| *** gyee has quit IRC | 00:30 | |
| renich | So, I've been trying to pinpoint where ssbench decides to use /v3/tokens and can't find it. | 00:39 |
|---|---|---|
| zaitcev_ | keystoneclient does it, I think. Although maybe it's swiftclient. And the algorithm is absurdly complicated and was updated several times. | 00:53 |
| zaitcev_ | Instead of letting swiftclient to default to legacy, they always try to make it default to "current" keystone. So it was v2 for a few years, then v3. But they want old configurations continue working when someone updates. So, the code tries to guess who old the Keystone is, mostly by looking at the endpoint path. | 00:55 |
| zaitcev_ | Of course it backfires. | 00:56 |
| *** tkajinam_ has joined #openstack-swift | 00:58 | |
| *** tkajinam has quit IRC | 01:01 | |
| renich | zaitcev_: OK... no idea where to fix that... man, this is hard... | 01:02 |
| zaitcev_ | renich: "``ssbench-master`` supports all the same authentication arguments, with similar semantics, as `python-swiftclient`_'s command-line tool, ``swift``." | 01:05 |
| zaitcev_ | sounds like it should be trivial, just give it -V 2 and done | 01:06 |
| renich | zaitcev_: yes, but it doesn't work when I want to use v3 auth | 01:08 |
| renich | zaitcev_: and, since I am using Rocky, I have no v2 | 01:08 |
| zaitcev_ | renich, I don't know what "yes, but" is about it. If -V is provided, there's no automatic discovery anymore, so just give it the right -A and it should be there. Well, maybe /tokens appended at the worst, but probably not even that. Still does not work? | 01:10 |
| renich | zaitcev_: let me try | 01:10 |
| renich | zaitcev_: http://paste.openstack.org/show/747667/ | 01:13 |
| renich | It's trying to find http://localhost:5000/v3/tokens and that endpoint doesn't exist. It has to be /v3/auth/tokens | 01:14 |
| zaitcev_ | Well then use -A http://localhost:5000/v3/auth - does that work? | 01:18 |
| renich | OK, will try | 01:19 |
| renich | zaitcev_: I got this: 2019-03-13 01:32:29,407:DEBUG:Making authentication request to http://localhost:5000/v3/auth/v2.0/tokens | 01:33 |
| renich | I mean, it tries authentication in that endpoint. | 01:33 |
| zaitcev_ | well, the client is not what's converting the args these days, so maybe it needs --os-auth-version=3 | 01:35 |
| zaitcev_ | holy shit | 01:36 |
| zaitcev_ | ssbench bundles its own client | 01:36 |
| zaitcev_ | which does not support v3 period | 01:37 |
| renich | zaitcev_: yep. | 01:38 |
| zaitcev_ | https://github.com/swiftstack/ssbench/blob/master/ssbench/swift_client.py#L296 | 01:38 |
| renich | zaitcev_: is there a way to re-enable v2 in rocky? | 01:38 |
| renich | zaitcev_: or, maybe, I can tweak that file? | 01:39 |
| renich | add v3 support or something? | 01:39 |
| zaitcev_ | Well, the bundled client looks like a clean clone of the normal client, so it should be trivial to hack over. | 01:39 |
| * renich isn't a python wizard, btw | 01:39 | |
| zaitcev_ | rm ssbench/swift_client.py seems like a good first step ... :-) | 01:39 |
| renich | lol | 01:39 |
| zaitcev_ | no, really | 01:39 |
| zaitcev_ | then when you run it, you see where imports are | 01:40 |
| renich | I, actually, am importing the native client (notmyname helped me do that) | 01:40 |
| zaitcev_ | oh | 01:40 |
| renich | #import ssbench.swift_client as client | 01:40 |
| renich | from swiftclient import client | 01:40 |
| renich | And I had to hardcode some timeouts | 01:40 |
| zaitcev_ | In that case, did --os-auth-version help? | 01:40 |
| renich | let me see. I am using -V 3 | 01:40 |
| renich | ssbench-master: error: unrecognized arguments: --os-auth-version 3 | 01:41 |
| renich | I will have to add it to line 327 in ssbench-master, right? | 01:42 |
| zaitcev_ | weird. Seems like it should work... https://github.com/swiftstack/ssbench/blob/master/bin/ssbench-master#L329 | 01:43 |
| zaitcev_ | I suppose one can do ST_AUTH_VERSION=3 ssbench-master and see what happens | 01:44 |
| renich | OK, let me try | 01:45 |
| renich | It's trying /v3/tokens still. It should try /v3/auth/tokens. Making authentication request to http://localhost:5000/v3/tokens | 01:46 |
| zaitcev_ | But this is with ST_AUTH=http://localhost:5000/v3/auth , right? | 01:46 |
| zaitcev_ | And you need to remove those additional command line arguments, just incase, if you use the variables. | 01:47 |
| *** kei-ichi has joined #openstack-swift | 02:09 | |
| *** psachin has joined #openstack-swift | 03:03 | |
| *** renich has quit IRC | 03:14 | |
| *** renich has joined #openstack-swift | 03:31 | |
| openstackgerrit | Matthew Oliver proposed openstack/swift master: Allow multple keystoneauths in the pipeline https://review.openstack.org/593977 | 03:35 |
| mattoliverau | Just a rebase | 03:35 |
| *** itlinux has quit IRC | 03:41 | |
| *** kei-ichi has quit IRC | 04:00 | |
| *** mahatic has joined #openstack-swift | 04:44 | |
| *** ChanServ sets mode: +v mahatic | 04:44 | |
| kota_ | oh, multiple keystoneauth | 05:25 |
| * kota_ found a typo "multple" -> "multiple" | 05:25 | |
| zaitcev_ | heh | 05:26 |
| *** zaitcev_ is now known as zaitcev | 05:26 | |
| openstackgerrit | Matthew Oliver proposed openstack/swift master: Add a basic partition diffing tool https://review.openstack.org/539466 | 05:41 |
| mattoliverau | kota_: ta :) | 05:41 |
| mattoliverau | ^ these are some old patches I thought I'd rebase while waiting for a cloud to build :) | 05:42 |
| mattoliverau | This last one I'm looking at making sure it's py3 compatible. Which it is mostly. But failed to open up rings in py3.. but otherwise works. I.e you can compare builders. Just not rings to rings or rings to builders. | 05:43 |
| *** pcaruana has joined #openstack-swift | 06:19 | |
| *** e0ne has joined #openstack-swift | 06:40 | |
| *** tonyb is now known as tonyb_gone | 07:20 | |
| *** tonyb_gone is now known as tonyb | 07:21 | |
| *** mikecmpb_ has quit IRC | 07:31 | |
| *** e0ne has quit IRC | 07:31 | |
| *** gkadam has joined #openstack-swift | 07:48 | |
| *** hseipp has joined #openstack-swift | 07:48 | |
| *** e0ne has joined #openstack-swift | 07:51 | |
| *** rcernin has quit IRC | 08:09 | |
| *** ccamacho has joined #openstack-swift | 08:23 | |
| *** e0ne has quit IRC | 08:36 | |
| alecuyer | kota_: sorry for the late reply, I was pulled to other things this week. I don't know what "dsvm" is, I will check and read your patch | 08:59 |
| kota_ | alecuyer: o/ | 09:00 |
| kota_ | dsvm is the shorter term of DevStack Virtual Machine | 09:00 |
| kota_ | then, the dsvm has been used for testing to keystone authentication for Swift. | 09:01 |
| kota_ | tempest also runs... maybe? | 09:01 |
| alecuyer | kota_: Thanks. I'll read the job logs to see what runs | 09:02 |
| *** mikecmpbll has joined #openstack-swift | 09:03 | |
| kota_ | alecuyer: thx | 09:03 |
| kota_ | the benefit of the patch is to make the dsvm gate appears on feature/losf branch | 09:03 |
| kota_ | that was missed even it runs on the master. | 09:03 |
| alecuyer | Oh I see, ok | 09:04 |
| alecuyer | I see a few "tempest" lines in job-output.txt but not sure it ran | 09:07 |
| alecuyer | http://logs.openstack.org/45/642645/11/check/swift-dsvm-functional/6a0046d/job-output.txt.gz#_2019-03-12_09_22_02_790285 | TASK [Link post-devstack tempest.log] | 09:07 |
| alecuyer | http://logs.openstack.org/45/642645/11/check/swift-dsvm-functional/6a0046d/job-output.txt.gz#_2019-03-12_09_22_02_851628 | controller | skipping: Conditional result was False | 09:07 |
| *** kukacz has quit IRC | 09:15 | |
| *** kukacz has joined #openstack-swift | 09:17 | |
| kota_ | let me check | 09:19 |
| kota_ | alecuyer: it looks like no tempest runs against to the swift in the dsvm, http://logs.openstack.org/45/642645/11/check/swift-dsvm-functional/6a0046d/testr_results.html.gz | 09:21 |
| kota_ | it looks just to use dsvm to setup swift + keystone environment | 09:21 |
| *** tkajinam__ has joined #openstack-swift | 09:22 | |
| *** FlorianFa has joined #openstack-swift | 09:23 | |
| kota_ | tools/playbooks/dsvm/run.yaml calls only swift functional tests so that I'm not assume tempest runs in the dir. | 09:23 |
| *** tkajinam_ has quit IRC | 09:25 | |
| kota_ | I think it's same with the Swift master, I'm not sure if it's nice or not but probably we don't have to worry about that "We may ought to have tempest in the losf branch" because if we need it, absolutely we need the gate on both the master and the feature/losf branch. | 09:25 |
| *** mikecmpbll has quit IRC | 09:35 | |
| alecuyer | kota_: thanks for the explanation. Something else I have to check, is the functests setup, I don't think they run with the losf diskfile (the default diskfile being the regular diskfile) | 09:36 |
| kota_ | alecuyer: yes. that's what I'd like to go next :) | 09:36 |
| kota_ | not yet started tho | 09:37 |
| kota_ | probably, we could have 1. create setup package of golang code in the functional gate, then, 2. enable losf diskfile as the backend for functional test config. | 09:38 |
| kota_ | probably it's a new gate job other than current jobs. | 09:38 |
| alecuyer | yes, for 2. we also need to start the index server ervice | 09:39 |
| alecuyer | s | 09:39 |
| kota_ | yes it is. | 09:39 |
| kota_ | those are the next thing we have to do in the losf branch. | 09:40 |
| alecuyer | ok | 09:41 |
| kota_ | after we setup the gate jobs, we could be able to change some internal design because the robust gate job keeps "we don't break anything in the Swift API level" | 09:41 |
| kota_ | that's my idea. | 09:42 |
| alecuyer | Yes that sounds good. Im not sure how to add that new job, but if there is anything I can do, let me know. (I intended to keep working on the grpc/http change, as soon as I can) | 09:44 |
| kota_ | thx alecuyer, the entry point is https://github.com/openstack/swift/blob/master/.zuul.yaml | 09:45 |
| kota_ | that's the definition of the gate jobs | 09:45 |
| kota_ | the jobs are listed https://github.com/openstack/swift/blob/master/.zuul.yaml#L376-L414 | 09:46 |
| kota_ | then, each item is detailed in the yaml | 09:46 |
| kota_ | some of items uses extra ansible playbook to setup the environment in the vm. the playbooks are gathered in the https://github.com/openstack/swift/tree/master/tools/playbooks | 09:47 |
| alecuyer | I see ! thanks a lot for these pointers, I'm reading | 09:48 |
| kota_ | alecuyer: my pleasure ;) | 09:49 |
| *** admin6 has joined #openstack-swift | 09:57 | |
| *** gkadam_ has joined #openstack-swift | 11:24 | |
| *** gkadam has quit IRC | 11:25 | |
| *** e0ne has joined #openstack-swift | 11:36 | |
| *** ybunker has joined #openstack-swift | 12:22 | |
| *** ybunker has quit IRC | 12:35 | |
| *** ybunker has joined #openstack-swift | 12:38 | |
| *** e0ne has quit IRC | 13:05 | |
| *** hseipp has quit IRC | 13:18 | |
| *** gkadam_ has quit IRC | 14:02 | |
| *** irclogbot_2 has quit IRC | 14:09 | |
| *** e0ne has joined #openstack-swift | 14:10 | |
| *** irclogbot_2 has joined #openstack-swift | 14:13 | |
| *** irclogbot_2 has quit IRC | 14:25 | |
| *** irclogbot_2 has joined #openstack-swift | 14:28 | |
| *** psachin has quit IRC | 15:07 | |
| *** irclogbot_2 has quit IRC | 15:36 | |
| *** irclogbot_2 has joined #openstack-swift | 15:40 | |
| *** irclogbot_2 has quit IRC | 15:49 | |
| *** irclogbot_2 has joined #openstack-swift | 15:52 | |
| *** irclogbot_2 has quit IRC | 15:52 | |
| *** irclogbot_2 has joined #openstack-swift | 15:57 | |
| *** ybunker has quit IRC | 16:14 | |
| notmyname | good morning | 16:24 |
| *** gyee has joined #openstack-swift | 16:25 | |
| *** FlorianFa has quit IRC | 16:41 | |
| *** FlorianFa has joined #openstack-swift | 16:54 | |
| *** e0ne has quit IRC | 17:01 | |
| *** gyee has quit IRC | 17:20 | |
| *** gyee has joined #openstack-swift | 17:20 | |
| *** gerardgine has joined #openstack-swift | 17:21 | |
| *** ccamacho has quit IRC | 17:29 | |
| openstackgerrit | Merged openstack/swift feature/losf: Enable swift-dsvm-functional tests at losf feature branch https://review.openstack.org/642645 | 17:34 |
| notmyname | yay! | 17:34 |
| *** gerardgine has quit IRC | 17:39 | |
| renich | zaitcev: hey man, thanks for all the help yesterday. I had to suddenly leave (was late for a gig), but many thanks. | 17:53 |
| *** ianychoi has quit IRC | 17:55 | |
| *** pcaruana has quit IRC | 18:29 | |
| *** gerardgine has joined #openstack-swift | 18:34 | |
| openstackgerrit | John Dickinson proposed openstack/swift master: crediting contributors to the un-landed hummingbird branch https://review.openstack.org/643134 | 18:41 |
| notmyname | timburke: ^^ thoughts? | 18:41 |
| notmyname | timburke: FWIW, there are a few other authors file updates that will need to happen before the next release tag. I'll work with you on that | 18:42 |
| *** gmann is now known as gmann_afk | 18:48 | |
| timburke | notmyname, makes sense to me | 18:49 |
| notmyname | timburke: I have not checked other branches (and there's the obvious gap of contributors only referenced via the commit message body), but those other branches, AFAIK, all ended up landing on master | 18:51 |
| *** e0ne has joined #openstack-swift | 19:01 | |
| zaitcev | I still have patch or maybe two outstanding for hummingbird. Some absurdly obvious bugs. | 19:16 |
| *** gerardgine has quit IRC | 19:23 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: Get functional/tests.py running under py3 https://review.openstack.org/642520 | 19:38 |
| *** gerardgine has joined #openstack-swift | 19:58 | |
| *** ggine has joined #openstack-swift | 20:03 | |
| tdasilva | ooohhh, time change | 20:03 |
| tdasilva | was trying to figure out why nobody had shown up yet for meeting | 20:04 |
| *** gerardgine has quit IRC | 20:04 | |
| *** ggine has quit IRC | 20:06 | |
| *** gerardgine has joined #openstack-swift | 20:09 | |
| notmyname | doh! sorry about that | 20:11 |
| notmyname | meeting is in 38 minutes at 2100utc (which doesn't change twice a year) | 20:12 |
| notmyname | personally, I set meetings like that for either iceland or ghana because they're both utc+0 with no daylight savings time change. (ie if you're calendar app doesn't support setting utc... *grrr* google) | 20:13 |
| tdasilva | notmyname: it does, the meeting in the calendar does show it correctly, i just had not looked at it yet today | 20:17 |
| notmyname | heh, ok :-) | 20:17 |
| notmyname | TBH, I looked down at my clock and had a mild panic that I'd missed it! | 20:17 |
| tdasilva | Ha! | 20:22 |
| *** gerardgine has quit IRC | 20:25 | |
| *** gerardgine has joined #openstack-swift | 20:31 | |
| *** e0ne has quit IRC | 20:38 | |
| *** alexlecuyer has joined #openstack-swift | 20:39 | |
| *** gmann_afk is now known as gmann | 20:41 | |
| *** e0ne has joined #openstack-swift | 20:41 | |
| *** e0ne has quit IRC | 20:45 | |
| mattoliverau | Morning | 20:55 |
| mattoliverau | notmyname Google calendar has a utc time one now. So no more Iceland :) | 20:56 |
| mattoliverau | *timezone | 20:56 |
| tdasilva | nice | 20:56 |
| notmyname | mattoliverau: oh nice | 20:57 |
| kota_ | morning | 20:57 |
| kota_ | oh, day light saving | 20:58 |
| notmyname | kota_: mattoliverau: have your countries changed yet? | 20:58 |
| * kota_ is wondering how mobile alarm works in the time change | 20:58 | |
| notmyname | I see 6am for japan and 8am for australia | 20:59 |
| kota_ | notmyname: no day light saving exist in Japan. Always same time for a year | 20:59 |
| notmyname | ah. how enlightened :-) | 20:59 |
| notmyname | IIRC I think EU (or UK) changes to daylight savings in a few weeks | 21:00 |
| mattoliverau | Not until April | 21:00 |
| notmyname | anyway, the meeting is now | 21:00 |
| notmyname | mattoliverau: (that's just a few weeks from now) | 21:00 |
| mattoliverau | Yeah, then it moves back an hour. | 21:00 |
| notmyname | zaitcev: rledisez: timburke: tdasilva: meeting ping | 21:00 |
| mattoliverau | So earlier meeting :( | 21:01 |
| mattoliverau | But used to it now :) | 21:01 |
| *** gerardgine has quit IRC | 21:43 | |
| notmyname | timburke: FWIW, I was assuming I'd pass off meeting chairing after the release is cut | 21:47 |
| timburke | i could also start chairing next meeting if you like. either way | 21:48 |
| timburke | w/e | 21:48 |
| notmyname | timburke: and also that you and I would take an hour or two next week to go over some logistics stuff (gerrit groups, launchpad, irc flags, irc bot, gerrit dashboards, community dashboards, etc) | 21:48 |
| timburke | oh yeah, speaking of releases -- i'd rather like to get https://review.openstack.org/#/c/636501/ landed and tag a pyeclib release, if only so i can run `tox -e py27` in my new dev env and have it actually work... | 21:51 |
| patchbot | patch 636501 - pyeclib - Try harder to find liberasurecode - 8 patch sets | 21:51 |
| timburke | i'll probably still need to do some hacking around upper-constraints (since we're past the library freeze, i don't expect the new pyeclib to get picked up until the train cycle starts), but i already have to do that for eventlet | 21:52 |
| *** rcernin has joined #openstack-swift | 21:57 | |
| *** gerardgine has joined #openstack-swift | 21:58 | |
| *** gerardgine has quit IRC | 22:14 | |
| *** gerardgine has joined #openstack-swift | 22:27 | |
| *** lifeless has quit IRC | 22:27 | |
| *** lifeless has joined #openstack-swift | 22:27 | |
| *** alexlecuyer has quit IRC | 22:37 | |
| *** gerardgine has quit IRC | 23:00 | |
| *** gerardgine has joined #openstack-swift | 23:01 | |
| *** ianychoi has joined #openstack-swift | 23:06 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: Get functional/tests.py running under py3 https://review.openstack.org/642520 | 23:06 |
| openstackgerrit | Tim Burke proposed openstack/swift master: Add non-voting py37 unit test job https://review.openstack.org/643183 | 23:20 |
| openstackgerrit | Tim Burke proposed openstack/swift master: Get functional/tests.py running under py3 https://review.openstack.org/642520 | 23:23 |
| *** renich has quit IRC | 23:24 | |
| openstackgerrit | Merged openstack/swift master: crediting contributors to the un-landed hummingbird branch https://review.openstack.org/643134 | 23:25 |
| openstackgerrit | Tim Burke proposed openstack/swift master: Fix py35 unit test job https://review.openstack.org/643184 | 23:30 |
| *** gerardgine has quit IRC | 23:32 | |
| timburke | so, ^^^ is kinda important... in retrospect, i guess having all our unit test jobs inherit from openstack-tox-py27 maybe isn't the greatest? idk | 23:33 |
| *** gerardgine has joined #openstack-swift | 23:35 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: Add non-voting py37 unit test job https://review.openstack.org/643183 | 23:45 |
| timburke | i still don't expect ^^^ to pass or anything (pretty sure we'll need at least https://review.openstack.org/#/c/638019/ to land first), but i figure it'll be good to see the failures | 23:47 |
| patchbot | patch 638019 - swift - Clean up how we walk through ranges in ECAppIter - 1 patch set | 23:47 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!