| *** chandan_kumar has joined #openstack-climate | 07:10 | |
| openstackgerrit | Swann Croiset proposed a change to stackforge/climate: Devstack : create the freepool used by reservation host plugin https://review.openstack.org/80023 | 07:25 |
|---|---|---|
| openstackgerrit | Swann Croiset proposed a change to stackforge/climate: Use nova client mixin in all plugins modules https://review.openstack.org/76830 | 07:29 |
| *** saju_m has joined #openstack-climate | 08:12 | |
| *** Nikolay_St has joined #openstack-climate | 08:59 | |
| openstackgerrit | Swann Croiset proposed a change to stackforge/climate: Devstack : create the freepool used by reservation host plugin https://review.openstack.org/80023 | 09:24 |
| DinaBelova | scroiset, I wanted to approve your 2nd patch, but suddenly approved your 3rd :) | 09:26 |
| DinaBelova | why are you moving create_aggregate from configuring if branch? | 09:26 |
| DinaBelova | I though that should be done in that part | 09:27 |
| *** chandan_kumar has quit IRC | 09:30 | |
| openstackgerrit | A change was merged to stackforge/climate: Change API v1 path to explicit module https://review.openstack.org/78135 | 09:39 |
| *** chandan_kumar has joined #openstack-climate | 09:46 | |
| scroiset | DinaBelova: yeah, I though too, but nova don't seems to be started at this step .. | 09:49 |
| scroiset | I got an error .. looking for logs | 09:50 |
| scroiset | http://logs.openstack.org/23/80023/2/check/gate-climate-devstack-dsvm/1343bbf/logs/devstacklog.txt.gz | 09:51 |
| scroiset | 2014-03-13 09:13:42.689 | ERROR: HTTPConnectionPool(host='127.0.0.1', port=8774): Max retries exceeded with url: /v2/1de5e616a1874c9abeca206a3111d9e2/os-aggregates (Caused by <class 'socket.error'>: [Errno 111] Connection refused) | 09:51 |
| scroiset | only horizon and keystone are UP according to logs .. http://logs.openstack.org/23/80023/2/check/gate-climate-devstack-dsvm/1343bbf/logs/ | 09:52 |
| DinaBelova | well, will it work on rejoin? | 09:57 |
| DinaBelova | I mean won't it try to create many-many freepools? | 09:58 |
| scroiset | hum .. don't think so... I'm checking | 10:00 |
| DinaBelova | thanks | 10:01 |
| scroiset | no .. since the rejoin script only launch screen, every things are already launch only once | 10:02 |
| scroiset | so I'm thing it's good :) | 10:02 |
| scroiset | thinking | 10:03 |
| DinaBelova | well, I guess you're right | 10:03 |
| DinaBelova | one moment, some final checks | 10:03 |
| scroiset | You should be sure before giving +2 ;) | 10:03 |
| DinaBelova | ;) | 10:03 |
| *** bauzas has joined #openstack-climate | 10:23 | |
| openstackgerrit | A change was merged to stackforge/climate: Start moving Climate docs from wiki to Sphinx docs https://review.openstack.org/76152 | 10:43 |
| openstackgerrit | A change was merged to stackforge/climate: Add installation guides to docs https://review.openstack.org/76153 | 10:44 |
| openstackgerrit | A change was merged to stackforge/climate: Fix REST API docs https://review.openstack.org/76493 | 10:46 |
| *** ppetit has joined #openstack-climate | 11:27 | |
| DinaBelova | bauzas, you here? | 11:44 |
| DinaBelova | I was reviewing https://review.openstack.org/#/c/71011/13 | 11:44 |
| DinaBelova | so I've checkouted that change and changed endpoints for climate in keystone | 11:44 |
| DinaBelova | but when typing climate lease-list I've got error in API - it looks like it's somehow connected to acl ... | 11:45 |
| DinaBelova | have you any ideas? | 11:45 |
| DinaBelova | bauzas, ^^ | 12:08 |
| bauzas | DinaBelova: sorry, was out to lunch | 12:23 |
| bauzas | DinaBelova: could you please explain your problem ? | 12:23 |
| bauzas | the python client is not yet delivered for API v2 | 12:24 |
| bauzas | so you have to request the API directly -thanks to Postman for instance - | 12:24 |
| DinaBelova | well, it looks like I've done smth wrong here even if I was curling | 12:25 |
| DinaBelova | let me check one more time | 12:25 |
| bauzas | sure | 12:31 |
| bauzas | I'm doing unittesting, so I have time for helping :D | 12:31 |
| bauzas | dear, I L-O-V-E unittest | 12:31 |
| DinaBelova | :) | 12:33 |
| DinaBelova | bauzas, btw, it seems to me that lease-list should work with current client | 12:37 |
| DinaBelova | as climate endpoint is got there via keystone | 12:38 |
| DinaBelova | using service type | 12:38 |
| DinaBelova | am I wrong? | 12:38 |
| bauzas | yup, provided it can deal with the output change | 12:38 |
| bauzas | I mean not having a dict, but rather either a dict or a list | 12:39 |
| DinaBelova | I've got error in climate-api | 12:39 |
| DinaBelova | so request came there | 12:39 |
| DinaBelova | but there was error | 12:39 |
| bauzas | ok, was it the error ? | 12:39 |
| bauzas | what is ? | 12:39 |
| bauzas | :) | 12:39 |
| DinaBelova | http://paste.openstack.org/show/JyzthKMo8SjvH2tAxQKL/ | 12:39 |
| DinaBelova | it came from acl | 12:39 |
| DinaBelova | as keystone middleware was used there | 12:40 |
| bauzas | lemme look | 12:40 |
| bauzas | interesting... | 12:41 |
| DinaBelova | yes, quite | 12:42 |
| bauzas | that's a pure keystone issue | 12:42 |
| DinaBelova | But keystone says NOTHING 0_o | 12:42 |
| bauzas | pretty strange | 12:42 |
| bauzas | if you look at the API code | 12:42 |
| bauzas | you'll see I'm basically wrapping the app using the middleware auth | 12:42 |
| *** pcargnel has joined #openstack-climate | 12:42 | |
| bauzas | the main difference is that I'm not using the same factory | 12:43 |
| bauzas | as for V1 | 12:43 |
| DinaBelova | well, I'll continue investigating | 12:43 |
| DinaBelova | really strange, I quite agree with u | 12:44 |
| *** cmart has joined #openstack-climate | 12:44 | |
| bauzas | oh gotcha ! | 12:44 |
| bauzas | it's due to the fact the API v1 auth middleware was too lazy | 12:44 |
| bauzas | you need to add some keystone conf params | 12:44 |
| bauzas | in the climate.conf file | 12:45 |
| bauzas | lemme give you these | 12:45 |
| bauzas | at least, try it | 12:45 |
| DinaBelova | ok, thanks | 12:45 |
| bauzas | warning, paste bomb | 12:45 |
| bauzas | [keystone_authtoken] | 12:45 |
| bauzas | auth_uri = http://192.168.122.237:5000 | 12:45 |
| bauzas | auth_strategy = keystone | 12:45 |
| bauzas | admin_user=admin | 12:45 |
| bauzas | admin_password=password | 12:45 |
| bauzas | auth_host=192.168.122.237 | 12:45 |
| bauzas | auth_port=5000 | 12:45 |
| bauzas | auth_protocol=http | 12:45 |
| DinaBelova | ok, I'll try it now | 12:45 |
| bauzas | sure | 12:46 |
| bauzas | there was a warning with API v1 if you don't fill in them | 12:46 |
| bauzas | so the middleware auth was probably working | 12:46 |
| bauzas | but not that way it should | 12:47 |
| DinaBelova | well, it did not help | 12:49 |
| DinaBelova | probably there is some more reason | 12:49 |
| DinaBelova | will continue checking | 12:49 |
| bauzas | anyway, keep these values | 12:52 |
| bauzas | that's worth having them | 12:52 |
| bauzas | but I'm proud to say I unittested the ACL side :D | 12:52 |
| bauzas | so that should/must work :p | 12:52 |
| bauzas | and btw. thanks to unittests I spotted one bug and fixed it | 12:53 |
| bauzas | so, the lesson is that unittests are useful. Awesome. :D | 12:53 |
| DinaBelova | :) | 12:54 |
| cmart | Hello guys! | 13:01 |
| bauzas | cmart: hi | 13:02 |
| cmart | I have some news regarding the https://blueprints.launchpad.net/climate/+spec/tenant-reservation-concept.. Do you remember that bp? | 13:03 |
| DinaBelova | yes, we remember :) | 13:03 |
| DinaBelova | what news? | 13:03 |
| cmart | I switched to Keystone in order to have one of the things that we need for our bp.. about the project's creation with start and end date.. | 13:04 |
| DinaBelova | great :) | 13:04 |
| DinaBelova | Have you discussed that only with Adam or with other Keystone guys too? | 13:05 |
| cmart | but... after talking with the keystone guys.. there are not so sure about putting those new columns and the associated functionality on their module | 13:05 |
| cmart | their argument is that Keystone is a general purpose service, and lots of people have proposed storing stuff in there that doesn't belong | 13:06 |
| cmart | and they told me to see if other projects have a need for the same feature, and build the case for it | 13:06 |
| DinaBelova | I've got smth terrible with internet connection... hate that :( As for Keystone guys argument, they are quite right really... That's overall discussion about Climate and other services | 13:07 |
| DinaBelova | as discussed on TC meeting | 13:07 |
| DinaBelova | so here are probplems, yes | 13:07 |
| bauzas | DinaBelova: cmart: indeed, maybe this usecase is a bit early | 13:08 |
| DinaBelova | that's why I was quite surprised when Adam suggested his solution | 13:08 |
| DinaBelova | cmart, sorry, but we can say for sure only after summit :( | 13:08 |
| bauzas | cmart: I mean, if you have chance to go at the Summit, that's something we should/could discuss with Keystone folks | 13:08 |
| DinaBelova | when Climate future will be decided | 13:08 |
| bauzas | +1 | 13:08 |
| cmart | ohhh :( | 13:09 |
| cmart | I think that I should start adding that stuff in Reserva (I really like that name! :() | 13:09 |
| bauzas | also, keep in mind Keystone is under FF | 13:09 |
| bauzas | nah, say hello to Blazar | 13:09 |
| DinaBelova | because now we've got direct phrase "we'll discuss that on summit. As this feature is interesting but we dunno where and how it should be imlemented" | 13:09 |
| cmart | Reserva, Reserva .. :) :) | 13:09 |
| cmart | Ok | 13:09 |
| bauzas | Blazar is the winner :) | 13:09 |
| DinaBelova | Blazar ;) | 13:09 |
| cmart | hahaha | 13:10 |
| bauzas | well, until it got validated | 13:10 |
| DinaBelova | I've sent email to OS foundation asking for help | 13:10 |
| DinaBelova | no answer yet | 13:10 |
| bauzas | no worries | 13:10 |
| DinaBelova | but they can be not really fast :) | 13:10 |
| bauzas | :) | 13:10 |
| bauzas | they are lawyers :p | 13:10 |
| bauzas | oh, interesting, got a 404 when looking at | 13:11 |
| cmart | DinaBelova, sent email asking for help.. on what subject? | 13:11 |
| bauzas | https://blueprints.launchpad.net/climate/+spec/tenant-reservation-concept. | 13:11 |
| cmart | not about the bp I mentioned, right? | 13:11 |
| bauzas | cmart: Keystone is under FF until Juno | 13:12 |
| bauzas | cmart: so, anyway, this bp shouldn't be merged until mid-May | 13:12 |
| DinaBelova | cmart, you may prepare change concept and leave it as draft | 13:13 |
| DinaBelova | if it'll be needed later | 13:13 |
| DinaBelova | but also we should think here about other variants | 13:13 |
| DinaBelova | if that won't work | 13:13 |
| bauzas | +1 | 13:14 |
| cmart | so next steps (for that bp) will be....Not to code it ? :) | 13:15 |
| DinaBelova | well, as said, you may have drafts... but that's the most logical thing we may do in such situation | 13:16 |
| bauzas | even if providing drafts, there is hard chance that it will get reviewed | 13:17 |
| bauzas | as core devs are focusing on bug fixing | 13:17 |
| bauzas | unless you got a sponsor on that BP | 13:17 |
| DinaBelova | to show on summit that Climate is a good idea we should 1/ modify lease-create process to have one more step "resource allocation" and 2/ provide not only compute resources reservation | 13:18 |
| *** casanch1 has joined #openstack-climate | 13:18 | |
| DinaBelova | that will give us more chances to win, really | 13:18 |
| DinaBelova | and to persuade than other community | 13:18 |
| bauzas | and we should potentially review how we schedule the resources in regards to quotas | 13:18 |
| bauzas | s/in/with | 13:18 |
| DinaBelova | bauzas, yes +1 | 13:19 |
| DinaBelova | so that are main priorities before summit | 13:19 |
| bauzas | +1 | 13:19 |
| bauzas | and do resilience on plugins | 13:19 |
| bauzas | I mean, think about scalability and recovery | 13:19 |
| DinaBelova | so really if there are people wanting to work i'd love to ask them working on volume reservation:) | 13:20 |
| bauzas | +1 | 13:20 |
| bauzas | once I'm finishing the Pecan move (should be a matter of days), I'll begin focusing on Climate workflows | 13:20 |
| DinaBelova | bauzas, what do you mean excatly? | 13:21 |
| bauzas | in order to see how we can take in account TC and community feedbacks | 13:21 |
| bauzas | that's currently a bit trouble in my mind | 13:21 |
| DinaBelova | I'd love to work on https://blueprints.launchpad.net/climate/+spec/new-resource-allocation-phase | 13:21 |
| bauzas | but I mean, focus on delivering a global identical workflow for all plugins | 13:21 |
| DinaBelova | it'll be cool | 13:21 |
| bauzas | missed that one :) | 13:22 |
| DinaBelova | new one | 13:22 |
| DinaBelova | :) | 13:22 |
| DinaBelova | that'll be the first step | 13:22 |
| DinaBelova | as it seems to me | 13:22 |
| bauzas | yup :) | 13:22 |
| bauzas | you have to ramp up on PR usecase | 13:23 |
| bauzas | in order to correctly understand what we did, and what bits of code can be reused | 13:23 |
| bauzas | I can help you | 13:23 |
| DinaBelova | +1 | 13:23 |
| casanch1 | hi guys | 13:23 |
| bauzas | meanwhile, I think I can provide some help on PR and VR usecase by hardening the way we deal with errors | 13:24 |
| DinaBelova | that'll be my main focus next week | 13:24 |
| DinaBelova | casanch1, \o | 13:24 |
| DinaBelova | bauzas, well, not bad :) | 13:24 |
| bauzas | DinaBelova: we definitely need a state model | 13:24 |
| bauzas | DinaBelova: as we previously discussed | 13:25 |
| bauzas | DinaBelova: that's huge work as well | 13:25 |
| bauzas | DinaBelova: is there any BP about it btw. ? | 13:25 |
| bauzas | DinaBelova: I can take that one | 13:25 |
| DinaBelova | https://blueprints.launchpad.net/climate/+spec/lease-state - I've assigned that to me, but I'm ready to share it:) | 13:25 |
| DinaBelova | it was a while ago | 13:25 |
| bauzas | got it https://blueprints.launchpad.net/climate/+spec/lease-state | 13:27 |
| bauzas | yup, please do | 13:27 |
| bauzas | assign it to me | 13:27 |
| bauzas | that doesn't mean we can mutually help ourselves | 13:27 |
| DinaBelova | well, it looks as assigned :? | 13:27 |
| DinaBelova | probably because I've done it already somewhen | 13:27 |
| scroiset | pcargnel: hi, what is composition relation ? | 13:27 |
| pcargnel | Hi!, I mean set the nova client wrapper as a property of the class. Better if is received in the init method. | 13:30 |
| DinaBelova | pcargnel, no, not better :) that's lazy nova client | 13:30 |
| DinaBelova | and it's needed not to call to nova on the __init__ step | 13:31 |
| DinaBelova | as there will be call to keystone | 13:31 |
| scroiset | DinaBelova: as Dina said .. need a lazy client | 13:31 |
| DinaBelova | and nobody guarantees there will be working keystone at the moment of climate manager start | 13:32 |
| DinaBelova | and we'll have nice devstack errors due to that | 13:32 |
| DinaBelova | like manager init -> create nova client -> go to keystone -> no keystone started -> ooooops | 13:33 |
| DinaBelova | so that property is our help to avoid that :) | 13:33 |
| pcargnel | Understood, as I say in the bug, my concern was just readability. I'm new at python and climate, if that is the better way to achieve this I will believe you guys :) | 13:34 |
| DinaBelova | pcargnel, np :) | 13:34 |
| scroiset | concerning the term 'mixin' .. it should provide a method, here it's a special case with a method becoming a property ... didn't find a good term for that | 13:36 |
| scroiset | Mixin seems to me a good name event so .. | 13:37 |
| scroiset | even | 13:37 |
| DinaBelova | scroiset, I'm ok with it :) | 13:38 |
| scroiset | great! | 13:38 |
| scroiset | so .. looking for some core reviewer to +2ed this one https://review.openstack.org/#/c/76830/ | 13:39 |
| *** pafuent has joined #openstack-climate | 13:43 | |
| DinaBelova | sorry, I'm testing iy | 13:44 |
| DinaBelova | it* | 13:44 |
| DinaBelova | not earlier :) | 13:44 |
| *** YorikSar_ has quit IRC | 13:45 | |
| bauzas | scroiset: about Mixin name | 13:48 |
| scroiset | np :D | 13:48 |
| scroiset | bauzas: yes? | 13:48 |
| bauzas | scroiset: I'm fine with delivering it now, but maybe we could rename it as NovaClientWrapper | 13:48 |
| bauzas | or Singleton | 13:48 |
| bauzas | but there is not need to say which design pattern it is | 13:49 |
| *** YorikSar has joined #openstack-climate | 13:49 | |
| bauzas | so I would vote for Wrapper | 13:49 |
| scroiset | not a singleton : no static stuff .. IMHO | 13:49 |
| bauzas | he he | 13:49 |
| scroiset | and for wrapper .. why not | 13:49 |
| bauzas | you challenge me on design patterns :p | 13:49 |
| scroiset | yep ;) | 13:50 |
| bauzas | this *is* a singleton, but also a Facade :p | 13:50 |
| scroiset | do you mean a ProxyObject ? | 13:50 |
| bauzas | nah | 13:50 |
| bauzas | https://github.com/sbauza/python-patterns | 13:51 |
| scroiset | interresting link thanks | 13:52 |
| DinaBelova | ++ for NovaClientWrapper | 13:52 |
| pcargnel | Just for my education, how you read the class declaration? xxxPlugin is built by BasePlugin and NovaClientWrapper? Something like that? | 14:02 |
| DinaBelova | well, for plugins using Nova -yes | 14:04 |
| *** YorikSar has quit IRC | 14:04 | |
| pcargnel | ok, thanks :) | 14:05 |
| bauzas | that's just multiple inheritance :) | 14:05 |
| DinaBelova | but while implementing for example volume plugin - we'll use smth like that for Cinder | 14:05 |
| DinaBelova | yes :) | 14:05 |
| *** YorikSar has joined #openstack-climate | 14:05 | |
| openstackgerrit | Diego Pamio proposed a change to stackforge/climate-nova: Unit tests for the ClimateFilter class https://review.openstack.org/79399 | 14:09 |
| DinaBelova | bauzas, I'm about https://review.openstack.org/#/c/79399/2..3/climatenova/scheduler/filters/climate_filter.py | 14:11 |
| DinaBelova | Diego sent new patch set | 14:11 |
| DinaBelova | so he listened to you and moved oslo.config back to the place it was in | 14:11 |
| DinaBelova | but really that's incorrect | 14:11 |
| DinaBelova | so I'd love to fix it here | 14:11 |
| DinaBelova | I've posted comment to change, but it looks like Diego moved it back anyway | 14:12 |
| *** saju_m has quit IRC | 14:22 | |
| openstackgerrit | Swann Croiset proposed a change to stackforge/climate: Use nova client wrapper in all plugins modules https://review.openstack.org/76830 | 14:51 |
| *** Nikolay_1t has joined #openstack-climate | 14:52 | |
| *** chandan_kumar has quit IRC | 15:02 | |
| *** Nikolay_St has quit IRC | 15:02 | |
| *** chandan_kumar has joined #openstack-climate | 15:03 | |
| *** YorikSar has quit IRC | 15:03 | |
| *** bauzas has quit IRC | 15:04 | |
| *** YorikSar has joined #openstack-climate | 15:05 | |
| *** ppetit has quit IRC | 15:09 | |
| *** ppetit has joined #openstack-climate | 15:09 | |
| *** ppetit has quit IRC | 15:10 | |
| *** YorikSar_ has joined #openstack-climate | 16:32 | |
| *** YorikSar has quit IRC | 16:36 | |
| openstackgerrit | Diego Pamio proposed a change to stackforge/climate-nova: Unit tests for the ClimateFilter class https://review.openstack.org/79399 | 16:43 |
| *** YorikSar_ is now known as YorikSar | 17:55 | |
| pafuent | Hi. I added trusts to physical host reservations and I need to test it (besides unit test). Which is the best way to do it? | 17:57 |
| *** YorikSar has joined #openstack-climate | 17:57 | |
| DinaBelova | pafuent, well, the simple thing is that physical reservations will work | 18:15 |
| pafuent | DinaBelova: But, If I did it wrong? | 18:16 |
| DinaBelova | so you need to test 1/ moving host to freepool by admin 2/ asking to reserve some hosts by user 3/ check VMs will be started on these hosts when lease starts | 18:16 |
| DinaBelova | pafuent, that simply won't work, to be honest | 18:17 |
| DinaBelova | if you did it wring | 18:17 |
| DinaBelova | also please share that change - as a draft if you wish - we all take a look on it | 18:17 |
| DinaBelova | if you'll use trusts correctly - keystone will work, if not - not | 18:17 |
| DinaBelova | trusts are like so :) | 18:17 |
| pafuent | DinaBelova: Yep, but I run only tox, and I would like to test at least once using the climate client | 18:18 |
| DinaBelova | that'll be great | 18:19 |
| pafuent | DinaBelova: So, to do the steps that you mention, I could use climate client? | 18:19 |
| DinaBelova | yes, sure | 18:19 |
| DinaBelova | I think that'll be the best testing here | 18:19 |
| DinaBelova | as trusts will be created/used/used... while this process | 18:20 |
| DinaBelova | so if somehow there will be errors - it's bad sign | 18:20 |
| pafuent | DinaBelova: I just realize that https://bugs.launchpad.net/climate/+bug/1285585/ and https://bugs.launchpad.net/climate/+bug/1285584 are dependant | 18:56 |
| pafuent | DinaBelova: I did the same changes than Swann plus remove the admin config options | 18:57 |
| pafuent | DinaBelova: I think I need to depend on this change | 19:01 |
| DinaBelova | ++, sorry, I thought you did so... | 19:04 |
| DinaBelova | :) | 19:04 |
| DinaBelova | and frankly speaking they are not really dependent, that's why I suppose we got this situation | 19:14 |
| pafuent | DinaBelova: Are dependent. In order to remove the admin I need to use trusts, and they are used by the NovaWrapper | 19:27 |
| *** dpamio has joined #openstack-climate | 19:27 | |
| pafuent | DinaBelova: Right? | 19:27 |
| DinaBelova | yees | 19:31 |
| DinaBelova | sorry | 19:31 |
| DinaBelova | it's quite late here... | 19:31 |
| DinaBelova | i'm going crazy :( | 19:31 |
| DinaBelova | it was a long-long day for me, sorry :( | 19:32 |
| DinaBelova | just one moment | 19:32 |
| DinaBelova | you should use create_ctx_from trust to get novaclient needed creds | 19:33 |
| DinaBelova | but I guess you know that :) | 19:33 |
| DinaBelova | but that's really interesting moment here: NovaWrapper creates long-running clieny | 19:35 |
| DinaBelova | I'm afraid that if we want to commit some actions using token get from trust | 19:36 |
| DinaBelova | we need to go away from this idea :? | 19:36 |
| DinaBelova | currently Swann's change works ok, because we create ctx from trust and I've checked only one request | 19:37 |
| DinaBelova | but it looks that probably it should be not long-running, but new for every request to nova client - not to keep old creds got from trust | 19:38 |
| DinaBelova | scroiset, ^^ | 19:38 |
| DinaBelova | scroiset, I'm interested if your wrapper will work when Pablo will move to trusts evrywhere | 19:39 |
| DinaBelova | I'm not sure, as I'm tired and sleepy, but it looks like that it might not work - as it's long-running client, so if there will be leases to start, for example, from different users with different token and client will be the same | 19:40 |
| DinaBelova | we will get unauthorised error | 19:40 |
| DinaBelova | scroiset, may you check this? | 19:40 |
| DinaBelova | pafuent, well, it looks like it's quite late for scroiset too :) | 19:42 |
| DinaBelova | we need to check this moment | 19:42 |
| pafuent | DinaBelova: Seems to be | 19:42 |
| pafuent | DinaBelova: I'll digest all this, and tomorrow talk with you and Swann | 19:43 |
| DinaBelova | pafuent, we'll check that tomorrow, but you may do that yourself - by adding your change above Swann's change | 19:43 |
| DinaBelova | ++ | 19:43 |
| DinaBelova | good idea | 19:43 |
| pafuent | DinaBelova: Just one more question | 19:43 |
| DinaBelova | yes? | 19:44 |
| pafuent | DinaBelova: Is NovaWrapper handling trust without changes, or I misunderstood something? | 19:44 |
| DinaBelova | now nova wrapper does not use trusts | 19:45 |
| DinaBelova | and it won't really | 19:45 |
| DinaBelova | in direct way | 19:45 |
| DinaBelova | the thing is to use trusts get from user who send request first time | 19:45 |
| DinaBelova | to commit actions on behalf of them | 19:46 |
| pafuent | DinaBelova: But If I don't pass a context, nova wrapper do a context.current() | 19:46 |
| pafuent | DinaBelova: Yes, so the current could be of other user | 19:46 |
| *** dpamio has quit IRC | 19:46 | |
| DinaBelova | yes | 19:46 |
| DinaBelova | so now there is specialadmin user | 19:46 |
| DinaBelova | climate_user in conf | 19:46 |
| DinaBelova | that does admin things with hosts | 19:47 |
| DinaBelova | but it should not be really | 19:47 |
| pafuent | DinaBelova: I missed the current thing in my first implementation | 19:47 |
| *** dpamio has joined #openstack-climate | 19:47 | |
| DinaBelova | there should be trust from admin user who placed host to freepool | 19:47 |
| DinaBelova | to be used for host-related ops | 19:47 |
| DinaBelova | that are done later now using climateuser | 19:47 |
| pafuent | DinaBelova: Ok. I'll recheck my patch and tomorrow I will check with you and Swann. Thanks for your comments. | 19:48 |
| DinaBelova | I'm not sure we're really speaking now about the same things... So it'll be better for you to publish first drafts above Swann's change when we'll understand how that'll work together, and we'll all take a look on it | 19:49 |
| DinaBelova | because now I guess we are speaking almost about the same, but I'm afraid not completely | 19:49 |
| DinaBelova | so now idea overall idea looks like following in little words: "remove climate_user, etc. and use admin trust for hosts related ops" | 19:50 |
| DinaBelova | and that's it | 19:50 |
| pafuent | DinaBelova: Yes | 19:50 |
| DinaBelova | cool | 19:50 |
| *** dpamio has quit IRC | 19:56 | |
| *** cmart has quit IRC | 20:08 | |
| *** pcargnel has quit IRC | 20:24 | |
| *** pafuent has left #openstack-climate | 20:55 | |
| *** casanch1 has quit IRC | 21:49 | |
| *** YorikSar_ has joined #openstack-climate | 23:40 | |
| *** YorikSar has quit IRC | 23:41 | |
| *** casanch1 has joined #openstack-climate | 23:56 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!