21:00:09 <notmyname> #startmeeting swift
21:00:10 <openstack> Meeting started Wed May 25 21:00:09 2016 UTC and is due to finish in 60 minutes.  The chair is notmyname. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:11 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:00:13 <openstack> The meeting name has been set to 'swift'
21:00:27 <notmyname> hello, everyone. who's here for the swift team meeting?
21:00:27 <clayg> LET'S DO THIS!
21:00:32 <mattoliverau> o/
21:00:36 <dmorita_> o/
21:00:37 <hurricanerix> \o/
21:00:38 <cschwede> o/
21:00:40 <jrichli> me
21:00:41 <hosanai_> o/
21:00:45 <timburke> o/
21:00:46 <tdasilva> hello
21:00:46 <cutforth> o/
21:00:51 <bkeller`> o/
21:01:06 <kota_> o/
21:01:13 <acoles> here
21:01:26 <notmyname> all right
21:01:28 <notmyname> agenda is at
21:01:33 <notmyname> #link https://wiki.openstack.org/wiki/Meetings/Swift
21:01:45 <notmyname> a few things to go over this week
21:01:52 <notmyname> #topic crypto update
21:02:01 <notmyname> jrichli: acoles: where are we this week?
21:02:11 <notmyname> ready to get a review branch to master yet? :-)(
21:02:23 <jrichli> we keep finding more to do :-)
21:02:24 <acoles> notmyname: well i keep moving cards on the trello board
21:02:34 <acoles> yeah, then more appear!
21:02:35 <notmyname> #link https://trello.com/b/63l5zQhq/swift-encryption
21:02:49 <acoles> but we are getting close, I hope
21:03:16 <notmyname> https://wiki.openstack.org/wiki/Swift/PriorityReviews is up to date with the crypto patches needing review? I still see some listed that were there last week
21:03:35 <acoles> been struggling to merge master to feature/crypto this week, hopefully we have nailed that as an other-requirements.txt issue
21:03:52 <notmyname> great. we should know something there as soon as the gate jobs run
21:04:01 <acoles> notmyname: yes priority reviews is up to date
21:04:09 <acoles> notmyname: yes some are same as last week
21:04:14 <notmyname> :-(
21:04:48 <jrichli> we got some feedback from one of the cyrptographers we work with: so far so good - small adjustments to make
21:04:50 <acoles> patch 316924 shouldn't be too hard, requires no knowledge of feature/crypto
21:04:50 <patchbot> acoles: https://review.openstack.org/#/c/316924/ - swift (feature/crypto) - crypto - make some probe tests compatible with crypto
21:04:58 <notmyname> jrichli: great
21:05:14 <notmyname> for everyone, if you're looking for something to review, please start with the stuff on https://wiki.openstack.org/wiki/Swift/PriorityReviews
21:05:47 <acoles> notmyname: it does feel like we're getting down to small tweaks (famous last words...)
21:06:07 <notmyname> acoles: jrichli: other than those reviews, anything you need from the rest of us?
21:07:07 <acoles> reviews would be great, and not just the priority ones
21:07:47 * acoles thanks mahatic and timburke for helping out
21:08:00 <timburke> happy to help
21:08:10 <notmyname> ok. we should all feel appropriately guilty for not reviewing ( mahatic and timburke excluded)
21:08:16 <jrichli> +2! and tdasilva
21:08:22 <jrichli> +2 was for the thanks
21:08:37 <acoles> hehe +2 that guilt ;)
21:08:39 <notmyname> jrichli: no, you're trying to lay down the guilt trip, right?
21:08:56 * jrichli shakes head
21:09:25 <notmyname> thanks for the crypto update and working on it. let's all do better this next week about reviewing the code
21:09:38 <notmyname> #topic rolling upgrade tests status update
21:09:41 <notmyname> cschwede: are you here?
21:09:47 <cschwede> sure!
21:09:57 <cschwede> nothing new - still waiting for reviews of https://review.openstack.org/#/c/304465/
21:09:58 <patchbot> cschwede: patch 304465 - openstack-infra/devstack-gate - Use subnodes for Swift storage nodes in a multinod...
21:10:20 <notmyname> and patch 297311 right?
21:10:21 <patchbot> notmyname: https://review.openstack.org/#/c/297311/ - openstack-infra/devstack-gate - Run swift services on subnode
21:10:55 <cschwede> no; as commented on that patch this is not required and in fact a different approach
21:11:03 <notmyname> ah, ok
21:11:36 <notmyname> I'll also start trying to bug people about patch 304465 since it's blocking us
21:11:36 <patchbot> notmyname: https://review.openstack.org/#/c/304465/ - openstack-infra/devstack-gate - Use subnodes for Swift storage nodes in a multinod...
21:11:44 <cschwede> thx!
21:11:47 <notmyname> thanks for the update
21:11:58 <mattoliverau> I'll poke the infra core on my team.
21:12:06 <mattoliverau> when he wakes up
21:12:07 <notmyname> great!
21:12:09 <notmyname> thanks
21:12:19 <notmyname> #topic pyeclib/liberasurecode migration updates
21:12:23 <notmyname> tdasilva: what's up with these?
21:12:40 <tdasilva> they've been moved to openstack domain: https://github.com/openstack/pyeclib
21:12:51 <tdasilva> but we are also waiting on a patch to add gate jobs
21:12:58 <tdasilva> https://review.openstack.org/#/c/317672/
21:12:58 <patchbot> tdasilva: patch 317672 - openstack-infra/project-config - add python jobs to pyeclib project
21:12:59 <notmyname> do they currently have the noop gate?
21:13:03 <tdasilva> correct
21:13:25 <notmyname> and can they accept patches right now? ie new code goes there instead of bitbucket?
21:13:33 <tdasilva> yes
21:13:53 <notmyname> has there been any update on the bitbucket side to point people to the openstack repo?
21:13:58 <tdasilva> but we would have to run tests on our machines
21:14:05 <notmyname> we should anyway ;-)
21:14:10 <tdasilva> no, I can work with Tushar on that
21:14:18 <notmyname> ok, thanks
21:14:43 <tdasilva> sorry, I meant, we would have to rely just on that in order to merge, as opposed to having jobs at the gate...
21:14:43 <notmyname> so it seems like we're pretty much ready to go on those, and the final cleanup work is to get the gate stuff set up
21:14:47 <notmyname> thanks
21:14:53 <tdasilva> right
21:14:53 <notmyname> right :-)
21:14:56 <clayg> right
21:15:20 <tdasilva> notmyname, notmyname: should we try to do a "small" release just to test that??
21:15:21 <notmyname> clayg was looking for a release there, I think, so we shoudl probably tag one with the bug fix and call it good!
21:15:28 <notmyname> yeah
21:15:29 <tdasilva> ok
21:16:12 <tdasilva> I'll work with clayg to see what he would like on that release
21:16:12 <notmyname> tdasilva: clayg: let's chat after the meeting about hwo to do a release and what needs to be in it
21:16:18 <tdasilva> ok
21:16:33 <tdasilva> sorry, ok let's talk after the meeting, but not immediately after ;)
21:16:49 <notmyname> ok
21:16:54 <notmyname> #topic symlinks
21:17:07 <notmyname> tdasilva: you're still up. what's up with symlinks (I think you added this one to the agenda)
21:17:43 <tdasilva> I just would like to ask if thereś anybody working on symlinks? I'd like to start contributing to it, but I don't want to step on someone's toes
21:17:49 <tdasilva> so I thought I'd ask first..
21:18:18 <notmyname> in Austin, jrichli stepped up to be the point person for it
21:18:34 <notmyname> and of course m_kazuhiro has done a lot of thinking (and code?) for it
21:18:37 <jrichli> tdasilva and I have had a few private chats
21:19:07 <tdasilva> ok...so I'll keep working with jrichli on that...
21:19:14 <acoles> tdasilva: last time we discussed it here there seemed to be more clarity needed about what use case(s) were to be addressed, did we get any further with that?
21:19:15 <notmyname> public as possible :-)
21:19:28 <tdasilva> and m_kazuhiro...
21:19:54 <jrichli> I tried to start listing the scenarios that need to be designed for the current POST strategy: but i need lots of help!
21:19:59 <jrichli> thats on the etherpad now
21:20:29 <tdasilva> right...I've been doing some thinking on the POST issue, but it's probably a good idea to focus on the use cases first
21:20:33 <jrichli> tdasilva has been working on the impl that is in review.  BTW - that should proably be a WIP
21:20:49 <tdasilva> yeah, I marked that as a WIP...
21:21:03 <acoles> btw, fast-post functional test is now *voting* in the gate ;)
21:21:03 <notmyname> would anyone like to work with jrichli and tdasilva on identifying those POST issues?
21:21:08 <notmyname> acoles: yay
21:21:49 <acoles> notmyname: I've thought a lot about the POST issues but I cannot commit time until we finish encryption
21:21:51 <clayg> acoles: wtg fast-post!
21:22:03 <notmyname> acoles: sounds reasonable :-)
21:22:24 <notmyname> and yeah, I think the symlink stuff is an "after crypto" priority overall
21:22:44 <tdasilva> acoles: I've been trying to write some stuff up and post it on the etherped. hopefully at some point we can get your feedback...
21:23:05 <tdasilva> notmyname: right, agreed
21:23:18 <notmyname> since it's written down, that's good. even if we don't get to it for a few more weeks, it's still something that we should be able to pick up quickly
21:23:45 <notmyname> tdasilva: anything else on symlinks for this meeting?
21:23:46 <mattoliverau> I'm happy to think POST issues :)
21:23:48 <tdasilva> is torgomatic still working on it?
21:24:03 <mattoliverau> sorry looks like internet is laggy
21:24:09 <tdasilva> he wrote the inital spec, not sure if he was still thinking about it
21:24:27 <notmyname> tdasilva: not sure if he's here. but no, not really, AFAIK
21:24:37 <torgomatic> let's see here...
21:24:38 <tdasilva> mattoliverau: https://etherpad.openstack.org/p/swift_symlinks
21:24:43 <torgomatic> "it has not been a priority recently"
21:24:47 <torgomatic> how'd I do? ;)
21:25:00 <tdasilva> torgomatic: hehe, ok, thanks!
21:25:30 <notmyname> ok, next topic
21:25:38 <notmyname> #topic bugs from translations
21:25:44 <notmyname> this one is interesting
21:25:48 <clayg> heh
21:25:50 <clayg> subjective
21:26:08 <notmyname> acoles: you put it on the agenda. want to give an overview?
21:26:14 <notmyname> https://bugs.launchpad.net/swift/+bug/1580678
21:26:15 <openstack> Launchpad bug 1580678 in OpenStack Object Storage (swift) "UnicodeDecodeError when rebalancing a ring" [Undecided,New]
21:27:48 <notmyname> not sure if I'm lagging or if acoles might be
21:28:02 <acoles> oh yes, I added this last week, the bug arises when a french locale is used and the translation has non-ascii characters, that causes a UnicodeError in our logging
21:28:13 * acoles was typing into wrong window!
21:28:16 <notmyname> heh
21:28:47 <acoles> just a question for the collective wisdom - any ideas on how we can test to prevent bugs creeping in with translations? anyone have any experience of that kind of test job?
21:28:52 <mattoliverau> unicode errors aren't funny notmyname :P
21:29:09 <clayg> acoles: I think the best solution is to stop wrapping strings in _()
21:29:10 <notmyname> they're fünny
21:29:36 <acoles> I suspect it's not confined to the french locale, in case anyone thinks I am picking on one nation :)
21:29:46 <acoles> clayg: right!
21:29:55 <notmyname> something like a job that runs tests in a different locale?
21:30:01 <clayg> timburke: you write it up and acoles and I will +2 it
21:30:03 <cschwede> maybe adding one test-run in the gte with a different locale set?
21:30:18 <notmyname> yeah, not translating the strings is one way to stop those errors...
21:30:26 <clayg> cschwede: if I change my local and run tests do they fail currently?
21:30:27 <notmyname> but I kinda think we probably need to do both
21:30:41 <acoles> or only allowing one language in a project ?? :P
21:30:49 <cschwede> clayg: iirc, yes (but it depends on the locale)
21:30:54 <torgomatic> yeah, testing with a non-english locale shouldn’t be too bad
21:31:11 <notmyname> acoles: :-)
21:31:26 <cschwede> while thinking about it - we would need to test all translated locales; some strings are translated in one locale, but not in others…
21:31:33 <acoles> cschwede: so is that something that can be configured in devstack, so easy enough to add a gate job?
21:31:50 <notmyname> can't we pick one of the 10-12 languages that are 100% translated?
21:31:50 <timburke> if we want an english-like locale that's certain to have lots of non-ascii characters in every translated string, try https://gist.github.com/tipabu/8b627bf99509188e274ede3978bd1149
21:32:10 <acoles> cschwede: yep, that's the problem, but a sample test would smoke out some bugs
21:32:15 <tdasilva> some other project must have run into this before...
21:32:25 <clayg> timburke: that script needs a docstring or something
21:32:38 <timburke> all my scripts need docstrings :P
21:32:49 <cschwede> acoles: i didn’t check this so far, but i assume it wouldn’t be too difficult to add it to devstack/devstack-gate
21:33:12 <notmyname> cschwede: would you have time to look into that this week?
21:33:30 <cschwede> not sure this week, but next Monday - yes
21:33:32 <acoles> of course, if we did have such a job, it would frustrate the translators if we didn't then fix the bugs!
21:33:45 <acoles> but that's better than frustrated ops
21:33:47 <notmyname> cschwede: ok, that's fine, I think
21:33:56 <acoles> cschwede: thank you
21:34:01 <acoles> cschwede: danke
21:34:08 <notmyname> great. we'll check back on this one next week
21:34:16 <notmyname> #topic priority reviews
21:34:25 <clayg> i didn't see anything in the bug report that indicated that LC_ALL=C.UTF-8 was an an acceptable work around
21:34:27 <notmyname> https://review.openstack.org/#/c/317475/ hasn't landed
21:34:27 <patchbot> notmyname: patch 317475 - swift - Send correct size in POST async update for EC object
21:34:57 <notmyname> torgomatic: you've already looked at it once. can you look again to give a review vote?
21:35:08 <torgomatic> notmyname: sure
21:35:14 <notmyname> can someone else also volunteer to review that patch please?
21:35:49 <kota_> I'm
21:36:09 <timburke> i probably will, too. i've looked at a similar change on crypto
21:36:25 <notmyname> kota_: timburke: thanks
21:36:49 <notmyname> I think that once patch 317475 lands I'll be able to tag a release
21:36:49 <patchbot> notmyname: https://review.openstack.org/#/c/317475/ - swift - Send correct size in POST async update for EC object
21:37:17 <notmyname> other than the crypto reviews that we already talked about, there's nothing else listed on the priority reviews page
21:37:25 <notmyname> anything that should be there?
21:37:31 <jrichli> notmyname: could we add container-sync?
21:37:38 <jrichli> patch 210099 and patch 225338
21:37:38 <patchbot> jrichli: https://review.openstack.org/#/c/210099/ - swift - Add process level concurrency to container sync
21:37:39 <patchbot> jrichli: https://review.openstack.org/#/c/225338/ - swift - Add thread level concurrency to container sync
21:37:52 <notmyname> hmm...they were there
21:37:57 <notmyname> jrichli: yeah, I can add them back
21:38:01 <jrichli> thx!
21:38:26 <notmyname> #topic open discussion
21:38:31 <notmyname> anything else to bring up this week?
21:38:47 <timburke> thanks torgomatic and notmyname for getting patch 311817 in!
21:38:47 <patchbot> timburke: https://review.openstack.org/#/c/311817/ - swift - Allow concurrent bulk deletes (MERGED)
21:39:15 <notmyname> I feel like I should try to give an update on the golang/TC stuff
21:39:31 <mattoliverau> notmyname: go ahead
21:39:48 <notmyname> i'm working on a reply to a couple of emails now, but it's gotten to be a rather confusing conversation
21:40:13 <notmyname> the discussion has moved way beyond "is a project allowed to use golang"
21:40:58 <notmyname> there's a lot more support for "if it's not python, it's not openstack" than I expected
21:41:33 <notmyname> so one current thread of the conversation is about what that would mean specifically for swift
21:41:36 <notmyname> and I'm working on that
21:42:03 <notmyname> another thread of the conversation is "is swift really part of openstack"
21:42:10 <clayg> lol
21:42:15 <kota_> lol
21:42:36 <notmyname> to be honest, both of those are really stressful and frustrating for me (as anyone who's in the same office as me can attest to)
21:42:48 <notmyname> i'm working on some replies to those now
21:43:32 <notmyname> this morning I got to dig into the annals of history and find the 2-year-old "gap analysis" questions from the TC and the "is swift required" discussions from defcore
21:44:15 <notmyname> nothing at all has been decided yet
21:44:45 <acoles> notmyname: the TC meeting yesterday seemed to be divided on 'allow no golang' vs 'allow golang with guidelines' - did I understand that they assigned champions to argue each side?
21:45:23 <notmyname> sort of. personally, I hate that model of discussion. but I've got an email waiting for a reply that's based off of that
21:45:39 <redbo> I demand trial by combat
21:46:05 <notmyname> asking me to describe the technical and community consequences of having swift split into pieces and have external dependencies that aren't in openstack
21:46:33 <notmyname> frankly I think that's is nearly completely untenable and overall a very very bad idea. so I need to write down all the reasons why
21:46:45 <mattoliverau> notmyname: wow, I think we all owe you a beer or something stronger at the mid-cycle! Thanks for all your effort on this.
21:47:10 <torgomatic> it's like golang has cooties
21:47:22 <clayg> srly, so good this is your job and like... not anyone else's in this channel
21:47:29 <dfg_> get him drunk and then get him to write some emails to openstack TC. great idea :)
21:47:34 <clayg> i can not imagine having enough %^&s to give
21:48:04 <acoles> torgomatic: lurgi in my locale
21:48:06 <clayg> i'm all like "yeah I don't know what you people are talking about, let's start rebasing a patch chain and see what it looks like"
21:48:35 <notmyname> us: "hey everyone, we're going to make things faster" tc: "but wait, are you sure you're even a part of openstack?"
21:48:58 <hurricanerix> notmyname reply with http://www.openstack.org/blog/2010/10/announcing-the-openstack-2010-1-austin-release/
21:49:03 <acoles> notmyname: it's payback for messing with the tables in Austin :)
21:49:18 <mattoliverau> rofl
21:49:19 <clayg> at somepoint they'd be like "you can't commit click +A in gerrit anymore" and I'd be like "yeah I don't know what you're talking about - let's just fork to this other github project and get back to work"
21:50:04 <notmyname> I'll keep everyone updated as the conversations keep going
21:50:23 <acoles> notmyname: thanks for the update
21:50:35 <mattoliverau> +1
21:50:41 <notmyname> anything else from anyone?
21:50:43 <kota_> +1
21:51:10 <tdasilva> looking forward to your email
21:51:21 <notmyname> tdasilva: that makes one of us! :-)
21:51:31 <clayg> lol
21:51:36 <redbo> as a hedge, I'll register openerstack.org
21:51:42 <notmyname> lol
21:51:51 <mattoliverau> redbo: lol
21:52:10 <torgomatic> stackforgery.com
21:52:16 <notmyname> thanks everyone for all your work on swift. regardless of openstack mailing list threads, swift's being used all over the world at large scale in production. and that's because of your work. thanks
21:52:24 <notmyname> #endmeeting