14:00:44 <tosky> #startmeeting sahara
14:00:45 <openstack> Meeting started Thu May 17 14:00:44 2018 UTC and is due to finish in 60 minutes.  The chair is tosky. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:46 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:49 <openstack> The meeting name has been set to 'sahara'
14:01:40 <jeremyfreudberg> o/
14:01:44 <jeremyfreudberg> not sure who else is here
14:02:24 <tosky> probably just us, but we can quickly sync
14:02:31 <tosky> at most, 45 minutes back
14:02:43 <tosky> #topic News/Updates
14:04:01 <tosky> you may have noticed a big rush of patches around the end of the weeks
14:04:03 <tosky> week*
14:04:21 <tosky> basically many regressions discovered in live testing for the image generation, backported to queens too
14:04:29 <tosky> we have a (more stable) 8.0.1 queens release
14:05:08 <jeremyfreudberg> yes, i saw that, thanks
14:05:14 <tosky> I worked on enabling wsgi by default, it was easier than expected, there are few patches pending
14:06:28 <jeremyfreudberg> meanwhile i have just been trying to get back in the saddle, i have a lot of halfway complete patches locally
14:06:35 <tosky> oh, nice to hear
14:07:05 <tosky> I've seen the patch a new version of the admin manage all, but given the WIP I didn't check it in depth
14:07:12 <tosky> apart from that, most of the patches in the queue should be either immediately meargable or require work
14:07:42 <jeremyfreudberg> it's not worth reviewing quite yet, we should discuss a bit about the next steps for it though
14:07:49 <jeremyfreudberg> and i'll take a look at the review queue
14:07:51 <tosky> I promise to recheck zchkun's patches (sorry!) more quickly; I was relying on the "Jeremy didn't look at it yet, I can wait" but that's not an excuse :)
14:08:43 <tosky> oh, I'm going to work on S3 verification, at least basic one, extending the existing tempest tests for job binaries and data sources
14:09:06 <jeremyfreudberg> tosky: that would be very nice
14:09:20 <tosky> there are two ways to deploy S3: with ceph, but I just discovered that no one tested radosgw with devstack-plugin-ceph; I'm working to try to fix it
14:09:46 <jeremyfreudberg> yes, i've seen devstack-plugin-ceph, it's a bit unmantained
14:09:49 <tosky> the other way (discovered today) is the usage of swift3, which is being merged into swift as s3api, and I may try that in the basic job to start the process
14:10:00 <tosky> but I'd like to have an alternative job with ceph too
14:10:21 <tosky> so let's see which one lands first
14:10:38 <jeremyfreudberg> there's definitely some value to the community if you put some hours into the ceph thing
14:10:42 <tosky> yep
14:10:51 <jeremyfreudberg> so even if swift3 is "easier", it's worth your digging around with ceph
14:11:12 <tosky> I completely agree
14:11:30 <tosky> I think it's all from my side; Telles is on PTO, and he will attend the Summit next week, so there could be some more news afterwards
14:11:49 <tosky> (if he meets people from community, etc etc, I always hope)
14:12:29 <jeremyfreudberg> that would be nice
14:12:36 <jeremyfreudberg> hi zchkun
14:12:53 <tosky> hi zchkun - I just wrote that I need to review your patches faster :)
14:13:03 <tosky> they are not forgotten
14:13:14 <zchkun> sorry, something wrong with my software
14:13:25 <jeremyfreudberg> zchkun: no problem
14:13:27 <zchkun> OK
14:13:40 <jeremyfreudberg> let's change the topic quickly to hadoop 3.0.x version
14:14:03 <zchkun> ok
14:14:09 <jeremyfreudberg> (tosky: i think the chair has to change it)
14:14:23 <tosky> #topic Hadoop 3.0.x
14:14:40 <tosky> jeremyfreudberg, zchkun: here you are
14:15:02 <tosky> so, there is an initial patch for sahara-extra
14:15:03 <zchkun> I select hadoop 3.0.0, and have completed the vanilla 3.0.0 upgrade test
14:16:07 <jeremyfreudberg> right, thanks for your work zchkun. i am looking at the release notes for 3.0.1, i have a question for you based on your experience with 3.0.0
14:16:12 <jeremyfreudberg> http://hadoop.apache.org/docs/r3.0.1/hadoop-project-dist/hadoop-common/release/3.0.1/RELEASENOTES.3.0.1.html
14:16:19 <jeremyfreudberg> HDFS-12990 | Critical | Change default NameNode RPC port back to 8020
14:16:43 <jeremyfreudberg> that was a bug from 3.0.0 fixed in 3.0.1 (or actually, it was a reverting of a feature)
14:16:51 <jeremyfreudberg> do you have any thoughts on it?
14:18:13 <zchkun> sorry, I fail to notice that
14:19:56 <jeremyfreudberg> zchkun: i don't think it is a big deal, we can look at how often we reference port 8020 in the sahara code
14:19:57 <zchkun> should we upgrade to 3.0.1?
14:20:18 <tosky> well, if we start fresh, maybe it's better to get the last version, if there are no known issues
14:20:27 <jeremyfreudberg> 3.0.1 may help to simplify the code, compared with 3.0.0
14:22:06 <tosky> so maybe 3.0.1 ?
14:22:37 <zchkun> no problem
14:22:43 <tosky> oki
14:22:49 <tosky> anything else about this point?
14:23:21 <jeremyfreudberg> nope, thanks zchkun
14:23:43 <tosky> #topic Open Discussion
14:23:51 <tosky> anything else to discuss today?
14:24:27 <tosky> remember that the meeting week is cancelled (Telles will attend the Summit, I will be on my own PTO)
14:25:20 <jeremyfreudberg> about the "admin manage all" feature, i do have a point to discuss
14:25:39 <tosky> oki, do you want a special topic or can we go quickly now?
14:25:49 <jeremyfreudberg> let's keep it in open discussion
14:26:09 <tosky> shoot
14:26:36 <jeremyfreudberg> anyway, right now it's pretty much magical, if you are admin, then you see a lot more of results in your GET requests, automatically
14:26:51 <jeremyfreudberg> i'm wondering if a toggle is needed
14:27:12 <jeremyfreudberg> perhaps in sahara config, to turn the feature off entirely
14:27:20 <jeremyfreudberg> or maybe something new in the api, to enable/disable per call
14:27:27 <tosky> to get all results?
14:27:30 <tosky> I may have said it already, but I would try to align on what other services do
14:27:47 <jeremyfreudberg> right, in nova it is explicit, in neutron it is implicit
14:27:58 <jeremyfreudberg> i think in heat it may be implicit as well
14:28:27 <tosky> request to the TC? Maybe there is some guidelines
14:29:20 <jeremyfreudberg> i can ask el miko (trying not to ping him right this second), as an api working group person, too
14:29:40 <tosky> oh, yes, the API working group is probably the best place
14:29:48 <jeremyfreudberg> but does anyone here have any thoughts? implicit vs explicit?
14:30:34 <tosky> uhm, I don't really know right now. I suspect it's one of those things were "right" or "wrong" depends on the person you ask to
14:32:02 <jeremyfreudberg> right, well i will ask around
14:32:35 <tosky> thanks!
14:32:50 <jeremyfreudberg> some other quick questions, about deprecation
14:32:54 <tosky> sure
14:33:07 <jeremyfreudberg> [keystone_authtoken] - is it time to remove yet?
14:34:08 <tosky> uh, we don't have the support neither in puppet nor in openstack-ansible-os_sahara; I should try to implement something at least for the latter (which is a bit easier)
14:34:13 <tosky> but always tons of other stuff to do
14:34:39 <jeremyfreudberg> tosky: oh, right, i forgot about puppet/ansible
14:35:48 <jeremyfreudberg> but in terms of time, i think we meet rule 4 here https://governance.openstack.org/tc/reference/tags/assert_follows-standard-deprecation.html#requirements no?
14:36:17 <jeremyfreudberg> https://review.openstack.org/#/c/524936/ was the patch
14:39:12 <tosky> jeremyfreudberg: yes, the minimum time is fine, we could remove it in rocky
14:39:43 <jeremyfreudberg> tosky: ok, great
14:39:53 <tosky> I think it's worth at least to try to have the deployment done correctly; luckily there is still time to think about it and cut it before rocky, if needed
14:40:13 <jeremyfreudberg> yup
14:41:49 <jeremyfreudberg> alright, that's all from me
14:42:15 <tosky> I guess we can close then
14:42:27 <tosky> thanks all, see you on the usual channel
14:42:51 * tosky sets a 3 minute last-chance alarm
14:46:52 <tosky> #endmeeting