15:00:57 <ricolin> #startmeeting heat
15:01:01 <openstack> Meeting started Wed May 24 15:00:57 2017 UTC and is due to finish in 60 minutes.  The chair is ricolin. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:02 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:04 <openstack> The meeting name has been set to 'heat'
15:01:08 <ricolin> #topic roll call
15:01:53 <therve> Hi
15:01:55 <zaneb> greetings
15:02:15 <ramishra> hi
15:02:24 <ricolin> therve, zaneb ramishra hi guys:)
15:03:33 <ricolin> #topic adding items to agenda
15:03:39 <ricolin> #link https://wiki.openstack.org/wiki/Meetings/HeatAgenda#Agenda_.282017-05-24_1500_UTC.29
15:05:02 <ricolin> #topic weekly report
15:05:15 <ricolin> only got six new medium bug this week
15:06:14 <ricolin> also it's closer to feature freeze so would be great if any help on reviewing specs
15:06:14 <ricolin> #link https://review.openstack.org/#/q/status:open+project:openstack/heat-specs
15:07:23 <ramishra> ricolin: Is there a patch in queue to implement that? Approving the spec would probably be of very little use.
15:08:43 <zaneb> can we approve https://review.openstack.org/#/c/267770/ now? the patch is close to landing
15:08:48 <ricolin> ramishra, yeah, agree with you. But do wondering any of those guys try to wait for spec approved before take action
15:09:25 <ricolin> ramishra, maybe we should leave them message to encourage doing implement and spec at some time
15:11:15 <ricolin> #topic seperators encoding https://review.openstack.org/#/c/462216/
15:11:46 <ricolin> this is about ramishra 's ML:)
15:11:48 <ricolin> #link http://lists.openstack.org/pipermail/openstack-dev/2017-May/117355.html
15:12:25 <ramishra> I did not get any solution from the devstack or infra team yet, little surprising
15:13:29 <ramishra> may be someone would look at it soon
15:13:39 <zaneb> ramishra: with the client change does everything work?
15:13:51 <ramishra> zaneb: I've responded to your comments
15:13:55 <ricolin> #link https://review.openstack.org/#/c/463510/
15:14:20 <zaneb> regardless of what else we do, we should stop %-encoding the / between the stack name and stack UUID
15:14:45 <ramishra> zaneb: It works locally for me, but don'te about that route not found error for waitconditions
15:14:52 <zaneb> ramishra: on the ML you mean? because you haven't responded on that review
15:15:09 <ramishra> zaneb: yeah, I responded on the ML
15:15:37 <zaneb> ok cool, I saw that
15:16:13 <therve> ramishra, Maybe the cfn client still does that encoding?
15:16:38 <zaneb> ramishra: I think we're in agreement, because I said the same thing in the very next paragraph after the one you replied to ;)
15:17:02 <ramishra> therve: it does not look like http://logs.openstack.org/16/462216/11/check/gate-heat-dsvm-functional-convg-mysql-lbaasv2-non-apache-ubuntu-xenial/e7d9e90/console.html#_2017-05-20_07_04_30_500696
15:17:30 <ricolin> so https://review.openstack.org/#/c/463510/ sounds like the solution to all, right?:)
15:17:57 <ramishra> I don't really understand why port 80 could not be reached on the host
15:18:15 <ramishra> so all tests where we signal from the vm are failing
15:18:53 <ramishra> You can see the AodhAlarm test passed.
15:19:23 <ramishra> I think autoscaling stuff is also working fine
15:19:38 <LanceHaig> o/
15:19:53 <ramishra> Btw, all the tests failing at the gate work fine for me locally
15:20:00 <ricolin> LanceHaig, o/
15:20:03 <ramishra> so it's surely some infra thing
15:21:58 <ricolin> ramishra, can't figure out too
15:22:52 <ricolin> okay, at least we can land the other fix:)
15:22:58 <ricolin> #topic py3.5 for heat-agents https://review.openstack.org/#/c/425057/
15:23:30 <ricolin> I can propose a gate for py35 in heat-agent
15:23:48 <therve> ramishra, It's failing with could not connect, it doesn't reach the cfn api
15:23:50 <ricolin> just think might be fine to let https://review.openstack.org/#/c/425057/ land first?
15:23:57 <ramishra> therve: yeah
15:24:12 <therve> ramishra, http://10.0.1.78/heat-api-cfn/ looks like a apache endpoint? And it's the non apache job
15:24:27 <ramishra> therve: I'm just testing one job
15:24:58 <ramishra> so I've just used the flag we aleady have HEAT_USE_MOD_WSGI
15:25:13 <ramishra> I'll change that later, once it works:)
15:25:21 <therve> ramishra, Sure, but it doesn't look like the issue is an encoding issue?
15:25:24 * therve is lost
15:26:02 <ramishra> therve: If the flag is not set then we're using apache+uwsgi
15:26:23 <ramishra> therve: yeah, the reamaining issue is not an encoding issue
15:26:25 <therve> ramishra, Which flag?
15:26:40 <therve> You're just setting heat_waitcondition_server_url incorrectly
15:26:40 <ricolin> therve, HEAT_USE_MOD_WSGI
15:27:07 <therve> But the logic in the patch looks backward
15:27:30 <therve> See http://logs.openstack.org/16/462216/11/check/gate-heat-dsvm-functional-orig-mysql-lbaasv2-ubuntu-xenial/e9e5a03/logs/etc/heat/heat.conf.txt.gz
15:27:35 <therve> It should use the apache endpoint
15:27:44 <ramishra> therve: yeah, I just mentioned, I wanted to just test one job
15:28:05 <ramishra> so the non-apache job is using apache+uwsgi in the last patch
15:28:12 <ramishra> I'll change that once it works
15:28:43 <therve> ramishra, It's not using apache? There is a cfn screen log
15:29:11 <ramishra> therve: yes, it's using systemd services+apache
15:29:18 <ramishra> and the proxy_pass
15:30:10 <ramishra> http://logs.openstack.org/16/462216/11/check/gate-heat-dsvm-functional-convg-mysql-lbaasv2-non-apache-ubuntu-xenial/e7d9e90/logs/apache_config/heat-wsgi-api.conf.txt.gz
15:30:40 <therve> OK but
15:30:43 <therve> ramishra, https://review.openstack.org/gitweb?p=openstack/heat.git;a=blob;f=devstack/lib/heat;h=bbbb9bab54e201772b3c813b94942984d5a1da8f;hb=539c24231570dc0031399fc6bd6e81eaa028a132#l132
15:30:48 <therve> That logic is backward
15:31:19 <ramishra> therve: yes, I just wanted to make it work first
15:31:25 <therve> We should connect to /heat-api-cfn when MOD_WSGI si true, not the other way around
15:31:45 <therve> Make what work? :)
15:31:51 <therve> OK I'll shut up
15:32:43 <ramishra> mod_wsgi and mod_proxy_uwsgi are different
15:33:10 <ramishra> so I'll add a new flag USE_UWSGI
15:33:52 <ramishra> so we've options 1. no apache 2. apache_mod_wsgi 3. apache+uwsgi(with mod_proxy_uwsgi)
15:34:10 <ramishra> We'll have different jobs for all three later
15:34:40 <ramishra> Does that help clear up the confusion?
15:35:23 <ramishra> therve: ?
15:35:30 <zaneb> let's say that it does and move on ;)
15:36:15 <ricolin> ramishra, therve said he will shut up:)
15:36:19 <ricolin> lol
15:36:37 <ricolin> so back to py35
15:37:20 <ricolin> Is it fine to let https://review.openstack.org/#/c/425057/ land first? And I will change the gate after
15:38:16 <ricolin> or we can insist to put the gate in first, and it's going to be a non voting period
15:38:56 <zaneb> I think it's fine to land it
15:39:22 <ricolin> cool
15:39:43 <ramishra> therve: I think it would be something like this https://review.openstack.org/#/c/462216/4/devstack/lib/heat
15:40:13 <ramishra> ricolin: sorry for continuing with the earlier topic
15:40:59 <ricolin> ramishra, np! we got to solve that issue:)
15:41:56 <zaneb> doesn't the actual encoding of stdout/stderr depend on the locale?
15:44:29 <therve> Yes
15:45:31 <zaneb> it seems like that decoding will be a possible source of exceptions
15:48:50 <ramishra> zaneb: are we stilltalking about the earlier topic?
15:49:04 <ricolin> ramishra, yes:)
15:49:30 <zaneb> ramishra: we're talking about https://review.openstack.org/#/c/425057/6 (py3.5)
15:49:41 <ramishra> zaneb: ok
15:49:57 <zaneb> I added a comment to the review
15:50:26 <zaneb> ricolin: next topic?
15:50:32 <zaneb> 10 mins left
15:50:44 <ricolin> #topic heat-template future
15:51:07 <ricolin> #link http://lists.openstack.org/pipermail/openstack-dev/2017-May/117080.html
15:51:42 <ricolin> first would like to see the fake client landing
15:51:43 <ricolin> #link https://review.openstack.org/#/c/425057/6
15:52:03 <ricolin> therve, also can we also add magnum in fake client?:)
15:52:43 <ricolin> wait wrong link
15:52:49 <therve> ricolin, Are there magnum templates?
15:53:04 <ricolin> #link https://review.openstack.org/#/c/465860/
15:53:29 <ricolin> #link https://review.openstack.org/#/c/455673/
15:53:43 <ricolin> not in current repo
15:53:58 <ricolin> but a patch for it:)
15:54:13 <therve> OK
15:54:33 <ricolin> also LanceHaig got some thread in ML
15:54:34 <ricolin> #link http://lists.openstack.org/pipermail/openstack-dev/2017-May/117080.html
15:55:00 <LanceHaig> Can someone explain to me how this fake client would work?
15:55:04 <ricolin> about some cool suggestion for heat
15:55:07 <LanceHaig> Or do w ego that on the list rather?
15:55:38 <LanceHaig> sorry for the typo's it has been a long day
15:55:54 <therve> LanceHaig, Btw, you can put openstack version names in template versions
15:56:01 <therve> Forgot to answer that
15:56:25 <LanceHaig> therve, thanks.
15:57:07 <ricolin> LanceHaig, I like what we have in heat-tutorial, would be greate to make those in heat repo. Also you might consider to use os-collect-config to replace cloud-init in tutorial, since that's more recommanded to do so.
15:57:38 <therve> Hum
15:57:52 <therve> If things work with cloud init, not sure there is a reason to use os collect config
15:58:04 <therve> occ requires a custom image, whereas cloud init is everywhere
15:58:24 <LanceHaig> what therve said
15:58:49 <LanceHaig> I am wanting to keep the examples as easy to use for all heat versions
15:59:09 <LanceHaig> Not everyone can run a new version of OS and some still have very old versions
15:59:59 <ricolin> therve, sure, but like Magnum team, they might facing metadata size limit though
16:00:05 <LanceHaig> I know if does not make sense to support old versions but at least we can get a set of working templates for the older versions and then mark it as unsupported so that they have something to work with
16:00:14 <LanceHaig> I am happy to put in the work to get this done
16:00:21 <LanceHaig> I just want to see that everyone agrees
16:00:27 <ricolin> LanceHaig, therve and keep it cloud-init might actually better:)
16:01:00 <ricolin> just thinking about the tutorial only
16:01:09 <LanceHaig> right
16:01:12 <LanceHaig> sorry
16:01:46 <therve> Alright, EOM
16:01:53 <ricolin> LanceHaig, I think we can make a set for that:)
16:01:58 <ricolin> therve, right!
16:02:10 <ricolin> shall we move back to heat:)
16:02:21 <LanceHaig> yes
16:02:22 <ricolin> #endmeeting