| *** bobh has joined #openstack-mistral | 00:02 | |
| *** jkilpatr_ has quit IRC | 00:21 | |
| *** jkilpatr_ has joined #openstack-mistral | 00:31 | |
| *** zhurong has joined #openstack-mistral | 00:39 | |
| *** jamielennox is now known as jamielennox|away | 02:07 | |
| *** gongysh has joined #openstack-mistral | 02:09 | |
| *** Kevin_Zheng has joined #openstack-mistral | 02:54 | |
| *** chlong has joined #openstack-mistral | 03:02 | |
| *** chlong_ has joined #openstack-mistral | 03:03 | |
| *** bobh has quit IRC | 03:11 | |
| *** gongysh has quit IRC | 03:21 | |
| *** jamielennox|away is now known as jamielennox | 03:23 | |
| *** gongysh has joined #openstack-mistral | 03:35 | |
| *** zhurong has quit IRC | 04:06 | |
| *** zhurong has joined #openstack-mistral | 04:34 | |
| openstackgerrit | Merged openstack/mistral-dashboard master: Updated from global requirements https://review.openstack.org/455937 | 04:56 |
|---|---|---|
| openstackgerrit | Merged openstack/mistral-lib master: Updated from global requirements https://review.openstack.org/455938 | 04:56 |
| *** chlong_ has quit IRC | 05:03 | |
| *** chlong has quit IRC | 05:03 | |
| *** jtomasek_ has joined #openstack-mistral | 05:04 | |
| *** sharatss has joined #openstack-mistral | 05:08 | |
| *** warface has joined #openstack-mistral | 05:10 | |
| *** jaosorior_away is now known as jaosorior | 05:13 | |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Change workflow lang schema to support advanced publishing https://review.openstack.org/455178 | 05:14 |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Change workflow lang schema to support advanced publishing https://review.openstack.org/455178 | 05:15 |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Advanced publishing: change workflow lang schema https://review.openstack.org/455178 | 05:33 |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Advanced publishing: add publishing of global variables https://review.openstack.org/456056 | 05:33 |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Advanced publishing: add publishing of global variables https://review.openstack.org/456056 | 05:35 |
| *** jkilpatr_ has quit IRC | 05:35 | |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Advanced publishing: add 'global' function to access global vars https://review.openstack.org/456059 | 05:43 |
| *** jtomasek_ has quit IRC | 05:44 | |
| *** mgershen1 has joined #openstack-mistral | 05:46 | |
| *** jkilpatr_ has joined #openstack-mistral | 05:46 | |
| *** mgershen has quit IRC | 05:46 | |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Advanced publishing: change workflow lang schema https://review.openstack.org/455178 | 05:47 |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Advanced publishing: add publishing of global variables https://review.openstack.org/456056 | 05:47 |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Advanced publishing: add 'global' function to access global vars https://review.openstack.org/456059 | 05:47 |
| *** jtomasek has quit IRC | 05:49 | |
| openstackgerrit | Merged openstack/python-mistralclient master: Updated from global requirements https://review.openstack.org/455998 | 06:04 |
| *** mgershen has joined #openstack-mistral | 06:15 | |
| *** mgershen1 has quit IRC | 06:17 | |
| *** mgershen1 has joined #openstack-mistral | 06:31 | |
| *** mgershen has quit IRC | 06:33 | |
| openstackgerrit | Merged openstack/mistral master: Updated from global requirements https://review.openstack.org/455936 | 06:45 |
| *** mgershen has joined #openstack-mistral | 06:47 | |
| *** mgershen1 has quit IRC | 06:48 | |
| *** brunograz has quit IRC | 06:54 | |
| *** jistr has quit IRC | 06:54 | |
| *** brunograz has joined #openstack-mistral | 06:55 | |
| *** jistr has joined #openstack-mistral | 06:56 | |
| *** mgershen1 has joined #openstack-mistral | 07:02 | |
| *** d0ugal has joined #openstack-mistral | 07:03 | |
| *** mgershen has quit IRC | 07:03 | |
| *** jpich has joined #openstack-mistral | 07:13 | |
| rakhmerov | d0ugal: hi, just reminding you about "engine commands" patch | 07:16 |
| rakhmerov | :) | 07:16 |
| *** mgershen1 has quit IRC | 07:19 | |
| *** mgershen has joined #openstack-mistral | 07:20 | |
| jaosorior | rakhmerov: Has there been any work done on the keystone bits you mentioned the other day? | 07:21 |
| sharatss | jaosorior, hi | 07:23 |
| jaosorior | sharatss: hey | 07:23 |
| sharatss | jaosorior, i have been looking into it from morning | 07:23 |
| *** shardy has joined #openstack-mistral | 07:24 | |
| jaosorior | sharatss: it honestly wasn't entirely clear to me what folks are supposed to be using instead of keystoneclient's Client class. I understand why folks should be using keystoneauth1's auth plugins and sessions. But the client? | 07:24 |
| jamielennox | jaosorior: the client is still the client and keystoneauth's sesssions/plugins work with it | 07:25 |
| jamielennox | just for most things you don't actually need to talk to keystone CRUD, just the auth bits | 07:25 |
| jaosorior | riiight | 07:25 |
| jaosorior | jamielennox: I see that it's only really being used to get the service catalog | 07:27 |
| jaosorior | jamielennox: how would you get the service catalog (in order to get mistral's URL) ? | 07:28 |
| jaosorior | oh wait, wouldn't that be using discovery? | 07:28 |
| jamielennox | jaosorior: the idea is that you shouldn't need to interact with the service catalog directly | 07:28 |
| jamielennox | jaosorior: is this for porting mistralclient in general? | 07:28 |
| jaosorior | jamielennox: yeah | 07:29 |
| jamielennox | ok | 07:29 |
| sharatss | jamielennox, jaosorior can refer https://github.com/openstack/python-zaqarclient/blob/master/zaqarclient/auth/keystone.py to get some idea | 07:29 |
| jamielennox | so at the session level when you request() there is an endpoint_filter parameter on which you set things like {'interface': 'public', 'service_type': 'mistral_type', 'version': (2,0)} | 07:30 |
| jamielennox | resp session.get('/path/to/resource', endpoint_filter={..}) | 07:30 |
| jamielennox | resp = session.get('/path/to/resource', endpoint_filter={..}) | 07:30 |
| jamielennox | so that you don't have to do that every time typically when you create the client object you create an adapter | 07:30 |
| jamielennox | https://github.com/openstack/keystoneauth/blob/master/keystoneauth1/adapter.py#L69 | 07:31 |
| jamielennox | typically you pass everything in __init__(self, **kwargs) to Adapter(**kwargs) | 07:32 |
| jamielennox | and for set the service_type | 07:32 |
| jamielennox | and force set the service_type | 07:32 |
| jaosorior | jamielennox: wouldn't the usage of Adapter tie the client to always use keystone? | 07:32 |
| jamielennox | no | 07:32 |
| jaosorior | jamielennox: they have a requirement that it shouldn't be tied to it | 07:32 |
| rakhmerov | jaosorior: didn't read all yet but what I meant is only "we shouldn't use keystoneclient for authentication", this is a deprecated routine as far as I understand | 07:33 |
| jamielennox | when you create the Adapter you pass things like (interface='public', service_type='mistral_type') | 07:33 |
| rakhmerov | for other things client, of course, should be used | 07:33 |
| jamielennox | that way when you do adapter.get('/path/to/resource') you don't need to specify that endpoint_filter | 07:33 |
| jaosorior | that sounds handy | 07:34 |
| jamielennox | so a session object is global, it is shared amongst a whole bunch of clients | 07:34 |
| *** jkilpatr_ has quit IRC | 07:34 | |
| jamielennox | an adapter is like one clients view, the things that only relate to that client instance | 07:34 |
| *** jkilpatr_ has joined #openstack-mistral | 07:35 | |
| jamielennox | so typically yea, you just create an adapter in Client.__init__ and then you can just do adapter.get('/path') and ignore all the auth and discovery stuff | 07:35 |
| rakhmerov | ok | 07:36 |
| jamielennox | rakhmerov: yea, keystoneclient should be only used for actually doing keystone CRUD, listing projects, creating users etc | 07:36 |
| rakhmerov | yep | 07:36 |
| jamielennox | for auth keystoneauth is sufficient | 07:36 |
| jamielennox | also keystoneauth (despite its name) doesn't tie you only to keystone | 07:37 |
| jamielennox | for example... | 07:37 |
| jamielennox | https://github.com/openstack/python-swiftclient/blob/master/swiftclient/authv1.py#L214 | 07:37 |
| rakhmerov | I guess this is the relevant doc to use: https://docs.openstack.org/developer/keystoneauth/index.html | 07:37 |
| jamielennox | to support swift v1 auth they made an (unecessarily complex) keystoneauth plugin | 07:38 |
| jamielennox | that doesn't talk to keystone, it just implements the keystoneauth1.BasePlugin interface | 07:38 |
| rakhmerov | yes, as far as I remember there are also some classes for OAuth2 and OpenID Connect | 07:39 |
| rakhmerov | that have nothing to do with keystone | 07:39 |
| jamielennox | well, typically those get a keystone token from oauth/oidc, but if you want to implement something outside that you can | 07:40 |
| jamielennox | and there's no reason you would have to get those plugins into keystoneauth if they only relate to mistral | 07:40 |
| rakhmerov | jamielennox: ooh, I guess they are designed with the goal to be used in federation | 07:41 |
| jamielennox | yep | 07:41 |
| rakhmerov | ok, got it | 07:42 |
| *** jkilpatr_ has quit IRC | 07:43 | |
| jamielennox | eeek, don't copy zaqarclient | 07:43 |
| *** jkilpatr_ has joined #openstack-mistral | 07:43 | |
| sharatss | jamielennox, zaqarclient does not use the generic plugins | 07:43 |
| jamielennox | sharatss: so i'm not sure how _get_keystone_session is invoked and there's always backwards compatibility to think of - but a large point of the auth plugins is so that services like zaqar don't need to handle every keystone auth paramater | 07:45 |
| jamielennox | if tomorrow keystone adds a new auth type, or new option to an existing type - it gets added once to the plugin and then all the clients can benefit from it | 07:45 |
| jamielennox | rather than what looks like zaqar still manually handling what type of auth it thinks you want | 07:46 |
| jamielennox | i'm going to assume that's just all backwards compatibility stuff | 07:46 |
| sharatss | jamielennox, yeah.. got it | 07:47 |
| rakhmerov | yes :) | 07:47 |
| jamielennox | but ideally zaqar would have replaced zaqarclient.auth.base with keystoneauth | 07:47 |
| rakhmerov | we do it pretty horribly now too | 07:48 |
| jamielennox | right, and that's part of the problem - everyone has so much backwards compat to handle that there's not even a really good clean implementation i can point you to | 07:48 |
| openstackgerrit | Merged openstack/mistral master: Add release note for resource RBAC feature https://review.openstack.org/454945 | 07:52 |
| rakhmerov | well, we're actually planning API v3 and at some point it'll be created (current plan is 2017) so for now maybe we could make some minimal changes in the client code just to be safe if keystone decides to get rid of deprecated stuff | 07:53 |
| openstackgerrit | Merged openstack/mistral master: Add Apache License Content in index.rst https://review.openstack.org/455271 | 07:53 |
| rakhmerov | and then in the new API we could fully use plugins and make it more flexible etc. w/o having to duplicate all params in our client | 07:53 |
| rakhmerov | sharatss, jamielennox: ^ not sure if it makes sense for you | 07:54 |
| rakhmerov | d0ugal: do we still need https://review.openstack.org/#/c/449215/ ? | 07:54 |
| rakhmerov | or can abandon it? | 07:54 |
| jamielennox | rakhmerov: i'm not sure what you mean on the API side, keystoneauth is completely client side | 07:55 |
| d0ugal | rakhmerov: I am still using it - why do you ask? that is a tripleo patch :) | 07:55 |
| rakhmerov | d0ugal: ooh, really? | 07:55 |
| rakhmerov | gosh... | 07:55 |
| rakhmerov | yes, sorry! | 07:55 |
| d0ugal | lol | 07:56 |
| rakhmerov | I was just going over your patches in gerrit to see if there's something that needs my attention | 07:56 |
| rakhmerov | :) | 07:56 |
| d0ugal | aha, thanks | 07:56 |
| d0ugal | I don't think there is anything | 07:56 |
| rakhmerov | jamielennox: I was talking about Mistral Client. Now we also pass huge number of auth params directly into Mistral client whereas it's not a good thing to do for the reasons you mentioned | 07:57 |
| rakhmerov | it needs to be offloaded to auth plugins as far as I understand | 07:57 |
| d0ugal | rakhmerov: but we could probably abandon some of the Mistral patches from 2015 and 2016 if you want to clean up ;) | 07:57 |
| jamielennox | rakhmerov: so the way i've handle both in the past is to check for the session param | 07:57 |
| jamielennox | rakhmerov: if session is present users are opting in to using complete session handling and everything else is ignore/invalid | 07:58 |
| rakhmerov | but we now can't get rid of that easily because use backwards compatibility | 07:58 |
| jamielennox | if session is not present you fall back to a deprecation warning and do it the old way | 07:58 |
| rakhmerov | so in our new client we can do it | 07:58 |
| jamielennox | well, more you fall back to a deprecation warning and create the session for htem | 07:58 |
| rakhmerov | ok | 07:59 |
| rakhmerov | d0ugal: yes, that's what I'd like to do | 07:59 |
| rakhmerov | will do it now I guess | 07:59 |
| d0ugal | We have been discussing how to handle old patches in TripleO - https://review.openstack.org/#/c/449332/ | 07:59 |
| d0ugal | which may be of interest :) | 08:00 |
| rakhmerov | d0ugal: that's a huge read ) | 08:01 |
| rakhmerov | what's the conclusion? | 08:01 |
| d0ugal | rakhmerov: the spec isn't that long, you can ignore the comments | 08:02 |
| rakhmerov | ah, ok | 08:02 |
| d0ugal | rakhmerov: but it is something like, patches that have not been touched for 180 days (and are not -1 WIP) should be abandoned | 08:02 |
| d0ugal | or if it has a -2 and hasn't been updated in 90 days it will be abandoned | 08:03 |
| d0ugal | rakhmerov: ^ | 08:05 |
| rakhmerov | ok | 08:05 |
| rakhmerov | yeah | 08:05 |
| rakhmerov | good stuff | 08:06 |
| rakhmerov | d0ugal: not sure if you saw my first message to you today, just reminding again about https://blueprints.launchpad.net/mistral/+spec/mistral-engine-commands-docs | 08:06 |
| d0ugal | yup | 08:06 |
| rakhmerov | it'd be nice if you could finish it today-tomorrow | 08:06 |
| d0ugal | Okay, I didn't realise it was so urgent | 08:07 |
| d0ugal | rakhmerov: I have removed myself from the assignee of https://blueprints.launchpad.net/mistral/+spec/mistral-os-install-docs | 08:09 |
| d0ugal | I don't have any plans to work on it at the moment. | 08:09 |
| rakhmerov | d0ugal: well, not so urgent but I just would like to finish what's possible to finish before P-1 release | 08:10 |
| rakhmerov | d0ugal: np, we'll find someone else | 08:10 |
| d0ugal | I might at some point, but it isn't on my current plans. | 08:10 |
| rakhmerov | d0ugal: on engine commands docs, if it's conflicting with you current plans then no worries, we'll move it on to P-2 | 08:11 |
| d0ugal | I guess I can do it | 08:11 |
| d0ugal | I'm not sure it makes sense to track documentation tasks like this, as they are "released" when they are merged - so don't really fit in with milestones very well | 08:12 |
| d0ugal | but I guess it mostly works. | 08:12 |
| rakhmerov | :) | 08:12 |
| rakhmerov | you're right but IMO this is mostly a discipline kind of thing, we need to track the progress somehow on these tasks too | 08:13 |
| rakhmerov | on publishing docs, I completely agree | 08:14 |
| *** shardy has quit IRC | 08:15 | |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Advanced publishing: change workflow lang schema https://review.openstack.org/455178 | 08:17 |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Advanced publishing: add publishing of global variables https://review.openstack.org/456056 | 08:17 |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Advanced publishing: add 'global' function to access global vars https://review.openstack.org/456059 | 08:17 |
| *** shardy has joined #openstack-mistral | 08:17 | |
| *** dmellado has joined #openstack-mistral | 08:23 | |
| rakhmerov | d0ugal: cleaned up gerrit a little bit | 08:50 |
| rakhmerov | 15-20 patches abandoned | 08:50 |
| *** shardy is now known as shardy_mtg | 08:57 | |
| *** jkilpatr_ has quit IRC | 09:06 | |
| *** jkilpatr_ has joined #openstack-mistral | 09:07 | |
| *** jkilpatr_ has quit IRC | 09:11 | |
| *** jkilpatr_ has joined #openstack-mistral | 09:12 | |
| d0ugal | nice | 09:29 |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Optimizing lang schema validation https://review.openstack.org/456147 | 10:11 |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Optimizing lang schema validation https://review.openstack.org/456147 | 10:22 |
| rakhmerov | d0ugal: do you mind if I approve https://review.openstack.org/#/c/414508/ ? | 10:24 |
| rakhmerov | assuming that we'll see possible challenges better once we have an impl | 10:25 |
| warface | hey, I was wondering if there is a python client available for interacting with Mistral Rest APIs ? | 10:29 |
| warface | I was trying to manage the workflow using django | 10:30 |
| rakhmerov | warface: yes, sure, https://github.com/openstack/python-mistralclient | 10:30 |
| rakhmerov | it has both Python API and CLI | 10:30 |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Changed the README.rst and added debug guide https://review.openstack.org/420687 | 10:32 |
| *** aspiers has quit IRC | 10:33 | |
| openstackgerrit | Merged openstack/mistral-specs master: Support region for openstack actions https://review.openstack.org/414508 | 10:34 |
| *** aspiers has joined #openstack-mistral | 10:36 | |
| warface | rakhmerov: thanks | 10:41 |
| warface | stumbled upon it. was hoping there might be something with docs :P | 10:42 |
| warface | it works great | 10:42 |
| warface | will try to document it as soon as possible | 10:43 |
| d0ugal | rakhmerov sure, approve it. I thought we wanted 3 +2s for specs? | 10:44 |
| warface | github PRs would be fine ? | 10:44 |
| openstackgerrit | Merged openstack/python-mistralclient master: Optimize the link address https://review.openstack.org/455272 | 10:44 |
| rakhmerov | warface: github PRs? No, we use review.openstack.org | 10:45 |
| warface | ok cool | 10:45 |
| rakhmerov | d0ugal: yeah, but it wasn't officially accepted yet | 10:45 |
| rakhmerov | anyway, I already did it | 10:45 |
| *** aspiers has quit IRC | 10:46 | |
| rakhmerov | d0ugal: btw, just FYI: seems like something has changed in dependencies (I guess oslo.messaging) and all our patches are failing now | 10:47 |
| rakhmerov | trying to investigate | 10:47 |
| d0ugal | fun | 10:48 |
| d0ugal | this seems to happen lots now | 10:48 |
| *** jkilpatr_ has quit IRC | 10:49 | |
| rakhmerov | yeah | 10:50 |
| rakhmerov | the thing is that if I run failing tests individually they pass | 10:51 |
| rakhmerov | they fail only if I run them all together | 10:51 |
| rakhmerov | it's a lot of fun to fix such issues ) | 10:52 |
| d0ugal | ha | 10:53 |
| d0ugal | That sounds like an issue thrash|g0ne seen last week. | 10:53 |
| rakhmerov | yeah | 10:56 |
| rakhmerov | even worse, on CI I see different errors than on my local env | 10:56 |
| rakhmerov | :) | 10:56 |
| *** aspiers has joined #openstack-mistral | 10:58 | |
| *** thrash|g0ne is now known as thrash | 11:03 | |
| *** shardy_mtg is now known as shardy | 11:13 | |
| *** shardy is now known as shardy_lunch | 11:19 | |
| *** jkilpatr_ has joined #openstack-mistral | 11:20 | |
| rakhmerov | d0ugal: found a serious bug in our tests, I'm surprised why we didn't hit it before.. | 11:21 |
| rakhmerov | fixing.. | 11:21 |
| *** zhurong has quit IRC | 11:26 | |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Fix serialization issue https://review.openstack.org/456180 | 11:37 |
| *** toure is now known as toure|biab | 11:38 | |
| *** toure|biab is now known as toure | 11:38 | |
| rakhmerov | d0ugal: I fixed the issue I mentioned but that's not what breaks our CI :( | 11:45 |
| rakhmerov | https://review.openstack.org/#/c/456180/ | 11:45 |
| *** chlong has joined #openstack-mistral | 11:52 | |
| *** chlong_ has joined #openstack-mistral | 11:52 | |
| *** chlong_ has quit IRC | 11:54 | |
| rakhmerov | d0ugal: yeah, oslo.messaging 5.21.0 was release yesterday | 12:00 |
| *** shardy_lunch is now known as shardy | 12:00 | |
| rakhmerov | and seems like it broke us: http://paste.openstack.org/show/606269/ | 12:00 |
| rakhmerov | ddeja: are you available by any chance? | 12:11 |
| d0ugal | ouch :( | 12:13 |
| *** gongysh has quit IRC | 12:15 | |
| rakhmerov | yep | 12:15 |
| rakhmerov | I don't understand yet if it's a problem on our side or not | 12:15 |
| *** gongysh has joined #openstack-mistral | 12:17 | |
| ddeja | rakhmerov: what's up? | 12:20 |
| rakhmerov | ddeja: trying to investigate a gate failure that started happening a couple of hours ago | 12:20 |
| rakhmerov | seems like it's related with new oslo.messaging | 12:21 |
| * ddeja gets a little deja-vu | 12:21 | |
| rakhmerov | :)) | 12:21 |
| rakhmerov | yes, that's why I decided to ping you | 12:21 |
| rakhmerov | http://logs.openstack.org/80/456180/1/check/gate-mistral-python35/9c1af7f/testr_results.html.gz | 12:22 |
| ddeja | I can take a look | 12:22 |
| rakhmerov | if you open this link and expand the first test you'll see that an option is duplicated in o.m | 12:22 |
| rakhmerov | executor_thread_pool_size | 12:22 |
| rakhmerov | the interesting thing is that we also register this option but in our KombuServer | 12:22 |
| ddeja | shit | 12:22 |
| rakhmerov | although it shouldn't be a problem because we don't create it for unit tests (no???) | 12:23 |
| rakhmerov | ddeja: you got an idea? | 12:23 |
| ddeja | we create it for some tests | 12:23 |
| rakhmerov | for some, yes | 12:23 |
| rakhmerov | do you think this might be a problem? | 12:24 |
| ddeja | I think so | 12:24 |
| ddeja | but hmm, on the other side | 12:24 |
| ddeja | even if we don't have the KomubServer, still this opt would be registered dosen of times | 12:25 |
| rakhmerov | yes | 12:25 |
| ddeja | give me a minute | 12:25 |
| rakhmerov | and seems like this is not a new option in o.m | 12:25 |
| rakhmerov | ddeja: yeah, so it seems like we should have had this problem before too but we didn't | 12:29 |
| rakhmerov | I don't understand yet what changed.. | 12:29 |
| ddeja | rakhmerov: unfortunatelly I don't have much time for this... | 12:37 |
| rakhmerov | ok, np | 12:37 |
| ddeja | What I'd like to propose is to | 12:37 |
| ddeja | comment those two lines: | 12:37 |
| ddeja | ./engine/rpc_backend/kombu/kombu_server.py: cfg.IntOpt('executor_thread_pool_size', | 12:37 |
| ddeja | ./engine/rpc_backend/kombu/kombu_server.py: self._executor_threads = CONF.executor_thread_pool_size | 12:37 |
| rakhmerov | ok | 12:38 |
| ddeja | (It is actually more then 2, but those are only 2 placeses where we using this varaible) | 12:38 |
| ddeja | and test if this fix this problem | 12:38 |
| rakhmerov | ok | 12:38 |
| ddeja | of course, it would make other tests to fails, but doesn't matter at this time | 12:39 |
| rakhmerov | yeah | 12:39 |
| rakhmerov | just to test the idea | 12:39 |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: TRYING TO FIX GATE FAILURE https://review.openstack.org/456201 | 12:39 |
| ddeja | since from looking into the code I don't see any place where we would be calling the Kombu server unless explicity said to do so in config | 12:39 |
| rakhmerov | ok, just sent a patch | 12:40 |
| ddeja | rakhmerov: also it is strange, since this test passes on my local env... | 12:43 |
| jenner_ | so is anyone running mistral using docker? I mean stretched over several containers for every component | 12:43 |
| rakhmerov | yeah, the problem is that these tasks pass individually | 12:44 |
| rakhmerov | they only fail on CI when they are run together | 12:44 |
| ddeja | great... | 12:44 |
| rakhmerov | ddeja: yeah, it's a lot of fun ) | 12:44 |
| rakhmerov | jenner_: I'm aware of some people, yes | 12:45 |
| jenner_ | rakhmerov: I'm basically trying to figure out if it's worth it to build an image for every component or if it's ok to just have a mistral-all and create extra start scripts | 12:46 |
| rakhmerov | ddeja: http://logs.openstack.org/01/456201/1/check/gate-mistral-python27-ubuntu-xenial/ea88a4f/testr_results.html.gz | 12:47 |
| rakhmerov | :))) | 12:47 |
| rakhmerov | but how???? | 12:47 |
| rakhmerov | I can't really understand.. | 12:47 |
| ddeja | rakhmerov: I have an idea... | 12:48 |
| ddeja | what is the current version of oslo? | 12:48 |
| rakhmerov | jenner_: well, I think it depends on what you need. Are you going to have a high load? | 12:48 |
| rakhmerov | ddeja: it became 5.21.0 yesterday | 12:48 |
| ddeja | I mean, that is in mistral | 12:48 |
| ddeja | ok | 12:48 |
| rakhmerov | ddeja: I think our CI just started pulling this new version and that's why it started happening | 12:49 |
| rakhmerov | but this option has been there for ages.. | 12:49 |
| rakhmerov | ddeja: yes, http://logs.openstack.org/01/456201/1/check/gate-mistral-python27-ubuntu-xenial/ea88a4f/console.html#_2017-04-12_12_45_56_165560 | 12:49 |
| rakhmerov | (just wanted to make sure) | 12:50 |
| jenner_ | rakhmerov: most likely, yes, also a couple of long running WFs (several hours) so at least the executors should be scaled horizontally 2+N times | 12:50 |
| *** dougshelley66 has left #openstack-mistral | 12:50 | |
| jenner_ | rakhmerov: we also thought about grouping the executors depending on our large projects, so we can update them independently | 12:50 |
| ddeja | rakhmerov: ok, I know | 12:51 |
| ddeja | (it's super stupid, to be honest) | 12:51 |
| rakhmerov | jenner_: ok, I'd probably recommend to start with some simple deployment configuration and see how it fits your needs. You can have, for example, one engine and many executors | 12:52 |
| rakhmerov | ddeja: fire away :) | 12:52 |
| rakhmerov | ddeja: I guess it's something stupid but I really don't see it | 12:52 |
| ddeja | rakhmerov: sooo, I'm not sure why, but in the same thread that is running this test | 12:53 |
| rakhmerov | jenner_: really sorry, fighting a fire now.. How about talking more detailed tomorrow? | 12:53 |
| rakhmerov | ddeja: yes.. | 12:54 |
| ddeja | we have previously run the KombuServer that registered the opt | 12:54 |
| ddeja | and now see this: | 12:54 |
| ddeja | in mistral: https://github.com/openstack/mistral/blob/master/mistral/engine/rpc_backend/kombu/kombu_server.py#L39 | 12:54 |
| ddeja | in oslo to 5.20: https://github.com/openstack/oslo.messaging/blob/5.20.0/oslo_messaging/server.py#L55 | 12:54 |
| ddeja | in oslo 5.21: https://github.com/openstack/oslo.messaging/blob/5.21.0/oslo_messaging/server.py#L55 | 12:54 |
| ddeja | that is why re-registering is failing | 12:54 |
| rakhmerov | looking.. | 12:55 |
| rakhmerov | you mean because of this "when executor is threading or eventlet" ? | 12:56 |
| rakhmerov | ddeja: ^ | 12:56 |
| jenner_ | rakhmerov: np, I'm still fighting with openshift anyway, takes some time to convince it to use an internal docker registry | 12:57 |
| rakhmerov | jenner_: ok, anyway, I'll be happy to talk more :) | 12:57 |
| ddeja | rakhmerov: description doesn't match | 12:58 |
| rakhmerov | oooooh!!! | 12:58 |
| rakhmerov | do you think it matters? | 12:58 |
| ddeja | I think so | 12:58 |
| rakhmerov | I thought you meant that they changed some threading behavior | 12:58 |
| rakhmerov | :) | 12:58 |
| rakhmerov | easy to check then... | 12:58 |
| ddeja | yes, just check the description in KombuServer | 12:59 |
| ddeja | but harder would be to fix this... | 12:59 |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: TRYING TO FIX GATE FAILURE https://review.openstack.org/456201 | 13:00 |
| rakhmerov | ddeja: why harder? If it's just a matter of changing a description | 13:01 |
| *** fultonj has quit IRC | 13:01 | |
| *** fultonj has joined #openstack-mistral | 13:01 | |
| ddeja | yes, but wht if oslo changes the description again? ;) | 13:03 |
| rakhmerov | true | 13:03 |
| ddeja | I'm thinking if it would be better to just import the whole thing just to have the registering kept in one place | 13:04 |
| rakhmerov | yeah, right, it deserves a filed bug. But for now we need to find a quick solution | 13:04 |
| ddeja | sure | 13:04 |
| rakhmerov | ddeja: I remember we argued with someone about all those opts spread out across the code. I'm actually for having them all in one place | 13:05 |
| ddeja | Yes, but this is slightly different place | 13:05 |
| ddeja | s/place/case | 13:05 |
| rakhmerov | I know, yes | 13:06 |
| ddeja | but I'm not sure how to properly handle it... | 13:06 |
| rakhmerov | ddeja: they really check if an option is registered like this: if opts[opt.dest]['opt'] != opt: | 13:09 |
| rakhmerov | and in Opt: | 13:09 |
| rakhmerov | def __ne__(self, another): | 13:09 |
| rakhmerov | return vars(self) != vars(another) | 13:09 |
| rakhmerov | :))) | 13:10 |
| rakhmerov | funny | 13:10 |
| rakhmerov | so they really compare all the properties | 13:10 |
| rakhmerov | anyway, that helped, the tests passed | 13:10 |
| ddeja | rakhmerov: I'm glad I could help | 13:11 |
| rakhmerov | ddeja: thanks a ton! | 13:11 |
| *** mgershen1 has joined #openstack-mistral | 13:13 | |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Fix a description of 'executor_thread_pool_size' option in Kombu RPC https://review.openstack.org/456201 | 13:14 |
| *** mgershen has quit IRC | 13:15 | |
| rakhmerov | ddeja, d0ugal: guys, please approve this patch asap to fix Py gates: https://review.openstack.org/#/c/456201/ | 13:15 |
| *** dprince has joined #openstack-mistral | 13:15 | |
| rakhmerov | kong, mgershen1: ^ | 13:16 |
| *** catintheroof has joined #openstack-mistral | 13:27 | |
| *** mgershen has joined #openstack-mistral | 13:29 | |
| *** mgershen1 has quit IRC | 13:32 | |
| *** zhurong has joined #openstack-mistral | 13:47 | |
| *** mgershen1 has joined #openstack-mistral | 13:52 | |
| *** mgershen has quit IRC | 13:54 | |
| *** zhurong_ has joined #openstack-mistral | 14:00 | |
| *** zhurong has quit IRC | 14:01 | |
| *** bobh has joined #openstack-mistral | 14:04 | |
| *** bobh has quit IRC | 14:08 | |
| *** bobh has joined #openstack-mistral | 14:08 | |
| *** mgershen has joined #openstack-mistral | 14:11 | |
| *** mgershen1 has quit IRC | 14:11 | |
| toure | d0ugal ping | 14:14 |
| *** mgershen1 has joined #openstack-mistral | 14:24 | |
| *** mgershen has quit IRC | 14:26 | |
| *** rbrady is now known as rbrady-afk | 14:35 | |
| *** tuan_ has joined #openstack-mistral | 14:36 | |
| d0ugal | toure: delayed pong | 14:37 |
| *** mgershen has joined #openstack-mistral | 14:41 | |
| *** mgershen1 has quit IRC | 14:43 | |
| *** sharatss has quit IRC | 14:47 | |
| *** zhurong_ has quit IRC | 14:57 | |
| *** mgershen1 has joined #openstack-mistral | 14:57 | |
| *** dougshelley66 has joined #openstack-mistral | 14:58 | |
| *** mgershen has quit IRC | 14:59 | |
| *** jkilpatr_ has quit IRC | 14:59 | |
| *** jkilpatr has joined #openstack-mistral | 15:00 | |
| *** jaosorior is now known as jaosorior_away | 15:02 | |
| *** fultonj has left #openstack-mistral | 15:04 | |
| *** d0ugal has quit IRC | 15:30 | |
| *** d0ugal has joined #openstack-mistral | 15:30 | |
| *** d0ugal has joined #openstack-mistral | 15:30 | |
| *** d0ugal has quit IRC | 15:31 | |
| *** d0ugal has joined #openstack-mistral | 15:31 | |
| *** d0ugal has quit IRC | 15:31 | |
| *** d0ugal has joined #openstack-mistral | 15:31 | |
| *** tuan_ has quit IRC | 15:31 | |
| *** warface has quit IRC | 15:35 | |
| *** jpich has quit IRC | 15:45 | |
| *** evgenyl has left #openstack-mistral | 15:52 | |
| *** srwilkers has joined #openstack-mistral | 15:52 | |
| openstackgerrit | Merged openstack/mistral master: Fix a description of 'executor_thread_pool_size' option in Kombu RPC https://review.openstack.org/456201 | 15:58 |
| *** Kevin_Zheng has quit IRC | 16:13 | |
| *** gongysh has quit IRC | 16:29 | |
| *** portdirect has joined #openstack-mistral | 16:49 | |
| *** rbrady-afk is now known as rbrady | 17:14 | |
| *** shardy has quit IRC | 17:15 | |
| *** jtomasek has joined #openstack-mistral | 17:45 | |
| *** jtomasek has quit IRC | 17:46 | |
| *** jtomasek has joined #openstack-mistral | 17:46 | |
| *** jtomasek has quit IRC | 17:47 | |
| *** jtomasek has joined #openstack-mistral | 17:49 | |
| *** jtomasek_ has joined #openstack-mistral | 17:50 | |
| *** jtomasek has quit IRC | 17:53 | |
| *** jtomasek_ has quit IRC | 17:56 | |
| *** srwilkers has quit IRC | 18:19 | |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Fix serialization issue https://review.openstack.org/456180 | 18:39 |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Advanced publishing: change workflow lang schema https://review.openstack.org/455178 | 18:39 |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Advanced publishing: add publishing of global variables https://review.openstack.org/456056 | 18:39 |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Advanced publishing: add 'global' function to access global vars https://review.openstack.org/456059 | 18:40 |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Optimizing lang schema validation https://review.openstack.org/456147 | 18:40 |
| *** jkilpatr_ has joined #openstack-mistral | 18:49 | |
| *** jkilpatr has quit IRC | 18:50 | |
| *** srwilkers has joined #openstack-mistral | 19:01 | |
| *** jkilpatr_ has quit IRC | 19:41 | |
| *** srwilkers has quit IRC | 19:51 | |
| *** jkilpatr_ has joined #openstack-mistral | 19:54 | |
| *** dprince has quit IRC | 20:11 | |
| *** thrash is now known as thrash|biab | 20:29 | |
| *** srwilkers has joined #openstack-mistral | 20:38 | |
| *** thrash|biab is now known as thrash | 20:55 | |
| *** toure is now known as toure|gone | 20:57 | |
| *** Qiming has quit IRC | 21:41 | |
| *** Qiming has joined #openstack-mistral | 21:46 | |
| *** thrash is now known as thrash|g0ne | 21:51 | |
| *** jkilpatr_ has quit IRC | 22:15 | |
| *** srwilkers has quit IRC | 22:46 | |
| *** catintheroof has quit IRC | 22:48 | |
| *** jkilpatr_ has joined #openstack-mistral | 22:49 | |
| *** bobh has quit IRC | 22:57 | |
| *** jkilpatr_ is now known as jkilpatr | 23:10 | |
| *** srwilkers has joined #openstack-mistral | 23:19 | |
| *** gongysh has joined #openstack-mistral | 23:41 | |
| *** bobh has joined #openstack-mistral | 23:46 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!