Thursday, 2019-02-21

*** itlinux has joined #openstack-swift00:05
*** itlinux has quit IRC00:06
*** [diablo] has quit IRC00:37
*** [diablo] has joined #openstack-swift00:39
*** mvkr has quit IRC00:40
*** mvkr has joined #openstack-swift00:42
*** mikecmpb_ has quit IRC00:49
*** mikecmpbll has joined #openstack-swift00:50
*** mikecmpbll has quit IRC01:13
*** gyee has quit IRC02:06
*** psachin has joined #openstack-swift02:52
openstackgerritPete Zaitcev proposed openstack/swift master: py3: port the container
*** spsurya has joined #openstack-swift03:34
openstackgerritPete Zaitcev proposed openstack/swift master: py3: port proxy container controller
*** ianychoi has quit IRC04:31
kota_alecuyer: thanks for checking the version conflict. I think it's UPPER CONSTRAINT so that we could use 1.3.3 unless the lower constraint requires higher version.04:37
kota_let me check04:37
kota_hmm... it looks like we don't require any lower constraint about grpcio. interesting.04:42
kota_I see, got it why it happened.05:18
kota_`-c constraint_file` forces the version rather than, requirements pin... I'm not sure it's expected or not tho.05:19
*** spsurya has quit IRC06:25
kota_thanks mattoliverau for the group E-mail!06:57
*** psachin has quit IRC07:10
*** rcernin has quit IRC07:13
*** spsurya has joined #openstack-swift07:17
*** pcaruana has joined #openstack-swift07:19
*** psachin has joined #openstack-swift07:32
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Add upper-constraints file
*** ccamacho has quit IRC08:18
*** tkajinam has quit IRC08:23
alecuyerkota_: yes that confused me (the "upper" name does not seem to match what it's doing)08:31
kota_alecuyer: alright, I'm trying to resolve it now and also asked at openstack-infra channel08:37
kota_i didn't get any answers in the channel tho.08:38
kota_ok, just `tox -e py27` doesn't reproduce the issue so that I have to look at openstack's tox-py27 job base. probably there is in somewhere around zuul setting.08:51
alecuyerok hmm08:57
*** mikecmpbll has joined #openstack-swift08:57
*** ccamacho has joined #openstack-swift09:09
*** ccamacho has quit IRC09:10
*** ccamacho has joined #openstack-swift09:11
*** e0ne has joined #openstack-swift09:14
*** hseipp has joined #openstack-swift09:15
kota_I found `tox_constraints_file` definition there, perhaps, the variable overwrite something?09:23
alecuyerI have to say I didn't even know what a pip constraint file was until today, so I have some reading to do09:27
kota_I read docs briefly about the constraint file, it says "it's for OpenStack" :/09:36
kota_it seems like, constraints is not a requirement. if a package exists in constraints and it doesn't present in requirements, the package won't be installed.09:37
alecuyerOK, but if it's in requirements, and in constraints, it will pick the version from constraints, right ?09:39
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: WIP: Add upper-constraints file
kota_alecuyer: it seems like current pip behaves so.09:43
kota_I cannot find the specification when the version conflicted with the constraints and the requirements. I looked for the tests for such a situation but that doesn't exist in the pip repo (
alecuyerWell, maybe we can upgrade grpcio version and the problem goes away, I'm testing that (also, I wonder if the constraint file would let us have a  higher version than 1.15.0 )09:49
kota_not sure but the current constraint looks intended by the other openstack project, so it might be difficult to bump up higher version...10:02
kota_ok, the last change affects the behavior like so that my direction seems correct.10:05
kota_I need to find the path to upper-constraint in the gate tho.10:05
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: WIP: Add upper-constraints file
kota_alright, it looks it's working as I intended
kota_gprcio==1.3.3 is in the gate!10:14
alecuyernice, well done !10:14
*** [diablo] has quit IRC10:15
kota_though I couldn't find the place the variable affected :P10:15
kota_alecuyer: helpful link is, we can see the status that currently are running in the gate.10:16
*** [diablo] has joined #openstack-swift10:17
alecuyerI see, thank you for the link10:17
kota_query, `swift` at left top box.10:17
kota_that shows only results for swift patches.10:18
kota_yey, at least, py35 gates succeeded!10:18
alecuyerok that's handy!10:18
kota_hmm... we got still a failure at lower constraints but we're going to progress step by step.10:19
kota_OMG, lower constraint was testing against to grpc 1.15.0, hah? Is it exactly *LOWER*?10:20
alecuyerAh right I see it in job-output :/10:20
*** spsurya has quit IRC10:22
kota_it inherits openstack-tox-lower-constraints directly so the change at swift-tox-base doesn't affect to the test.10:23
kota_hmm... nice almost all green except the lower-constraint so far.10:24
kota_alecuyer: can i squash the change into the LOSF patch?10:30
alecuyerkota_: sure10:30
kota_alright, let's do it.10:31
alecuyerfunctest passed it seems10:31
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Lots Of Small Files - alternate diskfile
*** psachin has quit IRC10:38
*** psachin has joined #openstack-swift10:42
kota_ok, I'm heading to my home today. If you find anything, please leave it here.10:42
alecuyeryes I will,10:42
alecuyerthanks for your work with this !10:43
kota_no worries. That is great job because all tests succeeded so far unless the version conflicts. AWESOME.10:43
*** mikecmpb_ has joined #openstack-swift11:12
*** mikecmpbll has quit IRC11:13
*** psachin has quit IRC11:22
*** FlorianFa has joined #openstack-swift11:55
*** FlorianFa has quit IRC11:56
*** FlorianFa has joined #openstack-swift11:56
*** godog has quit IRC11:58
*** godog has joined #openstack-swift12:02
*** e0ne has quit IRC12:24
*** e0ne has joined #openstack-swift12:31
alecuyerkota_: everything passed with your latest patch set12:49
*** mvkr has quit IRC15:00
*** jistr is now known as jistr|afk15:22
*** mvkr has joined #openstack-swift15:24
*** jistr|afk is now known as jistr15:32
alecuyerIt seems grpc (client, no streaming) and eventlet work together again from grpcio version 1.11.0 and later (have also superficially tested 1.18.0 as latest stable,  and 1.15.0 as present in openstack pip constraints file). there are 694 commits between 1.10.1 (last "broken" release) and 1.11.0. Among these are commits about gevent support but I don't see how these could affect eventlet. Still reading16:17
*** e0ne has quit IRC16:23
tdasilvaalecuyer: that sounds like good news...maybe :)16:34
alecuyertdasilva: yes, exactly, "maybe" :)16:36
*** gyee has joined #openstack-swift16:48
notmynamegood morning17:00
*** pcaruana has quit IRC17:07
openstackgerritTim Burke proposed openstack/python-swiftclient master: Make proper functions instead of assigning lambdas
*** hseipp has quit IRC17:29
*** mikecmpb_ has quit IRC17:35
*** ccamacho has quit IRC17:36
openstackgerritTim Burke proposed openstack/swift master: py3: port some more middleware tests
*** mvkr has quit IRC18:12
openstackgerritTim Burke proposed openstack/swift master: py3: Flag manage-shard-ranges subcommand as required
*** mikecmpbll has joined #openstack-swift18:22
openstackgerritClay Gerrard proposed openstack/swift master: Simplify empty suffix handling
claygI considered that I might try to put a probe test together to demonstrate the failure condition in the current suffix handling18:32
claygbut it requires setting reclaim age to something small... which might be useful in a probetest for a few reasons - but I don't think we currently do it18:33
*** mvkr has joined #openstack-swift18:59
*** mrjk_ has quit IRC19:00
*** mrjk has joined #openstack-swift19:02
*** ccamacho has joined #openstack-swift19:31
*** e0ne has joined #openstack-swift20:14
*** e0ne has quit IRC20:37
timburkethose log lines tests are getting on my nerves...
*** rcernin has joined #openstack-swift21:11
openstackgerritTim Burke proposed openstack/swift master: Clean up account-reaper a bit
zaitcevnotmyname: "My container has ~100 million small objects."  Time to back up slowly21:42
notmynameoh my!21:42
zaitcevI mean back away21:43
notmynamehave you heard the good news of our savior container sharding?21:43
zaitcevHe may mean the whole cluster, but ugh.21:45
zaitcevI don't have the privilege points to comment and ask.21:52
openstackgerritTim Burke proposed openstack/swift master: py3:  display help if no subcommand provided to manage-shard-ranges
openstackgerritTim Burke proposed openstack/swift master: manage-shard-ranges: nicer message if we can't get_info()
timburkeman, i'm really starting to see the value in us sticking with py27 for so long... idk that i want to have to go chasing the latest version all the time22:46
timburkethen discovering that the thing i want to use, the thing that's in the latest docs, isn't available on all the versions we "want" to support...22:46
*** tkajinam has joined #openstack-swift23:01
DHEZFS (prerelease version) has this feature where metadata can be put on a distinct disk from file payloads. so put the bulk of the file contents on spinning disks and metadata on an SSD. and I'm looking for something else that can provide that same feature because replication would just laugh off anything built like this23:19
DHEI assume nothing like this exists in other filesystems without a generic disk cache like bcache... :(23:19
*** gyee has quit IRC23:38
*** gyee has joined #openstack-swift23:52

Generated by 2.15.3 by Marius Gedminas - find it at!