19:02:39 <notreallymyname> #startmeeting swift
19:02:40 <openstack> Meeting started Wed Oct 30 19:02:39 2013 UTC and is due to finish in 60 minutes.  The chair is notreallymyname. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:02:41 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
19:02:43 <openstack> The meeting name has been set to 'swift'
19:03:00 <notreallymyname> welcome to the swift team meeting today
19:03:03 <notreallymyname> who's here?
19:03:05 <torgomatic> yo
19:03:07 <peluse> howdy
19:03:08 <portante> o/
19:03:09 <zaitcev> o/
19:03:29 * clayg isn't really here
19:03:36 * notreallymyname isn't either
19:03:50 <notreallymyname> agenda items https://wiki.openstack.org/wiki/Meetings/Swift
19:03:58 <notreallymyname> not much for me to bring up this week
19:03:59 <acoles> here
19:04:05 <notreallymyname> next week is the HK summit
19:04:13 <notreallymyname> should be fun
19:04:15 <lincolnt> Hi, Lincoln Thomas, HP, 1st time here, leading design session on Metadata Search in HK
19:04:30 <notreallymyname> lincolnt: great. looking forward to hearing from you next week
19:04:40 <lincolnt> thx
19:04:47 <notreallymyname> as for day-to-day stuff, I wanted to bring up the swift-bench separation
19:04:57 <portante> notreallymyname is usually known as notmyname, which is not his name
19:05:01 <swifterdarrell> o/
19:05:02 <notreallymyname> ah yes
19:05:21 <notreallymyname> I'm notmyname, but my bouncer is not working (or the guest wifi I'm on...)
19:05:22 <keving1> here
19:05:34 <notreallymyname> #topic swift-bench separation
19:05:40 <notreallymyname> it's pretty much ready to go
19:05:58 <notreallymyname> the goals here are (among other things) to remore the dependency of swift on python-swiftclient
19:06:13 <clayg> remore is a thing
19:06:14 <notreallymyname> so we need to make sure the rest of swift-bench is good to go and then extract it from swift itself
19:06:51 <portante> so the copy has been done, no other changes made to the swift-bench code on the swift repo
19:07:13 <portante> so the next thing to do is git rm the tree and change the requirements.txt file/
19:07:14 <portante> ?
19:07:21 <swifterdarrell> notreallymyname: last I checked (@ hackathon), the swift-bench repo needed a bit of work: setup.py, (pbr ANYONE?!), Changelog, etc, etc
19:07:23 <notreallymyname> right (sorry, had someone ask aquestion)
19:07:45 <notreallymyname> ya, swift-bench needs basic stuff like the testrunner, a readme, etc
19:07:54 <notreallymyname> setup.py
19:07:57 <notreallymyname> homefully not pbr ;-)
19:08:24 <notreallymyname> so we need someone to do it
19:08:34 * zaitcev hides
19:08:38 <notreallymyname> should take a day or two, mostly boilerplate stuff
19:08:47 <peluse> sounds so exciting
19:08:53 <notreallymyname> I won't be able to do it until after hong kong at the earliest
19:09:15 <notreallymyname> I guess we don't need a volunteer right now, but it's a thing that needs doing
19:09:18 <notreallymyname> moving on
19:09:29 <notreallymyname> #topic EC/policy status
19:09:38 <notreallymyname> torgomatic: peluse: keving1: any updates?
19:09:48 <peluse> so I've got one slide at the "Intel and OpenStack" session in HK covering EC - just FYI for everyone
19:10:01 <notreallymyname> cool
19:10:02 <peluse> ..and on the code front just waiting on you torgotmatic to review a few things :)
19:10:13 <torgomatic> sorry, I've been busy :|
19:10:24 <peluse> notmyname - do we want to do an impromptu polciies demo of any kind?
19:10:27 <torgomatic> too many things going on at once, and I don't multitask particularly well
19:10:28 <notreallymyname> I've got a tech session on it, and I'd like to do a high-level demo (ie two policies in one cluster, http only, no replication etc)
19:10:34 <notreallymyname> peluse: yes
19:10:39 <peluse> nice cross typing
19:10:42 <keving1> @togomatic nothing major…  pyeclib is up on pypi
19:10:45 <peluse> shall we jsut coordinate there?
19:10:58 <notreallymyname> peluse: or online before then :-)
19:11:07 <peluse> OK, lets sync up today or tomorrow then on the demo
19:11:17 <notreallymyname> peluse: tomorrow for me. I'm at a workshop all day
19:11:29 <peluse> torgomatic - back to your comment.  no huge hurry but I am holding off an onay more work expanding on the poolcieis until you take a quick look at the two patches I ahve up there
19:11:40 <notreallymyname> ok
19:11:42 <peluse> policies that is
19:12:11 <notreallymyname> #topic 410 response
19:12:17 <notreallymyname> torgomatic: tag
19:12:37 <torgomatic> so the question here is regarding a patch that would make some account GET/HEAD requests respond w/410
19:12:56 <torgomatic> right now, there's a bug where a deleted-but-not-reclaimed account yields 204 responses on GET/HEAD requests
19:13:03 <torgomatic> but you can't PUT/POST to it
19:13:06 <notreallymyname> torgomatic: can you see what the old behavior was and leave that as a comment in gerrit?
19:13:24 <portante> meaning, the account is marked as being created
19:13:30 <portante> but that created account can't be used
19:13:32 <torgomatic> notreallymyname: yes, I can do a little code archaeology
19:13:36 <portante> does it get reclaimed too?
19:13:37 <notreallymyname> portante: 204 -- no content
19:13:48 <notreallymyname> torgomatic: thanks
19:13:52 <torgomatic> oh, also this only happens when account autocreate is on
19:14:01 <peluse> is this a reported bug?
19:14:02 <portante> but does it change the account so that it is actually created again?
19:14:27 <torgomatic> portante: I don't think so
19:14:44 <portante> it seems that autocreate on should create the account and not return a 410
19:14:52 <portante> seems counter to the notion of auto-create, no?
19:14:56 <notreallymyname> why did we add the autocreate magic responses? ;-)
19:15:08 <portante> or am I missing something else about the nature of the API?
19:15:10 <torgomatic> there's an interval after an account is deleted in which it cannot be recreated
19:15:15 <torgomatic> that's to let the reaper do its job
19:15:37 <portante> and you are saying that is true today before this change
19:15:39 <torgomatic> accounts are weird in that you can delete them when nonempty
19:15:48 <portante> oh
19:15:48 <torgomatic> portante: that's always been true AFAIK
19:15:57 <notreallymyname> "you" being a cluster operator (ie reseller admin)
19:16:44 <torgomatic> anyway, I can investigate and see what the behavior was before the fake-account-listing stuff happened
19:16:56 <portante> so by adding the 410 response behavior, we are acknowledging what is actually happening on the backend
19:17:04 <notreallymyname> portante: ya
19:17:14 <torgomatic> I just want to give people a heads-up that this might be happening, so if you have any insight or opinions, go add yourself as a watcher on that review
19:17:31 <torgomatic> because we don't currently emit 410 responses, but we might be about to start
19:17:58 <notreallymyname> but clients will properly respond to classes of responses like they should, right? ;-)
19:18:10 <torgomatic> and that's worth bringing up at a meeting so interested parties can go look, and uninterested parties can ignore it :)
19:18:34 <notreallymyname> indeed. thanks for doing so
19:18:41 <portante> can you post the review id here?
19:18:45 <notreallymyname> anything else on that for today (ie until we know the history)
19:18:45 <torgomatic> notreallymyname: well, existing behavior is 204 on HEAD plus 401 (403?) on PUT/POST, so that's clearly wrong and has to be fixed
19:18:54 <portante> https://review.openstack.org/#/c/54449/
19:18:54 * torgomatic has nothing else
19:18:55 <notreallymyname> yes
19:18:58 <portante> done
19:19:11 <notreallymyname> portante: and the author is otherjon in #openstack-swift
19:19:20 <notreallymyname> #topic open discussion
19:19:26 * notreallymyname wants to get to lunch
19:19:31 <notreallymyname> what else is going on?
19:19:47 <notreallymyname> anything needing discussion or about HK next week?
19:20:12 <portante> when is the next target for a swift release?
19:20:16 <portante> dec or jan?
19:20:18 <notreallymyname> FWIW, there were about 60 new swift "clusters" deployed this morning at the workshop
19:20:27 <portante> nice
19:20:49 <portante> and what are we targetting for that release?
19:21:01 <notreallymyname> portante: we don't do time-based releases. it would make sense to me to release after swift-bench is separated and to check with the other features under dev
19:21:34 <notreallymyname> portante: since 1.10 I haven't had a chance to look, but I'd like to see the diskfile stuff cleaned up before a release (in an ideal world)
19:21:42 <portante> redhat would like to see the account and container server API refactoring land so that we can complete the gluster layering on supported APIs
19:21:42 <notreallymyname> portante: that would be grounds for a release IMO
19:22:01 <notreallymyname> kk
19:22:14 <portante> zaitcev is working through those changes, so we'd like to get eyes on them
19:22:34 <portante> notreallymyname: okay, thanks
19:22:42 <notreallymyname> anything else?
19:22:51 <peluse> not from my end
19:22:59 <portante> nothing here
19:23:03 <torgomatic> nope
19:23:20 <notreallymyname> ok, have a good rest of the day. thanks for coming
19:23:22 <notreallymyname> #endmeeting