| *** markvoelker_ has quit IRC | 01:34 | |
| *** markvoelker has joined #refstack | 01:38 | |
| openstackgerrit | megan guiney proposed openstack/refstack master: Subunit Data upload API https://review.openstack.org/530681 | 01:55 |
|---|---|---|
| openstackgerrit | Chandan Kumar proposed openstack/python-tempestconf master: Increase compute.build_timeout 600 https://review.openstack.org/569034 | 06:40 |
| openstackgerrit | Chandan Kumar proposed openstack/python-tempestconf master: Allow non admin users to upload image to glance https://review.openstack.org/568252 | 06:41 |
| *** kopecmartin has joined #refstack | 07:06 | |
| *** tosky has joined #refstack | 07:55 | |
| openstackgerrit | Chandan Kumar proposed openstack/python-tempestconf master: Allow non admin users to upload image to glance https://review.openstack.org/568252 | 08:32 |
| openstackgerrit | Chandan Kumar proposed openstack/python-tempestconf master: Allow non admin users to upload image to glance https://review.openstack.org/568252 | 08:32 |
| openstackgerrit | Chandan Kumar proposed openstack/python-tempestconf master: Allow non admin users to upload image to glance https://review.openstack.org/568252 | 08:35 |
| chandankumar | tosky: kopecmartin arxcruz ^^ exception updated | 08:36 |
| tosky | chandankumar: with unit test? | 08:36 |
| chandankumar | tosky: oh shit, I need to add unit test, I will do that | 08:36 |
| tosky | eh, I know, but it's for your own safety | 08:37 |
| kopecmartin | tosky, chandankumar arxcruz , it passed \o/ , please have a look https://review.openstack.org/#/c/562672/ | 08:41 |
| chandankumar | kopecmartin: I should have improve the unit test also for the same to consume it | 08:43 |
| chandankumar | instead of dependent on main method | 08:43 |
| tosky | kopecmartin: uhm, uhm, can't we really avoid a duplication of parsing cloud data? | 08:44 |
| tosky | the value returned by get_arg_parser, which means the value returned by parse_argument, already knows the cloud data | 08:46 |
| kopecmartin | tosky are you saying about this call: os_client_config.OpenStackConfig(), which is there twice? | 08:48 |
| kopecmartin | it's just an object instantiation , not parsing of a cloud | 08:49 |
| chandankumar | kopecmartin: is this https://github.com/openstack/python-tempestconf/blob/master/etc/default-overrides.conf#L117 taken care by tempestconf? | 08:54 |
| chandankumar | bootable=true | 08:54 |
| chandankumar | volume_services=true | 08:54 |
| kopecmartin | chandankumar, AFAIK no | 08:55 |
| chandankumar | kopecmartin: if it is possible by some magic then we can get it done | 08:56 |
| tosky | kopecmartin: I remember that you documented the way to use this new public API, was it in this review or just on IRC or I just dreamt about it? | 08:56 |
| tosky | like: call get_cloud_config, call config_tempest, etc | 08:57 |
| tosky | I'm still not totally convinced that another instance of os_client_config is needed; I thought that parsing the parameters could be isolated completely | 08:58 |
| kopecmartin | tosky, i get it this way, parse_arguments deals with CLI only .. get_cloud_creds needs arguments for querying the cloud in order to obtain its credentials | 08:59 |
| kopecmartin | the object is needed twice, because it's two scopes, and we need two independent methods from that library | 09:00 |
| kopecmartin | tosky, which new public API? :D | 09:00 |
| tosky | kopecmartin: this API :) | 09:02 |
| tosky | tempestconf for any user (starting with refstack) | 09:03 |
| kopecmartin | tosky, i have a coffee now :D, let's go one by one | 09:15 |
| kopecmartin | ok, so this patch https://review.openstack.org/#/c/562672/ | 09:15 |
| kopecmartin | as i explained above, i think those two methods have totally different usage, but both of them needs some method from os-.. * library, I think it's ok this way | 09:16 |
| tosky | ok, fine, I won't block that | 09:18 |
| kopecmartin | I'm still gonna need to edit the commit message there, because we don't have to be dependent on the turning the ceilometer off | 09:18 |
| kopecmartin | tosky, btw, do we still want to do it? | 09:19 |
| kopecmartin | turn the ceilometer off on packstack? | 09:19 |
| tosky | kopecmartin: I was hoping that it could save resources, but if it does not do anything, it's pointless | 09:19 |
| kopecmartin | tosky, it didn't solve the timeout issues, so there is enough resources which are allocated either way | 09:20 |
| kopecmartin | ok, I'll edit the commit message | 09:20 |
| openstackgerrit | Martin Kopec proposed openstack/python-tempestconf master: Make tempestconf easier to use as an library https://review.openstack.org/562672 | 09:21 |
| kopecmartin | arxcruz, thanks for review, I've just edited the commit message only, so you can +2 again :) https://review.openstack.org/#/c/562672/ | 09:22 |
| arxcruz | kopecmartin: second +2 in the same patch it's 2 beers | 09:22 |
| arxcruz | krusovice | 09:22 |
| tosky | even the choice! | 09:22 |
| kopecmartin | arxcruz, :D done | 09:23 |
| arxcruz | ;) | 09:23 |
| kopecmartin | but you have to call me out too, when you are in Brno next time | 09:23 |
| kopecmartin | arxcruz, fhubik says, that only u suka ;) | 09:24 |
| arxcruz | kopecmartin: done | 09:24 |
| kopecmartin | tosky, ok, this patch now https://review.openstack.org/#/c/568695/ | 09:28 |
| kopecmartin | you're saying it's fine, but we need to figure out a generic way, how to deal with multiple versions in the future (if we will have multiple versions for a service again) | 09:29 |
| kopecmartin | is that correct? | 09:29 |
| tosky | kopecmartin: yes; it works now, but it removes some more general code which could be improved for sure, but it may be needed (for example, I would split SERVICE_VERSIONS so that each service provides all the versions supported) | 09:32 |
| kopecmartin | ok, agree | 09:33 |
| kopecmartin | tosky, we can still have the SERVICE_VERSIONS containing only one version for a service right now | 09:34 |
| kopecmartin | but I'm thinking if there is a nicer way, how to do it | 09:34 |
| tosky | do you mean keep SERVICE_VERSIONS for now, but empty (as there are no more multiple versions), while thinking about a better solution? | 09:35 |
| kopecmartin | yes | 09:36 |
| tosky | that's a possibility to not block the rest of the content of the review (the code which checks for deprecated versions, which is good) | 09:37 |
| tosky | maybe a TODO near the definition of SERVICE_VERSIONS could help | 09:37 |
| kopecmartin | ok, I'll add it there, we will create a task for another sprint which will be dealing with this, i have a few ideas, but it requires more time | 09:38 |
| tosky | sure | 09:38 |
| kopecmartin | tosky, chandankumar one big question, which is blocking my mind, where/ how /what are we gonna backport? for which versions of openstack? | 09:39 |
| kopecmartin | we made a huge amount of changes | 09:39 |
| kopecmartin | are they gonna be packaged for older versions? | 09:39 |
| tosky | kopecmartin: if we are going to have this working as refstack-client, we will need to support at least the same version of openstack supported at the time of the integration | 09:42 |
| tosky | which may be a problem with the removal of old versions | 09:42 |
| tosky | checking for deprecated is fine, because that is discovered, but totally removing cinder v2, or identity v2, at this time, not so much | 09:44 |
| kopecmartin | tosky, this removing of old version is applicable only since 13 and higher | 09:44 |
| tosky | kopecmartin: s/13/queens/ :) | 09:44 |
| tosky | I think that refstack-client supports more than that | 09:45 |
| kopecmartin | tosky, 13 is consists of two characters, it's faster :D | 09:45 |
| kopecmartin | yes, and here I'm not sure, how it works for older versions , can we just backport only some of the commits? if yes, it's fine | 09:46 |
| *** xinliang has quit IRC | 09:46 | |
| tosky | backports which commits to which version of what? | 09:47 |
| tosky | at the time when refstack-client starts using tempestconf, that version of tempestconf (so future one) should support all versions of openstack that refstack-client supports (so likely older) | 09:48 |
| kopecmartin | commits of python-tempestconf, which i want to rpm packages for older versions of OpenStack | 09:49 |
| kopecmartin | oh, i see, so i can't chose which commits from the history I will backport, i can backport all commits , before a certain commit, or none | 09:49 |
| tosky | if you are talking about "which version of tempestconf can I use with RDO", we may decide (I think we discussed it with chandankumar) to still pin to 1.1 and create a 1.1 branch for bugfixes there, if the changes are too invasive | 09:51 |
| chandankumar | kopecmartin: tosky currently for ocata-1.1.3 and pike, queens - 1.1.4 is working | 09:55 |
| chandankumar | tosky: kopecmartin if we want to support older releases i mean pike, queens also we need to find what is broken and fix it | 09:56 |
| chandankumar | it might needs to keep hardcoded stuff to support previous releases | 09:56 |
| tosky | well, if we don't drop for old versions, as needed by refstack-client, it would help for using tempestconf directly on older versions | 09:57 |
| *** xinliang has joined #refstack | 09:58 | |
| chandankumar | kopecmartin: tempestconf is branchless, we donot suppport backport for that, a particular tag should satisy the need, but we can work on supporting older release also but we need to find out what is getting broken for previous releases | 09:59 |
| tosky | chandankumar: again, my question is: you wrote "we can work", but I think it is more "we must work", if refstack-client needs to support more releases | 10:00 |
| tosky | my point, more than my question | 10:00 |
| chandankumar | tosky: yup we must work but let me get some rdoci jobs to find the breakage from pike release onwards as ocata support is getting EOL soon | 10:01 |
| tosky | chandankumar: also for refstack-client? | 10:02 |
| tosky | chandankumar: if you can execute the last guidelines with the last refstack-client, even Newton is supported: https://refstack.openstack.org/#/guidelines | 10:03 |
| chandankumar | tosky: for refstack-client we can just run the tempest provided refstack tests for each release and find out what is broken | 10:03 |
| chandankumar | till newton? It would be too much costlier | 10:04 |
| tosky | chandankumar: well, then wait for mergining until the next version of the guidelines are released | 10:06 |
| tosky | if you select version "Next", it starts from Ocata | 10:07 |
| tosky | but again, is it true that you can run refstack-client/master and expect to execute all those tests? If it is, yes, we need to keep the compatibility | 10:07 |
| *** kopecmartin is now known as mkopec|lunch | 10:22 | |
| *** mkopec|lunch is now known as mkopec | 11:21 | |
| *** mkopec has quit IRC | 11:21 | |
| *** kopecmartin has joined #refstack | 11:22 | |
| openstackgerrit | Martin Kopec proposed openstack/python-tempestconf master: Filter deprecated versions out https://review.openstack.org/568695 | 11:32 |
| kopecmartin | tosky, chandankumar this way it shouldn't break anything ^^ | 11:33 |
| chandankumar | kopecmartin: you can proactively verify it https://trello.com/c/sOUB7PeX/749-enable-logging-for-tempest-container | 11:42 |
| openstackgerrit | Chandan Kumar proposed openstack/python-tempestconf master: Allow non admin users to upload image to glance https://review.openstack.org/568252 | 12:10 |
| *** tosky has quit IRC | 12:32 | |
| *** tosky has joined #refstack | 12:36 | |
| openstackgerrit | Martin Kopec proposed openstack/python-tempestconf master: Store volumev2 and v3 extensions as well https://review.openstack.org/568876 | 12:57 |
| kopecmartin | tosky, this is longer, but more generic solution .. would it be better this way? ^^ | 12:58 |
| tosky | kopecmartin: commented :) | 12:59 |
| tosky | thanks for working on it | 12:59 |
| openstackgerrit | Martin Kopec proposed openstack/python-tempestconf master: Store volumev2 and v3 extensions as well https://review.openstack.org/568876 | 13:12 |
| openstackgerrit | Chandan Kumar proposed openstack/python-tempestconf master: Allow non admin users to upload image to glance https://review.openstack.org/568252 | 13:49 |
| *** kopecmartin_ has joined #refstack | 13:57 | |
| *** kopecmartin has quit IRC | 14:00 | |
| *** kopecmartin_ has quit IRC | 14:20 | |
| *** kopecmartin_ has joined #refstack | 14:33 | |
| *** kopecmartin_ has quit IRC | 16:26 | |
| *** eglute has quit IRC | 22:48 | |
| *** eglute has joined #refstack | 22:49 | |
| *** tosky has quit IRC | 23:36 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!