Earlier  
Later
New Search
Posted Nick Remark
#openstack-sdks - 2019-09-17
15:03:43 gtema where basically the devstack installation itself failed
15:04:47 johnsom As did all of ours when OSC merged the breaking change.
15:05:08 gtema aaah, it was the OSC change, got it
15:06:37 johnsom Yeah, the client output for neutron ports significantly changed in a non-backwards compatible way. This broke our devstack plugin script.
15:06:53 gtema ok
15:07:40 johnsom We wrote a big awk script to handle both the old and new output formats to work around it.
15:08:40 gtema yeah, does not look like a perfect approach ;-)
15:16:42 frickler maybe for such complex data manipulation a python script that uses the sdk would be a better solution
17:09:17 openstackgerrit Merged openstack/openstacksdk master: Fix image create with tags https://review.opendev.org/682454
21:06:59 mordred frickler, johnsom: clarkb did some experiments with using sdk in a python script in devstack for some other things. looking at that octavia chunk, it might be a more pleasant set of things to do as an sdk script
21:07:41 johnsom mordred You are proposing changing the devstack plugins from bash to python?
21:08:09 mordred johnsom: not generally - but it might be easier to just add a python script that the bash plugin calls to do the interaction
21:08:24 mordred like - there's a bunch of bash you're having to do there
21:08:25 johnsom Would work... Not sure what all we would need to bring in for some of that other system stuff, or have a bunch of command calls.
21:09:50 johnsom It's a pretty simple awk script really. My guess is faster than starting up a python environment. Really it's just a bummer that the -f value output format changed such that we need to deal with both.
21:11:24 johnsom Would be an interesting exercise to compare our OSC calls to non-OSC calls. If it's OSC heavy then just changing the devstack plugin to python might make sense.
21:18:24 mordred johnsom: yeah - I think it would only really be a win if you were converting a bunch of calls into a single script - not just the awk line - a single awk invocation is almost certainly cheaper
21:19:03 mordred but a single python script vs 8 osc invocations *might* be a win. otoh - it adds another layer
21:19:25 johnsom Yep
22:30:35 lifeless starting python is slowish
22:30:39 lifeless importing python code is slowish
22:30:44 lifeless running python code is slowish
22:30:57 lifeless doing all three is about the same time as running 10 or so awks
#openstack-sdks - 2019-09-18
04:55:51 openstackgerrit Duc Truong proposed openstack/openstacksdk master: [WIP] Add clustering update_action https://review.opendev.org/682787
04:58:25 openstackgerrit Duc Truong proposed openstack/openstacksdk master: [WIP] Add clustering update_action https://review.opendev.org/682787
06:21:22 openstackgerrit inspurericzhang proposed openstack/shade master: fix "How To Contribute" url https://review.opendev.org/682801
06:34:37 openstackgerrit inspurericzhang proposed openstack/openstacksdk master: fix "How To Contribute" url https://review.opendev.org/682803
07:57:55 openstackgerrit Merged openstack/openstacksdk master: Build PDF docs https://review.opendev.org/601659
08:03:32 frickler a python script might be easier to understand, debug and test than awk, even if it is slower
08:53:11 openstackgerrit Merged openstack/openstacksdk master: Cleanup doc/source/conf.py https://review.opendev.org/679910
09:10:51 mordred yay the pdf docs patches actually landed
09:11:17 gtema yes, it took quite some dozens of rechecks
09:54:20 openstackgerrit inspurericzhang proposed openstack/shade master: fix "How To Contribute",bugs,review url https://review.opendev.org/682801
14:28:30 efried Can openstacksdk image do fancy things like download at this point?
14:29:41 efried https://docs.openstack.org/openstacksdk/latest/user/proxies/image_v2.html#openstack.image.v2._proxy.Proxy.download_image looks like it -- does this work?
14:31:38 openstackgerrit Bram proposed openstack/openstacksdk master: support restore from volume backup in cloud layer https://review.opendev.org/682909
15:41:14 openstackgerrit Duc Truong proposed openstack/openstacksdk master: Add clustering update_action https://review.opendev.org/682787
15:56:52 mriedem dtroyer: i'm +1 on this osc change https://review.opendev.org/#/c/673725/ but found a way to tighten up one of the tests which was a false positive before - i left the diff inline and was wondering if you'd want that in the same change or in a follow up?
15:57:15 mriedem i'm thinking follow up...
15:59:27 dtroyer mriedem: follow-up is fine, the usual worry is it never getting done, I trust you will see to that :)
15:59:38 mriedem i've got the patch locally, pushing
15:59:39 dtroyer er, see to it getting done
16:01:10 openstackgerrit Matt Riedemann proposed openstack/python-openstackclient master: Provide stderr in exception when check_parser fails https://review.opendev.org/682939
16:06:37 openstackgerrit Artem Goncharov proposed openstack/python-openstackclient master: WIP: Switch image to use SDK https://review.opendev.org/650374
17:24:35 openstackgerrit Merged openstack/openstacksdk master: baremetal-introspection: add manage_boot argument to start_introspection https://review.opendev.org/681225
#openstack-sdks - 2019-09-19
11:57:17 openstackgerrit Andreas Florath proposed openstack/python-openstackclient master: Fix osc-lib interface change: catch osc-lib Forbidden https://review.opendev.org/683118
11:58:52 openstackgerrit Andreas Florath proposed openstack/osc-lib master: Fix osc-lib interface change: throw osc-lib Forbidden https://review.opendev.org/683119
11:59:43 openstackgerrit Andreas Florath proposed openstack/python-openstackclient master: Fix osc-lib interface change: catch osc-lib Forbidden https://review.opendev.org/683118
16:00:10 elmiko API SIG office hour is now starting
16:00:25 mordred you're an office hour
16:01:09 elmiko haha
16:04:39 dtantsur o/
16:04:52 dtantsur I'm here for a short time since I'm still on a crazy post-relocation schedule
16:05:37 mordred dtantsur: I have missed things - where did you relocation?
16:05:50 dtantsur mordred: Düsseldorf (west Germany)
16:06:03 mordred dtantsur: oh wow
16:06:13 elmiko dtantsur: ack, hope the move/post-move are going well
16:06:40 dtantsur it goes so-so :) finding a flat is never fun in Germany
16:06:44 elmiko by way of business, i did start triaging our backlog of reviews
16:06:52 mordred dtantsur: one of these days I really want to come see the hanging train next door to you in Wuppertal
16:06:57 elmiko it's slow moving, but i think i can make some progress over the next few weeks
16:07:10 dtantsur mordred: come and see :) I haven't seen it as well yet
16:07:17 dtantsur great job elmiko!
16:07:23 elmiko hanging train sounds fun
16:07:41 dtantsur https://en.wikipedia.org/wiki/Wuppertal_Suspension_Railway
16:08:07 elmiko neat
16:43:44 openstackgerrit Michael Johnson proposed openstack/python-openstackclient master: Switch to using osc_lib.utils.tags https://review.opendev.org/662864
17:01:16 elmiko take care all o/
#openstack-sdks - 2019-09-20
06:46:47 openstackgerrit zhurong proposed openstack/openstacksdk master: Fix the wrong doc use oslo_conf param https://review.opendev.org/683296
16:26:49 openstackgerrit OpenStack Release Bot proposed openstack/openstacksdk master: Update master for stable/train https://review.opendev.org/683478
16:27:21 openstackgerrit OpenStack Release Bot proposed openstack/os-client-config master: Update master for stable/train https://review.opendev.org/683481
16:28:58 openstackgerrit OpenStack Release Bot proposed openstack/osc-lib master: Update master for stable/train https://review.opendev.org/683493
#openstack-sdks - 2019-09-23
03:48:50 openstackgerrit Steve Baker proposed openstack/openstacksdk master: Support vendor data in configdrive building https://review.opendev.org/683842
05:22:05 openstackgerrit Steve Baker proposed openstack/openstacksdk master: Support vendor data in configdrive building https://review.opendev.org/683842
08:10:27 openstackgerrit pengyuesheng proposed openstack/openstacksdk master: Update the constraints url https://review.opendev.org/683862
08:15:13 openstackgerrit pengyuesheng proposed openstack/osc-lib master: Update the constraints url https://review.opendev.org/683865
12:13:28 openstackgerrit Merged openstack/openstacksdk master: Update master for stable/train https://review.opendev.org/683478
12:35:41 openstackgerrit Dmitry Tantsur proposed openstack/openstacksdk master: Make the bifrost job non-voting https://review.opendev.org/683946
12:36:01 dtantsur mordred: pretty urgent ^^
12:54:11 enriquetaso dtroyer, o/ Cinder team would like to use the current cinder-pythonclient plugin inside the OSC. is any chance that the OSC support multiversion or accept some work in this area to support this approach.
12:55:02 dtantsur enriquetaso: not quite the answer to your question, but you can look at the ironicclient plugin, it does support microversion
12:55:21 dtantsur we've also made the mistakes you're probably about to make (like requiring explicit OS_VOLUME_API_MICROVERSION)
12:56:33 frickler dtantsur: do you have an example for that bifrost failure?
12:56:48 enriquetaso dtantsur++
12:57:15 enriquetaso dtantsur: Great, with this information I'll start working on the plugin
12:57:19 dtantsur frickler: any patch on SDK, I think.
12:57:29 dtantsur frickler: https://zuul.opendev.org/t/openstack/build/ce636fd8f50d4998ab2a5e838949fb39
13:00:44 frickler dtantsur: ah, thx, I didn't look closely enough. +3
13:06:04 enriquetaso mordred o/ ^ any thoughts? :)
13:29:46 dtroyer enriquetaso: I'm not sure exactly what you are asking, OSC _does_ use the current python-cinderclient, however any changes should consider switching to the SDK and using that to support microversions properly. Now if we just had a stable 1.0 release… /me ducks
13:32:03 dtantsur heh
13:33:44 enriquetaso dtroyer, oh
13:52:53 enriquetaso thanks dtroyer
15:38:01 openstackgerrit Merged openstack/cliff master: Stop wildcard importing argparse https://review.opendev.org/611889
16:53:03 openstackgerrit Merged openstack/openstacksdk master: Make the bifrost job non-voting https://review.opendev.org/683946
#openstack-sdks - 2019-09-24
09:16:36 openstackgerrit Vishakha Agarwal proposed openstack/keystoneauth master: Generate pdf documentation https://review.opendev.org/682272
11:38:46 mordred enriquetaso: I think dtroyer got you - but I agree with dtantsur and dtroyer - and I think the best outcome would be switching to sdk
11:42:09 enriquetaso mordred, thanks for answering me. This is my first time working with openstacksdk. Is openstack cli going to deprecate?
11:53:26 mordred oh no - openstack cli is the future. sdk is what we want to use for REST interactions with services instead of the per-service python-*client libraries
11:54:36 mordred but the idea is that we put shared needed logic for things into sdk - then it's available to python users, as well as other sdk consumers like ansible and salt, and we can build the osc commands from that same logic

Earlier   Later