Earlier  
Later
New Search
Posted Nick Remark
#openstack-sdks - 2019-08-09
18:17:30 Shrews hrm, seems so. wow, none of the object code dealing with metadata allows for this
18:26:11 Shrews also, the autocreate header we put on the objects has a different header value???
18:26:19 Shrews HOW IS ANYTHING WORKING?!?!?
18:26:21 Shrews :)
18:29:08 Shrews returned header is: x-object-meta-x-object-meta-x-sdk-autocreated
18:29:23 Shrews the header we save is: x-object-meta-x-sdk-autocreated
18:29:47 Shrews an extra "x-object-meta" is prepended somewhere
18:30:02 Shrews doesn't seem to happen on the checksum headers though
18:30:27 Shrews can open... worms everywhere
18:32:27 mordred Shrews: see - I look away for a few minutes and you discover that the entire world is broken
18:32:52 Shrews mordred: stop letting me do things
18:33:11 mordred Shrews: no- do more things! the worms - they're everywhere?!
18:34:07 Shrews serves me right for trying to do a real-world test on my object cleanup code
18:35:53 Shrews i guess this means we are also missing functional tests of all this code
18:36:24 efried Shrews: Would you please have a look at https://review.opendev.org/#/c/674675/ when you get a chance? TIA
18:36:55 efried mordred: I think your patch may get rid of the warnings, but IMO we still want that ^ yah?
18:47:38 Shrews efried_pto: yeah, i'm gonna tackle this swift problem first while i have it paged in
18:51:25 mordred efried_pto: yeah - might as well
19:23:11 Shrews AHA! found the header bug
19:31:40 openstackgerrit David Shrewsbury proposed openstack/openstacksdk master: Add option to lowercase object meta keys https://review.opendev.org/675682
19:31:41 openstackgerrit David Shrewsbury proposed openstack/openstacksdk master: bug: avoid unnecessary object meta prefix https://review.opendev.org/675683
19:31:41 openstackgerrit David Shrewsbury proposed openstack/openstacksdk master: bug: Consider case when comparing object values https://review.opendev.org/675684
19:31:49 Shrews mordred: enjoy all the things
19:32:19 Shrews mordred: two of those probably need tests
19:32:28 timburke Shrews, mordred fwiw, swift will title case header names going out, but those are case-insensitive per the rfc. i know on py2, something (mimetools? rfc822?) would lower-case them when reading the response, while py3 preserves the casing from the wire
19:32:56 timburke header values should preserve whatever casing you sent with the PUT/POST
19:33:04 Shrews timburke: well, looks like we get them back in camelcase, even though we put in lower
19:33:15 Shrews i'm not sure where or why that's happening
19:34:22 Shrews unless we camel case our debug output just to be weird? /me checks
19:35:55 openstackgerrit Merged openstack/openstacksdk master: Skip most service_description for unknown services https://review.opendev.org/675135
19:39:08 openstackgerrit David Shrewsbury proposed openstack/openstacksdk master: Retry large object manifest upload https://review.opendev.org/675134
19:39:10 mordred Shrews: so ...
19:39:40 mordred Shrews: the headers object on the requests.Response here: https://review.opendev.org/#/c/675682/1/openstack/cloud/_object_store.py ... should be of type CaseInsensitiveDict
19:40:25 mordred Shrews: that's super weird that you're having to lowercase them (although I totally believe you)
19:42:49 Shrews mordred: oh? i not sure that i've even heard of CaseInsensitiveDict. If that's the case, my failed dict.get() calls were probably just the extra prefix thing
19:43:17 mordred Shrews: it's an object from the requests library
19:43:30 Shrews mordred: YOU'RE an object from the requests library
19:43:31 mordred because http headers are theoretically case insensitive
19:43:42 mordred Shrews: some days I certainlky feel that way
19:44:11 Shrews mordred: ok. i'll get rid of the lowercase stuff then
19:47:21 openstackgerrit David Shrewsbury proposed openstack/openstacksdk master: bug: avoid unnecessary object meta prefix https://review.opendev.org/675683
19:47:22 openstackgerrit David Shrewsbury proposed openstack/openstacksdk master: Retry large object manifest upload https://review.opendev.org/675134
19:48:35 Shrews mordred: ok, that should be what we need now
20:04:12 mordred Shrews: that looks fantastic!
20:09:03 Shrews mordred: oh, oops. found a nit to fix in that
20:09:12 mordred doh
20:09:13 openstackgerrit David Shrewsbury proposed openstack/openstacksdk master: Retry large object manifest upload https://review.opendev.org/675134
20:10:06 Shrews just minor cleanup i forgot to do
23:30:21 openstackgerrit Merged openstack/openstacksdk master: Allow limiting Connection service_types from oslo.config https://review.opendev.org/674675
#openstack-sdks - 2019-08-10
06:16:46 openstackgerrit Shubham Potale proposed openstack/openstacksdk master: Add support for nfv-orchestration service https://review.opendev.org/675604
06:40:50 openstackgerrit Merged openstack/openstacksdk master: bug: avoid unnecessary object meta prefix https://review.opendev.org/675683
07:06:20 openstackgerrit Shubham Potale proposed openstack/python-openstackclient master: Add command for upload vnf package API https://review.opendev.org/675612
09:37:57 openstackgerrit Shubham Potale proposed openstack/openstacksdk master: Add support for nfv-orchestration service https://review.opendev.org/675604
#openstack-sdks - 2019-08-11
06:02:02 openstackgerrit Adrian Turjak proposed openstack/keystoneauth master: add support for auth_receipts and multi-method auth https://review.opendev.org/675049
#openstack-sdks - 2019-08-12
03:01:27 openstackgerrit Brin Zhang proposed openstack/python-openstackclient master: Add server migration list CLI https://review.opendev.org/675304
04:09:22 openstackgerrit Brin Zhang proposed openstack/python-openstackclient master: Add server migration list CLI https://review.opendev.org/675304
09:29:32 dtantsur mordred, morning, have a second for https://review.opendev.org/#/c/674298/ ?
09:32:28 frickler mordred: kmalloc: cmurphy: I finally found the location where ksc hardcodes using the admin endpoint: https://opendev.org/openstack/python-keystoneclient/src/branch/master/keystoneclient/httpclient.py#L251
09:35:11 frickler we can do a workaround in heat, but fixing in ksc would seem better to me in the long run. seems that a lot of tests depend on the current default though
10:41:03 openstackgerrit Bence Romsics proposed openstack/openstacksdk master: Add router add/remove route operations https://review.opendev.org/674324
10:55:00 openstackgerrit Bence Romsics proposed openstack/python-openstackclient master: Add command: router add/remove route --route https://review.opendev.org/674325
14:15:55 Shrews dtantsur: https://review.opendev.org/675134 would really help openstack infra avoid a pretty bad object leak we are seeing, if you have some time to review it
14:17:57 dtantsur Shrews: found a few minor issues, wanna fix them now or follow-up?
14:19:49 Shrews dtantsur: How does re-raising the exception lose the traceback?
14:20:15 dtantsur Shrews: when you do 'raise e', the traceback will start from this instruction. when you do 'raise', the traceback is reused from the original exception.
14:20:18 dtantsur python fun
14:20:41 Shrews dtantsur: ah. so should all of those just be "raise" then ?
14:21:12 dtantsur Shrews: likely yes. In most case it's an antipattern to do 'except Something as e... raise e'.
14:21:26 dtantsur (python 3 has 'raise from', but let's ignore it for now)
14:21:34 Shrews dtantsur: ok. lemme fix that up then
14:23:51 openstackgerrit David Shrewsbury proposed openstack/openstacksdk master: Retry large object manifest upload https://review.opendev.org/675134
14:24:26 Shrews dtantsur: ok, that look more sane?
14:24:53 dtantsur yep
14:25:27 Shrews dtantsur: thx for the review. will bug mordred when he has internet again
14:37:58 Shrews mordred: since you are back, can you re-examine 675134 por favor?
14:38:16 mordred Shrews: I can!
14:39:12 mordred Shrews: +A - woot
14:40:21 mordred dtantsur: wow, that's fun with user_data
14:40:37 dtantsur yeah..
14:45:16 efried dtantsur: I'm not convinced you've actually done anything there.
14:45:30 dtantsur where?
14:46:01 efried I'm playing around with it locally, but I think 't' vs 'b' only affects the data type you get *back*
14:46:15 efried I'm not sure it affects the data going in.
14:46:56 efried and since you're not returning anything here...
14:47:16 dtantsur I was under impression 't' vs 'b' affects both directions
14:47:58 dtantsur efried: http://paste.openstack.org/show/756340/
14:48:18 dtantsur (this is python 3, I'm not sure there is a difference in python 2)
14:49:17 efried okay, I was trying to come up with a repro like that --^
14:49:31 efried and yeah, I meant to do it in py3, but I'm using py2 instead.
14:49:46 efried so, I stand corrected
14:49:55 efried <waves hand like jedi> I was never here
14:50:26 dtantsur hehe
16:20:56 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Allow services to override proxy creation. https://review.opendev.org/674110
16:20:57 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Fix discovery cache sharing https://review.opendev.org/675121
16:20:57 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Add support for global_request_id https://review.opendev.org/674807
16:20:58 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Rationalize endpoint_for and get_endpoint_from_catalog https://review.opendev.org/675130
16:20:58 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Validate that connect_as connects as the project https://review.opendev.org/675178
16:20:59 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Replace catalog-v3.json with keystoneauth fixture https://review.opendev.org/675187
16:23:22 mordred dtantsur: I had to rebase the stack, so I went ahead and squashed the cloud_region rename patch into https://review.opendev.org/#/c/674807
16:24:31 openstackgerrit Merged openstack/openstacksdk master: baremetal.configdrive: tolerate user_data as a string https://review.opendev.org/674298

Earlier   Later