08:03:20 <therve> #startmeeting heat
08:03:21 <openstack> Meeting started Wed Sep 21 08:03:20 2016 UTC and is due to finish in 60 minutes.  The chair is therve. Information about MeetBot at http://wiki.debian.org/MeetBot.
08:03:22 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
08:03:24 <openstack> The meeting name has been set to 'heat'
08:03:29 <therve> #topic Roll call
08:03:33 <ramishra> hi
08:03:37 <stevebaker> \o
08:03:46 <ricolin> o/
08:04:33 <huangtianhua> hi
08:04:51 <duvarenkov> hi
08:05:30 <elynn> o/
08:06:25 <therve> #topic Adding items to agenda
08:06:32 <therve> #link https://wiki.openstack.org/wiki/Meetings/HeatAgenda#Agenda_.282016-09-21_0800_UTC.29
08:06:38 <skraynev> 0/
08:07:25 <therve> #topic RC2 status
08:07:36 <therve> #link https://review.openstack.org/#/q/status:open+project:openstack/heat+branch:stable/newton
08:07:47 <therve> This is some remaining backports for RC2
08:08:24 <therve> Do we have any master patches that still need to go in?
08:08:52 <stevebaker> I'm not aware of any that haven't been proposed
08:08:59 <stevebaker> to stable
08:09:25 <stevebaker> like these https://review.openstack.org/#/c/373615/ https://review.openstack.org/#/c/373518/
08:10:21 <therve> I'm interested in https://review.openstack.org/#/c/372804/
08:10:32 <therve> If someone could have a look that'd be great
08:11:44 <therve> No other request?
08:12:07 <stevebaker> I've just proposed a python-heatclient-1.5.0, for tripleo
08:12:15 <therve> OK cool
08:12:56 <therve> It looks like we're in okay shape then. We can propose rc2 early next week.
08:13:08 <therve> #topic Gate status
08:13:23 <therve> ramishra, Let me guess: neutron breakage
08:13:37 <ramishra> It's been ok since a few hours back.
08:13:51 <ramishra> therve: actually openstacksdk change.
08:14:01 <ramishra> I mean new version
08:14:18 <therve> ramishra, Is https://review.openstack.org/#/c/373850/ necessary?
08:14:33 <stevebaker> I didn't even realise openstackclient was using openstacksdk
08:14:50 <ramishra> therve: getting a new osc release would take time I think
08:15:02 <ramishra> if we are ok with it, then we can wait for it.
08:15:06 <therve> stevebaker, turtles all the way down
08:15:19 <therve> Let's just approve that
08:15:23 <stevebaker> I mean, I'm not against it
08:15:49 <therve> I don't give a crap if it works :)
08:16:09 <therve> Anyway
08:16:20 <therve> Regarding the gate, it feels we had a couple of good days
08:16:37 <therve> But intermittent failures got back it seems
08:16:45 <therve> Maybe it was just infra being in slightly better shape
08:17:35 <therve> Moving on
08:17:37 <ramishra> I've raised a new bug for SDG test, stevebaker if you have some clue what's going on.
08:17:42 <ramishra> https://bugs.launchpad.net/heat/+bug/1625921
08:17:43 <openstack> Launchpad bug 1625921 in heat "SoftwareDeploymentGroupTest fails at times with TimeoutException" [Medium,New]
08:18:12 <stevebaker> ramishra: OK, I'll take a look tomorrow unless you get there first
08:18:43 <therve> I'll have a quick look too
08:18:48 <ramishra> stevebaker: I had a look this morning. No idea why signals to resources are ignored.
08:18:51 <ramishra> no errors
08:19:10 <therve> #topic Quota validation spec
08:19:19 <therve> #link https://review.openstack.org/#/c/371910/
08:19:44 <therve> duvarenkov, for you I suppose
08:19:55 <duvarenkov> yeah
08:20:18 <ricolin_afk> duvarenkov: for cinder quota, we already have a version of quota validation merged
08:20:20 <therve> So like all validation, I'm concerned about duplicating info from other services
08:20:28 <therve> How do you retrieve quota info?
08:20:43 <ricolin_afk> duvarenkov: is that that what you refering?
08:21:01 <duvarenkov> I'll look about cinder quota validation
08:21:10 <duvarenkov> Didnt know about it
08:21:51 <duvarenkov> therve, I suppose, from service client
08:22:28 <therve> ricolin_afk, Doesn't that just validation quota update?
08:22:39 <therve> I mean, does it check we create a volume for example?
08:22:50 <ricolin_afk> therve: yes it is
08:23:19 <prazumovsky> ricolin_afk: could you share, where we have quota validation? Or you mean quota resource validation? In that case, no, it's like separate pre-create validation, which calculate, is stack exceeds quota or not
08:23:44 <ramishra> IMO, the idea is good, use the quota api from services to check requirement vs availability during validation.
08:24:00 <ricolin_afk> prazumovsky: I only mean a quota validation for cinder::quota resource
08:24:34 <therve> ricolin_afk, Right, here we're talking about checking quota when you create a resource
08:24:34 <ricolin_afk> that's why I'm asking what we're looking at for quota
08:24:39 <duvarenkov> <ricolin_afk>: No, thats not the idea of the spec
08:24:53 <duvarenkov> Its whole stack validation, not just one resource
08:24:53 <shardy> does pre-create validation of quota really mean anything?  You've not got any way to know if the same quota exists when the stack is actually instantiated?
08:25:16 <shardy> seems like a runtime thing to me, so you can fail-fast before starting the stack create
08:25:16 <ricolin_afk> duvarenkov: got it
08:25:24 <ramishra> shardy: +1, good pint
08:26:05 <therve> Yeah, it's highly racy
08:26:24 <duvarenkov> Don't quite understand. Some quota is always existent.
08:26:37 <duvarenkov> By cinder quota resource we can just change it
08:26:40 <therve> duvarenkov, Once you got the quota, it may have already changed
08:26:49 <therve> There is no transactional information
08:27:19 <therve> You may prevent from creating stacks that would succeed, and let validate stacks that will fail
08:27:37 <shardy> Even on create it's racy, if you create multiple stacks concurrently
08:27:49 <shardy> because quota doesn't allow for reservations
08:28:02 <duvarenkov> Yes, I think you are right
08:28:22 <shardy> so you could argue it's best to just to even try to validate, and just let the last unlucky stack fail when the quota is exceeded
08:28:31 <ricolin_afk> shardy: agree
08:28:32 <shardy> not even try to validate I mean
08:28:53 <therve> It may be interesting for preview
08:29:06 <shardy> yeah, perhaps for preview it could be useful
08:29:14 <duvarenkov> Can you, please, share your thoughts in comments for review? We'll think about pros and cons
08:29:22 <shardy> most of the preview results aren't accurate anyway ;)
08:29:32 <therve> Right :)
08:29:52 <therve> OK, moving on
08:29:57 <therve> #topic Open discussion
08:30:24 <shardy> I had one question, following on from the validation topic above
08:30:32 <therve> Shoot
08:30:58 <shardy> Has anyone given any further thought to how we solve the problem of nested validation, where a required parameter in a nested template can't know if any value will/would be passed from the parent?
08:31:07 <shardy> ramishra: I remember we talked about it fairly recently
08:31:24 <shardy> I'm not sure if there's a bug, but it was a known issue from when I worked on the implementation with jdob
08:31:31 <shardy> and was somewhat discussed in austin:
08:31:36 <shardy> https://etherpad.openstack.org/p/heat-newton-validation-improvements
08:31:41 <ricolin_afk> shardy: for https://bugs.launchpad.net/heat/+bug/1559807
08:31:43 <openstack> Launchpad bug 1559807 in heat "Some intrinsic functions fail when referencing get_attr" [Medium,Confirmed] - Assigned to Steven Hardy (shardy)
08:31:51 <shardy> It's become something of a headache for TripleO
08:31:52 <ricolin_afk> shardy: I'm target about it
08:32:02 <ricolin_afk> shardy: include nested or not
08:32:28 <ramishra> shardy: I had the impression that someone is working on it after the last summit discussion.
08:32:42 <therve> shardy, I don't remember talking about that specific issue during the session
08:32:45 <ricolin_afk> shardy: and for nested stack, I have some poc for considering centerlize validate for resource group
08:33:37 <shardy> ricolin_afk: Yeah, it's quite closely related to that
08:33:38 <ricolin_afk> shardy: Is that what you needed?
08:33:50 <ramishra> therve: I aslso assumed it's part of the overall validation improvement topic.
08:33:55 <shardy> therve: basically my problem will be fixed when we pass around a placeholder object, instead of None during validation
08:34:01 <shardy> so I think it was indirectly discussed
08:34:05 <therve> Ah, okay
08:34:12 <shardy> I'll check the bugs and raise one specific to validation tho
08:35:03 <therve> That'd be cool, not sure I understand the specifics of your problem
08:35:43 <therve> I don't think anyone started working on the placeholder thing
08:35:53 <ricolin_afk> We simply seperate validate and real creation by passing place holder
08:36:02 <shardy> therve: ack, that's basically what I wanted to confirm
08:36:04 <therve> "simply" :)
08:36:16 <shardy> be good to figure out a way to fix that for Ocata
08:36:34 <therve> I think we figured out the way, just need to push the idea
08:36:55 <therve> Hopefully once release craziness settles down we can find some time
08:37:02 <shardy> +1
08:37:04 <shardy> thanks
08:37:29 <therve> ramishra, BTW, when do you take over? :)
08:37:46 <stevebaker> ramishra: congrats by the way
08:37:51 <ramishra> therve: you can continue as long as you want:)
08:37:57 <shardy> +1 thanks for volunteering ramishra
08:38:08 <therve> ramishra, So yesterday? :p
08:38:14 <ramishra> thanks shardy stevebaker
08:38:22 <therve> But yeah congrats!
08:38:39 <therve> I think I'll do next meeting, and rc2, and give you the keys to the castle
08:38:46 <ramishra> I think this is the election week, once the results are out;)
08:39:26 <therve> Sounds good :)
08:39:29 <therve> Anything else?
08:40:25 <therve> Alright, thanks all!
08:40:27 <therve> #endmeeting heat