| *** bobh has quit IRC | 00:38 | |
| *** bobh has joined #openstack-mistral | 00:38 | |
| *** bobh has quit IRC | 00:43 | |
| *** weshay has quit IRC | 02:30 | |
| *** weshay has joined #openstack-mistral | 02:32 | |
| openstackgerrit | Andras Kovi proposed openstack/mistral master: Allow noop as task name https://review.openstack.org/582966 | 04:39 |
|---|---|---|
| *** hardikjasani has joined #openstack-mistral | 05:10 | |
| *** jenner_ has left #openstack-mistral | 05:50 | |
| *** akovi has joined #openstack-mistral | 06:07 | |
| hardikjasani | How do I create alembic migration script? Should I copy existing one and modify it, or use some tool to generate one? | 06:12 |
| mcdoker181818 | alembic revision -m 'blablabla' | 06:15 |
| mcdoker181818 | I create manually :D | 06:16 |
| akovi | Create manually ++ | 06:16 |
| rakhmerov | hardikjasani: hi | 06:16 |
| rakhmerov | yes :) | 06:16 |
| rakhmerov | you can also look at http://alembic.zzzcomputing.com/en/latest/tutorial.html#create-a-migration-script | 06:17 |
| hardikjasani | ^all: Thanks, I will go with tool. Fewer chances of mistakes :) | 06:19 |
| hardikjasani | Well this took more time than expected :D | 06:59 |
| hardikjasani | Turns out migration file layout in mistral is not what alembic expects, we need to use mistral-db-manage wrapper. | 06:59 |
| hardikjasani | mistral-db-manage revision -m "slug goes here" --rev-id "revision-id" | 07:00 |
| hardikjasani | like mistral-db-manage revision -m "add a column" --rev-id 030 | 07:00 |
| akovi | yeah, I'd try not to waste time with tools | 07:01 |
| *** shardy has joined #openstack-mistral | 07:06 | |
| rakhmerov | d0ugal: hey | 07:08 |
| rakhmerov | ping me when you have time | 07:08 |
| mcdoker181818 | What do you think to check a "Authozation: Bearer token" here too https://github.com/openstack/mistral/blob/master/mistral/auth/keycloak.py#L42 ? | 07:13 |
| mcdoker181818 | `X-Auth-Token` is only used by Openstack services | 07:13 |
| mcdoker181818 | Not very convenient | 07:13 |
| mcdoker181818 | rakhmerov: akovi: | 07:14 |
| rakhmerov | mcdoker181818: hi | 07:14 |
| rakhmerov | what's the context? What are you trying to solve | 07:14 |
| rakhmerov | ? | 07:14 |
| rakhmerov | mcdoker181818: I'm just not so strong at this security stuff so I'm asking to explain a little bit ) | 07:16 |
| mcdoker181818 | rakhmerov: My microservices use a Authorization header because it's a standard https://tools.ietf.org/html/rfc6750 Not very convenient to change header only for Mistral | 07:19 |
| rakhmerov | aah, I see | 07:19 |
| rakhmerov | then yes, sure | 07:20 |
| rakhmerov | I guess we should have even ignored OpenStack specific header here for the Keycloak integration | 07:20 |
| rakhmerov | but we can check both, yes | 07:20 |
| mcdoker181818 | guess we already can't because backward compatible :) | 07:21 |
| mcdoker181818 | Ok, good | 07:21 |
| *** AlexeyAbashkin has joined #openstack-mistral | 07:22 | |
| rakhmerov | yep | 07:41 |
| openstackgerrit | Andras Kovi proposed openstack/mistral master: Allow noop as task name https://review.openstack.org/582966 | 07:56 |
| d0ugal | rakhmerov: Hi, I have time now. | 08:09 |
| akovi | hi d0ugal, I'd like to discuss https://review.openstack.org/582966 a bit | 08:16 |
| akovi | I think this change can be extended to not restrict task names at all | 08:16 |
| akovi | We should document that the priority of name resolution is 1st:"tasks", 2nd:"engine commands" | 08:18 |
| akovi | But at least noop should be allowed as it severely breaks BC | 08:18 |
| akovi | d0ugal: ping :) | 08:19 |
| d0ugal | akovi: breaks BC? | 08:21 |
| akovi | backward comp | 08:21 |
| d0ugal | oh, was noop allowed previously? When did we disallow it? | 08:21 |
| akovi | yes, everything was allowed actually | 08:21 |
| d0ugal | I see | 08:21 |
| d0ugal | Oh, you have a related change | 08:21 |
| * d0ugal looks | 08:22 | |
| akovi | this is kind of nice as it allows to override engine commands if needed | 08:22 |
| d0ugal | akovi: I don't even think I seen the original change. Why do we need it? Could we just revert that? | 08:22 |
| d0ugal | That change landed when I was on paternity leave :) | 08:22 |
| akovi | I put some improvement in this one that enhances consistency :) | 08:23 |
| d0ugal | akovi: but yes, I agree, we shouldn't have changed this | 08:23 |
| akovi | Ok, I'll exend with the other commands too | 08:24 |
| d0ugal | akovi: Thanks | 08:24 |
| d0ugal | We didn't even include a release note for the change :( | 08:24 |
| *** shardy has quit IRC | 08:24 | |
| d0ugal | akovi: the backwards incompatible change landed in queens! | 08:26 |
| d0ugal | so it has even been release and we didn't get any bug reports for it... | 08:26 |
| rakhmerov | :) | 08:33 |
| rakhmerov | I don't mind reverting it | 08:33 |
| *** shardy has joined #openstack-mistral | 08:37 | |
| *** gkadam has joined #openstack-mistral | 08:41 | |
| d0ugal | rakhmerov: but I guess since it was released as part of queens and we have had no complaints it wasn't that bad... | 08:43 |
| d0ugal | well, we had our first complaint now from akovi :-D | 08:43 |
| akovi | :) | 08:46 |
| rakhmerov | yeah :) | 08:50 |
| rakhmerov | I assume there could be people who just preferred to be silent :) | 08:50 |
| d0ugal | True | 08:50 |
| d0ugal | Should we also then backport the fix to Queens? | 08:50 |
| rakhmerov | and may be even throw their attempts to use Mistral :) | 08:50 |
| rakhmerov | aah... | 08:50 |
| rakhmerov | good question | 08:50 |
| d0ugal | I guess we should, because if we fix it for Rocky then it would work in Pike and Rocky but not Queens. That seems trange. | 08:51 |
| rakhmerov | I'm wondering if it would be complaint with stable branch policy | 08:51 |
| d0ugal | strange | 08:51 |
| rakhmerov | yeah | 08:51 |
| rakhmerov | you're making a point | 08:51 |
| d0ugal | rakhmerov: I think it would be okay, because we are fixing a backwards compatability issue | 08:51 |
| rakhmerov | yep, ok | 08:52 |
| rakhmerov | agree | 08:52 |
| rakhmerov | akovi: seems like you have a green light ) | 08:52 |
| d0ugal | btw, I don't know if you know, but apetrich is on vacation until august 1st | 08:53 |
| d0ugal | (He is moving to Berlin) | 08:53 |
| d0ugal | rakhmerov: Can you help me understand something in the rpc clients? | 08:58 |
| d0ugal | rakhmerov: What does this call? https://github.com/openstack/mistral/blob/master/mistral/rpc/clients.py#L425-L433 | 08:58 |
| d0ugal | I guess it call the "notify" method somewhere, but where? :) | 08:58 |
| d0ugal | https://github.com/openstack/mistral/blob/master/mistral/notifiers/notification_server.py#L59 | 08:59 |
| d0ugal | Here? | 08:59 |
| mcdoker181818 | yep | 08:59 |
| d0ugal | How does it know "notify" is that notify method? where are they connected? | 09:00 |
| akovi | d0ugal: I think you're looking for this https://github.com/openstack/mistral/blob/master/mistral/notifiers/notification_server.py#L59 | 09:03 |
| d0ugal | akovi: Yeah, I found that. I just don't understand how the two are linked | 09:04 |
| d0ugal | How does it know the "notify" string maps to that method | 09:04 |
| mcdoker181818 | d0ugal: You can take a look on kombu implementation | 09:04 |
| d0ugal | ah ok | 09:05 |
| d0ugal | so it isn't a simple answer :) | 09:05 |
| mcdoker181818 | https://github.com/openstack/mistral/tree/master/mistral/rpc/kombu | 09:05 |
| d0ugal | Thanks, I'll take a read :) | 09:06 |
| mcdoker181818 | It's simple answer. But my english is harder to understand than the Mistral or oslo messaging code :D | 09:07 |
| mcdoker181818 | You can debug the example https://github.com/openstack/mistral/tree/master/mistral/rpc/kombu/examples | 09:08 |
| d0ugal | mcdoker181818: oh, I didn't know we had examples. That is useful, thanks. | 09:13 |
| rakhmerov | d0ugal: hey.. | 09:26 |
| rakhmerov | reading.. | 09:26 |
| d0ugal | rakhmerov: it's okay, I think I got the answers I need :) | 09:26 |
| rakhmerov | so, https://github.com/openstack/mistral/blob/master/mistral/rpc/clients.py#L427 is the key | 09:26 |
| rakhmerov | it forms a topic for Rabbit | 09:26 |
| rakhmerov | and when we initialize an RPC server we subscribe to that topic | 09:27 |
| rakhmerov | here: https://github.com/openstack/mistral/blob/master/mistral/notifiers/notification_server.py#L46-L47 | 09:28 |
| rakhmerov | it's the place where we register the current object (instance of NotifierServer) in the Oslo Messaging RPC server | 09:29 |
| rakhmerov | so that its methods (including "notify") are automatically linked to the corresponding MQ topics | 09:29 |
| d0ugal | Right | 09:29 |
| d0ugal | that makes sense | 09:29 |
| rakhmerov | yep | 09:29 |
| rakhmerov | how exactly? Well, you need to drill down to o.m | 09:30 |
| rakhmerov | :) | 09:30 |
| d0ugal | haha, indeed | 09:30 |
| rakhmerov | yep | 09:30 |
| d0ugal | okay, so I think I understand the mistral side of it at least. | 09:30 |
| mcdoker181818 | Our kombu implementation is simple version of o m | 09:32 |
| mcdoker181818 | rakhmerov: Why you dind't contribute at least once to om? | 09:32 |
| rakhmerov | mcdoker181818: we tried to | 09:36 |
| rakhmerov | spent huge amount of time and their team decided that it doesn't fall into the project scope really (which is still weird to me) | 09:36 |
| rakhmerov | there was even a suggestion to start a new project | 09:37 |
| rakhmerov | mcdoker181818: all in all, there is a long story behind it | 09:38 |
| d0ugal | rakhmerov: Can you look at this again? https://review.openstack.org/581059 | 09:42 |
| rakhmerov | sure | 09:42 |
| d0ugal | Added tests for the action itself and one for the engine calling async actions. | 09:43 |
| mcdoker181818 | rakhmerov: Got it | 09:43 |
| rakhmerov | d0ugal: looks good (+2ed), just one minor comment to address if you happen to send another ps | 09:47 |
| rakhmerov | mcdoker181818: yeah, we're still hoping to get back to Kombu RPC and make it more production ready | 09:48 |
| rakhmerov | because for us it's more beneficial architecturally | 09:48 |
| mcdoker181818 | I agree. Except for executor | 09:59 |
| *** shardy_ has joined #openstack-mistral | 09:59 | |
| *** shardy has quit IRC | 10:02 | |
| *** d0ugal has quit IRC | 11:15 | |
| *** d0ugal has joined #openstack-mistral | 11:18 | |
| *** d0ugal has quit IRC | 11:18 | |
| *** d0ugal has joined #openstack-mistral | 11:18 | |
| openstackgerrit | Hardik Jasani proposed openstack/mistral master: [WIP] Add namespace support for workbooks and actions https://review.openstack.org/582482 | 11:21 |
| openstackgerrit | Hardik Jasani proposed openstack/mistral master: Fix rev-id parameter in mistral-db-manage https://review.openstack.org/583214 | 11:44 |
| jlejeune | hello @all | 11:47 |
| jlejeune | I have some questions for you :) | 11:47 |
| jlejeune | I want to deploy a global mistral server in "front of" several openstack regions | 11:48 |
| jlejeune | I don't want that my mistral service be a part of these openstack regions | 11:49 |
| jlejeune | so I configure my mistral service with that option : Pecan’s auth_enable = False | 11:50 |
| jlejeune | and I declare os-target-* ENV variable matching my first openstack region (just an example) | 11:51 |
| jlejeune | but when I want to call some openstack actions like cinder.services_list for example, I've got a 400 http error on POST call over /v2/action_executions url with no error message except ERROR (app) | 11:53 |
| jlejeune | (using pike release) | 11:53 |
| jlejeune | first question: is my use case can work with that configuration ? | 11:54 |
| jlejeune | second question: do I miss something ? ;) | 11:54 |
| thrash | jlejeune: I think if you auth_enable = False that would disable any of the auth processing. | 12:07 |
| thrash | including target. | 12:07 |
| thrash | But I could be wrong. | 12:08 |
| jlejeune | I can try using keystone auth | 12:12 |
| openstackgerrit | Hardik Jasani proposed openstack/mistral master: [WIP] Add namespace support for workbooks and actions https://review.openstack.org/582482 | 12:27 |
| akovi | jlejeune: the target-* are the params you are looking for | 12:33 |
| akovi | however, we never tried with multiple regions so it is possible that something is missing there | 12:34 |
| jlejeune | akovi: yep I actually try to use them :) | 12:37 |
| jlejeune | I need to define OS_TARGET_AUTH_URL right ? | 12:38 |
| akovi | yes | 12:39 |
| jlejeune | but as soon I define this variable, a simple std.echo action doesn't work anymore | 12:39 |
| akovi | can you send an example? | 12:39 |
| jlejeune | OS_TARGET_AUTH_URL is mapped to our global keystone endpoint | 12:39 |
| akovi | you can paste it on http://paste.openstack.org/ | 12:39 |
| akovi | some logs or some more info would be useful to diagnose the issue | 12:44 |
| openstackgerrit | Andras Kovi proposed openstack/mistral master: Allow engine commands as task name https://review.openstack.org/582966 | 12:46 |
| jlejeune | akovi: doing | 12:54 |
| jlejeune | akovi: http://paste.openstack.org/show/726109/ | 13:04 |
| openstackgerrit | Andras Kovi proposed openstack/mistral master: Allow engine commands as task name https://review.openstack.org/582966 | 13:04 |
| akovi | jlejeune: can you please check the server logs? | 13:06 |
| akovi | the client logs seem to be useless | 13:06 |
| jlejeune | it's a client error, and I have nothing more in server logs | 13:08 |
| akovi | is your API running on localhost? | 13:09 |
| jlejeune | yep | 13:10 |
| jlejeune | but I don't see the call in api log | 13:11 |
| jlejeune | like the error occured before | 13:11 |
| jlejeune | the POST | 13:11 |
| openstackgerrit | Brad P. Crochet proposed openstack/mistral master: Install mistral-lib as a sibling for tox jobs https://review.openstack.org/583232 | 13:11 |
| openstackgerrit | Brad P. Crochet proposed openstack/mistral master: Add user creds trusts to db https://review.openstack.org/462056 | 13:12 |
| openstackgerrit | Brad P. Crochet proposed openstack/mistral master: Install mistral-lib as a sibling for tox jobs https://review.openstack.org/583232 | 13:14 |
| d0ugal | rakhmerov: The stable branch is created when the RC1 is released. | 13:22 |
| akovi | seemingly the authentication succeeds | 13:27 |
| akovi | can you run: mistral --debug run-action std.noop | 13:28 |
| akovi | without the target stuff | 13:28 |
| jlejeune | of course | 13:30 |
| akovi | can you show the logs? | 13:30 |
| jlejeune | DEBUG (extension) found extension EntryPoint.parse('table = cliff.formatters.table:TableFormatter') │ | 13:31 |
| jlejeune | DEBUG (extension) found extension EntryPoint.parse('json = cliff.formatters.json_format:JSONFormatter') │ | 13:31 |
| jlejeune | DEBUG (extension) found extension EntryPoint.parse('shell = cliff.formatters.shell:ShellFormatter') │ | 13:31 |
| jlejeune | DEBUG (extension) found extension EntryPoint.parse('value = cliff.formatters.value:ValueFormatter') │ | 13:31 |
| jlejeune | DEBUG (extension) found extension EntryPoint.parse('yaml = cliff.formatters.yaml_format:YAMLFormatter') │ | 13:32 |
| jlejeune | DEBUG (command) run(Namespace(columns=[], fit_width=False, formatter='table', input=None, max_width=0, name='std.noop', noindent=False, prefix='', print_empty=False, run_sync=False, save_result=False, target=None, variables=[])) │ | 13:32 |
| jlejeune | DEBUG (connectionpool) Starting new HTTP connection (1): localhost │ | 13:32 |
| jlejeune | DEBUG (connectionpool) http://localhost:8989 "POST /v2/action_executions HTTP/1.1" 201 290 │ | 13:32 |
| jlejeune | DEBUG (httpclient) HTTP POST http://localhost:8989/v2/action_executions 201 │ | 13:32 |
| jlejeune | {"result": null} | 13:32 |
| jlejeune | sorry for the flood :/ | 13:32 |
| jlejeune | 2018-07-17 13:31:29.504 3089 INFO mistral.api.controllers.v2.action_execution [req-72b4b495-3ec7-4e4b-8ae3-d65af91b3dcb - - - - -] Create action_execution [action_execution=ActionExecution [id='Unset', workflow_name='Unset', workflow_nam│ | 13:32 |
| jlejeune | espace='Unset', task_name='Unset', task_execution_id='Unset', state='Unset', state_info='Unset', tags='Unset', name='std.noop', description='Unset', accepted='Unset', input='Unset', output='Unset', created_at='Unset', updated_at='Unset',│ | 13:32 |
| jlejeune | params='Unset']] | 13:32 |
| jlejeune | api-log ^ | 13:32 |
| akovi | the HTTP 400 must be returned by the API | 13:34 |
| jlejeune | agree | 13:34 |
| akovi | and there should be some error that is not logged | 13:35 |
| akovi | maybe in the executor log? | 13:36 |
| akovi | the auth parameters are used only during the action execution | 13:36 |
| akovi | though, the std.echo or noop does not used it at all | 13:37 |
| jlejeune | that's why it's weird | 13:37 |
| akovi | I think there may be something in the context creation | 13:38 |
| akovi | the api should be run with debug logging | 13:39 |
| jlejeune | it is | 13:41 |
| akovi | I know | 13:42 |
| akovi | is it possible that the service catalog that is returned by keystone is large? | 13:43 |
| akovi | can it be that the json is not sent fully to the API? | 13:43 |
| jlejeune | it's large, yes | 13:46 |
| akovi | then this is the issue | 13:46 |
| jlejeune | hm | 13:47 |
| akovi | we have https://bugs.launchpad.net/mistral/+bug/1699248 and https://bugs.launchpad.net/python-mistralclient/+bug/1702324 as open issues | 13:47 |
| openstack | Launchpad bug 1699248 in Mistral "The X-Target-Service-Catalog header can get too big" [High,In progress] - Assigned to Andras Kovi (akovi) | 13:47 |
| openstack | Launchpad bug 1702324 in python-mistralclient "mistralclient sends too big header" [Undecided,New] - Assigned to Andras Kovi (akovi) | 13:47 |
| akovi | the service catalog is sent in the http header which is quite limited | 13:47 |
| akovi | if it is too large, the server will return an error without processing the request | 13:47 |
| jlejeune | ok | 13:49 |
| akovi | I have been trying to come up with fixes for this issue but I haven't found a clean solution yet | 13:50 |
| jlejeune | maybe I can test your patch | 13:50 |
| akovi | https://review.openstack.org/#/q/topic:service-catalog-issue+(status:open+OR+status:merged) | 13:53 |
| *** bobh has joined #openstack-mistral | 13:53 | |
| akovi | I haven't played with them for a while | 13:53 |
| jlejeune | :) | 13:54 |
| *** bobh has quit IRC | 13:58 | |
| jlejeune | akovi: your patch works well ! | 14:16 |
| akovi | :) | 14:23 |
| jlejeune | I hope it will be quickly approved and merged in next releases | 14:23 |
| akovi | I'll try to work on it. We need it as well | 14:23 |
| jlejeune | no doubt :) | 14:24 |
| jlejeune | I come back | 14:24 |
| *** bobh has joined #openstack-mistral | 14:28 | |
| *** d0ugal has quit IRC | 14:37 | |
| *** d0ugal has joined #openstack-mistral | 14:45 | |
| *** josecastroleon has joined #openstack-mistral | 14:58 | |
| *** akovi has quit IRC | 14:59 | |
| *** josecastroleon has quit IRC | 14:59 | |
| *** d0ugal has quit IRC | 15:03 | |
| *** hardikjasani has quit IRC | 15:10 | |
| *** d0ugal has joined #openstack-mistral | 15:15 | |
| *** d0ugal has quit IRC | 15:26 | |
| *** jistr is now known as jistr|afk | 15:38 | |
| *** d0ugal has joined #openstack-mistral | 15:57 | |
| *** gkadam has quit IRC | 15:59 | |
| openstackgerrit | Brad P. Crochet proposed openstack/mistral master: Install mistral-lib as a sibling for tox jobs https://review.openstack.org/583232 | 16:14 |
| *** d0ugal has quit IRC | 16:18 | |
| *** harlowja has joined #openstack-mistral | 16:18 | |
| *** d0ugal has joined #openstack-mistral | 16:30 | |
| *** AlexeyAbashkin has quit IRC | 16:59 | |
| *** bobh has quit IRC | 17:13 | |
| *** livelace2 has joined #openstack-mistral | 17:19 | |
| *** jistr|afk is now known as jistr | 17:30 | |
| *** harlowja has quit IRC | 17:54 | |
| *** shardy_ has quit IRC | 17:54 | |
| *** bobh has joined #openstack-mistral | 18:10 | |
| *** jistr is now known as jistr|off | 18:25 | |
| *** harlowja has joined #openstack-mistral | 18:32 | |
| *** bobh has quit IRC | 18:54 | |
| openstackgerrit | Hardik Jasani proposed openstack/mistral master: [WIP] Add namespace support for workbooks and actions https://review.openstack.org/582482 | 19:07 |
| *** AlexeyAbashkin has joined #openstack-mistral | 20:22 | |
| openstackgerrit | Brad P. Crochet proposed openstack/mistral master: Add user creds trusts to dbA new table is created to store the user creds trusts.Change-Id: Ia2a19a3fcd8808475a16d4d439e085e62a00dfdcPartial-Bug: #1595084Depends-On: If6ff496fe562c1f7ec76cb7e38e89300b4880034Depends-On: I1f42836ec549a3e56681af29738868bf589 https://review.openstack.org/583353 | 20:34 |
| openstack | bug 1595084 in Mistral "Workflow execution lifespan is limited by auth token expire time" [High,In progress] https://launchpad.net/bugs/1595084 - Assigned to Brad P. Crochet (brad-9) | 20:34 |
| openstackgerrit | Brad P. Crochet proposed openstack/mistral master: Add user creds trusts to db https://review.openstack.org/462056 | 20:35 |
| *** AlexeyAbashkin has quit IRC | 20:54 | |
| *** d0ugal has quit IRC | 20:58 | |
| *** d0ugal has joined #openstack-mistral | 21:00 | |
| *** d0ugal has quit IRC | 21:00 | |
| *** d0ugal has joined #openstack-mistral | 21:00 | |
| *** jtomasek has quit IRC | 21:01 | |
| *** jtomasek has joined #openstack-mistral | 21:06 | |
| openstackgerrit | Brad P. Crochet proposed openstack/mistral master: Add user creds trusts to db https://review.openstack.org/462056 | 21:20 |
| *** bobh has joined #openstack-mistral | 21:44 | |
| *** toure is now known as toure|gone | 22:04 | |
| *** bobh has quit IRC | 22:05 | |
| *** apetrich has quit IRC | 22:08 | |
| *** apetrich has joined #openstack-mistral | 22:12 | |
| *** jtomasek has quit IRC | 22:36 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!