Tuesday, 2017-06-13

openstackgerritSamuel de Medeiros Queiroz proposed openstack-infra/shade master: De-client-ify Project Update  https://review.openstack.org/47362900:03
*** yolanda has quit IRC00:14
*** yolanda has joined #openstack-shade00:37
openstackgerritSamuel de Medeiros Queiroz proposed openstack-infra/shade master: De-client-ify Project Delete  https://review.openstack.org/47362801:23
openstackgerritSamuel de Medeiros Queiroz proposed openstack-infra/shade master: De-client-ify Project Create  https://review.openstack.org/47360901:24
openstackgerritSamuel de Medeiros Queiroz proposed openstack-infra/shade master: De-client-ify Project Update  https://review.openstack.org/47362901:24
*** yolanda has quit IRC01:37
*** gouthamr has quit IRC01:42
*** dfflanders has joined #openstack-shade01:47
*** gouthamr has joined #openstack-shade02:52
openstackgerritMonty Taylor proposed openstack-infra/shade master: RESTify create_server  https://review.openstack.org/47355503:06
openstackgerritMerged openstack-infra/shade master: Improve grant docs on when and how use domain arg  https://review.openstack.org/47358903:50
openstackgerritMerged openstack-infra/shade master: Don't remove top-container for stack and zone REST API calls  https://review.openstack.org/47351703:53
*** htaccess has joined #openstack-shade04:00
htaccesshello, commit ba0e9451e664743dbc5d24307ca5a3829ee3e91b released in shade==1.21.0 breaks ansible instance creation using os_server on our cloud04:03
htaccesshttps://paste.fedoraproject.org/paste/GadXSOVxLBY2xBp4Gu6bkA04:05
htaccessexpecting ports to have a created_at key seems reasonable but for whatever reason ours don't04:06
*** gkadam has joined #openstack-shade04:06
htaccesswhere is the right place to file a bug report?04:07
htaccesshttps://github.com/openstack-infra/shade does not seem to have issues enabled04:10
htaccessit would be nice to point the github readme at the openstack gerrit as github is higher in google search04:12
htaccessonly breaks instances that need a floating ip04:15
*** slaweq has joined #openstack-shade05:08
*** slaweq has quit IRC05:08
*** gouthamr has quit IRC05:23
*** slaweq has joined #openstack-shade05:36
*** slaweq has quit IRC05:41
*** yolanda has joined #openstack-shade05:54
*** ioggstream has joined #openstack-shade06:05
*** slaweq has joined #openstack-shade06:37
*** slaweq has quit IRC06:42
*** yolanda has quit IRC07:28
*** gkadam is now known as gkadam-afk07:51
*** gkadam-afk is now known as gkadam08:53
*** yfried has joined #openstack-shade09:02
*** yfried has quit IRC09:05
*** dfflanders has quit IRC09:16
*** cdent has joined #openstack-shade09:59
openstackgerritRosario Di Somma proposed openstack-infra/shade master: Don't remove top-container element for sec group REST API calls  https://review.openstack.org/47377010:04
fricklerhtaccess: https://storyboard.openstack.org/#!/project/76010:05
*** slaweq has joined #openstack-shade10:43
*** slaweq has quit IRC10:48
*** cdent has quit IRC10:51
*** nokes has joined #openstack-shade11:03
*** noshankus has quit IRC11:04
*** nokes is now known as noshankus11:04
*** noshankus has quit IRC11:14
Shrewshtaccess: looks like that's fixed in the latest version of shade11:28
Shrewsbut not yet released11:29
openstackgerritSamuel de Medeiros Queiroz proposed openstack-infra/shade master: Enforce API version in functional tests  https://review.openstack.org/47379511:32
samueldmqmorning shade11:32
*** cdent has joined #openstack-shade11:38
mordredhtaccess: yah - as Shrews says, it's fixed in master. we'll try to get a release cut today11:44
mordredShrews, samueldmq: morning!11:44
*** slaweq has joined #openstack-shade11:44
mordredShrews: the stack at https://review.openstack.org/#/c/473555/ passes tests now if you have a sec to look at it (create_server is kinda heavily used, so eyes on it would be nice)11:45
Shrewsmordred: sure, after i breakfast11:45
*** slaweq has quit IRC11:49
*** openstackgerrit has quit IRC12:18
*** cdent has quit IRC12:22
*** olaph has quit IRC12:27
*** cdent has joined #openstack-shade12:34
rodsmordred is this safe `'delete_on_termination': True`? I'm wondering if would be safer to explicitly let the user delete the volume12:36
*** slaweq has joined #openstack-shade12:45
mordredrods: it's a nova feature12:46
mordredrods: when you create a server using boot-from-volume, you can tell it to delete the volume when you delete the server - this is helpful if you're using bfv but you're doing it with a normal image workflow12:47
mordredso - "boot me a server from the ubuntu-xenial image but make me a volume out of the image first and boot using that"12:47
*** slaweq has quit IRC12:50
rodsmordred ya, I know the feature, my concern is that in our case is not an option, is just set to True. Worried someone could have a bad surprise12:52
mordredOH! that12:54
Shrewsmordred: default for delete_on_termination seems to be False12:54
Shrewsfwiw12:54
mordredsorry - this is a whole different thing12:54
mordredso when you boot from an _IMAGE_ normally but attach a volume at boot time, nova requires you to pass an entry into the block_device_mapping for the image12:55
mordredit's not actually a volume12:55
mordredu'destination_type': u'local', is the key12:55
mordrednovaclient does this behind the scenes for the user, so the REST translation is the first I learned of it :)12:56
mordredhttps://github.com/openstack/python-novaclient/blob/master/novaclient/v2/servers.py#L737-L74512:57
rodsmordred oh I see, thx12:58
Shrewsmordred: ugh, i forgot how much i hate these api docs12:59
*** slaweq has joined #openstack-shade13:01
mordredI also learned that apparently one is supposed to POST to /os-volumes_boot instead of /servers13:02
mordredeven though that's not in the API docs13:02
mordredin fact, I should file a bug on that13:02
*** slaweq has quit IRC13:06
Shrewsmordred: where is the server normalization happening now?13:07
mordredShrews: in _list_servers13:09
Shrewsmordred: but create_server() also returns a server13:10
Shrewsdict13:10
mordredShrews: yes - but it calls get_server_by_id first which also does the normalization13:10
Shrewsah, ok13:11
Shrewsmordred: didn't test it, but no glaring problems jump out13:12
mordredShrews: woot! thanks13:12
Shrewsand i'm not about to validate all of the rest api stuff  :)13:12
Shrewsbut it looks reasonable13:13
mordredShrews: that's what the tests are for13:13
Shrewsyup13:13
mordredShrews: fwiw, the functional tests do a very good job for us here13:13
Shrewsyup13:13
*** openstackgerrit has joined #openstack-shade13:20
openstackgerritMonty Taylor proposed openstack-infra/shade master: Convert get_server_by_id  https://review.openstack.org/47383813:20
openstackgerritMonty Taylor proposed openstack-infra/shade master: Convert delete server mocks to requests_mock  https://review.openstack.org/47383913:20
openstackgerritMonty Taylor proposed openstack-infra/shade master: Convert delete server calls to REST  https://review.openstack.org/47384013:20
mordredShrews: there's get_server_by_id and delete calls13:20
mordredShrews: only list remains now. I'm going to see if I cna get that done in the next 40 minutes or so and if so, get it landed and ina release- if not, we can just release what we have now to unbreak htaccess13:20
*** gouthamr has joined #openstack-shade13:36
*** cdent has quit IRC13:51
*** slaweq has joined #openstack-shade14:02
*** gkadam has quit IRC14:02
*** slaweq has quit IRC14:06
*** slaweq has joined #openstack-shade15:03
*** cdent has joined #openstack-shade15:04
*** slaweq has quit IRC15:07
*** slaweq has joined #openstack-shade15:24
*** slaweq has quit IRC15:26
*** slaweq has joined #openstack-shade15:47
*** slaweq has quit IRC15:55
*** cdent has quit IRC16:01
* morgan goes to look for a new phone16:11
morganphone was lost today :(16:12
openstackgerritSamuel de Medeiros Queiroz proposed openstack-infra/shade master: Enforce API version in functional tests  https://review.openstack.org/47379516:12
mordredmorgan: GROSS16:14
morganslipped walking the dog, heard phone skitter... and gone (not online anymore)16:16
morganso... in a drain? in water? who knows.16:16
morganRIP Pixel XL16:16
morganalso... lots of Ibprofen16:16
morgancause... ouch.16:17
morganmordred: i still have my Nexus 5x, i might just order a new sim for it16:17
morganand run with it for a bit (as underpowered as it is)16:17
mordredmorgan: fwiw, I did that with my one plus when I left my pixel in a cab in berlin a few weeks before the barcelona summit16:18
mordredmorgan: fi sim in non-fi phone is REALLY bad16:18
morganyeah the 5x is thankfully a Fi phone16:18
mordredoh right16:18
morganhehe16:18
mordredthat's much better16:18
morganright?!16:19
mordredjlk: speaking of ansible - I've got patches up leading up to https://review.openstack.org/#/c/473838 which RESTify create_server ... once they're in, I'm going to add support for network tags for yolanda16:31
mordredjlk: I'm pinging you because network tags are a microversion-protected feature (and will be our first) ... so I thought maybe we should think real quick about the 'right' way to expose that in ansible?16:31
jlkneat16:31
jlkI'll peruse the stack.16:32
mordredjlk: like, I'm thinking, always accept tag as input, but if it's given and the cloud doens't support it, return an error about microversion mismatch?16:32
jlkI have to page back in openstack knowledge bits.16:32
jlkare there some clouds that don't allow getting an instance by iD? Or should I be looking at a different patch in teh stack?16:33
mordredoh - sorry - I'm talking about hte next patches I'm going to write16:33
mordredjlk: that's the last patch in a stack of 5 that convert create_server and get_server_by_id to rest16:34
mordredthose are all good and shouldn't break nothing16:34
mordredjlk: the question comes in that basically, in general, shade isn't planning on *directly* exposing microversion things to its users. rather it will use the best microversion it knows how to use after negotiating such with the cloud in question16:34
jlkah16:35
mordredbut for some things, like "the user wants to express network tags" - the feature just flat doesn't work on older clouds16:35
mordredI _think_ we just treat that the same as any feature that may or may not work depending on cloud age, yeah?16:35
jlkyeah that seems like time to throw an error.16:35
mordredcool16:35
*** slaweq has joined #openstack-shade16:56
*** slaweq has quit IRC17:01
openstackgerritMonty Taylor proposed openstack-infra/shade master: Convert delete server calls to REST  https://review.openstack.org/47384017:04
*** yfried has joined #openstack-shade17:51
*** slaweq has joined #openstack-shade17:57
*** slaweq has quit IRC18:01
*** ioggstream has quit IRC18:15
*** cdent has joined #openstack-shade18:25
*** slaweq has joined #openstack-shade18:58
samueldmqmordred: how can I get the "final" url inside the shade adapter?19:02
samueldmqI guess it isnt always available in endpoint_override19:03
*** slaweq has quit IRC19:03
samueldmqre: https://review.openstack.org/#/c/47379519:03
*** yfried has quit IRC19:04
mordredsamueldmq: get_endpoint should always return the correct value19:04
openstackgerritSamuel de Medeiros Queiroz proposed openstack-infra/shade master: Enforce API version in functional tests  https://review.openstack.org/47379519:05
samueldmqmordred: thanks19:05
mordredsamueldmq: however, it's worth noting that self.operator_cloud.cloud_config.get_api_version('identity') will return the detected version19:06
mordredsamueldmq: if you look at it after the adapter has been constructed19:06
samueldmqmordred: so I can match self.cloud_config.get_api_version('identity') with what's in self.get_endpoint()19:06
mordredsamueldmq: we update the value there after discovery so that the rest of shade knows which logic path to use19:06
samueldmqfrom inside the adapter code19:06
mordredsamueldmq: you don't need to call get_endpoint at all - you can just check self.cloud_config.get_api_version('identity')19:07
mordredas long as you've referenced the adapter - so if you do "self.operator_cloud._identity_client" - that'll cause the adapter to be created which will update the api_version if needed19:07
mordredthen you can check self.cloud_config.get_api_version('identity')19:07
samueldmqmordred: yes, but what I want to do is to check no v3 calls are made in the legacy tests, for instance19:08
samueldmqfirst, I do that inside the request() method of the adapter19:08
samueldmqand I can check the version the test says (2.0) against self.cloud_config.get_api_version('identity')19:09
samueldmqsince we want to short-circuit if v3 is called in the -legacy gate19:10
samueldmqsee https://review.openstack.org/#/c/47379519:10
*** slaweq has joined #openstack-shade19:11
samueldmqmordred: ah I got it, if discovery gets v3 (even if I ask v2.0 because only v3 is available) it will update the version number anyways19:14
samueldmqso I can check against that19:14
mordredsamueldmq: yes indeed. and if that's broken a _ton_ of other stuff will break :)19:15
samueldmqneat19:16
samueldmqmordred: and self.cloud_config.get_api_version('identity') is '3' or '2.0' (str), correct?19:17
mordredsamueldmq: yes. well- it'll actually be "2" after discovery if we find 2 iirc ...19:17
mordredsamueldmq: self.cloud_config.get_api_version('identity')[0] is always safe to check against either "3" or "2"19:18
mordred:)19:18
samueldmq++19:29
*** cdent has quit IRC19:56
openstackgerritMonty Taylor proposed openstack-infra/shade master: Convert delete server mocks to requests_mock  https://review.openstack.org/47383920:27
openstackgerritMonty Taylor proposed openstack-infra/shade master: Convert delete server calls to REST  https://review.openstack.org/47384020:27
openstackgerritMonty Taylor proposed openstack-infra/shade master: Cleanup volumes in functional tests in parallel  https://review.openstack.org/47396920:27
openstackgerritMonty Taylor proposed openstack-infra/shade master: Update tests for server calls that aren't list  https://review.openstack.org/47397020:27
openstackgerritMonty Taylor proposed openstack-infra/shade master: Migrate non-list server interactions to REST  https://review.openstack.org/47397120:27
openstackgerritMonty Taylor proposed openstack-infra/shade master: Properly expand server dicts after rebuild and update  https://review.openstack.org/47397220:27
*** slaweq has quit IRC20:41
*** slaweq has joined #openstack-shade20:42
*** slaweq has quit IRC20:46
*** gouthamr has quit IRC21:03
*** slaweq has joined #openstack-shade21:20
*** slaweq has quit IRC21:24
*** gouthamr has joined #openstack-shade21:28
openstackgerritMonty Taylor proposed openstack-infra/shade master: Cleanup volumes in functional tests in parallel  https://review.openstack.org/47396921:37
*** slaweq has joined #openstack-shade22:33
*** slaweq has quit IRC22:38
*** slaweq has joined #openstack-shade23:34
*** slaweq has quit IRC23:38

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!