| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-sdks - 2020-07-13 | |||
| 17:03:37 | openstackgerrit | Felix Yan proposed openstack/python-openstackclient master: Fix compatibility issue in 5.3 https://review.opendev.org/740730 | |
| 17:41:18 | openstackgerrit | Felix Yan proposed openstack/python-openstackclient master: Fix compatibility issue in 5.3 https://review.opendev.org/740730 | |
| 21:58:27 | TheJulia | so... I'm wondering if stevedore 3.1.0 is also problematic... | |
| 21:59:09 | TheJulia | http://paste.openstack.org/show/795899/ | |
| 21:59:48 | TheJulia | same error on https://374eb4aa1fb8d5a8fb5d-25632cacac12665a0cc8d10743d2803a.ssl.cf2.rackcdn.com/731180/16/gate/ironic-tempest-ipa-partition-uefi-pxe-grub2/45a0e01/job-output.txt logging stevedore 3.1.0 is being installed | |
| 22:07:44 | TheJulia | without a clear mind, it looks like openstackclient is going to need to change... ultimately. | |
| #openstack-sdks - 2020-07-14 | |||
| 03:25:21 | openstackgerrit | zhangboye proposed openstack/keystoneauth master: migrate testing to ubuntu focal https://review.opendev.org/740856 | |
| 05:37:18 | openstackgerrit | zhangboye proposed openstack/openstacksdk master: migrate testing to ubuntu focal https://review.opendev.org/740870 | |
| 12:26:39 | frickler | smcginnis: TheJulia: still seeing things fail on focal, too https://zuul.opendev.org/t/openstack/build/1f2b05ef79454adbaa9f284965663ecd | |
| 12:36:22 | openstackgerrit | Mark Chappell proposed openstack/ansible-collections-openstack master: Add support for setting the Flavor when creating a load balancer https://review.opendev.org/740948 | |
| 12:38:11 | openstackgerrit | Mark Chappell proposed openstack/ansible-collections-openstack master: Add support for setting the Flavor when creating a load balancer https://review.opendev.org/740948 | |
| 12:42:26 | mordred | TheJulia, frickler, smcginnis: https://review.opendev.org/#/c/739327/7/openstackclient/common/clientmanager.py - line 155 of clientmanager is explicitly trying to handle this AttributeError case ... why is the except not catching the execption? :( | |
| 12:47:37 | mordred | oh - wait - I'm dumb (still coffeeing) - looking further | |
| 12:47:40 | frickler | mordred: because the except path also does not work: AttributeError: 'EntryPoint' object has no attribute 'module' | |
| 12:48:18 | mordred | yeah - I'm now seeing that | |
| 12:48:22 | mordred | I'm not reproducing this locally yet | |
| 12:49:42 | frickler | mordred: I've only seen this on focal, might be py3.8 or 3.7 vs. 3.6 | |
| 12:50:07 | mordred | nod. I'll try 3.8 real quick | |
| 12:50:09 | mordred | (also wat?) | |
| 12:50:54 | mordred | like - why should this be different on 3.8 ... but I guess let's find out! | |
| 12:51:39 | mordred | ok. reproduced | |
| 12:51:51 | mordred | so - for another data point - it works locally on 3.7 but not on 3.8 | |
| 13:00:16 | smcginnis | There is a difference with 3.8. They added to the std lib. | |
| 13:00:41 | openstackgerrit | Monty Taylor proposed openstack/python-openstackclient master: Fix plugin loading for 3.8 https://review.opendev.org/740950 | |
| 13:00:49 | mordred | smcginnis: ^^ that is a bad fix but is a fix | |
| 13:01:46 | smcginnis | mordred: Felix has https://review.opendev.org/#/c/740730/ up. Should we update that one? | |
| 13:02:08 | smcginnis | mordred: pep8 error on line 168, BTW. | |
| 13:02:27 | mordred | oh - great. if felix's works, let's do that | |
| 13:02:28 | mordred | mine is more complex | |
| 13:02:57 | smcginnis | Actually looks like yours handles thing better. So he had part of the fix. | |
| 13:03:13 | smcginnis | Maybe just acknowledge and thank, but point to your more complete solution. | |
| 13:03:23 | mordred | I confirm felix's fix works | |
| 13:03:44 | mordred | smcginnis: no - I think his does what I tried to first - I didn't find "value" | |
| 13:04:03 | mordred | so I went with doing ep.load() instead - so I think his is actually better | |
| 13:04:09 | smcginnis | Cool. | |
| 13:04:31 | smcginnis | He also went through the trouble of creating a story and everything, so I would feel bad to just abandon it. ;) | |
| 13:04:57 | mordred | smcginnis: we might want to check topic:osc-performance for other places that do module vs module_name in a try/except | |
| 13:04:59 | mordred | they'll need this same fix | |
| 13:05:26 | mordred | smcginnis: I've +A'd felix's patch - I think we can cut a bugfix for osc as soon as it lands | |
| 13:05:32 | smcginnis | Want to abandon https://review.opendev.org/#/c/740950/ ? | |
| 13:05:38 | mordred | oh - you know - we should really add a release note | |
| 13:06:15 | TheJulia | ++ | |
| 13:06:33 | smcginnis | I think Doug was still working on some patches, but I'm sure it's not his top priority. | |
| 13:07:00 | openstackgerrit | Monty Taylor proposed openstack/python-openstackclient master: Fix compatibility issue in 5.3 https://review.opendev.org/740730 | |
| 13:34:50 | nightmare_unreal | hello I want to add more selectable fields in "openstack server list -c ". Currently columns are limited ('ID', 'Name', 'Status', 'Networks', 'Image', 'Flavor'). | |
| 13:35:07 | nightmare_unreal | I am assuming it's more related to cliff rather than openstackclient | |
| 14:14:19 | amotoki | nightmare_unreal: if you would like to specify more fields in -c, you also need --long option. Is this what you want? | |
| 14:24:06 | nightmare_unreal | amotoki: I know about --long but I was hoping is there a way to add columns iwithout specifying --project or --user | |
| 14:34:06 | amotoki | nightmare_unreal: originally you talked about -c option. how are --project/--user options related? | |
| 14:34:29 | amotoki | nightmare_unreal: do you want to see project/user information in the output of "openstack server list"? | |
| 15:27:06 | openstackgerrit | Merged openstack/openstacksdk master: Functional tests to explicitly use ML2/OVS https://review.opendev.org/740706 | |
| 15:27:08 | openstackgerrit | Merged openstack/openstacksdk master: Add block_storage find functions https://review.opendev.org/740303 | |
| 16:04:03 | openstackgerrit | Monty Taylor proposed openstack/js-openstack-lib master: [WIP] Getting Started documentation https://review.opendev.org/351875 | |
| 16:17:08 | openstackgerrit | Merged openstack/python-openstackclient master: Fix compatibility issue in 5.3 https://review.opendev.org/740730 | |
| 17:21:41 | TheJulia | does it make sense to cut an osc release? | |
| 17:22:28 | mordred | TheJulia: yup! I will do that right now | |
| 17:23:14 | TheJulia | <3 | |
| 17:23:42 | mordred | TheJulia, smcginnis: remote: https://review.opendev.org/741009 Release 5.3.1 of python-openstackclient | |
| 17:25:13 | TheJulia | Thanks mordred | |
| #openstack-sdks - 2020-07-15 | |||
| 07:36:14 | nightmare_unreal | amito: yes . I basically want openstack server list output with different columns selection | |
| 07:36:56 | nightmare_unreal | amotoki: i was hoping to get something like this : "openstack server list -c project_id " | |
| 07:42:12 | amotoki | nightmare_unreal: I thought something similar before. Is your point to specify a (non-default) field in -c option without --long option, right? | |
| 07:42:47 | amotoki | nightmare_unreal: "openstack server list -c project_id " only shows project IDs of servers, so I am a bit confused on what you would like to do. | |
| 07:44:24 | amotoki | nightmare_unreal: or would you like to add columns by -c option? it might be another discussion but if so we cannot change -c option behavior (as it affects existing users) | |
| 07:44:43 | nightmare_unreal | yes amotoki | |
| 07:44:57 | nightmare_unreal | exactly | |
| 07:45:03 | amotoki | nightmare_unreal: yes for which one? | |
| 07:45:09 | nightmare_unreal | you got the point 1st point | |
| 07:45:22 | nightmare_unreal | "I thought something similar before. Is your point to specify a (non-default) field in -c option without --long option, right?" | |
| 07:45:25 | nightmare_unreal | this amotoki ^ | |
| 07:45:57 | amotoki | nightmare_unreal: thanks for clarification. It sounds a good improvement. | |
| 07:46:27 | nightmare_unreal | yeah, that's where I am stuck amotoki . I am not sure where to start | |
| 07:46:41 | nightmare_unreal | should I look into cliff because it shows output or the CLI server.py file | |
| 07:46:47 | nightmare_unreal | where the data is generated | |
| 07:48:11 | nightmare_unreal | i have submitted story here : https://storyboard.openstack.org/#!/story/2007925 | |
| 07:51:21 | amotoki | nightmare_unreal: IIRC fields aree filtered at https://opendev.org/openstack/python-openstackclient/src/branch/master/openstackclient/compute/v2/server.py#L1539 | |
| 07:52:58 | amotoki | nightmare_unreal: "columns" are defined conditionally based on --long option https://opendev.org/openstack/python-openstackclient/src/branch/master/openstackclient/compute/v2/server.py#L1381-L1445 | |
| 07:54:03 | nightmare_unreal | yeah that's where I am confused. If we don't specify --long option how can we even get the output for the columns .because "openstack server list" has limited output | |
| 07:56:03 | amotoki | perhaps it needs some refactoring so that columns specified in -c option are picked up (including a map between a real field name and a display name) | |
| 07:57:14 | amotoki | we also need to consider the consistency across commands.... | |
| 07:57:22 | nightmare_unreal | yeah that too | |
| 07:57:42 | nightmare_unreal | thing is -c option is parsed by cliff | |
| 07:58:09 | nightmare_unreal | https://opendev.org/openstack/cliff/src/branch/master/cliff/display.py#L69 | |
| 08:02:15 | amotoki | nightmare_unreal: so the question is whether cliff app (like OSC command) can access options defined in cliff. | |
| 08:03:00 | nightmare_unreal | yeah , because i don't think -c is handled/accessed by osc cli . ' | |
| 08:03:01 | amotoki | nightmare_unreal: another approach is that cliff provides a convenient function to handle columns. | |
| 08:04:39 | nightmare_unreal | amotoki: not sure this is right approach but can we add argument to parse (-c ) in server.py and search the result based on that ? | |
| 08:05:44 | amotoki | nightmare_unreal: -c option is already parsed. I think you can access the parsed result by parsed_args.columns, can't you? | |
| 08:08:06 | nightmare_unreal | True https://opendev.org/openstack/cliff/src/branch/master/cliff/display.py#L98 but if the column doesn't exist in "openstack server list" or any other command . it will output no recognized column | |
| 08:09:53 | amotoki | nightmare_unreal: I didn't get your point. I think you need to adjust 'columns' in ListServer in server.py to take into account parsed_args.columns. | |
| 08:11:03 | nightmare_unreal | got it amotoki . I understand now. I was assuming the parsed_args.column was not accessible in server.py but it is | |
| 08:11:07 | nightmare_unreal | that makes sense | |
| 08:19:39 | amotoki | nightmare_unreal: :) | |
| 08:19:58 | nightmare_unreal | Thanks :) | |
| 13:12:56 | openstackgerrit | jayaditya gupta proposed openstack/python-openstackclient master: Add a few selectable fields to the "openstack server list" output https://review.opendev.org/741181 | |
| 14:10:23 | openstackgerrit | Lance Bragstad proposed openstack/python-openstackclient master: Bypass user and group verification in RemoveRole https://review.opendev.org/740380 | |
| 18:13:45 | openstackgerrit | Artom Lifshitz proposed openstack/openstacksdk master: WIP: Test and record keypair max microversion https://review.opendev.org/741295 | |
| 19:41:02 | openstackgerrit | Artom Lifshitz proposed openstack/openstacksdk master: Add tests for compute microversion 2.2 and 2.10 https://review.opendev.org/741295 | |
| 19:48:39 | openstackgerrit | Artom Lifshitz proposed openstack/openstacksdk master: Add tests for compute microversion 2.2 and 2.10 https://review.opendev.org/741295 | |
| 19:49:40 | openstackgerrit | Artom Lifshitz proposed openstack/openstacksdk master: Add tests for compute microversion 2.2 and 2.10 https://review.opendev.org/741295 | |