| *** renich has quit IRC | 00:35 | |
| *** renich has joined #openstack-swift | 00:38 | |
| *** gyee has quit IRC | 01:16 | |
| openstackgerrit | Merged openstack/swift master: Fix lower-constraint deps handling https://review.openstack.org/645624 | 01:43 |
|---|---|---|
| openstackgerrit | Pete Zaitcev proposed openstack/swift master: py3: port bulk middleware https://review.openstack.org/619303 | 02:37 |
| *** renich has quit IRC | 02:42 | |
| *** renich_ has joined #openstack-swift | 02:43 | |
| *** psachin has joined #openstack-swift | 02:52 | |
| *** renich has joined #openstack-swift | 03:38 | |
| *** renich_ has quit IRC | 03:38 | |
| *** rcernin has quit IRC | 04:47 | |
| *** rcernin has joined #openstack-swift | 05:02 | |
| *** rcernin has quit IRC | 06:03 | |
| *** tdasilva has quit IRC | 06:15 | |
| *** ccamacho has quit IRC | 06:23 | |
| *** rcernin has joined #openstack-swift | 06:48 | |
| *** e0ne has joined #openstack-swift | 07:17 | |
| *** e0ne has quit IRC | 07:19 | |
| *** pcaruana has joined #openstack-swift | 07:20 | |
| *** joeljwright has quit IRC | 07:28 | |
| *** joeljwright has joined #openstack-swift | 07:29 | |
| *** ChanServ sets mode: +v joeljwright | 07:29 | |
| *** rcernin has quit IRC | 07:48 | |
| *** rdejoux has joined #openstack-swift | 08:01 | |
| *** rcernin has joined #openstack-swift | 08:05 | |
| *** ccamacho has joined #openstack-swift | 08:05 | |
| *** tkajinam has quit IRC | 08:09 | |
| *** mikecmpbll has joined #openstack-swift | 08:49 | |
| *** hseipp has joined #openstack-swift | 08:52 | |
| *** e0ne has joined #openstack-swift | 09:47 | |
| *** e0ne has quit IRC | 10:31 | |
| *** e0ne has joined #openstack-swift | 10:36 | |
| *** hoonetorg has quit IRC | 10:39 | |
| *** hseipp has quit IRC | 10:50 | |
| *** hoonetorg has joined #openstack-swift | 10:52 | |
| *** e0ne has quit IRC | 10:58 | |
| *** e0ne has joined #openstack-swift | 11:05 | |
| *** mvkr has joined #openstack-swift | 11:09 | |
| kota_ | rledisez: looking at https://review.openstack.org/648245 it looks like the result clearly says "the environment doesn't have py35 interpreter" so I don't think we could fix it with "recheck" | 11:53 |
| patchbot | patch 648245 - swift (feature/losf) - Merge remote-tracking branch 'remotes/origin/maste... - 1 patch set | 11:53 |
| kota_ | maybe we may be missing a piece of something... | 11:54 |
| kota_ | and also I found we don't have py37 non-voting that is covered at the master so the diff could be in .zuul.yaml conf... | 11:55 |
| kota_ | agh ubuntu package repository is under Internal Server Error :/ https://packages.ubuntu.com/ja/bionic/python3-dev | 12:02 |
| kota_ | but it looks like "= 3.6.5-3" from the cache and bindep just saying as "python3" ... | 12:03 |
| kota_ | idk why the master is still work with that... ah... it looks like the master is working with xenial, maybe? | 12:06 |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift feature/losf: DNM: test to enable python3.5 interpreter at the gate https://review.openstack.org/648381 | 12:11 |
| kota_ | if i understand the issue correctly, that should work but IMO it's wrong way because the root cause should be from the difference on the vm (the master running on the xenial and losf on the bionic) so the better way should fix the difference... | 12:13 |
| rledisez | kota_: yeah, i tried it once, I was going to ask you what you think about it, but you already looked at it :) | 12:15 |
| kota_ | rledisez: yeah, just pushed my comment to the gerrit before I'm heading to my home. | 12:15 |
| kota_ | sorry, currently no idea where the vm difference comes from but probably it's sort of the playbooks setting in .zuul.yaml or zuul setting in out of our repository... | 12:17 |
| kota_ | it might be fast to ask at #openstack-infra channel | 12:17 |
| kota_ | anyway thanks for updating the branch, rledisez!! | 12:17 |
| kota_ | looking at the http://logs.openstack.org/45/648245/1/check/swift-tox-py35/45dd925/job-output.json.gz, zuul looks to try to get missing dependency by `bindep bindep.txt test py35` | 12:20 |
| kota_ | I think the latter two (test py35) is for collecting the dependency for the env but we don't have any py35 specific bindep for now. | 12:20 |
| kota_ | so bindep doesn't collect the py35 package, then use py36 that is the default package of bionic | 12:21 |
| kota_ | that's my thought. | 12:21 |
| *** e0ne has quit IRC | 12:41 | |
| *** e0ne has joined #openstack-swift | 12:43 | |
| kota_ | huh, ok. http://logs.openstack.org/81/648381/1/check/swift-tox-py35/e56191d/job-output.json.gz says "msg": "No package matching 'python3.5' is available" | 12:47 |
| * kota_ need more time to know but no time I have in my office today. | 12:48 | |
| *** e0ne has quit IRC | 13:20 | |
| *** itlinux has quit IRC | 13:26 | |
| *** mvkr has quit IRC | 14:12 | |
| *** mvkr has joined #openstack-swift | 14:37 | |
| *** e0ne has joined #openstack-swift | 14:38 | |
| *** itlinux has joined #openstack-swift | 14:47 | |
| *** renich has quit IRC | 14:51 | |
| *** renich has joined #openstack-swift | 14:53 | |
| *** tdasilva has joined #openstack-swift | 15:06 | |
| *** ChanServ sets mode: +v tdasilva | 15:06 | |
| *** e0ne has quit IRC | 15:14 | |
| *** e0ne has joined #openstack-swift | 15:20 | |
| clayg | in-process s3api tests start strong, get slow, then every subsequent test fails - but the rest of test modules after that still pass! https://www.irccloud.com/pastebin/uvOIGxsO/failed-test.out | 15:25 |
| clayg | http://logs.openstack.org/17/647817/1/gate/swift-tox-func-s3api/1b2ff73/job-output.txt.gz from p 647817 | 15:26 |
| patchbot | https://review.openstack.org/#/c/647817/ - swift - Make reconciler test more stable in the gate - 1 patch set | 15:26 |
| timburke | good morning | 15:27 |
| clayg | the slow test must have also hit some errors/retries - but it eventually passed... I don't know how s3api reset/authentication works with in-process tests... if memcache failed that would probably effect other non-s3api tests 🤔 | 15:28 |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: Allow "harder" symlinks https://review.openstack.org/633094 | 15:28 |
| timburke | kota_, rledisez: fwiw, i fixed that error on master with https://review.openstack.org/#/c/643184/ | 15:33 |
| patchbot | patch 643184 - swift - Fix py35 unit test job (MERGED) - 1 patch set | 15:33 |
| timburke | ...but the merge should include that commit... | 15:33 |
| rledisez | timburke: thx. i'll redo the merge | 15:34 |
| timburke | rledisez, idk that it'll help... the only extra patch on master (as compared to your merge) is https://review.openstack.org/#/c/645624/ (at least, that i see) | 15:36 |
| patchbot | patch 645624 - swift - Fix lower-constraint deps handling (MERGED) - 2 patch sets | 15:36 |
| timburke | i kinda sorta seem to remember some issue with zuul job caching and feature branch merges causing trouble before... | 15:37 |
| *** e0ne has quit IRC | 15:40 | |
| clayg | timburke: while you're thinking about making the gate better - any idea on my s3api Access Denied thing? | 16:04 |
| clayg | https://github.com/openstack/swift/blob/master/test/functional/s3api/s3_test_client.py#L97 | 16:05 |
| clayg | Do you think any of the other exceptions would have been more interesting/useful - or do you think they were all 403's? | 16:05 |
| timburke | first exception is the most likely to be the cause of trouble -- i'm not *sure*, but i'd guess the others (if any) would also have been 403s | 16:07 |
| timburke | i wonder if it was a matter of some acl data not making it to all replicas or something? and then we happened to hit (and cache) the out-of-date replica | 16:08 |
| *** gyee has joined #openstack-swift | 16:12 | |
| clayg | the reset code is just trying to list all buckets - oh maybe it was one of the other requests... so preserving stack might help! | 16:23 |
| *** e0ne has joined #openstack-swift | 16:27 | |
| *** admin6 has quit IRC | 16:30 | |
| *** hseipp has joined #openstack-swift | 16:49 | |
| *** renich has quit IRC | 16:54 | |
| *** renich has joined #openstack-swift | 17:01 | |
| *** psachin has quit IRC | 17:05 | |
| timburke | rledisez, so i found http://eavesdrop.openstack.org/irclogs/%23openstack-infra/%23openstack-infra.2018-01-24.log.html#t2018-01-24T23:39:37 | 17:07 |
| *** e0ne has quit IRC | 17:07 | |
| timburke | there was an idea that seemed to work out in the next day's logs: http://eavesdrop.openstack.org/irclogs/%23openstack-infra/%23openstack-infra.2018-01-25.log.html#t2018-01-25T00:13:27 | 17:07 |
| timburke | and it kicked out a bug: https://storyboard.openstack.org/#!/story/2001496 | 17:07 |
| *** e0ne has joined #openstack-swift | 17:08 | |
| timburke | ...which i *think* is still unresolved? i'm still not familiar enough with storyboard to know at a glance | 17:09 |
| timburke | might be able to do a whitespace change to .zuul.yaml to make it realize the file changed | 17:09 |
| *** e0ne has quit IRC | 17:17 | |
| openstackgerrit | Tim Burke proposed openstack/swift feature/losf: Merge remote-tracking branch 'remotes/origin/master' into merge-master https://review.openstack.org/648245 | 17:19 |
| *** e0ne has joined #openstack-swift | 17:21 | |
| *** ccamacho has quit IRC | 17:23 | |
| *** hseipp has quit IRC | 17:24 | |
| *** mvkr has quit IRC | 17:31 | |
| *** mikecmpbll has quit IRC | 17:36 | |
| *** e0ne has quit IRC | 17:45 | |
| *** itlinux has quit IRC | 17:46 | |
| *** itlinux has joined #openstack-swift | 17:47 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: WIP: s3api: Make multi-deletes async https://review.openstack.org/648263 | 19:08 |
| timburke | clayg, ok, so what can we do to make ^^^ give some advantage to swift's bulk delete, too? i guess we could buffer (say) 1k items at a time, group by container, and do the replicate trick if the number of objects for a given container is over some threshold? if it's under, just fire off deletes like we always have | 19:12 |
| *** rdejoux has quit IRC | 19:13 | |
| timburke | thing is, bulk as it exists today is very stream-oriented -- which is generally the right way to go about it... | 19:13 |
| clayg | orly? I thought the body was json | 19:13 |
| clayg | oh, wow, no it's line oriented :'( | 19:14 |
| clayg | what the literal f$%^ - we stream in the entire thing to build up a list | 19:15 |
| timburke | yeah: https://github.com/openstack/swift/blob/2.21.0/swift/common/middleware/bulk.py#L329-L365 | 19:15 |
| clayg | i.. don't... | 19:15 |
| timburke | oh god! we do, don't we! | 19:15 |
| timburke | 🤦 | 19:16 |
| clayg | well that's fine - i don't understand why it needs to be so big anyway | 19:16 |
| *** e0ne has joined #openstack-swift | 19:33 | |
| *** rcernin has quit IRC | 19:36 | |
| tdasilva | timburke: I think we agreed to +A p 647656, correct? | 19:52 |
| patchbot | https://review.openstack.org/#/c/647656/ - pyeclib - Prep for 1.6.0 release - 1 patch set | 19:52 |
| clayg | About 38% of teams changed their leads, including Swift where Tim Burke takes the helm in replacement for John Dickinson who had been leading that team since the Diablo cycle! | 19:53 |
| clayg | ^ from the foundation email! | 19:53 |
| timburke | wait -- so who was in charge for Cactus? | 19:53 |
| timburke | tdasilva, yeah, i think so | 19:53 |
| *** e0ne has quit IRC | 20:00 | |
| clayg | IIRC the notmyname's reign pre-dates the title PTL, so maybe that's part of it... | 20:03 |
| *** pcaruana has quit IRC | 20:08 | |
| openstackgerrit | Merged openstack/pyeclib master: Prep for 1.6.0 release https://review.openstack.org/647656 | 20:12 |
| *** itlinux has quit IRC | 20:59 | |
| *** openstackgerrit has quit IRC | 21:07 | |
| *** [diablo] has quit IRC | 21:09 | |
| *** openstackgerrit has joined #openstack-swift | 21:12 | |
| openstackgerrit | Merged openstack/swift feature/losf: Merge remote-tracking branch 'remotes/origin/master' into merge-master https://review.openstack.org/648245 | 21:12 |
| *** mikecmpbll has joined #openstack-swift | 21:16 | |
| *** rcernin has joined #openstack-swift | 21:44 | |
| *** [diablo] has joined #openstack-swift | 21:53 | |
| *** rcernin has quit IRC | 21:55 | |
| *** rcernin has joined #openstack-swift | 22:04 | |
| *** jeremyfr has joined #openstack-swift | 22:25 | |
| jeremyfr | Hello, is objects listing from a container supposed to be consistent ? I have a container with ~17000 objects, most of the time I request the full list, I do not get all the objects. | 22:27 |
| *** renich has quit IRC | 22:32 | |
| *** renich_ has joined #openstack-swift | 22:32 | |
| timburke | jeremyfr, no, listings are eventually consistent -- the object layer is responsible for updating listings following a PUT or DELETE, and if the container server is busy, it may take a bit for the object-updater to get around to it | 22:38 |
| jeremyfr | Thanks for the explanation timburke | 22:45 |
| mattoliverau | morning | 22:47 |
| *** tkajinam has joined #openstack-swift | 22:59 | |
| *** renich_ has quit IRC | 23:12 | |
| *** gyee has quit IRC | 23:57 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!