay> the name of the cluster template has to be something that can be a valid hostname 14:30:43 <elmiko> i'm thinking we will need to store the password in the keystore and keep a reference in the cluster object. but obviously, this will add a bunch to our internal model (i'm assuming we won't expose it through rest) 14:31:12 <tmckay> If the name of the cluster is really "Cluster template" it won't work (unless I'm missing something) 14:31:28 <Poornima> tmckay, so no space 14:31:57 <Poornima> using this guideline http://docs.openstack.org/developer/sahara/horizon/dashboard.user.guide.html 14:36:41 <Poornima> tmckay++ that helped the cluster was created 14:37:09 <egafford> Poornima: Hooray! 14:37:21 <tmckay> yay! 14:37:28 <Poornima> egafford, clap clap :D 14:39:47 <egafford> elmiko: Are you thinking about a generalized keystore reference feature, or something really specific (do you think there's likely to be reuse?) 14:40:12 <elmiko> oh, i just meant the external key manager stuff 14:40:58 <elmiko> like, we store the random oozie password with the key manager, record the id in the cluster object, then use that password when the create oozie mysql script is deployed 14:43:15 <egafford> Right. That doesn't seem like a huge model change, though it does seem really specific. 14:43:34 <elmiko> yea, and we might need to repeat this pattern for other hardcoded pws 14:43:53 <elmiko> i guess this may require a spec to address the fix correctly 14:43:57 <egafford> Yeah, seems likely. 14:44:10 <elmiko> too much work, not enough time ... le sigh 14:44:32 <egafford> "This is a secret that is not personally a top-level field." does seem like it could come up again for sure. 14:45:19 <elmiko> right, maybe some field in the cluster model for "passwords", which doesn't get exposed in the rest 14:45:43 <egafford> Sure sure; and then anything that needed it could just pull its key from the map. 14:46:03 <egafford> It'd be a bit of a doc challenge to ensure folks used the right keys through the CLI, but doable. 14:46:53 <elmiko> it may not even need to be exposed from the cli side. these are mostly internal passwords that are used to setup the services on the nodes. 14:47:05 <elmiko> eventually we probably want to provide a feature for users to get access to those passwords 14:48:43 <egafford> elmiko: Probably, but eventually may be a ways off. 14:48:57 <elmiko> yea... 14:49:17 <egafford> Anything else for this meeting? 14:49:39 <egafford> Going once... 14:50:02 <egafford> Going twice... 14:50:21 <egafford> #endmeeting