Earlier  
Later
New Search
Posted Nick Remark
#openstack-sdks - 2017-04-05
13:42:48 openstackgerrit David Rabel proposed openstack/python-openstackclient master: Add --network and --port to server create https://review.openstack.org/444924
14:06:40 openstackgerrit Brian Curtin proposed openstack/python-openstacksdk master: Add get_stack_* methods to documentation https://review.openstack.org/453646
14:25:25 openstackgerrit Brian Curtin proposed openstack/python-openstacksdk master: Clean up some errant doc warnings/errors https://review.openstack.org/453661
14:30:15 openstackgerrit Rodolfo Alonso Hernandez proposed openstack/python-openstackclient master: Add "qos-policy" option to "port create" & "port set" https://review.openstack.org/419236
14:44:25 reedip ZZelle_ yes hello
14:48:00 openstackgerrit David Rabel proposed openstack/python-openstackclient master: Add --network and --port to server create https://review.openstack.org/444924
14:49:13 reedip rabel_b1 : hi
14:49:56 rabel_b1 reedip: hi. arg, i just made it worse... :D
14:50:55 reedip rabel_b1 : I saw ZZelle_ 's comments. I agree, lets discuss this in detail here or on the bug . dtroyer 's input would alsio be great here :)
14:51:08 reedip s/bug/patch/
14:51:32 openstackgerrit David Rabel proposed openstack/python-openstackclient master: Add --network and --port to server create https://review.openstack.org/444924
14:52:06 openstackgerrit Merged openstack/python-openstacksdk master: Add get_stack_* methods to documentation https://review.openstack.org/453646
14:53:50 rabel_b1 i would prefer not to discuss that in the change. here is fine, launchpad also. but on the change there is already enough, if it's only about the implementation. so i would like to keep the "what to implement?" question separate
14:55:13 rabel_b1 also in the launchpad bug we already started talking about that.
14:55:38 ZZelle_ rabel_b1, reedip, it's a mess to jump between launchpad and review.o.o ... let's discuss in one place
14:55:49 reedip agreed ZZelle_
14:56:17 reedip but where ?? :)
14:56:48 ZZelle_ reedip, review.o.o :)
14:57:03 ZZelle_ reedip, the code won't go in launchpad :)
14:57:08 reedip rabel_b1 : is that ok with u ??
14:57:15 reedip ZZelle_ hehe.. yeah
14:58:19 rabel_b1 ZZelle_, reedip: ok, review then.
14:58:31 openstackgerrit Monty Taylor proposed openstack/os-client-config master: Add ability to pass in user_agent https://review.openstack.org/452550
14:58:48 rabel_b1 but maybe we can also talk about it here first
14:59:39 rabel_b1 since we three are here right now and we can post the outcome as comment on the review?
15:00:33 ankur-gupta-f4 Which patchset are you discussing
15:00:53 reedip ankur-gupta-f4 : I was waiting for you to chime in :)
15:01:03 reedip https://review.openstack.org/444924
15:01:38 ankur-gupta-f4 Oh yea. Fun
15:01:51 reedip :D
15:03:12 reedip I will be Back ... ( in a jiffy )
15:03:16 ZZelle_ rabel_b1, reedip, dtroyer, the main point to agree is the format
15:03:48 ZZelle_ rabel_b1, reedip, dtroyer, imo, there is no reason to have 3 options --nic, --network, --port which supports complex values
15:04:34 ZZelle_ rabel_b1, reedip, let's use --network --port for simplex cases: --network $network or --port $port
15:05:15 ZZelle_ rabel_b1, reedip, if you need a finer interface description then use --nic $key=$value,...
15:07:48 rabel_b1 ZZelle_, reedip, ankur-gupta-f4, dtroyer: sounds reasonable to me. i think i first had in mind that --nic could be deprecated in the future in favour of --network and --port. but since those are not going to implement the --nic auto and --nic none functionality either, we will probably stay with having three parameters.
15:08:19 reedip ZZelle_ ok, so you mean with --network and --port we do not need to specify nic
15:08:21 rabel_b1 so, since --nic will not be deprecated, i think you are right, ZZelle_.
15:08:30 ZZelle_ rabel_b1, what means --nic none or --nic auto?
15:08:54 reedip if that is so ZZelle_ , rabel_b1 , we need to be pretty thorough in the Help Message
15:09:16 rabel_b1 ZZelle_: afaiu --nic none means having a nic without a port connected to it and --nic auto means looking for a network that fits it needs. but i'm not sure here.
15:09:29 ZZelle_ reedip, yes, --network $network enables to specify a nic on a specific network, --port port a nic using a specific port and --nic is for other cases
15:09:40 ZZelle_ rabel_b1, ok
15:10:38 ankur-gupta-f4 If --nic will be deprecated then --network and --port need to fulfill all the options that --nic supports
15:10:42 ankur-gupta-f4 Imho
15:11:05 rabel_b1 ankur-gupta-f4: yes, but it very probably won't
15:11:17 reedip rabel_b1 : IIUC currently --nic should have default none if nothing is specified, you might need to handle that with --network and --port as well ( i.e. if no option is specified, treat --nic as none in the body )
15:11:31 reedip auto option should be with --nic only , not with --network/--port
15:11:55 reedip ankur-gupta-f4 ; it would be difficult to replacte the auto behavior with --network/--port
15:12:58 ankur-gupta-f4 It shouldn't be that complicated if both nic and network options are handled in similar ways
15:13:07 rabel_b1 yes, dtroyer said something similar: none and auto should only be available for --nic, not for --network/--port
15:14:11 reedip ankur-gupta-f4 : it shouldnt be, but ideally it need not be required if we have --nic :) anyways, I think rabel_b1 has reached a satisfying conclusion , thanks to ZZelle_ :)
15:15:44 ankur-gupta-f4 Okay
15:16:18 ankur-gupta-f4 Will review it later. Still seems like it is would be confusing for end user.
15:16:27 ankur-gupta-f4 Lets see how it turnsout
15:17:11 rabel_b1 ankur-gupta-f4, reedip, ZZelle_: so, everyone seems to be fine with that solution. i will implement that in a new patch set.
15:17:39 rabel_b1 reedip: i am not quite sure yet how to do the help message the best way.
15:18:51 openstackgerrit Sindhu Devale proposed openstack/python-openstackclient master: "floating ip set/unset port" for OSC https://review.openstack.org/383025
15:19:22 reedip rabel_b1 : hmm ... for help message, you can simply mention like what we have mentioned here
15:19:27 rabel_b1 ZZelle_: i use network= instead of net-id= in that place to distinguish between --nic and --network later for the error message. see line 610+
15:20:26 rabel_b1 ZZelle_: maybe it is better to change the error message to something more generic?
15:20:26 ZZelle_ rabel_b1, i will wait for the next PS as it's not really meaningful to review the current one more
15:21:20 rabel_b1 ok, i will first create a new patch set now, then we can discuss on that. :)
15:21:24 sindhu dtroyer: hi, https://review.openstack.org/#/c/383025/ is ready. Have added a detail commit message
15:24:05 dtroyer I don't mind having multiple ways to do things when they fit particular use cases, I think —port and —network do as long as they are simple options. in PS9 they are nearly aliases to —nic which I do not think adds much here.
15:28:05 openstackgerrit Sindhu Devale proposed openstack/python-openstackclient master: "floating ip set/unset port" for OSC https://review.openstack.org/383025
15:28:58 reedip dtroyer : missing your opinion on this : ^^ ( I am going to bug you a bit for this , hope you dont mind :) )
15:31:11 dtroyer reedip: you mean 383025?
15:31:20 reedip dtroyer : YES :)
15:31:33 dtroyer what specifically?
15:31:46 reedip well, its ready ..
15:32:14 dtroyer not quite: https://review.openstack.org/#/c/383025/74/doc/source/command-objects/floating-ip.rst@156
15:32:40 reedip ohh ...
15:32:56 reedip 3rd April , will look into it.. thanks
15:33:16 dtroyer also, what I was looking for with regard to the SDK is when that dependent review gets released.
15:33:32 dtroyer ie, it's merged, what is the next release expected to be?
15:33:42 dtroyer but that's just a bit of leg owrk to track down
15:34:53 reedip dtroyer: thats what sindhu has mentioned in the commit message and in the file
15:35:28 ankur-gupta-f4 reedip make note in commit saying 0.9.15 SDK release needed
15:35:40 dtroyer it only points to a (now merged) review. it is more helpful to say "is expected to be relased in 0.9.xx", but that's just a bit of research to figure out if that has happened
15:35:43 openstackgerrit Brian Curtin proposed openstack/python-openstacksdk master: Enable warnings_as_errors in doc enforcer https://review.openstack.org/453676
15:36:01 dtroyer you don't always know, that's fine
15:36:08 reedip dtroyer : ankur-gupta-f4 : ok...
15:37:04 reedip dtroyer: I think I and sindhu agree that that type of work could be tracked down by us... but yes mentioning a ToDO with the approporate SDK release is imp
15:37:43 briancurtin if you guys need an SDK release, we’re probably close to one. should probably wait on a couple of these load_balancer ones since they’re close to being done and they’re all related
15:38:31 sindhu reedip: for some reason I thought u addressed https://review.openstack.org/#/c/383025/74/doc/source/command-objects/floating-ip.rst@156, are u doing it or shall I do it
15:39:00 reedip sindhu : be my guest if you want to .. I was searching for the patch that merged on 3rd april :)
15:39:06 dtroyer thanks briancurtin. I think there are a couple of things in OSC that go with new/fixed SDK bits. We're cutting a release today or tomorrow for OSC and osc-lib so we're not really blocked on the SDK
15:40:14 sindhu reedip: oh ok
15:40:25 dtroyer sindhu, reedip: actually, it isn't just released, but that version of the SDK needs to be in the minimum in global requirements (requirements.txt) before we can assume it will be present
15:40:57 reedip dtroyer : yeah, thats what is bliocking us and is our requirement ( the SDK needs to be in ou requirements )
15:41:52 sindhu dtroyer: Oh ok ... but we don't have to be blocked by that now right?
15:41:58 ankur-gupta-f4 briancurtin: for the load balancer ones we are concerned over the functional tests not being run. And we haven't yet been able to verify whether the calls/functionally are working correctly yet. I don't know if you want to wait upon those to release. Else, we can push those all though and deal with fixing them later
15:42:44 briancurtin ankur-gupta-f4: yeah, actually it does look like the functional tests might take a bit so we could be waiting a while. i’ll take a look in about an hour and see if we have anything else necessary to get a release out
15:44:08 openstackgerrit Nakul Dahiwade proposed openstack/python-openstacksdk master: [WIP] Introduce Member for Octavia (loadbalancing) https://review.openstack.org/451488
15:44:30 openstackgerrit Nakul Dahiwade proposed openstack/python-openstacksdk master: Introduce Member for Octavia (loadbalancing) https://review.openstack.org/451488
15:44:31 ankur-gupta-f4 that seems best. I would prefer to hold off on merging the LBs till I/we can verify its working as expected locally before merging.
15:44:54 dtroyer sindhu: if an OSC review requires an SDK review to be present to work, yes, you must either handle both cases (with and without the SDK review present) or you must wait until the SDK release that includes that dependent review is in global-requirements.
15:45:10 reedip ankur-gupta-f4 : feel free to let me know if something needs to be done for octaviaclient
15:45:32 ankur-gupta-f4 yea review the initial. Figure out how to deal with that function that just "pass"es

Earlier   Later