| *** openstack has joined #openstack-mistral | 05:58 | |
| ddeja | rakhmerov: no problem | 06:10 |
|---|---|---|
| openstackgerrit | Marcos Fermín Lobo proposed openstack/mistral: Release note for magnum actions support https://review.openstack.org/329894 | 06:39 |
| openstackgerrit | Marcos Fermín Lobo proposed openstack/mistral: Release note for magnum actions support https://review.openstack.org/329894 | 06:40 |
| *** shardy has joined #openstack-mistral | 07:30 | |
| d0ugal | "YaqlEvaluationException: Can not evaluate YAQL expression: $.status or 'SUCCESS' , data=....." | 08:02 |
| d0ugal | What would caust that YAQL to fail? | 08:02 |
| d0ugal | Is it because the 'status' key can't be found? | 08:02 |
| d0ugal | It does say "error: u'status' " at the end, after the data - which is a bit big to paste here | 08:03 |
| d0ugal | I am looking at http://paste.openstack.org/show/516493/ | 08:03 |
| d0ugal | and I can't see status in the data, so I guess that must be in | 08:04 |
| d0ugal | s/in/it/ | 08:04 |
| d0ugal | Is there a way to determine where in the workflow that is happening? That YAQL is used in a few places. | 08:05 |
| therve | d0ugal, I don't know that, but I think $.get('status', 'SUCCESS') is what you want | 08:15 |
| d0ugal | therve: aha, thanks! I'll try that :) | 08:23 |
| ddeja | rakhmerov: are you there? I have one question regrding RPC patches | 08:25 |
| rakhmerov | ddeja: here | 08:25 |
| ddeja | cool | 08:25 |
| rakhmerov | d0ugal: wait a sec, I'll take a look.. | 08:25 |
| ddeja | rakhmerov: kombu message object has 'type' property | 08:26 |
| ddeja | if on server side something wrong happens during message proccesing, server would send message to client with exception as a message body and woudl set type to 'error' | 08:27 |
| rakhmerov | d0ugal: I'd like to look at your WF, seems like the problem in task "send_message" | 08:28 |
| ddeja | than, on client side, if we receive message with type error, we are raising exception, which woudl go directly to kombu, which doesn't seems like desire behaviour, but I may miss something | 08:28 |
| ddeja | rakhmerov: what do you think? | 08:29 |
| rakhmerov | ddeja: reading.. | 08:29 |
| ddeja | link to code https://review.openstack.org/#/c/194117/20/mistral/engine/rpc_direct/kombu/kombu_client.py L109 | 08:30 |
| rakhmerov | what do you call client and server here? | 08:30 |
| ddeja | client and server in RPC manner | 08:30 |
| ddeja | client is sending a message to server to do something and wait for response | 08:30 |
| rakhmerov | ok | 08:31 |
| rakhmerov | ok, what do you mean directly to kombu? | 08:32 |
| rakhmerov | don't 100% follow | 08:32 |
| ddeja | _on_response method is called by kombu library | 08:33 |
| ddeja | so if we raise exception in it, flow goes back to kombu | 08:33 |
| ddeja | and we can't controll it | 08:33 |
| rakhmerov | ok, yeah | 08:36 |
| rakhmerov | I see now, so _on_response() is just a callback that we provide for kombu | 08:36 |
| ddeja | rakhmerov: yup | 08:36 |
| rakhmerov | well, I guess there should be a defined contract for this type of callback | 08:37 |
| rakhmerov | that defines how callbacks should behave | 08:37 |
| rakhmerov | usually, it's should be explained, particularly, what exceptions are handled and what are bubbled up to kombu users | 08:38 |
| ddeja | rakhmerov: OK | 08:38 |
| rakhmerov | try to look at the doc or their source code | 08:38 |
| rakhmerov | it might be explained | 08:38 |
| rakhmerov | but generally, yes, this is a concern | 08:38 |
| ddeja | ok, will do | 08:38 |
| ddeja | thanks | 08:38 |
| rakhmerov | if, for example, it handles exceptions somehow then we need to understand what we'll get as a result in this case | 08:39 |
| rakhmerov | yeah, it requires some research | 08:39 |
| ddeja | rakhmerov: yup. I wrote some tests for kombu and find a lot, unfortunately | 08:40 |
| rakhmerov | :) | 08:40 |
| ddeja | places that needs some research ;) | 08:40 |
| rakhmerov | I see | 08:40 |
| d0ugal | rakhmerov: https://review.openstack.org/#/c/300200/17/workbooks/baremetal.yaml | 08:41 |
| rakhmerov | d0ugal: yeah, it's send_message | 08:41 |
| rakhmerov | <% $.status or 'SUCCESS' %> | 08:41 |
| rakhmerov | it's not a valid YAQL | 08:41 |
| d0ugal | aha :) | 08:42 |
| d0ugal | so is $.get what I want? | 08:42 |
| rakhmerov | not sure, let me check | 08:42 |
| rakhmerov | d0ugal: ok, seems like YAQL function coalesce() will work for you | 08:49 |
| rakhmerov | this is how I made this conclusion: https://github.com/openstack/yaql/blob/master/yaql/tests/test_branching.py#L51 | 08:50 |
| rakhmerov | :) | 08:50 |
| rakhmerov | the doc is still really pure for YAQL unfortunately | 08:50 |
| d0ugal | rakhmerov: oh, interesting. I didn't know about that function, I'll try it | 08:52 |
| rakhmerov | d0ugal: ok | 08:53 |
| d0ugal | Yeah, I was considering trying to improve the docs but I am still learning | 08:53 |
| d0ugal | Thanks! | 08:53 |
| rakhmerov | me too ) | 08:53 |
| rakhmerov | at some point we'll have to improve the doc ) | 08:53 |
| rakhmerov | otherwise we'll be spending most of our time sitting in IRC and answering YAQL syntax questions | 08:54 |
| *** dmk0202 has joined #openstack-mistral | 09:02 | |
| kong | mflobo: ping | 09:20 |
| openstackgerrit | Lingxian Kong proposed openstack/mistral: Release note for magnum actions support https://review.openstack.org/329894 | 09:23 |
| d0ugal | rakhmerov: I get the same error with coalesce btw, so I think it is because the status key doesn't exist. | 09:24 |
| rakhmerov | d0ugal: did you change this YAQL everywhere in WF? | 09:25 |
| d0ugal | rakhmerov: Yeah, search and replace | 09:25 |
| d0ugal | rakhmerov: The new error was the same but with coalesce in it :) | 09:25 |
| rakhmerov | hm.. as far as I remember, it shouldn't curse if a key doesn't exist | 09:25 |
| d0ugal | using .get like therve suggested seemed to work | 09:25 |
| rakhmerov | should just return None | 09:25 |
| d0ugal | <% $.get('status', 'SUCCESS') %> | 09:25 |
| rakhmerov | ooh, ok | 09:26 |
| rakhmerov | there might have been some changes recently in semantics of $.key_name thing | 09:26 |
| rakhmerov | ok | 09:26 |
| d0ugal | Right, possibly. | 09:26 |
| d0ugal | I think the current approach did work at some point for us too | 09:26 |
| ddeja | rakhmerov: one short question regarding RPC: all implementation should follow oslo.messaging patterns? I mean, all clients should implements two methods: call_sync and call_async and can I assume that they should behave similar to oslo.messaging call and cast? http://docs.openstack.org/developer/oslo.messaging/rpcclient.html#oslo_messaging.RPCClient.call | 09:36 |
| ddeja | rakhmerov: And another one: is there a document describing how engine communicates with executor? | 09:38 |
| ddeja | that would help me a lot | 09:38 |
| rakhmerov | ddeja: on the first one: absolutely no | 09:39 |
| rakhmerov | we have our own protocol which is expressed by our RPC interfaces | 09:40 |
| rakhmerov | and whatever implements them can be used | 09:40 |
| rakhmerov | regardless of what oslo.messaging provides | 09:40 |
| ddeja | rakhmerov: OK, but our kombu driver should be similar to oslo.messaging right? (despite ACK-then-process vs process-then-ACK feature) | 09:41 |
| rakhmerov | on the second question: there's no good doc I think, it's basically just classes in rpc.py that tell how communications look like | 09:41 |
| rakhmerov | EngineClient, ExecutorClient | 09:41 |
| rakhmerov | this is all our protocol | 09:42 |
| rakhmerov | ddeja: why similar? | 09:42 |
| ddeja | rakhmerov: on second - oh, OK, I was wondering if there is something else ;) | 09:42 |
| rakhmerov | ddeja: basically the idea is the following: we have several methods in EngineClient and ExecutorClient | 09:42 |
| rakhmerov | for example, ExecutorClient.run_action() | 09:43 |
| ddeja | rakhmerov: I mean - kombu async_call should behave like oslo.messaging cast - send a message and return imidiately. And same kombu sync_call should behave like oslo.messaging call - send a message and wait for response | 09:43 |
| rakhmerov | this is a part of our high-level protocol | 09:43 |
| rakhmerov | yeah, but you're talk about impl details | 09:43 |
| ddeja | rakhmerov: yup, that's what I wanted to confirm :) | 09:44 |
| rakhmerov | underneath it can be implemented in different ways, but maybe yes, you're right | 09:44 |
| rakhmerov | whenever we have to deal with sync calls we have to do some waitings | 09:44 |
| ddeja | OK, cool | 09:44 |
| rakhmerov | although those waitings are evil, IMO | 09:44 |
| ddeja | thanks, I was a little confused after going deep into code and need some clarification so I can sort everything out in my mind :) | 09:45 |
| ddeja | but now everything is clear for me | 09:45 |
| szaher | rakhmerov: Hi, Yes I am using liberty | 09:52 |
| rakhmerov | szaher: then you pain is explainable | 09:52 |
| rakhmerov | ..your pain.. | 09:52 |
| rakhmerov | is there a possibility to switch to a newer version? | 09:53 |
| szaher | I am Ok as long as this has no effect/conflict on my liberty env | 09:53 |
| rakhmerov | ddeja: yeah, don't stick too much to existing Nikolay's code | 09:53 |
| rakhmerov | szaher: I see, that stack trace is a known problem which is caused by oslo.messaging and the patch has been backported into Mitaka | 09:54 |
| rakhmerov | but may be pretty hard to backport it to Liberty | 09:54 |
| ddeja | rakhmerov: I helped him with this yesterday ;) | 09:55 |
| ddeja | but there was still some authentication issue | 09:55 |
| rakhmerov | ooh, ok | 09:55 |
| ddeja | that didn't show in liberty | 09:55 |
| ddeja | with same mistral.conf | 09:55 |
| *** saphi has joined #openstack-mistral | 09:56 | |
| szaher | by the way according to the log mistral was able to list images but he failed in getting the requested Image ! | 09:56 |
| ddeja | szaher: One thing, can you show your workflow? | 09:58 |
| szaher | ddeja rakhmerov: http://paste.openstack.org/show/516531/ | 09:58 |
| szaher | ddeja: This is the action that I would like to run https://github.com/openstack/mistral-extra/blob/stable/liberty/examples/v2/openstack/glance_actions.yaml | 09:58 |
| rakhmerov | yeah, I remember something like this... | 09:59 |
| rakhmerov | nmakhotkin: hi Nikolay, can you please look at this printout? As far as I remember you worked on fixing this or similar problem | 10:00 |
| *** cheneydc has quit IRC | 10:01 | |
| szaher | rakhmerov: any idea about the authentication error ? | 10:11 |
| rakhmerov | not yet, I want to wait for nmakhotkin, he should know | 10:12 |
| szaher | Ok. Please, let me know if you got any update. | 10:12 |
| openstackgerrit | hardik proposed openstack/mistral: Refactored tempest tests https://review.openstack.org/330438 | 10:20 |
| rakhmerov | szaher: ok, sorry for this | 10:20 |
| nmakhotkin | hi, szaher, rakhmerov | 10:23 |
| nmakhotkin | as far as I remember, this error was related to inproper initialization of API client | 10:23 |
| nmakhotkin | or it also might be due to the trust-scoped token (but as I see, the token is not trust-scoped) | 10:24 |
| nmakhotkin | the similiar thing I saw in novaclient when requesting servers_list using trusted token | 10:25 |
| nmakhotkin | but here it seems it is a bit another issue | 10:26 |
| szaher | nmakhotkin: http://paste.openstack.org/show/516536/ | 10:26 |
| szaher | nmakhotkin: Hi. I pasted my keystone auth configuration in this paste | 10:28 |
| szaher | could you please, take a look and let me know what am doing wrong ? | 10:29 |
| nmakhotkin | szaher, could you please try the same workflow using your admin credentials from this conf? | 10:29 |
| nmakhotkin | is it succeed? | 10:29 |
| szaher | mistral user is admin | 10:29 |
| szaher | nmakhotkin: http://paste.openstack.org/show/516531/ This is the log for this action | 10:30 |
| nmakhotkin | szaher: what credentials are you using when you start workflow? | 10:31 |
| nmakhotkin | the same as in conf or another? | 10:31 |
| szaher | nmakhotkin: I didn't provide additional credentials and I don't know even how to do that | 10:34 |
| nmakhotkin | hmm | 10:34 |
| nmakhotkin | szaher: Do you use mistral CLI for starting workflows? | 10:35 |
| nmakhotkin | then you should have in your env variables named 'OS_USERNAME', 'OS_PASSWORD' and so on | 10:35 |
| szaher | Yes | 10:36 |
| szaher | I am using different credentials | 10:36 |
| nmakhotkin | and these variables are used when you start workflow | 10:36 |
| nmakhotkin | try to provide admin credentials and start workflow | 10:36 |
| szaher | OS_USERNAME=admin | 10:36 |
| nmakhotkin | I mean credentials which are in mistral.conf | 10:36 |
| szaher | instead of mistral user (which is admin) I should provide the admin credentials ? | 10:37 |
| nmakhotkin | no no | 10:37 |
| nmakhotkin | Try to use credentials from mistral.conf | 10:38 |
| nmakhotkin | for you it would be OS_USERNAME=mistral | 10:38 |
| szaher | Ok, I need to export the same credentials in my client env | 10:38 |
| szaher | rakhmerov: by the way you have this bug in mitaka branch http://paste.openstack.org/show/516538/ | 10:38 |
| *** saphi has quit IRC | 10:43 | |
| openstackgerrit | Merged openstack/mistral: Release note for magnum actions support https://review.openstack.org/329894 | 10:49 |
| szaher | nmakhotkin: I don't why I am keep getting this error even in mitaka branch http://paste.openstack.org/show/516542/ | 10:56 |
| szaher | rakhmerov: I would say mistral mitaka is broken as well ! | 10:57 |
| szaher | stable/mitaka is broken | 10:57 |
| szaher | nmakhotkin: I used the same credentials and it's not working | 11:03 |
| rakhmerov | szaher: ok, I see | 11:05 |
| rakhmerov | we'll look at it | 11:06 |
| szaher | rakhmerov: Could you please, suggest one single branch that I can use to install mistral ? | 11:06 |
| szaher | rakhmerov: till now I spent 3 days installing mistral and till now it's causing a lot of problems and it's not working ! | 11:07 |
| openstackgerrit | Andras Kovi proposed openstack/mistral: Use client credentials to retrieve service list https://review.openstack.org/330469 | 11:09 |
| szaher | rakhmerov: are you here ? | 11:14 |
| rakhmerov | on a meeting | 11:14 |
| rakhmerov | szaher: people use both Liberty and Mitaka, Mitaka is more stable apparently | 11:15 |
| rakhmerov | the issue you have is probably not an installation issue | 11:16 |
| rakhmerov | it's a problem with some OpenStack actions | 11:16 |
| rakhmerov | we'll have to look at it | 11:16 |
| rakhmerov | szaher: please file bugs at Launchpad, we'll be addressing them | 11:17 |
| rakhmerov | bugs are possible :) | 11:17 |
| szaher | rakhmerov: I would say if people are using mistral so this means at least something should work ! | 11:21 |
| szaher | The basic test which is getting the first image id or name is not working | 11:22 |
| rakhmerov | szaher: I got it | 11:22 |
| rakhmerov | no need to repeat | 11:22 |
| rakhmerov | we'll look at it | 11:22 |
| rakhmerov | these OpenStack actions is not a core part of Mistral btw, it's rather an extension | 11:23 |
| rakhmerov | and it's hard to maintain for a huge number of reasons | 11:23 |
| rakhmerov | something might not work, so we need to look at it and fix it | 11:24 |
| szaher | \q | 11:24 |
| rakhmerov | szaher: it's also Open Source and you can help us fix it | 11:25 |
| rakhmerov | resources are kind of limited | 11:25 |
| rakhmerov | szaher: do you have any other problems except not working Glance action? | 11:34 |
| rakhmerov | where we could try to help | 11:34 |
| rakhmerov | szaher: btw, these OpenStack actions are not being completely reworked since we realized some time ago that they are not designed well enough | 11:35 |
| rakhmerov | they will be changed almost completely in this cycle | 11:36 |
| rakhmerov | in Newton | 11:36 |
| openstackgerrit | Dawid Deja proposed openstack/mistral: New RPC layer implementation https://review.openstack.org/194117 | 11:48 |
| openstackgerrit | Dawid Deja proposed openstack/mistral: Integrating new RPC layer with Mistral https://review.openstack.org/197563 | 11:48 |
| openstackgerrit | Dawid Deja proposed openstack/mistral: Make RPC implementation configurable https://review.openstack.org/197562 | 11:48 |
| openstackgerrit | Dawid Deja proposed openstack/mistral: Adding OsloRPC server and client https://review.openstack.org/197561 | 11:48 |
| openstackgerrit | Dawid Deja proposed openstack/mistral: Add tests for Kombu driver https://review.openstack.org/330483 | 11:48 |
| mflobo | kong, pong | 11:48 |
| *** dprince has joined #openstack-mistral | 11:54 | |
| openstackgerrit | Dawid Deja proposed openstack/mistral: Integrating new RPC layer with Mistral https://review.openstack.org/197563 | 12:00 |
| openstackgerrit | Dawid Deja proposed openstack/mistral: Add tests for Kombu driver https://review.openstack.org/330483 | 12:00 |
| *** shardy has quit IRC | 12:04 | |
| *** bobh has joined #openstack-mistral | 12:19 | |
| *** bobh has quit IRC | 12:47 | |
| *** bobh has joined #openstack-mistral | 13:19 | |
| szaher | rakhmerov: How can I debug the messages to fix this https://bugs.launchpad.net/mistral/+bug/1593227 ? | 14:05 |
| openstack | Launchpad bug 1593227 in Mistral "mistral messaging error " [Undecided,New] | 14:05 |
| szaher | nmakhotkin: still getting the auth problem | 14:20 |
| szaher | How can I solve it, any ideas ? | 14:20 |
| *** jistr is now known as jistr|mtg | 14:28 | |
| nmakhotkin | szaher: Try to use glance CLI | 14:30 |
| nmakhotkin | command glance image-list | 14:30 |
| nmakhotkin | it seems even in this case you should get auth problem | 14:31 |
| szaher | working fine | 14:31 |
| szaher | nmakhotkin: do you want it from server side or client side ? | 14:34 |
| *** rrecio has joined #openstack-mistral | 15:05 | |
| *** rrecio_ has joined #openstack-mistral | 15:12 | |
| *** rrecio has quit IRC | 15:12 | |
| *** jistr|mtg is now known as jistr | 15:19 | |
| *** rbrady has quit IRC | 15:20 | |
| *** rbrady has joined #openstack-mistral | 15:31 | |
| *** openstackgerrit has quit IRC | 15:34 | |
| *** openstackgerrit has joined #openstack-mistral | 15:34 | |
| *** dprince has quit IRC | 16:01 | |
| *** dprince has joined #openstack-mistral | 16:02 | |
| *** gyee has joined #openstack-mistral | 16:06 | |
| *** dprince has quit IRC | 16:20 | |
| *** dprince has joined #openstack-mistral | 16:21 | |
| *** dmk0202 has quit IRC | 16:42 | |
| openstackgerrit | Andras Kovi proposed openstack/mistral: Use client credentials to retrieve service list https://review.openstack.org/330469 | 18:11 |
| *** gyee has quit IRC | 18:25 | |
| *** tonytan4ever has joined #openstack-mistral | 18:43 | |
| *** dmk0202 has joined #openstack-mistral | 18:44 | |
| *** dmk0202 has quit IRC | 18:45 | |
| openstackgerrit | Andras Kovi proposed openstack/mistral: Example Mistral docker container broke due to oslo.policy update https://review.openstack.org/330748 | 18:57 |
| *** dprince has quit IRC | 19:27 | |
| *** dmk0202 has joined #openstack-mistral | 19:40 | |
| *** dmk0202 has quit IRC | 19:43 | |
| *** tonytan4ever has quit IRC | 20:19 | |
| *** dmk0202 has joined #openstack-mistral | 20:23 | |
| *** gyee has joined #openstack-mistral | 20:24 | |
| *** rrecio_ has quit IRC | 21:23 | |
| *** dmk0202 has quit IRC | 21:38 | |
| *** bobh has quit IRC | 21:39 | |
| *** dmk0202 has joined #openstack-mistral | 22:04 | |
| *** dmk0202 has quit IRC | 22:39 | |
| *** chlong has quit IRC | 23:42 | |
| *** bobh has joined #openstack-mistral | 23:49 | |
| *** bobh has quit IRC | 23:53 | |
| *** bobh has joined #openstack-mistral | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!