| *** Nikolay_St has joined #openstack-climate | 03:44 | |
| *** chandankumar_ has joined #openstack-climate | 05:26 | |
| *** DinaBelova_ is now known as DinaBelova | 05:40 | |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 05:40 | |
| *** YorikSar has quit IRC | 05:48 | |
| *** DinaBelova has quit IRC | 06:28 | |
| *** DinaBelova has joined #openstack-climate | 06:31 | |
| *** SergeyLukjanov is now known as SergeyLukjanov_a | 06:59 | |
| *** SergeyLukjanov_a is now known as SergeyLukjanov_ | 07:00 | |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 07:08 | |
| *** SergeyLukjanov is now known as SergeyLukjanov_ | 07:18 | |
| *** YorikSar has joined #openstack-climate | 07:22 | |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 07:53 | |
| *** bauzas has joined #openstack-climate | 08:12 | |
| bauzas | hi folks | 08:14 |
|---|---|---|
| bauzas | DinaBelova, SergeyLukjanov, Nikolay_St ? | 08:16 |
| DinaBelova | bauzas, o/ | 08:16 |
| bauzas | waiting for SergeyLukjanov | 08:17 |
| SergeyLukjanov | o/ | 08:17 |
| bauzas | ok, lemme explain my thoughts | 08:17 |
| bauzas | and then you'll comment | 08:17 |
| Nikolay_St | o/ | 08:17 |
| bauzas | sounds like there was indeed a misunderstanding, but not about what you think | 08:17 |
| bauzas | first, lemme briefly explain again how PR usecase works | 08:18 |
| bauzas | admin provisions hosts | 08:18 |
| bauzas | into an aggregate | 08:18 |
| bauzas | user asks for a reservation set | 08:18 |
| bauzas | Climate then creates an aggregate, but *not visible* to the user | 08:18 |
| bauzas | (currently, it is, because for POC, we needed AZes until the filter got merged) | 08:19 |
| bauzas | at the lease start, Climate takes the hosts from the freepool to the user aggregate | 08:19 |
| bauzas | the user still doesn't know that he's having a dedicated agg | 08:20 |
| bauzas | the goal of all that crap is that, when the user will *effectively* ask for a VM, the Climate filter will pick the right hosts from the pool | 08:20 |
| bauzas | so, long story short, with PR, we don't need to inpersonate actions on behalf of the user | 08:21 |
| bauzas | that said | 08:21 |
| bauzas | and that's exactly what I said, I do understand and reckon the need of trusts | 08:21 |
| bauzas | and you're right, trusts are also for inpersonate actions | 08:22 |
| bauzas | but the thing is : trusts can only be created by API v3 | 08:22 |
| bauzas | and consequently, it creates a dependency on a Keystone V3 endpoint | 08:23 |
| SergeyLukjanov | bauzas, it was added by default in Havana | 08:23 |
| SergeyLukjanov | bauzas, so, that's not a problem | 08:23 |
| bauzas | SergeyLukjanov: indeed I knew it :) | 08:23 |
| SergeyLukjanov | bauzas, but what's your concern than? | 08:24 |
| bauzas | I'm not against using the trusts, | 08:24 |
| bauzas | I'm just saying this should be optional | 08:24 |
| SergeyLukjanov | why? | 08:24 |
| bauzas | because for example, DevStack doesn't create a V3 endpoint... | 08:25 |
| DinaBelova | we don't need it | 08:25 |
| SergeyLukjanov | bauzas, bah | 08:25 |
| DinaBelova | as it's done in Heat | 08:25 |
| DinaBelova | they do not use 3rd endpoint | 08:25 |
| SergeyLukjanov | bauzas, take a look on how Heat works | 08:25 |
| DinaBelova | Nick should rewrite trusts, nobody is trying to deny it | 08:25 |
| DinaBelova | in Heat style | 08:25 |
| DinaBelova | but still I believe that actions that shoulld be done by some user should be done with his name, token, etc. | 08:26 |
| SergeyLukjanov | v2.0 api is still the default keystone api and all projects using v3 on-demand for stuff like trysts | 08:26 |
| SergeyLukjanov | trusts( | 08:26 |
| SergeyLukjanov | * :0 | 08:26 |
| bauzas | SergeyLukjanov: that's exactly what I'm saying :) | 08:27 |
| bauzas | v2 is stil the default api | 08:27 |
| SergeyLukjanov | bauzas, but v3 is always enabled since the H release | 08:27 |
| bauzas | so we need to make sure we can run in such what I'm saying a degraded mode | 08:27 |
| DinaBelova | but that's not OS style to use things that won't be a default way of doing things next release! | 08:27 |
| SergeyLukjanov | bauzas, there are several facts about current code/reviews - service user creds shouldn't be used only for tokens checks and trusts usage, probably for some read-only ops | 08:29 |
| SergeyLukjanov | bauzas, but not for managing aggregates | 08:29 |
| DinaBelova | bauzas because that's not really a good idea | 08:29 |
| SergeyLukjanov | bauzas, admin user creds should be used for it (admin user how creates such hidden aggregates) | 08:30 |
| DinaBelova | devstack (with climate) will create climate user | 08:30 |
| bauzas | that's not what Nova thought about pclouds | 08:30 |
| bauzas | I'm looking at how Heat is dealing with that :) | 08:31 |
| bauzas | the only pointer I have is https://github.com/openstack/heat/blob/master/heat/common/heat_keystoneclient.py | 08:31 |
| bauzas | do you have others ? | 08:31 |
| scroiset | used here : heat/engine/clients.py | 08:32 |
| SergeyLukjanov | trusts was created for the projects like climate | 08:33 |
| SergeyLukjanov | service account usage is totally incorrect for any write ops | 08:33 |
| SergeyLukjanov | if you'd like to have a centralized user to make such ops | 08:33 |
| SergeyLukjanov | you just need to add special admin user to confs | 08:34 |
| SergeyLukjanov | for making it | 08:34 |
| SergeyLukjanov | like in Heat before the trusts were impletemted | 08:34 |
| DinaBelova | bauzas, yes, it is the main one | 08:34 |
| SergeyLukjanov | but I dont think any objections why trusts shouldn't be used everywhere | 08:34 |
| DinaBelova | but we have trusts now :) | 08:34 |
| DinaBelova | yep | 08:34 |
| SergeyLukjanov | so, my point is that aggregates should be created by specific admins to keep track on em | 08:35 |
| SergeyLukjanov | deleted/removed too | 08:35 |
| SergeyLukjanov | so, just by storing trusts ctx in db for aggregates | 08:35 |
| SergeyLukjanov | it'll be much more secure and trackable | 08:35 |
| SergeyLukjanov | I'm ok with using service user for read-only ops | 08:36 |
| bauzas | ok, I think I got your ideas | 08:36 |
| SergeyLukjanov | like hypervisors query | 08:37 |
| DinaBelova | bauzas, that will be just more secure | 08:38 |
| bauzas | the thing is, we need to document in the README.md that Climate does require the trusts | 08:38 |
| bauzas | so, here is what I propose | 08:38 |
| bauzas | regarding to the 0.1 release dataz | 08:39 |
| bauzas | date | 08:39 |
| bauzas | hold on, I'm running out of battery | 08:39 |
| bauzas | 1. OK for creating a trust by default at the lease creation | 08:39 |
| bauzas | 2. we should amend our very poor README.md or document somewhere that we need to enable API v3 | 08:40 |
| SergeyLukjanov | agreed | 08:40 |
| DinaBelova | +1 | 08:40 |
| Nikolay_St | +1 | 08:40 |
| DinaBelova | we should also define new release date | 08:40 |
| bauzas | 3. we log a bug for PR asking to use trusts instead of service creds that we target to 0.1.1 or 0.2 | 08:41 |
| DinaBelova | because I'm quite sure we won't be able to do all this in 2 days | 08:41 |
| bauzas | I don't think we need to postpone | 08:41 |
| bauzas | provided README.md file says 'enable API v3', it still does work with service accounts | 08:41 |
| Nikolay_St | may be it's a good thing to postpone for a week, and may be I can help you with places you want to use trusts? | 08:41 |
| SergeyLukjanov | #3 I'm absolutely disagree that service user could be used for write ops | 08:41 |
| DinaBelova | #3 - that's just unappropriate to use smth like climate user for write ops in nova.... | 08:42 |
| DinaBelova | that's quite incorrect imho | 08:42 |
| bauzas | DinaBelova: SergeyLukjanov: we live with High priority bugs | 08:42 |
| bauzas | even for the 0.1 release | 08:42 |
| SergeyLukjanov | but it's totally incorrect | 08:43 |
| DinaBelova | that's not a bug | 08:43 |
| SergeyLukjanov | it's not a bug | 08:43 |
| SergeyLukjanov | :) | 08:43 |
| DinaBelova | that's incorrect behaviour | 08:43 |
| DinaBelova | that just should not exist, I believe :D | 08:43 |
| bauzas | folks, you reviewed it ;) | 08:43 |
| SergeyLukjanov | so, if you prefer to use dedicated user - use it or use trusts instead | 08:43 |
| bauzas | anyway, I can do an emergency patch today for using trusts | 08:43 |
| DinaBelova | yep.And I'm really sorry i missed that. as said in mail | 08:44 |
| bauzas | or I amend the README.md file for saying to create a dedicated account for climate | 08:44 |
| bauzas | I think we can still reach the date | 08:44 |
| SergeyLukjanov | bauzas, due to tha admin-manier of creating aggregates I'm ok now with both dedicated user and trusts for them | 08:44 |
| bauzas | SergeyLukjanov: so that's only a matter of documenting it ? | 08:45 |
| SergeyLukjanov | dedicated but not service | 08:45 |
| SergeyLukjanov | so, small patch for replacing usage of service acc with special one + docs will be ok for me | 08:45 |
| DinaBelova | in conf file we have now smth like climate user | 08:45 |
| DinaBelova | not admin special created for that | 08:45 |
| DinaBelova | like it was in heat before trusts | 08:46 |
| bauzas | ok, handling it for this morning | 08:46 |
| SergeyLukjanov | heat was used heat_admin or smth like that | 08:46 |
| bauzas | can do that | 08:46 |
| SergeyLukjanov | ok, thx | 08:46 |
| SergeyLukjanov | looks like we'll need to re-iterate on it after the 0.1 | 08:46 |
| DinaBelova | bauzas, thank you for starting this discussion | 08:46 |
| bauzas | will discuss with f_rossigneux for making sure its patch does contain the fix | 08:46 |
| SergeyLukjanov | to cleanup that stuff | 08:46 |
| bauzas | we need to create a bug for asking to use trusts | 08:46 |
| DinaBelova | bauzas, I'm just happy we found this moment now | 08:47 |
| DinaBelova | not later | 08:47 |
| DinaBelova | bauzas, +1 | 08:47 |
| bauzas | ok, I'm on it | 08:47 |
| *** SergeyLukjanov is now known as SergeyLukjanov_ | 08:47 | |
| *** DinaBelova is now known as DinaBelova_ | 08:48 | |
| bauzas | DinaBelova_: https://bugs.launchpad.net/climate/+bug/1270765 | 09:29 |
| bauzas | please triage it | 09:29 |
| bauzas | while I'm writing the patch :) | 09:29 |
| Nikolay_St | so | 09:37 |
| Nikolay_St | I lost the point | 09:37 |
| Nikolay_St | what's the final decision about trusts patchset? | 09:37 |
| bauzas | 1. ok for using trusts | 09:38 |
| bauzas | 2. we need to document it somewhere (probably in readme file) that we need keystone api v3 | 09:38 |
| bauzas | 3. I'm fixing the issue with reservation pools and inventory by creating a separate admin user (flagset) | 09:39 |
| Nikolay_St | fine | 09:42 |
| bauzas | about 1., we need to make sure that if trusts won't be available, we will gracefully be able to create the lease thou | 09:42 |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 09:43 | |
| *** SergeyLukjanov is now known as SergeyLukjanov_ | 09:48 | |
| Nikolay_St | bauzas: about vm plugin and 'only logging' | 09:58 |
| Nikolay_St | we prefer not to break service if we have some troubles with vms | 09:59 |
| Nikolay_St | that's why we use logging and don't raise any exceptions | 09:59 |
| Nikolay_St | we've discussed it with Dina previous Wednesdat, I think | 10:00 |
| Nikolay_St | DinaBelova_: ping | 10:00 |
| Nikolay_St | bauzas: ping | 10:00 |
| bauzas | sorry, was on my patch | 10:00 |
| bauzas | I'm about to send it | 10:00 |
| Nikolay_St | btw, I prefer log a bug about 'cannot create lease without v3 api' | 10:01 |
| bauzas | YorikSar already gave -1 for some previous identical patterns | 10:01 |
| Nikolay_St | not to fix it now | 10:01 |
| bauzas | ie. having an except: clause with a log in it | 10:01 |
| Nikolay_St | bauzas: I can't get about what are your latest words :( | 10:03 |
| bauzas | ok | 10:04 |
| bauzas | I'm just saying that I already saw YorikSar giving a -1 for having a try:except: condition with the except: clause catching the exception and only logging it | 10:04 |
| bauzas | as far as I remember, it was for the reservation pool code | 10:05 |
| bauzas | Nikolay_St: the main issue you currently have is that the Manager exceptions are not handled by the API | 10:05 |
| bauzas | this is due to the fact that AMQP is raising a different Exception | 10:05 |
| bauzas | https://review.openstack.org/#/c/63445/ | 10:06 |
| bauzas | this one is waiting for a second +2/A | 10:06 |
| bauzas | if the Manager raises an Exception, the API will catch it | 10:06 |
| bauzas | and correctly render the exception to the user | 10:07 |
| bauzas | so, related to your patch | 10:07 |
| bauzas | I would prefer having a way to catch the exception, process what it needs to process, and then raise the exception that API will catch and show to the user | 10:11 |
| bauzas | because if not, it will fail silently | 10:11 |
| bauzas | Nikolay_St: you see my point ? | 10:13 |
| Nikolay_St | well | 10:14 |
| Nikolay_St | I try to get it | 10:14 |
| openstackgerrit | Sylvain Bauza proposed a change to stackforge/climate: Fix service user creating objects https://review.openstack.org/67801 | 10:16 |
| bauzas | Nikolay_St: looking at your code | 10:17 |
| Nikolay_St | bauzas: on which patch? | 10:17 |
| bauzas | https://review.openstack.org/#/c/49142/ | 10:17 |
| bauzas | L109 | 10:19 |
| bauzas | oops | 10:19 |
| bauzas | L89 sorrty | 10:19 |
| bauzas | oh | 10:20 |
| bauzas | gotcha | 10:20 |
| bauzas | this code is executed by Climate, not coming from the API | 10:21 |
| bauzas | it was late in the evening yesterday, so missed it | 10:21 |
| Nikolay_St | yeap | 10:21 |
| Nikolay_St | so, what's comment are correct, and which not? | 10:22 |
| bauzas | amending it now | 10:22 |
| Nikolay_St | Ok, I'll wait for it | 10:24 |
| Nikolay_St | so | 10:24 |
| Nikolay_St | finally what about: | 10:25 |
| Nikolay_St | 14:01 < Nikolay_St> btw, I prefer log a bug about 'cannot create lease without v3 api' | 10:25 |
| Nikolay_St | bauzas: ? | 10:25 |
| Nikolay_St | DinaBelova_: ? | 10:25 |
| Nikolay_St | SergeyLukjanov_: ? | 10:25 |
| bauzas | Nikolay_St: could you please be more explicit ? | 10:26 |
| bauzas | I missed your point | 10:26 |
| Nikolay_St | ok | 10:27 |
| Nikolay_St | so, I prefer not to make some changes in my patches right know, but just test them 'as is'(close to last comment, of courcse). And after that log the bug about | 10:28 |
| Nikolay_St | 'lease creatin failed if we don't use keystone api v3' | 10:28 |
| bauzas | ok with it | 10:28 |
| Nikolay_St | fine | 10:29 |
| bauzas | once your utils patch got merged, I also need to create a bug for PR using your clients and use trusts | 10:30 |
| Nikolay_St | well | 10:35 |
| Nikolay_St | bauzas: we will have code freeze tomorrow? | 10:36 |
| bauzas | code freeze, but that means you can still produce new patchsets | 10:36 |
| bauzas | provided it's bug fix | 10:36 |
| bauzas | Nikolay_St: I have a bug with my venv | 10:40 |
| Nikolay_St | bauzas: ? | 10:40 |
| bauzas | (climate)bauzas@climate-devstack-ctrl:~$ cd climate && python setup.py install && cd - | 10:40 |
| bauzas | running install | 10:40 |
| bauzas | Wheel installs require setuptools >= 0.8 for dist-info support. | 10:40 |
| bauzas | pip's wheel support requires setuptools >= 0.8 for dist-info support. | 10:40 |
| bauzas | Storing debug log for failure in /tmp/tmpSdLwux | 10:40 |
| bauzas | error: ['/opt/climate/bin/python', '-m', 'pip.__init__', 'install', 'pbr>=0.5.21,<1.0', 'Babel>=1.3', 'eventlet>=0.13.0', 'Flask>=0.10,<1.0', 'iso8601>=0.1.8', 'kombu>=2.4.8', 'oslo.config>=1.2.0', 'python-novaclient>=2.15.0', 'netaddr>=0.7.6', 'python-keystoneclient>=0.4.1', 'Routes>=1.12.3', 'SQLAlchemy>=0.7.8,<=0.7.99', 'stevedore>=0.12', 'WebOb>=1.2.3,<1.3'] returned 1 | 10:40 |
| Nikolay_St | bauzas: about keystoneclient version? | 10:40 |
| bauzas | yep, upgrading keystoneclient also upgraded pip to 1.5 | 10:41 |
| bauzas | and consequently it breaks | 10:41 |
| Nikolay_St | wow | 10:41 |
| bauzas | how did you fix it ? | 10:41 |
| bauzas | because I guess you already deployed keystoneclient-0.4.2 | 10:42 |
| Nikolay_St | I test all my changes on vm with devstack directly | 10:44 |
| Nikolay_St | bauzas: so, I haven't this problems | 10:44 |
| Nikolay_St | bauzas: https://review.openstack.org/#/c/52296/33/climate/exceptions.py is comment right? I think no, am I wrong? | 10:45 |
| bauzas | Nikolay_St: if these exceptions are raised by the Manager, by methods called by API | 10:50 |
| bauzas | then if you want them to appear on the API side, you need a rebase to the patch you just approved | 10:50 |
| bauzas | ideally, all Manager defined exceptions should reside in a separate Manager exceptions file | 10:51 |
| bauzas | Nikolay_St: what's your pip version ? | 10:53 |
| bauzas | pip -V ? | 10:53 |
| Nikolay_St | pip 1.4.1 | 10:53 |
| bauzas | when I asked to upgrade to keystoneclient==1.4.2, it pulled me a lot of new deps | 10:54 |
| bauzas | incl. pip-1.5 | 10:54 |
| bauzas | I'll just amend my reqs.txt file, delete and recreate my venv and see what's wront | 10:55 |
| bauzas | btw, I still don't see requirements.txt updated in the utils patch | 10:57 |
| bauzas | while it should | 10:57 |
| Nikolay_St | I'll update in next patchset | 10:58 |
| bauzas | cool | 10:58 |
| bauzas | about ClimateKeystoneClient | 10:58 |
| Nikolay_St | so, it's ready I think | 10:58 |
| bauzas | I was a fan of try: except: | 10:58 |
| bauzas | with the ctx stuff | 10:58 |
| bauzas | the thing you did with Nova | 10:58 |
| Nikolay_St | well, I rewrite it like nova client | 10:58 |
| bauzas | ok, cool | 10:59 |
| Nikolay_St | with one specific hack | 10:59 |
| Nikolay_St | I guess everything will be okay | 10:59 |
| bauzas | your reply was speaking about trusts | 10:59 |
| bauzas | yep, think so :) | 10:59 |
| bauzas | the ideal is that for Nova, if we don't specify a context, the client will handle it anyway :) | 10:59 |
| Nikolay_St | I've seen the way (pathetic) | 11:00 |
| Nikolay_St | well, I think the good way for me is to upload next patchsets for trust and utils | 11:01 |
| Nikolay_St | and then begin testing all this stuff + plugin | 11:01 |
| bauzas | yup | 11:01 |
| bauzas | I'm currently installing my venv | 11:01 |
| bauzas | I kept the same reqs.txt file, I only updated python-keystoneclient to 1.4.2 | 11:01 |
| bauzas | let's see if it goes with pip-1.5 | 11:02 |
| openstackgerrit | Sylvain Bauza proposed a change to stackforge/climate: Fix service user creating objects https://review.openstack.org/67801 | 11:05 |
| bauzas | it goes with pip-1.1 ... | 11:06 |
| Nikolay_St | bauzas: https://review.openstack.org/#/c/49142/32/climate/plugins/instances/vm_plugin.py about line 80 | 11:08 |
| Nikolay_St | why do you think we can get a loop here? and what exit clause are you thinking of? | 11:09 |
| bauzas | it can potentially iterate over and over with _check_active returning False | 11:10 |
| bauzas | and btw., it will return None if it goes into the LOG.exception() | 11:11 |
| bauzas | which also reiterates the loop | 11:11 |
| Nikolay_St | mm... | 11:27 |
| Nikolay_St | are you sure that nova haven't any exits itself? I mean, if task status will be the same for a long time, isn't it nova-side problem? | 11:28 |
| bauzas | even if it's a Nova problem, we should not trust that Nova can handle it :) | 11:30 |
| Nikolay_St | I can't see decision for that problem | 11:32 |
| bauzas | at least, provide an exit way | 11:33 |
| bauzas | iterate until a certain number of times for example... | 11:33 |
| bauzas | or, consider that the exception wil return True | 11:33 |
| bauzas | anything but that infinite loop | 11:34 |
| openstackgerrit | Nikolay Starodubtsev proposed a change to stackforge/climate: Implement keystone trust support https://review.openstack.org/48002 | 11:39 |
| openstackgerrit | Nikolay Starodubtsev proposed a change to stackforge/climate: Add utils to work with openstack components https://review.openstack.org/52296 | 11:39 |
| openstackgerrit | Nikolay Starodubtsev proposed a change to stackforge/climate: Implement basic plugin for VM management https://review.openstack.org/49142 | 11:39 |
| Nikolay_St | bauzas: are you here? | 12:12 |
| Nikolay_St | bauzas: http://paste.openstack.org/show/0uRPk8OWfXdvIgnSDCTU/ any suggestion about it? | 12:12 |
| Nikolay_St | bauzas: fix it myself | 12:17 |
| bauzas | sorry, was out for lunch | 12:22 |
| bauzas | Nikolay_St: you need to create a policy.json file as given in the etc/ directory | 12:25 |
| bauzas | just copy it | 12:25 |
| bauzas | etc/policy.json -> /etc/climate/policy.json | 12:26 |
| bauzas | Nikolay_St: you there ? | 12:41 |
| Nikolay_St | bauzas: | 12:41 |
| Nikolay_St | ye | 12:41 |
| bauzas | the utils review is currently on the zuul pipeline | 12:41 |
| bauzas | but we already have results | 12:41 |
| bauzas | it fails with the requirements check | 12:42 |
| bauzas | https://jenkins02.openstack.org/job/gate-climate-requirements/26/console | 12:42 |
| bauzas | this is a big issue : | 12:42 |
| bauzas | https://jenkins02.openstack.org/job/gate-climate-requirements/26/console | 12:43 |
| bauzas | oops | 12:43 |
| bauzas | Requirement python-keystoneclient>=0.4.2 does not match openstack/requirements value python-keystoneclient>=0.4.1 | 12:43 |
| Nikolay_St | yeap | 12:43 |
| Nikolay_St | I see it | 12:43 |
| Nikolay_St | that's funny, I think :D | 12:43 |
| bauzas | dunno if python-keystoneclient==0.4.2 would match | 12:44 |
| bauzas | I don't think so | 12:44 |
| Nikolay_St | I'll fix it later in the evening | 12:47 |
| Nikolay_St | I need to go unfortunately | 12:47 |
| bauzas | okay, | 12:47 |
| bauzas | that's bad, I was about giving a +2 to utils :( | 12:47 |
| bauzas | SergeyLukjanov_: all the zuul pipes are currently queued with climate-devstack-dsvm | 12:48 |
| bauzas | SergeyLukjanov_: | 12:48 |
| bauzas | SergeyLukjanov_: http://status.openstack.org/zuul/ | 12:48 |
| bauzas | SergeyLukjanov_: :( | 12:48 |
| openstackgerrit | A change was merged to stackforge/climate: Fix Manager exceptions rendering for end-users https://review.openstack.org/63445 | 12:55 |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 12:58 | |
| *** DinaBelova_ is now known as DinaBelova | 12:58 | |
| DinaBelova | bauzas, guys, I'm finally here.... | 13:01 |
| bauzas | DinaBelova: hi | 13:01 |
| DinaBelova | I had to go afk for an hour, and this hour somehow become much more... | 13:01 |
| DinaBelova | I hate this day :D | 13:02 |
| SergeyLukjanov | bauzas, heh, looks like lack of the devstack nodes | 13:02 |
| bauzas | DinaBelova: no pb | 13:02 |
| bauzas | that's bad, the utils patch got a -1 for the requirements | 13:03 |
| bauzas | as for the trusts, I'm amending my comments based on what we discussed today | 13:03 |
| bauzas | but there are still rewrites to do :( | 13:03 |
| DinaBelova | Yes, I.ve read all conversation | 13:04 |
| SergeyLukjanov | bauzas, all the problems are because i2 freeze tomorrow, everyone want to push something before the release | 13:04 |
| SergeyLukjanov | btw, I think queue time is about 2-2.5h, not more | 13:04 |
| DinaBelova | luckly, we are not official OS project now :D so we may postpone release on one day if it will be needed and nobody will notice :D | 13:05 |
| bauzas | yup | 13:05 |
| bauzas | fyi, I suddently noticed that f_rossigneux is on vacation this week... | 13:05 |
| bauzas | so scroiset will handle his patches | 13:06 |
| bauzas | he's about committing new patchsets for the PR | 13:06 |
| DinaBelova | bauzas, ok | 13:07 |
| DinaBelova | also | 13:07 |
| DinaBelova | gate job for utils is failing because of that 4.2 keystoneclient version.... that has needed change, but is not defined in global reqs | 13:08 |
| DinaBelova | we need Nick to do that this evening :D | 13:08 |
| bauzas | DinaBelova: yup, I already discussed this with Nick | 13:08 |
| bauzas | he told me he had to go | 13:08 |
| DinaBelova | yes, I'm just sad he missed fact we need exact req like in global ones... | 13:09 |
| bauzas | https://review.openstack.org/63445 is needed now | 13:09 |
| bauzas | and the gate is around 2h30 mins | 13:09 |
| bauzas | of delay | 13:09 |
| DinaBelova | I see it's merged already | 13:09 |
| bauzas | how possible ? | 13:10 |
| DinaBelova | dunno | 13:10 |
| DinaBelova | 0_0 | 13:10 |
| DinaBelova | StatusMerged | 13:10 |
| DinaBelova | :) | 13:10 |
| bauzas | gotcha... | 13:10 |
| DinaBelova | I may send you screenshot :D | 13:10 |
| bauzas | no no, seeing it | 13:10 |
| bauzas | but zuul is telling me that the non-voting devstack gate is still queued for this patch | 13:11 |
| bauzas | zuul is lying | 13:11 |
| bauzas | :p | 13:11 |
| DinaBelova | it might not, really... | 13:11 |
| DinaBelova | merging may be done without non-voting stuff | 13:11 |
| bauzas | lemme check github | 13:11 |
| DinaBelova | simetimes it happens with such queues | 13:11 |
| bauzas | sure, that's possible | 13:11 |
| bauzas | but that's a terrible waste of resources | 13:11 |
| DinaBelova | I remember sometimes I see Jenkins comments about devstack job AFTER actual merge | 13:12 |
| DinaBelova | yep, it is.. | 13:12 |
| bauzas | about Manager exceptions, we really need now to put them in manager/exceptioins.py | 13:13 |
| bauzas | even if it's not API manageable | 13:14 |
| DinaBelova | As I see you've get such change now | 13:14 |
| DinaBelova | I remember giving you +2 | 13:14 |
| DinaBelova | for that | 13:15 |
| DinaBelova | bauzas, smaaaal fix needed https://review.openstack.org/#/c/67801/ | 13:15 |
| bauzas | sure, say it | 13:16 |
| DinaBelova | I've publishe small comment | 13:16 |
| bauzas | oh dear | 13:16 |
| DinaBelova | yeep | 13:16 |
| bauzas | I'm on it | 13:16 |
| DinaBelova | ok | 13:16 |
| bauzas | do you know any good WYSIWYG editor for .md files ? | 13:16 |
| bauzas | we could btw. move it to .rst no ? | 13:17 |
| bauzas | what do you think of it ? | 13:17 |
| bauzas | as I need to publish another patchset | 13:17 |
| DinaBelova | I love more rst, really | 13:17 |
| DinaBelova | so i think that's ok | 13:17 |
| bauzas | ok, doing it now then | 13:17 |
| DinaBelova | for changing .md -> .rst | 13:18 |
| *** SergeyLukjanov is now known as SergeyLukjanov_ | 13:26 | |
| openstackgerrit | Sylvain Bauza proposed a change to stackforge/climate: Fix service user creating objects https://review.openstack.org/67801 | 13:40 |
| bauzas | folks, could I please have a quick review for https://review.openstack.org/#/c/67801/ ? | 13:41 |
| bauzas | scroiset is about to create a dependency on this patch | 13:41 |
| bauzas | and as the PR reviews are 4 to be chained... :) | 13:41 |
| bauzas | we create a 5th level of dep | 13:42 |
| openstackgerrit | Nikolay Starodubtsev proposed a change to stackforge/climate: Add utils to work with openstack components https://review.openstack.org/52296 | 14:01 |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 14:04 | |
| openstackgerrit | Nikolay Starodubtsev proposed a change to stackforge/climate: Add utils to work with openstack components https://review.openstack.org/52296 | 14:23 |
| openstackgerrit | Sylvain Bauza proposed a change to stackforge/climate: Fix service user creating objects https://review.openstack.org/67801 | 14:24 |
| bauzas | sorry, sent a change without testing it | 14:25 |
| bauzas | blame me or flame me :D | 14:25 |
| bauzas | I do have pep8 compliancy, but that's doesn't cover all the cases :) | 14:25 |
| DinaBelova | :D | 14:27 |
| DinaBelova | missed that too | 14:27 |
| DinaBelova | :D | 14:27 |
| DinaBelova | bauzas, I guess you saw my link about keystoneclient req to Nick's change | 14:28 |
| DinaBelova | :) | 14:28 |
| bauzas | DinaBelova: indeed, gave a +1 | 14:28 |
| DinaBelova | yeeep... | 14:29 |
| bauzas | sorry, I'm not part of the infra gang | 14:29 |
| DinaBelova | also we may look on related bug to that req change | 14:29 |
| DinaBelova | because we might have the same error | 14:29 |
| DinaBelova | don't know now, really | 14:29 |
| DinaBelova | will see how Jenkins will run jobs | 14:29 |
| DinaBelova | I mean this bug https://bugs.launchpad.net/nova/+bug/1268614 | 14:29 |
| bauzas | don't think so | 14:30 |
| bauzas | we don't have a gate for this | 14:30 |
| DinaBelova | oh, yep | 14:31 |
| DinaBelova | in my dreams we have :D | 14:31 |
| openstackgerrit | Sylvain Bauza proposed a change to stackforge/climate: Fix service user creating objects https://review.openstack.org/67801 | 14:46 |
| bauzas | DinaBelova: could you please +2 again ? Swann needed the patch to be rebased | 14:46 |
| bauzas | thanks | 14:47 |
| DinaBelova | :) | 14:47 |
| bauzas | Nikolay_St: https://review.openstack.org/#/c/52296/ | 14:48 |
| bauzas | you reverted to keystoneclient-0.4.1 | 14:48 |
| bauzas | so, I can propose to vote +2 for it | 14:48 |
| bauzas | but we create a bug | 14:49 |
| bauzas | for keeping track of the need | 14:49 |
| bauzas | you OK ? | 14:49 |
| bauzas | DinaBelova: Nikolay_St ? | 14:49 |
| DinaBelova | bauzas, the problem here is | 14:49 |
| DinaBelova | that we have no real opportunity to merge that unlit that patch to global reqs will be merged | 14:50 |
| DinaBelova | Nick uses 4.2 specific things | 14:50 |
| DinaBelova | so we NEED to be sure 4.2 will be installed, not 4.1 | 14:50 |
| DinaBelova | so we need to write 4.2 in reqs really | 14:50 |
| DinaBelova | but we won't pass gate until that change to global reqs will be merged | 14:51 |
| DinaBelova | that's some kind of circle here... | 14:51 |
| bauzas | well, here are my thoughts | 14:51 |
| *** SergeyLukjanov is now known as SergeyLukjanov_a | 14:51 | |
| bauzas | the code is almost functional, except that keystoneclient | 14:52 |
| bauzas | we are blocking a patch because of another | 14:52 |
| bauzas | so, maybe Nick can gracefully handle the miss of keystoneclient.Client() and provides a clear error message | 14:52 |
| *** SergeyLukjanov_a is now known as SergeyLukjanov_ | 14:52 | |
| bauzas | so that there is a clear need | 14:53 |
| bauzas | and once the requirements gate got updated, we both upgrade reqs.txt and we remove the try:exception: clause | 14:53 |
| bauzas | I'm particularly concerned about when this requirements gate would be updated | 14:54 |
| bauzas | and I don't want to be stuck because of it | 14:54 |
| DinaBelova | I suppose you're right here | 14:54 |
| openstackgerrit | Swann Croiset proposed a change to stackforge/climate: Physical host reservation plugin https://review.openstack.org/54285 | 14:54 |
| openstackgerrit | Swann Croiset proposed a change to stackforge/climate: Physical host reservation (manager and base plugin) https://review.openstack.org/64706 | 14:54 |
| openstackgerrit | Swann Croiset proposed a change to stackforge/climate: Physical host reservation (DB related things) https://review.openstack.org/64705 | 14:54 |
| DinaBelova | will you write Nick comments about that try/except clause for using 4.2 specific moments? | 14:55 |
| bauzas | will do | 14:55 |
| DinaBelova | ok | 14:55 |
| DinaBelova | https://review.openstack.org/#/c/52296/34/climate/utils/openstack/keystone.py, line 76 is that moment | 14:58 |
| bauzas | indeed I know | 15:01 |
| bauzas | figure out that'me who asked for that line ;) | 15:01 |
| bauzas | comments left | 15:07 |
| DinaBelova | ok, tnx | 15:10 |
| *** DinaBelova is now known as DinaBelova_ | 15:16 | |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 15:19 | |
| *** DinaBelova_ is now known as DinaBelova | 15:20 | |
| bauzas | folks, I'm currently testing the DB query filter per tenant_id | 15:22 |
| bauzas | there are chances to have a new patchset now that we have policies in place | 15:23 |
| bauzas | I also need to make sure the patch can be merged with latest patches | 15:23 |
| *** oblako has joined #openstack-climate | 15:33 | |
| DinaBelova | ok | 15:46 |
| *** oblako has quit IRC | 15:53 | |
| *** SergeyLukjanov is now known as SergeyLukjanov_ | 16:09 | |
| *** DinaBelova is now known as DinaBelova_ | 16:17 | |
| *** DinaBelova_ is now known as DinaBelova | 16:39 | |
| *** exaggeration has joined #openstack-climate | 17:03 | |
| bauzas | DinaBelova: ping ? | 17:04 |
| DinaBelova | bauzas, pong | 17:04 |
| bauzas | I finished up testing filtering out tenants | 17:04 |
| bauzas | that workds | 17:04 |
| bauzas | works | 17:04 |
| bauzas | there is a corollary, still :D | 17:05 |
| bauzas | admins can only leases from their own tenant | 17:05 |
| bauzas | so, I can propose 2 things : | 17:05 |
| bauzas | 1. leave as it is, but add a flag --all-tenants | 17:05 |
| bauzas | that requires 2 patches, one for the API (will be in the same CR as for the filter), and one for the client | 17:06 |
| bauzas | 2. idendifies that's an admin and provide him all results | 17:06 |
| bauzas | what are your preference ? | 17:06 |
| bauzas | s/are/is | 17:06 |
| bauzas | (sorry, a bit tired) | 17:06 |
| bauzas | DinaBelova: ? | 17:07 |
| DinaBelova | bauzas, sorry, needed to go for 5 mins :) | 17:09 |
| DinaBelova | I'm here now :) | 17:09 |
| DinaBelova | reading your messages | 17:09 |
| bauzas | no pb, I'm about to leave | 17:09 |
| bauzas | will work this evening | 17:09 |
| bauzas | I want to finish this patch | 17:09 |
| bauzas | and btw, the way I filter is not the best, IMHO | 17:10 |
| bauzas | but we're running out of time | 17:10 |
| DinaBelova | I think second one is now ok | 17:10 |
| *** exaggeration has quit IRC | 17:10 | |
| bauzas | ok, will use the policy then | 17:10 |
| DinaBelova | ok | 17:10 |
| bauzas | and check the is_admin policy | 17:11 |
| DinaBelova | bauzas, thank you for your efforts :) | 17:11 |
| bauzas | we all are on the same boat :) | 17:11 |
| bauzas | and I really *want* Climate to get its 0.1 soon | 17:12 |
| DinaBelova | :D | 17:12 |
| DinaBelova | you are not alone :) | 17:13 |
| *** bauzas has quit IRC | 17:13 | |
| *** chandankumar_ has quit IRC | 17:42 | |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 17:52 | |
| *** chandankumar_ has joined #openstack-climate | 18:03 | |
| *** SergeyLukjanov is now known as SergeyLukjanov_a | 18:24 | |
| *** SergeyLukjanov_a is now known as SergeyLukjanov_ | 18:25 | |
| openstackgerrit | A change was merged to stackforge/climate: Fix service user creating objects https://review.openstack.org/67801 | 18:39 |
| *** chandankumar_ has quit IRC | 18:44 | |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 19:02 | |
| *** SergeyLukjanov is now known as SergeyLukjanov_a | 19:31 | |
| *** SergeyLukjanov_a is now known as SergeyLukjanov_ | 19:32 | |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 19:35 | |
| *** malchevsky has joined #openstack-climate | 19:51 | |
| *** callow has joined #openstack-climate | 20:02 | |
| *** callow has quit IRC | 20:15 | |
| *** SergeyLukjanov is now known as SergeyLukjanov_ | 20:22 | |
| *** bauzas has joined #openstack-climate | 20:28 | |
| *** germesfinans has joined #openstack-climate | 20:43 | |
| *** germesfinans has quit IRC | 20:44 | |
| *** malchevsky has quit IRC | 20:57 | |
| *** DinaBelova is now known as DinaBelova_ | 21:00 | |
| *** sviblovo has joined #openstack-climate | 21:05 | |
| *** sviblovo has quit IRC | 21:17 | |
| *** fnmath has joined #openstack-climate | 21:23 | |
| *** cherok has joined #openstack-climate | 21:35 | |
| *** fnmath has quit IRC | 21:52 | |
| *** cherok has quit IRC | 22:16 | |
| openstackgerrit | Sylvain Bauza proposed a change to stackforge/climate: Filter DB queries per project_id https://review.openstack.org/59109 | 23:25 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!