Posted	Nick	Remark
#openstack-sdks - 2020-05-25
20:21:06	openstackgerrit	Monty Taylor proposed openstack/js-openstack-lib master: Remove jsdoc https://review.opendev.org/730678
21:20:37	openstackgerrit	Steve Baker proposed openstack/keystoneauth master: Implement HTTP Basic client support in keystoneauth1 https://review.opendev.org/727562
21:53:40	openstackgerrit	Ghanshyam Mann proposed openstack/keystoneauth master: Add in-tree upper-constraints for py35 job https://review.opendev.org/730687
22:05:28	openstackgerrit	Ghanshyam Mann proposed openstack/keystoneauth master: Add in-tree upper-constraints for py35 job https://review.opendev.org/730687
22:17:37	openstackgerrit	Ghanshyam Mann proposed openstack/keystoneauth master: Add in-tree upper-constraints for py35 job https://review.opendev.org/730687
22:31:47	openstackgerrit	Ghanshyam Mann proposed openstack/keystoneauth master: Add in-tree upper-constraints for py35 job https://review.opendev.org/730687
22:41:41	openstackgerrit	Ghanshyam Mann proposed openstack/keystoneauth master: Add in-tree upper-constraints for py35 job https://review.opendev.org/730687
#openstack-sdks - 2020-05-26
00:13:19	openstackgerrit	Ghanshyam Mann proposed openstack/keystoneauth master: Add in-tree upper-constraints for py35 job https://review.opendev.org/730687
00:59:51	openstackgerrit	Ghanshyam Mann proposed openstack/keystoneauth master: Add in-tree upper-constraints for py35 job https://review.opendev.org/730687
04:31:19	openstackgerrit	subham rai proposed openstack/openstacksdk master: baremetal: support for volume connectors API https://review.opendev.org/718746
06:06:26	adriant	mordred, I may have yet another weird Swift issue/bug, but this time because of RGW!
06:07:16	adriant	and it's a bug both the sdk and swiftclient share :/
06:12:49	adriant	I think I'll bug the swift people directly for this one...
07:51:44	openstackgerrit	Andreas Jaeger proposed openstack/shade master: Remove publish-to-pypi template https://review.opendev.org/730722
08:34:31	openstackgerrit	Hervé Beraud proposed openstack/openstacksdk master: Refresh python versions https://review.opendev.org/730729
08:35:10	openstackgerrit	Hervé Beraud proposed openstack/openstacksdk master: Refresh python versions https://review.opendev.org/730729
08:36:33	openstackgerrit	Hervé Beraud proposed openstack/openstacksdk master: Refresh python versions https://review.opendev.org/730729
08:55:51	openstackgerrit	Daniel Bengtsson proposed openstack/osc-lib master: Stop configuring install_command in tox and stop use pip. https://review.opendev.org/694510
12:05:30	mordred	adriant: neat! is it actually a bug in RGW?
12:05:48	mordred	adriant: (happy for us to handle it in sdk for the users if it is)
12:59:30	adriant	mordred, nah, it's a url thing that is kind of awkward to solve nicely
12:59:39	mordred	adriant: yay
12:59:51	mordred	my favorite things
13:00:02	adriant	yep...
13:00:37	adriant	it's just the handling of how the '/info' url is built from the entry in the keystone catalog
13:01:41	adriant	because by default as documented by ceph peeps you'd have the 'Swift' (RGW) endpoing in keystone as http://radosgw.example.com:8080/swift/v1
13:02:23	adriant	while the info endpoint under RGW is actually at: http://radosgw.example.com:8080/swift/info
13:02:33	adriant	and under normal swift is at http://radosgw.example.com:8080/info
13:03:00	adriant	so the standard sane logic is to strip back to the base url and just add '/info' which... doesn't work for RGW
13:03:14	adriant	unless you deploy it without the '/swift/' part somehow sanely
13:03:46	adriant	and we just happen to write logic that relies on the 'info' API... and is now busted on our only RGW cloud...
13:15:45	mordred	adriant: that's ... honestly very broken
13:35:23	dtantsur	omg
13:45:36	mordred	it's especially fun since swift doesn't have version discovery
13:45:53	mordred	adriant: this is going to be a stupid suggestion ...
13:46:20	dtantsur	try all the URLs \o/
13:46:31	mordred	adriant: but what if we pop things from the base url ... unless url_parts[-1] == 'swift'
13:48:04	mordred	a "swift" isn't likely to be legitimagely in the url path in a way that _should_ be stripped back
13:48:21	mordred	dtantsur: yup
13:48:29	mordred	dtantsur: I'm reminded of similar code in keystoneauth
13:48:29	dtantsur	mordred: unless you have an object or a container named "swift" :)
13:48:32	dtantsur	heh
13:48:42	mordred	dtantsur: well - but no - those wouldn't be in teh keystone catalog url ;)
13:48:51	dtantsur	I hope :D
13:49:09	mordred	and even if you had a project id of "swift" - swift puts project id in the url with an AUTH_ prefix
13:49:31	mordred	so even being crazy I can't come up with any circumstances in which a "swift" in a keystone catalog url should be stripped back
13:49:47	dtantsur	fair enough
13:51:41	openstackgerrit	Sagi Shnaidman proposed openstack/ansible-collections-openstack master: Redesign OpenstackModule class https://review.opendev.org/722283
14:56:35	openstackgerrit	Sagi Shnaidman proposed openstack/ansible-collections-openstack master: Redesign OpenstackModule class https://review.opendev.org/722283
17:21:35	timburke	mordred, swift *usually* has a prefix like AUTH_ in account names ;-) like so many things, it's configurable: https://github.com/openstack/swift/blob/2.25.0/etc/proxy-server.conf-sample#L414-L419
17:22:46	openstackgerrit	subham rai proposed openstack/openstacksdk master: baremetal: support for volume connectors API https://review.opendev.org/718746
17:29:26	openstackgerrit	Merged openstack/openstacksdk master: Refresh python versions https://review.opendev.org/730729
17:29:28	mordred	timburke: \o/
17:33:58	timburke	fwiw, though, it should be safe to just always pop off two path components from the (rstrip('/')ed) url from the catalog
17:39:12	mordred	timburke: apparently the RGW people put something different into the catalog - http://radosgw.example.com:8080/swift/v1 - which just seems incorrect
17:41:21	mordred	yeah - just looked at a cloud I know to be using rgw instead of swift, and they do not just have /v1 in the catalog
17:42:06	mordred	adriant: I think those ceph docs are just wrong. I thnk you want ttp://radosgw.example.com:8080/swift/v1/{ project_id } in the catalog
17:42:22	mordred	adriant: that's what I'm seeing from working clouds that are using rgw
18:57:22	openstackgerrit	Sagi Shnaidman proposed openstack/ansible-collections-openstack master: Add setup.py for install with pip https://review.opendev.org/730400
20:03:19	openstackgerrit	Hervé Beraud proposed openstack/os-client-config master: Cap jsonschema 3.2.0 as the minimal version https://review.opendev.org/730963
21:33:13	openstackgerrit	Sagi Shnaidman proposed openstack/ansible-collections-openstack master: Add setup.py for install with pip https://review.opendev.org/730400
23:53:56	openstackgerrit	João Vale proposed openstack/openstacksdk master: Fix deleting stacks by id when waiting for result https://review.opendev.org/717423
#openstack-sdks - 2020-05-27
00:05:08	openstackgerrit	João Vale proposed openstack/openstacksdk master: Fix deleting stacks by id when waiting for result https://review.opendev.org/717423
00:05:08	openstackgerrit	João Vale proposed openstack/openstacksdk master: Fix deleting stacks by id when waiting for result https://review.opendev.org/717423
09:19:24	openstackgerrit	Pavlo Shchelokovskyy proposed openstack/python-openstackclient stable/ussuri: Resolve PEP8 https://review.opendev.org/731122
09:48:41	openstackgerrit	Sagi Shnaidman proposed openstack/ansible-collections-openstack master: Add setup.py for install with pip https://review.opendev.org/730400
09:59:55	openstackgerrit	Sagi Shnaidman proposed openstack/ansible-collections-openstack master: Add setup.py for install with pip https://review.opendev.org/730400
11:05:49	openstackgerrit	Sagi Shnaidman proposed openstack/ansible-collections-openstack master: Add setup.py for install with pip https://review.opendev.org/730400
12:23:56	openstackgerrit	Sagi Shnaidman proposed openstack/ansible-collections-openstack master: Add setup.py for install with pip https://review.opendev.org/730400
12:37:56	openstackgerrit	Pavlo Shchelokovskyy proposed openstack/keystoneauth stable/ussuri: Fix E741 pep8 failure https://review.opendev.org/731188
14:19:31	openstackgerrit	wu.chunyang proposed openstack/cliff master: Use unittest.mock instead of third party mock https://review.opendev.org/731222
14:46:26	openstackgerrit	Sagi Shnaidman proposed openstack/ansible-collections-openstack master: Add setup.py for install with pip https://review.opendev.org/730400
15:34:15	openstackgerrit	Sean McGinnis proposed openstack/python-openstackclient master: Add py38 package metadata https://review.opendev.org/730227
15:59:37	openstackgerrit	Dmitry Tantsur proposed openstack/openstacksdk master: baremetal: use proxy methods in unregister_machine https://review.opendev.org/731255
17:35:05	openstackgerrit	Merged openstack/cliff master: Use unittest.mock instead of third party mock https://review.opendev.org/731222
17:38:10	openstackgerrit	Merged openstack/cliff master: Add Python3 victoria unit tests https://review.opendev.org/719014
17:40:39	openstackgerrit	Sagi Shnaidman proposed openstack/ansible-collections-openstack master: Add setup.py for install with pip https://review.opendev.org/730400
19:12:41	openstackgerrit	Sagi Shnaidman proposed openstack/ansible-collections-openstack master: Add setup.py for install with pip https://review.opendev.org/730400
23:24:52	adriant	mordred: We likely need a similar fix in the SDK to this swiftclient bug: https://review.opendev.org/#/c/724247/1 as I cna confirm the bug is also present in the openstacksdk when talking to a 'Swift' cluster that's really RGW
#openstack-sdks - 2020-05-28
01:39:18	openstackgerrit	Sagi Shnaidman proposed openstack/ansible-collections-openstack master: Add setup.py for install with pip https://review.opendev.org/730400
03:22:22	openstackgerrit	Pete Zaitcev proposed openstack/python-openstackclient master: Allow to list containers of other accounts https://review.opendev.org/731395
08:24:49	nightmare_unreal	stephenfin: hello, this one is waiting for a long time now . When you get time can you please review it : https://review.opendev.org/#/c/713881/
08:24:49	nightmare_unreal	Thanks
09:13:51	alistarle	Hi, this is business hour for SDK right ?
09:15:06	alistarle	Can you have a look at this patch https://review.opendev.org/#/c/726621/ ? It is a small one, but a dependency for a more global one about implementation of keystone unified limit in glance/nova
11:15:20	nightmare_unreal	hello can someone leave their insight on this : https://storyboard.openstack.org/#!/story/2007727 .
11:38:02	dtantsur	lack of proper microversioning support is one of the known OSC problems
11:38:48	nightmare_unreal	so how should I proceed/look if I want to fix/work on that ?
11:39:20	dtantsur	nightmare_unreal: attend the PTG next week, convince people it's a priority, convince people "just specify the microversions" is not an adequate answer
11:39:30	dtantsur	sorry, I don't have an easy answer
11:39:51	nightmare_unreal	Okay :) thanks.
11:40:09	dtantsur	I hope when OSC fully switches to openstacksdk, the problem will go away
11:40:12	dtantsur	not 100% sure though
11:40:37	nightmare_unreal	1 more thing on opendev there are 2 osc python-openstackclient and openstackclient
11:40:44	dtantsur	in the ironic plugin we went ahead and hacked together semi-decent microversion support
11:40:58	dtantsur	nightmare_unreal: ugh. let mordred know.
11:41:44	nightmare_unreal	Okay dtantsur . I was thinking one is old and just maintained for sometime
11:42:18	dtantsur	I'm not sure, better ask him
11:42:21	gtema	python-openstackclient is like a "core" or the OSC, while openstackclient is a bunch of packages included. Technically "openstackclient" includes "python-openstackclient" and lots of additional "plugins"
11:43:15	nightmare_unreal	okay now I understand

Earlier   Later