openstackgerrit | huanghe proposed openstack/valence master: Updated from global requirements https://review.openstack.org/448435 | 08:08 |
---|---|---|
openstackgerrit | Anusha Ramineni proposed openstack/valence master: Add API Validation framework for flavors https://review.openstack.org/422547 | 09:19 |
*** ananth_n has joined #openstack-valence | 14:48 | |
*** ramineni_ has joined #openstack-valence | 14:51 | |
*** ntpttr__ has joined #openstack-valence | 14:53 | |
*** ntpttr__ is now known as ntpttr_laptop | 14:53 | |
*** HuBian has joined #openstack-valence | 14:57 | |
*** shuquan has joined #openstack-valence | 14:58 | |
ananth_n | hi everyone | 14:59 |
shuquan | hi | 14:59 |
ntpttr_laptop | hi | 14:59 |
ramineni_ | hi | 14:59 |
ananth_n | any agenda items to discuss - we needed some info from @ramineni_ last week IIRC | 15:00 |
ananth_n | i believe ntpttr_laptop wanted some clarification | 15:01 |
ntpttr_laptop | I think madhuri mentioned we should have a quick discussion on the oslo library additions | 15:01 |
mkrai | Hi | 15:01 |
*** yufei has joined #openstack-valence | 15:01 | |
HuBian | hi | 15:01 |
ntpttr_laptop | ananth_n: we talked about it a bit over email, it had to do with the generic driver framework | 15:01 |
ramineni_ | ananth_n: ntpttr_, sure, for adding generic controller ? | 15:01 |
ntpttr_laptop | ramineni_: yeah that was the one | 15:02 |
ramineni_ | ananth_n: ntpttr_laptop: yes, will work on the spec as discussed | 15:02 |
ntpttr_laptop | I guess I do have one follow on question | 15:02 |
ananth_n | thanks. I might have missed the email | 15:02 |
mkrai | ramineni_: What's this generic controller? | 15:02 |
ntpttr_laptop | do you have plans for any southbound drivers other than redfish in the immediate future? | 15:02 |
mkrai | Can you explain briefly or we have any spec for it? | 15:02 |
ntpttr_laptop | what other frameworks do you think we'd like to add? | 15:03 |
*** Jinxing has joined #openstack-valence | 15:03 | |
ramineni_ | ntpttr_laptop: yes, we have ExpressEther driver to add, working on the PoC part of that as well | 15:04 |
*** HuBian has left #openstack-valence | 15:04 | |
*** HuBian has joined #openstack-valence | 15:04 | |
ntpttr_laptop | awesome, that's good to hear! | 15:04 |
ntpttr_laptop | looking forward to it | 15:04 |
ramineni_ | ntpttr_laptop: thanks, will update once we complete that | 15:05 |
ntpttr_laptop | mkrai: briefly the spec will cover adding a generic driver that covers all required functions that a southbound driver must satisfy, and will pass things along to things like redfish or now expressether to talk to the rsd systems | 15:05 |
ramineni_ | mkrai: its for supporting other drivers than redifsh | 15:05 |
mkrai | Ok got it | 15:06 |
mkrai | Thanks | 15:06 |
ntpttr_laptop | okay, now on to the oslo libraries | 15:06 |
yufei | hi,erveryone, I want to ask a question about multi-pod-managers, currenlty we can create/delete podm in valence, but when we can’t compose node in the podm that was created in valence, valence now just compose it in the podm configured in valence.conf. | 15:06 |
mkrai | Ok so what's team thought on it? | 15:06 |
yufei | Is there any plan or spec for this? | 15:06 |
ntpttr_laptop | yufei: yes HuBian is working on this | 15:07 |
mkrai | yufei: If you don't mind, we will get back to your question after oslo discussion | 15:07 |
*** HuBian has left #openstack-valence | 15:07 | |
yufei | ok, thanks | 15:07 |
ntpttr_laptop | mkrai: I think it's a good idea, we can get more functionality from them | 15:07 |
ramineni_ | yufei: HuBian: i have one request on multi podmanager , in the mockup, there is not doc the request format incase of adding new podmanager | 15:07 |
ntpttr_laptop | and they don't add real dependencies on openstack services, just pip requirements | 15:07 |
ramineni_ | https://github.com/openstack/valence/tree/master/api-ref/source/mockup | 15:07 |
ntpttr_laptop | guys one second, we're on oslo libraries | 15:08 |
mkrai | ntpttr_laptop: Right | 15:08 |
*** HuBian has joined #openstack-valence | 15:08 | |
*** HuBian has quit IRC | 15:08 | |
ntpttr_laptop | we can talk about multipodmanager next | 15:08 |
ananth_n | folks, let's discuss on oslo | 15:08 |
ramineni_ | ntpttr_laptop: ok, sory | 15:08 |
ntpttr_laptop | does anyone have any objections to it? If not I propose we go ahead with adding them | 15:08 |
mkrai | Until unless we are not depending on any OS services, then that should be fine I guess | 15:08 |
ntpttr_laptop | ramineni_: no problem :) | 15:08 |
ananth_n | i don't have objections to using oslo libraries as long as it doesn't result in having to start additional services - say amqp | 15:09 |
ananth_n | since these libs would be well tested, no issues using them as libraries | 15:09 |
*** hubian has joined #openstack-valence | 15:09 | |
ramineni_ | mkrai: im ok too, if that counts :) , but i remember discussing this before, so gave a just heads up | 15:09 |
ntpttr_laptop | yeah there's no requirement for rabbit or anything | 15:09 |
mkrai | ramineni_: What's your thought? | 15:09 |
ntpttr_laptop | ramineni_: yeah we don't want to add dependencies on any other services that would need to be installed, like rabbitmq or keystone | 15:10 |
ntpttr_laptop | oslo libraries should be all right, for things like logging and config files | 15:10 |
ntpttr_laptop | okay I think it looks like we're good to go with that | 15:10 |
ananth_n | for some items where we have native python libraries and oslo - i think we should evaluate the native ones before oslo | 15:10 |
ananth_n | ah.. :) ntpttr_laptop gave an example | 15:11 |
mkrai | Cool I think we have agreed for it. I will go ahead with revision of the patch | 15:11 |
mkrai | I request you all to review the patches | 15:11 |
ramineni_ | ok | 15:11 |
mkrai | Thanks! | 15:11 |
ntpttr_laptop | ananth_n: I agree, the oslo libraries for configuration and logging are pretty flushed out and mkrai did you say they're needed for the ironic integration? | 15:11 |
mkrai | Yes | 15:12 |
ananth_n | ok - that said, there was one patch for switching to oslo logging. it was abandoned, should we resurrect it? | 15:12 |
mkrai | But in general also its good to have these libraries | 15:12 |
ramineni_ | https://review.openstack.org/#/c/392416/7/valence/redfish/redfish.py | 15:12 |
ramineni_ | :) | 15:12 |
ramineni_ | thats the reason gave the comment :) | 15:12 |
mkrai | I think we should | 15:13 |
ntpttr_laptop | ramineni_: ah yes, you're remembering right. It's my opinion that the slo libraries for logging and config are not really that big of a dependency | 15:13 |
mkrai | These are just python packages | 15:13 |
ntpttr_laptop | they can be added to requirements.txt and installed quickly | 15:13 |
ananth_n | we could keep oslo logging to sections that are openstack specific - does that make sense? | 15:13 |
ananth_n | maybe my question should be "does that make any sense?" :D | 15:14 |
ntpttr_laptop | I think that would be a bit confusing | 15:14 |
mkrai | ananth_n: I guess it should be oslo logging for valence. | 15:15 |
ntpttr_laptop | I think logging would be better if it were the same across the project | 15:15 |
ramineni_ | ananth_n: if we are using it, i think consistency would be better? | 15:15 |
mkrai | Means we will log the details of what valence is doing | 15:15 |
ramineni_ | ntpttr_laptop: +1 | 15:15 |
ntpttr_laptop | any other thoughts going once | 15:16 |
ntpttr_laptop | going twice | 15:16 |
ntpttr_laptop | okay let's talk about multipodm :) | 15:16 |
ntpttr_laptop | thanks mkrai | 15:17 |
mkrai | Thanks everyone! | 15:17 |
ntpttr_laptop | ramineni_: what was your question about multi pod manager? | 15:17 |
hubian | for the mutlepodm there would be a little refactor | 15:17 |
ramineni_ | i have one request on multi podmanager , in the mockup, doc is missing on the request format incase of adding new podmanager | 15:18 |
ntpttr_laptop | hubian: sure, what would it entail? | 15:18 |
ramineni_ | https://github.com/openstack/valence/tree/master/api-ref/source/mockup | 15:18 |
ntpttr_laptop | hubian: and do you think you could add the details ramineni_ is looking for to the api_spec when you do that :)? | 15:18 |
ramineni_ | hubian: there is only response part | 15:18 |
ramineni_ | hubian: it would help in adding validation for podmanager | 15:18 |
ramineni_ | API validation** | 15:19 |
hubian | ok, i lost the part that how to add podm into valence | 15:19 |
hubian | i will add this then | 15:19 |
ramineni_ | hubian: great , thanks a lot :) | 15:20 |
mkrai | is multi-podm fully supported? | 15:20 |
ntpttr_laptop | hubian: one more question - do you think it would be possible to add podm info to the database entries for composed nodes? | 15:20 |
mkrai | I guess we need to do the changes in nodes API also. We would now need to pass the name of podm in nodes APIs. Right? | 15:20 |
hubian | for the refactor i should have write a spec before this meeting | 15:20 |
ntpttr_laptop | I think it would be useful to be able to grab which pod manager a node belongs to, for instance if pod managers are managed by different drivers redfish ones would be supported by an ironic plugin but some others might not be | 15:20 |
hubian | but i have not.finish | 15:21 |
mkrai | ntpttr_laptop: Yes same question | 15:21 |
ntpttr_laptop | hubian: no problem, just let us know when it's up and we'll review it :) | 15:21 |
ananth_n | hubian: no problem, please check into valence-specs when you complete | 15:21 |
hubian | so please give me a little more time, i will finish the spec by the end of this week | 15:21 |
ntpttr_laptop | awesome | 15:22 |
yufei | looking forward to your spec | 15:22 |
mkrai | I think the solution is to make a one-to-many relationship b/w podm and nodes | 15:22 |
hubian | yes actually we need to add relations for nodes and podm in the database | 15:23 |
ntpttr_laptop | cool, in fact maybe not just nodes but any other resources that we keep in the database, systems etc. | 15:24 |
hubian | yeah | 15:24 |
ramineni_ | mkrai: Is it possible to have a spec before for ironic plugin aswell, it will help to understand the flow /requirements better? | 15:25 |
yufei | If we compose a node in one podm, should we allow to delete this podm now? | 15:25 |
ananth_n | yufei: no, we will have to wait till all the nodes under the podm are released | 15:26 |
mkrai | I have a desgin document for that. Is it fine if I post the detail on bp whiteboard? | 15:26 |
ananth_n | else, we will have a composed node with no means to manage it | 15:26 |
mkrai | ananth_n: +1 | 15:27 |
ramineni_ | mkrai: it would be better in spec? for any discussion or comments on that | 15:27 |
mkrai | Ok will try to post it this week | 15:27 |
ramineni_ | mkrai: thanks, that would be great | 15:27 |
ntpttr_laptop | anyone have any other topics they'd like to discuss? | 15:28 |
ananth_n | nope | 15:30 |
ntpttr_laptop | looks like it :) | 15:30 |
ananth_n | i guess we can call it a meeting | 15:30 |
ntpttr_laptop | okay, thank you everyone | 15:30 |
*** shuquan has quit IRC | 15:31 | |
*** yufei has left #openstack-valence | 15:31 | |
*** ananth_n has quit IRC | 15:31 | |
hubian | bye , see you next week | 15:31 |
*** hubian has quit IRC | 15:32 | |
*** ramineni_ has left #openstack-valence | 15:35 | |
*** Jinxing has quit IRC | 15:35 | |
*** ntpttr_laptop has quit IRC | 16:07 | |
*** lin_yang has joined #openstack-valence | 18:07 | |
openstackgerrit | Nate Potter proposed openstack/valence master: Implement TripleO Configuration Plugin [WIP] https://review.openstack.org/448807 | 22:28 |
openstackgerrit | Nate Potter proposed openstack/valence master: Implement TripleO Configuration Plugin [WIP] https://review.openstack.org/448807 | 22:29 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!