Friday, 2014-08-29

notmynametsg: oh nice00:00
notmynametsg: what's the outstanding patch that requires that?00:00
notmynametsg: or is it yet to be written?00:01
*** marcusvrn has quit IRC00:01
notmynamein order to do gate testing on that functionality, it will have to be added to our requirements.txt. in order for that to happen, it will need to be in global-requirements00:01
tsgnotmyname: the footer patch - WIP ..00:02
notmynameand so since we're getting to the end of the juno cycle, global requirements may go into a freeze next week (well, there's a feature freeze for other projects starting next week--not sure if that affects requirements)00:02
pelusenotmyname: how long is the freeze for?  I assume we're not concerned about the juno cutoff, just delays in getting it in because of the freeze?00:03
tsgnotmyname: I will ping Sergey  (eventlet author) again tonight00:03
notmynametsg: basically, distros look to global requirements to build their packages. so if we need this for testing, even if it doesn't land until after juno, we'll need to get it in soon. otherwise we may have to wait until after juno to land that patch00:03
notmynamepeluse: right. we (swift) arent' concerned about the freeze other that we want to make sure we can test things00:04
pelusegot it, thx00:04
tsgnotmyname: will do my best to get it in.  I am not sure about the process to get it uploaded to pypi (hope Sergey does that as part of the release)00:05
notmynamethanks00:06
notmynameI'm asking in -infra00:06
notmynameto clarify feature freezes etc00:06
notmynamethis would be more straightforward if we all just pulled of some massive effort to get ec done by juno ;-)00:07
peluseha!00:07
peluseno plans this weekend I hope for anyone...00:08
* torgomatic loves time-based releases00:08
torgomaticno, wait not love... the other one00:08
notmynamethe other fun part is if we say this is "required" or not :-)00:08
acorwintorgomatic: ive heard the opposite of love isn't hate - it's apathy00:08
notmynametsg: torgomatic: or is this actually required?00:09
torgomaticacorwin: meh00:09
notmynameacorwin: nec00:09
notmynamemeh00:09
torgomaticwell, EC can't work without the 100-continue headers, so... go from there?00:09
notmynamegot it00:10
notmynamejust making sure I say the right thing :-)00:10
pelusetsg slipped in some other fixes to eventlet to make it more attractive to do a 0.15.2 also - nice!00:11
notmyname:-)00:12
tsgpeluse: gotta show some committment :)00:12
notmynametsg: clearly, you've been a slacker until then /s ;-)00:12
*** Edward-Zhang has quit IRC00:12
tsgheh00:12
*** shri has left #openstack-swift00:13
notmynamehmm...now they're saying that since the current thing is >= then we just get it automatically and don't need to bump the global reqs until it lands in master00:13
notmynamesince they just install the latest version that matches00:13
tsgnotmyname: yeah that should work00:13
tsgnotmyname: its just that bandersnatch takes forever (!) to update pypi cache00:14
notmynameok, we can't add comments to the requirements file00:15
* peluse chuckles... "bandersnatch"00:15
tsglast time we pushed pyeclib, it never happened and I finally had clarkb kick bandersnatch to make it pull new version00:16
notmynameoh actually we _can_ add comments00:17
notmynamemaybe we can deal with the global reqs patches and then add a big comment section at the top to say the actually required versions00:18
*** kevinc_ has quit IRC00:18
*** shakamunyi has joined #openstack-swift00:18
notmynametsg: thanks. I gotta run. thanks for jumping in and for getting it done :-)00:23
tsgnp00:23
*** morganfainberg is now known as morganfainberg_Z00:32
*** morganfainberg_Z is now known as morganfainberg00:33
*** shakamunyi has quit IRC00:35
*** dmorita has joined #openstack-swift00:38
*** tumf has joined #openstack-swift00:40
*** foexle has quit IRC00:42
*** shakamunyi has joined #openstack-swift00:49
*** morganfainberg is now known as morganfainberg_Z00:53
*** gyee has quit IRC00:54
*** kota_ has joined #openstack-swift00:55
*** dmsimard is now known as dmsimard_away00:57
*** nosnos has joined #openstack-swift01:03
*** echevemaster has joined #openstack-swift01:03
*** shakamunyi has quit IRC01:09
*** addnull has joined #openstack-swift01:11
*** mwstorer has quit IRC01:12
*** tumf has quit IRC01:13
*** shakamunyi has joined #openstack-swift01:21
*** shakamunyi has quit IRC01:29
*** tumf has joined #openstack-swift02:10
openstackgerritKota Tsuyuzaki proposed a change to openstack/swift: Small Fix for FakeServerConnection  https://review.openstack.org/11767902:30
*** tumf has quit IRC02:43
*** nosnos has quit IRC02:48
*** nosnos has joined #openstack-swift02:49
*** nosnos has quit IRC02:53
*** haomaiw__ has quit IRC03:01
*** haomaiwang has joined #openstack-swift03:01
*** chandankumar has joined #openstack-swift03:11
*** haomaiw__ has joined #openstack-swift03:13
*** haomaiwang has quit IRC03:13
*** morganfainberg_Z is now known as morganfainberg03:15
*** haomaiwang has joined #openstack-swift03:17
*** mrsnivvel has joined #openstack-swift03:19
*** haomaiw__ has quit IRC03:20
*** haomaiwang has quit IRC03:23
*** haomaiwang has joined #openstack-swift03:23
*** chandankumar has quit IRC03:25
*** anticw_ is now known as anticw03:35
*** tumf has joined #openstack-swift03:40
*** haomai___ has joined #openstack-swift03:41
*** haomaiwang has quit IRC03:44
*** nosnos has joined #openstack-swift03:44
*** chandankumar has joined #openstack-swift04:13
*** tumf has quit IRC04:13
*** sanchit has joined #openstack-swift04:14
sanchitHi, Just like we can give Read permissions to all users from all referring domains by using .r:*, .rlistings. Can we do something similar for giving Container Write permissions?04:16
sanchitCurrently I am giving -H "X-Container-Read: .r:*,.rlistings"04:17
sanchitWhat can be the parameters for X-Container-Write?04:17
kota_sanchit: swift doesn't allow the referrer ACL format for write.04:32
*** chandankumar has quit IRC04:32
kota_sanchit: I don't know the reason but swift clearly deny to set such a ACL. (e.g. X-Container-Write: .r:*)04:34
*** shakamunyi has joined #openstack-swift04:38
*** tsg has quit IRC04:38
*** mahatic has joined #openstack-swift04:40
*** shakamunyi has quit IRC04:58
*** chandankumar has joined #openstack-swift05:01
*** nosnos has quit IRC05:01
*** nosnos has joined #openstack-swift05:01
*** nosnos_ has joined #openstack-swift05:03
*** nosnos has quit IRC05:04
sanchit@kota_: Thank you for the information.05:07
*** tumf has joined #openstack-swift05:08
*** chandankumar has quit IRC05:08
sanchitI was trying to set X-Container-Write:* thinking that all the users of all the account can write to that container but that also isn't allowed I guess05:08
*** tumf has quit IRC05:08
*** tumf has joined #openstack-swift05:09
*** nosnos_ has quit IRC05:11
*** nosnos has joined #openstack-swift05:11
*** chandankumar has joined #openstack-swift05:15
*** haomai___ has quit IRC05:18
*** haomaiwang has joined #openstack-swift05:19
*** tumf has quit IRC05:19
*** haomai___ has joined #openstack-swift05:20
*** haomaiwang has quit IRC05:20
*** chandankumar has quit IRC05:21
*** chandankumar has joined #openstack-swift05:22
*** Edward-Zhang has joined #openstack-swift05:25
*** occup4nt has joined #openstack-swift05:27
*** sanchit has quit IRC05:28
*** occupant has quit IRC05:29
*** occupwnt has joined #openstack-swift05:31
*** occup4nt has quit IRC05:34
*** haomaiwang has joined #openstack-swift05:35
*** haomai___ has quit IRC05:38
*** nosnos has quit IRC05:42
*** nosnos has joined #openstack-swift05:44
*** sungju has quit IRC05:47
*** dkwon has joined #openstack-swift05:48
*** occupwnt has quit IRC05:48
*** occupwnt has joined #openstack-swift05:49
*** nosnos has quit IRC05:49
*** nosnos has joined #openstack-swift05:49
*** tumf has joined #openstack-swift05:49
*** dkwon has quit IRC05:50
*** nosnos has quit IRC05:53
*** tsg has joined #openstack-swift05:54
*** occup4nt has joined #openstack-swift05:54
*** chandankumar has quit IRC05:56
*** occupwnt has quit IRC05:56
*** nosnos_ has joined #openstack-swift05:57
*** StevenK has quit IRC06:12
*** StevenK has joined #openstack-swift06:13
*** tumf has quit IRC06:23
*** joeljwright has joined #openstack-swift06:25
*** shakamunyi has joined #openstack-swift06:25
*** chandankumar has joined #openstack-swift06:25
*** shakamunyi has quit IRC06:29
*** chandankumar has quit IRC06:55
*** yuanz has joined #openstack-swift07:03
*** bvandenh has joined #openstack-swift07:10
*** nosnos_ has quit IRC07:19
*** tumf has joined #openstack-swift07:19
*** nosnos has joined #openstack-swift07:20
*** nosnos has quit IRC07:24
*** shakamunyi has joined #openstack-swift07:26
*** shakamunyi has quit IRC07:30
openstackgerritMatthew Oliver proposed a change to openstack/swift: Add concurrent reads option to proxy  https://review.openstack.org/11771007:35
*** nosnos has joined #openstack-swift07:42
*** tumf has quit IRC07:53
*** kota_ has quit IRC08:04
*** mkollaro has joined #openstack-swift08:06
*** tumf has joined #openstack-swift08:08
*** chandankumar has joined #openstack-swift08:13
*** tsg has quit IRC08:15
*** shakamunyi has joined #openstack-swift08:26
*** shakamunyi has quit IRC08:31
*** tumf has quit IRC08:40
*** tumf has joined #openstack-swift08:40
*** mkollaro has quit IRC08:54
*** Edward-Zhang has quit IRC08:57
*** mkollaro has joined #openstack-swift09:05
*** mkollaro has quit IRC09:05
*** mkollaro has joined #openstack-swift09:05
*** mkollaro has quit IRC09:06
*** mkollaro has joined #openstack-swift09:06
*** mkollaro has quit IRC09:07
*** mkollaro has joined #openstack-swift09:07
*** haomaiwang has quit IRC09:09
*** chandankumar has quit IRC09:09
*** haomaiwang has joined #openstack-swift09:09
*** chandankumar has joined #openstack-swift09:11
*** bvandenh has quit IRC09:16
*** peluse has quit IRC09:16
*** bvandenh has joined #openstack-swift09:17
*** peluse has joined #openstack-swift09:17
*** sendak.freenode.net sets mode: +v peluse09:17
*** yuanz has quit IRC09:17
*** mahatic has quit IRC09:17
*** mikehn has quit IRC09:17
*** annegentle has quit IRC09:17
*** lpabon has quit IRC09:17
*** jd__ has quit IRC09:17
*** dfg_ has quit IRC09:17
*** notmyname has quit IRC09:17
*** evanjfraser has quit IRC09:17
*** Alex_Gaynor has quit IRC09:19
*** kragniz has quit IRC09:20
*** Alex_Gaynor has joined #openstack-swift09:21
*** mikehn has joined #openstack-swift09:22
*** haomai___ has joined #openstack-swift09:24
*** haomaiwang has quit IRC09:26
*** shakamunyi has joined #openstack-swift09:27
*** yuanz has joined #openstack-swift09:27
*** mahatic has joined #openstack-swift09:27
*** annegentle has joined #openstack-swift09:27
*** lpabon has joined #openstack-swift09:27
*** jd__ has joined #openstack-swift09:27
*** evanjfraser has joined #openstack-swift09:27
*** notmyname has joined #openstack-swift09:27
*** sendak.freenode.net sets mode: +v notmyname09:27
*** dfg has joined #openstack-swift09:27
*** kragniz has joined #openstack-swift09:29
*** shakamunyi has quit IRC09:32
*** aix has joined #openstack-swift09:40
*** haomai___ has quit IRC09:45
*** AbyssOne_ has joined #openstack-swift09:46
*** AbyssOne has quit IRC09:46
*** anteaya has quit IRC09:46
*** anteaya has joined #openstack-swift09:48
*** marzif_ has joined #openstack-swift09:48
*** jeblair_ has joined #openstack-swift09:48
*** otherjon_ has joined #openstack-swift09:48
*** mhu1 has joined #openstack-swift09:50
*** haomaiwa_ has joined #openstack-swift09:50
*** jeblair has quit IRC09:50
*** marzif has quit IRC09:50
*** hugokuo has quit IRC09:50
*** otherjon has quit IRC09:50
*** cebruns has quit IRC09:50
*** mhu has quit IRC09:50
*** acorwin has quit IRC09:50
*** mhu1 is now known as mhu09:50
*** otherjon_ is now known as otherjon09:50
*** cebruns has joined #openstack-swift09:51
*** sileht has quit IRC09:51
*** acorwin has joined #openstack-swift09:51
*** hugokuo has joined #openstack-swift09:51
*** Anticimex has quit IRC09:52
*** Anticimex has joined #openstack-swift09:53
*** sileht has joined #openstack-swift09:54
*** rpedde_ has joined #openstack-swift09:55
*** matt__ has joined #openstack-swift09:58
*** zhiyan_ has joined #openstack-swift09:58
*** CJM has joined #openstack-swift09:59
*** CJM is now known as Guest113110:00
*** mariusv_ has joined #openstack-swift10:01
*** EmilienM_ has joined #openstack-swift10:01
*** pconstantine_ has joined #openstack-swift10:02
*** mitz_ has joined #openstack-swift10:03
*** sudorandom_ has joined #openstack-swift10:03
*** erlon has quit IRC10:09
*** pconstantine has quit IRC10:09
*** zhiyan has quit IRC10:09
*** early has quit IRC10:09
*** sudorandom has quit IRC10:09
*** rpedde has quit IRC10:09
*** mitz- has quit IRC10:09
*** mattoliverau has quit IRC10:09
*** EmilienM has quit IRC10:09
*** CrackerJackMack has quit IRC10:09
*** sudorandom_ is now known as sudorandom10:09
*** erlon has joined #openstack-swift10:10
*** rpedde_ is now known as rpedde10:10
*** EmilienM_ is now known as EmilienM10:10
*** early has joined #openstack-swift10:10
*** zhiyan_ is now known as zhiyan10:11
*** nosnos has quit IRC10:16
*** nosnos has joined #openstack-swift10:17
*** dmorita has quit IRC10:20
*** nosnos has quit IRC10:21
*** addnull_ has joined #openstack-swift10:22
*** echevemaster has quit IRC10:26
*** addnull_ has quit IRC10:27
*** dosaboy_ has joined #openstack-swift10:27
*** shakamunyi has joined #openstack-swift10:28
*** haomaiwa_ has quit IRC10:29
*** akp_ has joined #openstack-swift10:29
*** mkollaro has quit IRC10:30
*** jokke__ has joined #openstack-swift10:30
*** shakamunyi has quit IRC10:32
*** mariusv_ has quit IRC10:33
*** mariusv_ has joined #openstack-swift10:33
*** akp has quit IRC10:34
*** addnull has quit IRC10:34
*** jokke_ has quit IRC10:34
*** dosaboy has quit IRC10:34
*** openstackgerrit has quit IRC10:34
*** openstackgerrit has joined #openstack-swift10:41
*** mariusv_ has quit IRC10:44
*** mariusv has joined #openstack-swift10:46
*** mariusv has quit IRC10:46
*** mariusv has joined #openstack-swift10:46
*** joeljwright has quit IRC10:47
*** jdag__ has joined #openstack-swift10:49
*** swills` has joined #openstack-swift10:50
*** preeto_ has joined #openstack-swift10:51
*** clayg_ has joined #openstack-swift10:51
*** rpedde_ has joined #openstack-swift10:54
*** ujjain2 has joined #openstack-swift10:54
*** joearnold has quit IRC10:55
*** preeto has quit IRC10:55
*** ondergetekende has quit IRC10:55
*** swills has quit IRC10:55
*** clayg has quit IRC10:55
*** EmilienM has quit IRC10:55
*** rpedde has quit IRC10:55
*** dvorkbjel has quit IRC10:55
*** jdag_ has quit IRC10:56
*** mtreinish has quit IRC10:56
*** ujjain has quit IRC10:56
*** rpedde_ is now known as rpedde10:56
*** joearnold_ has joined #openstack-swift10:56
*** joearnold_ is now known as joearnold10:56
*** EmilienM has joined #openstack-swift10:56
*** ondergetekende has joined #openstack-swift10:57
*** mtreinish has joined #openstack-swift10:57
*** jdag__ is now known as jdag_10:58
*** dvorkbjel has joined #openstack-swift10:59
*** addnull has joined #openstack-swift11:00
*** mariusv__ has joined #openstack-swift11:02
*** ondergetekende has quit IRC11:03
*** ondergetekende has joined #openstack-swift11:04
*** dvorkbjel has quit IRC11:05
*** mariusv__ has quit IRC11:06
*** mariusv__ has joined #openstack-swift11:07
*** dvorkbjel has joined #openstack-swift11:08
*** akp_ has quit IRC11:10
*** akp has joined #openstack-swift11:10
*** Dafna has joined #openstack-swift11:12
*** addnull has quit IRC11:12
*** addnull has joined #openstack-swift11:13
*** mkollaro has joined #openstack-swift11:15
*** mkollaro has quit IRC11:16
*** mkollaro has joined #openstack-swift11:16
*** addnull has quit IRC11:17
*** torgomatic has quit IRC11:20
*** torgomatic has joined #openstack-swift11:20
*** ChanServ sets mode: +v torgomatic11:20
*** addnull has joined #openstack-swift11:21
*** TaiSHi has quit IRC11:23
*** mkollaro1 has joined #openstack-swift11:27
*** mkollaro has quit IRC11:27
*** joearnold has quit IRC11:27
*** anderstj has quit IRC11:27
*** TaiSHi has joined #openstack-swift11:27
*** shakamunyi has joined #openstack-swift11:28
*** preeto has joined #openstack-swift11:29
*** j_king_ has joined #openstack-swift11:29
*** preeto_ has quit IRC11:30
*** glange has quit IRC11:30
*** j_king has quit IRC11:30
*** dosaboy_ has quit IRC11:30
*** dosaboy has joined #openstack-swift11:30
*** mariusv__ has quit IRC11:31
*** chandankumar has quit IRC11:32
*** shakamunyi has quit IRC11:33
*** anderstj has joined #openstack-swift11:34
*** joearnold has joined #openstack-swift11:35
*** glange has joined #openstack-swift11:35
*** ChanServ sets mode: +v glange11:35
*** chandankumar has joined #openstack-swift11:44
*** jokke__ is now known as jokke_11:47
*** developer has joined #openstack-swift11:58
developerHas anyone ever tried installing/configuring stunnel with Swift?11:59
developerAnd know what's the best way to do so?11:59
ahale_we tried that but ended up using haproxy 1.5 ssl termination, can't remember the details but it was kinda sucky12:01
*** marcusvrn has joined #openstack-swift12:03
developerOk, but were you able to get stunnel set up successfully?12:07
developerand if so, is there a specific set of instructions for setting it up for use with Swift?12:09
*** HenryG has joined #openstack-swift12:12
*** jokke_ has quit IRC12:14
*** jokke_ has joined #openstack-swift12:14
ahale_oh yeah, it was able to get setup and working fine - nothing specific i can remember about it though12:15
*** peluse has quit IRC12:16
*** peluse has joined #openstack-swift12:16
*** TaiSHi has quit IRC12:18
*** TaiSHi has joined #openstack-swift12:23
*** mahatic has quit IRC12:24
*** shakamunyi has joined #openstack-swift12:29
*** shakamunyi has quit IRC12:34
*** chandankumar has quit IRC12:49
*** tumf has quit IRC13:00
*** swills` is now known as swills13:26
*** swills has quit IRC13:28
*** swills has joined #openstack-swift13:28
*** shakamunyi has joined #openstack-swift13:30
*** tumf has joined #openstack-swift13:31
*** shakamunyi has quit IRC13:35
*** annegent_ has joined #openstack-swift13:45
*** bvandenh has quit IRC13:52
*** dmsimard_away is now known as dmsimard13:55
*** kevinc_ has joined #openstack-swift14:02
*** tumf has quit IRC14:04
*** addnull has quit IRC14:05
*** addnull has joined #openstack-swift14:05
*** addnull has quit IRC14:09
*** addnull has joined #openstack-swift14:18
*** shakamunyi has joined #openstack-swift14:30
*** tdasilva has joined #openstack-swift14:33
*** homegrown has joined #openstack-swift14:36
*** jergerber has joined #openstack-swift14:37
*** ZackaryDoyle has quit IRC14:38
*** shakamunyi has quit IRC14:42
*** astellwag has quit IRC14:43
*** j_king_ has quit IRC14:43
*** j_king has joined #openstack-swift14:45
*** astellwag has joined #openstack-swift14:45
*** tsg has joined #openstack-swift14:45
*** shakamunyi has joined #openstack-swift15:00
*** tumf has joined #openstack-swift15:01
*** samwebster has joined #openstack-swift15:03
*** PsionTheory has joined #openstack-swift15:10
*** addnull has quit IRC15:13
*** jeblair_ is now known as jeblair15:17
*** Midnightmyth has joined #openstack-swift15:19
homegrownwhen browsing to a container with around 500k objects in it, it hangs for around 15 seconds - any tips on speeding this up?15:19
*** haomaiwa_ has joined #openstack-swift15:26
*** kevinc_ has quit IRC15:34
*** tumf has quit IRC15:34
*** mrsnivvel has quit IRC15:35
*** haomaiwa_ has quit IRC15:37
*** judd7 has joined #openstack-swift15:37
*** kevinc_ has joined #openstack-swift15:40
*** addnull has joined #openstack-swift15:50
*** addnull has quit IRC15:52
*** homegrown has left #openstack-swift15:58
*** annegent_ has quit IRC16:06
*** pberis has quit IRC16:07
*** clayg_ is now known as clayg16:14
*** ChanServ sets mode: +v clayg16:14
torgomatichomegrown: define "browsing"16:26
clayglol @ torgomatic16:27
torgomaticwell, Swift doesn't really give you a browser-consumable view of a container, which probably means there's some piece of intermediate software that's translating to HTML.16:29
claygtorgomatic: yeah, but whatever that thing is it's making probably 50 big 'ol container listing requests16:29
pelusetorgomatic:  see http://en.wikipedia.org/wiki/Browsing  :)16:29
torgomaticI mean, technically the text/plain response will render in a browser, but...16:30
claygpeluse: that page does not include anything about swift16:30
claygtorgomatic: we have json too - and browsers love xml16:30
peluseclayg:  that was sorta the point, guess my humor is off on a Friday...16:30
*** tumf has joined #openstack-swift16:31
claygpeluse: lol - no just playing dumb and dishing it back out16:31
claygpeluse: irc friday fail16:31
claygpoor homegrown - he didn't even stick around to see all this fun16:31
torgomatic"love" and "xml" are two words that never belong together :)16:31
* peluse hears the "wooonk wooonk" sound somehwere in the background16:31
torgomaticThis is what I get for trying to be helpful :)16:32
pelusetough crowd...16:32
pelusehometown: we really will try to help though... honest.  Just provide a bit more context if you can.  Also, are your acct/container DB's on SSDs?16:33
claygtorgomatic: do you think the status code for container listings should be 203 - because of eventual consistency?16:33
claygpeluse: yeah I really think he left?16:33
torgomaticProbably gone. I'm on my phone atm and it stinks at telling me who's in a channel16:34
peluseyou know, my IRC client doesn't provide any timestamps - does anyone else's?  I can't tell how long ago anyone said anything16:34
claygpeluse: irssi does, but I think i had to ask it to - i hate configuring irssi - i wouldn't recommend it16:35
pelusebah, not worth it :)16:36
*** gyee has joined #openstack-swift16:36
*** Dafna has quit IRC16:36
peluseman, and why do I keep losing my + sign?  Sheesh, I'm ready for the 3 day weekend...16:36
pelusetorgomatic:  reviewing your 0 copy stuff... PFC.  What drove this effort, just random though "I bet I could..."?16:37
*** echevemaster has joined #openstack-swift16:38
torgomaticpeluse: I wanted to make dense storage servers work better, and I saw redbo's sendfile thing and wondered if there were a way to do that without losing quarantine16:39
torgomaticand there was16:39
pelusecool16:40
*** judd7 has quit IRC16:42
*** mkollaro1 has quit IRC16:47
*** zul has quit IRC16:48
*** samwebster has quit IRC16:49
*** samwebster has joined #openstack-swift16:52
*** samwebster has quit IRC16:58
notmynamegood morning16:59
*** mahatic has joined #openstack-swift16:59
*** developer has quit IRC17:01
*** tsg has quit IRC17:02
*** tsg has joined #openstack-swift17:03
*** tumf has quit IRC17:04
*** cdelatte has joined #openstack-swift17:04
*** mahatic has quit IRC17:07
pelusegood morning!17:07
notmynameI'll be your Swift social media maven this morning....17:10
notmynameI've got a big customer meeting this morning. big enough that I'm wearing nice shoes and a sport coat. good times17:11
pelusea sport coat??? wow cool cat17:11
*** echevemaster has quit IRC17:12
*** fifieldt has joined #openstack-swift17:12
tdasilvapeluse: I've been using XChat for a few weeks now and it shows timestamps...17:14
notmynamelimechat does too17:14
notmynamegotta run...17:16
pelusetdasilva:  thanks, I think I'll try that17:16
pelusenotmyname:  enjoy your dog and pony show!17:16
*** peluse has quit IRC17:17
*** mahatic has joined #openstack-swift17:18
*** zul has joined #openstack-swift17:19
*** chandankumar has joined #openstack-swift17:20
*** peluse has joined #openstack-swift17:21
*** mahatic has quit IRC17:22
openstackgerritDavid Goetz proposed a change to openstack/swift: Combine acc/cont put_* methods and fix their lock problem.  https://review.openstack.org/11784217:23
pelusetdasilva, : OK, using it now.  Looks pretty good17:25
tdasilvapeluse: yeah, I had been using Quassel before but notifications never worked well for me...XChat seems to have a simpler UI but works pretty well17:26
dfgnotmyname: what does SwiftStack Cluster CI mean?17:27
dfgdoes it just fail it if its from me? that would kinda make sense17:27
pelusedfg:  heh, that's funny. he stepped away but we were involved a little with that.  its a small rack of systems that have our latest system on a chip in them (Avoton) and some big drives from HGST I think to suplement VM testing with real HW17:30
pelusedfg:  its all housed and managed by the guys at SwiftStack though so I can't tell you much more about it than that :)17:31
dfgpeluse: ok cool. thx. it puts the console.log into RAX's CDN? thats kinda funny. i get a 404 though17:32
pelusedfg: I'm not sure that its totally live and ready for use yet, could be that its visible but not quite ready.  obviously notmyname will know when he gets back17:34
*** mariusv has quit IRC17:34
*** annegent_ has joined #openstack-swift17:35
*** shri has joined #openstack-swift17:37
*** tsg has quit IRC17:37
*** tsg has joined #openstack-swift17:39
*** annegent_ has quit IRC17:39
*** nunizacu has joined #openstack-swift17:46
nunizacuhi, i got a problem - swift list returns  400 Bad Request   Token exceeds maximum length17:47
nunizacuswauth 50017:47
nunizacuor forbidden17:48
nunizacuis all openstack channels dead?17:51
*** mahatic has joined #openstack-swift17:52
tdasilvanunizacu: swauth was developed by gholt, who is not here at the moment, I think...if other people are not responding is probably because they can't help17:53
tdasilvanunizacu: you could also try posting a new issue on swauth github page: https://github.com/gholt/swauth/issues17:54
*** aix has quit IRC17:54
*** judd7 has joined #openstack-swift17:56
nunizacui dont really have an account17:56
nunizacuproblem is i set swpasswd but it is unusable, i se no idiot-proof guide how to do it for programer17:57
nunizacui even had to change ports from 6000 to 7000 couse they caused ssh disconnection17:59
*** tumf has joined #openstack-swift18:01
claygdfg: where did you see "SwiftStack Cluster CI"18:05
peluseclayg, :  its in the list:  https://wiki.openstack.org/wiki/ThirdPartySystems18:08
tdasilvatorgomatic: Do you have time to talk about the constraints middleware patch? 113325...18:12
torgomatictdasilva: sure18:15
tdasilvatorgomatic: was just trying to understand your comments :-) I'm still a little bit confused about the distinction between "global" vs "artificial" restrictions...18:16
torgomatictdasilva: it's more about restrictions that break certain storage modes vs. ones that don't18:17
torgomaticlike if you store a 6 GiB file on an object server, then drop the restriction to 5 GiB, that 6er doesn't break replication or anything18:17
torgomaticbut if you have a diskfile with certain naming restrictions, and there exists an object in the cluster with a name outside those restrictions, you can't ever use that diskfile18:18
torgomatics/cluster/policy/g nowadays, but the point remains18:18
*** judd7 has quit IRC18:20
tdasilvaI think this is the part where my confusion is...for example, again using the example of swiftonfile because that is easy...are saying that one should not add a swiftonfile as a separate policy to an existing cluster because it tithens the restrictions?18:20
*** judd7 has joined #openstack-swift18:20
torgomatictdasilva: say you have policy N, and it's got a bunch of plain old obj servers behind it... just stuff on XFS in the usual way18:21
torgomaticand someone comes along and says PUT /v1/AUTH_me/stuff/some///obj///name18:21
torgomaticnow you can't ever add a swift-on-file obj server to that policy, because if replication tries to move /v1/AUTH_me/stuff/some///obj///name onto it, then the operation will fail18:22
tdasilvaYes, I agree...but it would be fine in a separate policy, correct?18:23
tdasilvamy point being that I see the restrictions being per policy18:23
tdasilvaand my expectation is that disk files are the "same" in a specific policy18:24
torgomatictdasilva: I don't think I agree with that one-diskfile-per-policy premise18:24
torgomaticI don't think operators want to run different types of storage in parallel in their clusters for long periods of time, but during migrations, they'll have to18:24
torgomaticso someone switching out XFS for swift-on-file, for example18:25
dfgclayg: it added itself as a reviewer to my patch and then said that it failed18:25
dfgor I failed. idk18:25
torgomatictdasilva: I guess what I want to know is this: can swift-on-file use an alternate mapping of object-name to on-disk-name such that it doesn't have extra restrictions above those of XFS?18:26
torgomaticlike, collapse /+ to / before saving18:26
torgomatic(and by "xfs" I mean "the diskfile that currently saves stuff on a POSIX filesystem that is typically XFS")18:26
tdasilvawell, the reason for keeping the current mapping is to make it easy for providing access to those files through other protocols. Still, I think we could solve the problem to remove that restriction18:29
tdasilvaand I'm not really disagreeing with the points you made, more just trying to understand...18:30
torgomatictdasilva: I guess my main worry is that you can't migrate to a diskfile with stricter constraints/restrictions than what you have now18:30
tdasilvayes...that is very much true...but I think that is a problem that will always be there18:31
tdasilvausing another example from John was the min_size for EC18:31
torgomaticwhich is another thing I don't like ;)18:32
torgomaticand for essentially the same reasons18:32
tdasilvahumm...interesting...the way I was seeing is that it is sort of inevitable18:32
tdasilvawhen you have different storage backends18:32
tdasilvaunless you have a global set of restrictions that intersect them all18:33
tdasilvaas you suggested, but then you are never able to "extend" your cluster18:33
torgomaticeh, not necessarily... only if those different storage backends are less capable than the base one18:33
tdasilvawith new storage backaends18:33
*** tumf has quit IRC18:34
*** samwebster has joined #openstack-swift18:35
tdasilvaright...but it is a true possibility18:35
tdasilvaagain..the idea of having a min_obj_size, if that was the case, you would never extend an existing cluster to add EC18:35
*** annegent_ has joined #openstack-swift18:36
torgomatictdasilva: exactly right, which is why a backend that can't store 0-byte files is worse, from the user's perspective, than one that can18:37
torgomaticin particular, if an operator were to set their default policy to be EC on a new cluster *and* EC couldn't store 0-byte files, then existing clients would break18:37
torgomatic(for the record, I think EC *can* store 0-byte files, so this is hypothetical)18:38
tdasilva:-) understood18:38
torgomaticit's sort of the same thing with the constraints we have today... they all limit the size of a thing to be in the range [0, N] units18:40
*** annegent_ has quit IRC18:40
torgomaticand if you set up your cluster with [0, M] where M < N, you'll break some clients that have the old limits hard-coded (from back before they were configurable)18:40
*** Trixboxer has joined #openstack-swift18:41
tdasilvai understand this point, but still not sure what you would propose18:42
torgomaticwell, if swift-on-file really can't accept certain names, then swift-on-file should have such a middleware with the extra constraints in it18:42
torgomaticbut I don't like that Swift's example of non-XFS diskfile has extra restrictions on it; I feel like that sets a bad precedent18:42
torgomaticclients end up getting an API that takes most, but not all, of the requests that other Swift clusters take18:43
tdasilvaok...yes, that's also what I was thinking...for example...keep the restrictions that are there in the proxy today18:44
tdasilvaif a new storage policy with a new diskfile is added with special constraints, they can provide their own middleware18:44
tdasilvawith no changes to proxy code18:44
tdasilvadid I understand you correctly?18:45
torgomatictdasilva: yes, that sounds good to me18:45
openstackgerritDavid Goetz proposed a change to openstack/swift: Combine acc/cont put_* methods and fix their lock problem.  https://review.openstack.org/11784218:45
tdasilvain that specific scenario, you would still have the issue you described with the clients18:46
*** annegent_ has joined #openstack-swift18:46
torgomatictdasilva: yes, but if the storage backend can't be fixed, then you're stuck with it18:46
tdasilvaexactly18:47
tdasilvaagreed18:47
torgomaticI'd argue that the first approach should always be to fix the storage backend, and only fall back on a restrictions middleware as a last resort18:47
tdasilvado you mean to fix the storage backend to try to remove the extra restrictions or so that it (storage node) returns the correct error in case of those restrictions18:48
tdasilva?18:48
torgomatictdasilva: the former; remove the extra restrictions18:49
tdasilvato a certain extent I agree, but we also have to expect that in some cases it won't be possible...for example, in the case of swiftonfile we could try to solve the issue with /, but we would still have the proble with the max_length of the "parts" (i.e. sub-directories) of the object name18:51
tdasilvaunless we also tried to truncate them18:51
torgomatictdasilva: so in that case it makes sense for swift-on-file to provide a middleware that implements the extra checking18:51
nunizaculots of talk, maybe somebody help me?18:51
torgomaticnunizacu: have you tried turning it off and on again?18:52
nunizacui till 12 pm18:52
tdasilvatorgomatic: the other issue we ran into was that when the object server returns a 400 to the proxy, the proxy in turn returns a 50318:53
torgomaticnunizacu: in all seriousness, it's better to state the problem and see who pops up than to ask for a volunteer to help18:53
nunizacuand i talk like that to customerst :D18:53
tdasilvatorgomatic: maybe we could fix that...18:53
nunizacui stated already18:53
nunizacuno wway to login into swift18:53
torgomatictdasilva: that'd be an interesting patch... I'd think a quorum of 400s would result in a 400 going out, but I could be wrong18:53
*** occup4nt is now known as occupant18:54
torgomaticnunizacu: what auth system? tempauth? swauth? keystone?18:54
nunizacu14.04, tried keystone and swatuth18:54
*** developer has joined #openstack-swift18:55
nunizacuswift gives double 8080 with too long token18:55
nunizacuswauth premission denied18:55
nunizacuand im sure i put the same password both places and restart18:56
tdasilvain swiftonfile (which has only one diskfile, so quorum of one) we get the 503 right here: https://github.com/openstack/swift/blob/master/swift/proxy/controllers/obj.py#L69918:56
tdasilvatorgomatic: ^^18:56
nunizacui get rather 4xx and 50018:56
developerHow does one see previous conversations before one has joined the conversation?18:56
torgomaticnunizacu: the double 8080 is, I think, a bug in that version of swiftclient, so you might try upgrading... token too long is because keystone PKI tokens are fantastically huge, and you have to bump up a config value so eventlet can eat them18:56
torgomaticdeveloper: IRC bouncers are handy for that18:57
tdasilvadeveloper: there's a log: http://eavesdrop.openstack.org/irclogs/%23openstack-swift/18:57
torgomatic(except when your darn phone IRC client consumes the backlog and crashes, leaving you with no idea what was said... dang it)18:57
nunizacutell me where to look, im programer for 18 years18:57
developerAll right, I see18:58
nunizacuand 16 sysadmin18:58
torgomaticnunizacu: I'm trying to see where it is18:58
*** chandankumar has quit IRC18:58
nunizacubut swauth dont work also18:58
torgomaticnunizacu: does swauth also result in token-too-long errors?18:58
torgomaticnunizacu: ah, here we go... in swift.conf, set max_header_size = 16384 (or whatever; default is 8192)18:59
* torgomatic thought it was the proxy conf and couldn't find it18:59
nunizacuoh i see the problem19:00
nunizacuat least one19:00
torgomatictdasilva: that's an odd place to see that; I'd think len(conns) == 1 and min_conns == 1, so that shouldn't happen19:01
tdasilvawell..but the conn fails19:01
tdasilvathat's why19:01
tdasilvabecause the call to check_constraints in the obj server happens when trying to setup the connection19:01
torgomatictdasilva: apparently I need to go digging in that code some more; I thought conns didn't fail unless it was a TCP-level failure, but I believe you19:02
tdasilvatorgomatic: I could mistaken too, so I will check again, but from what I understood before when looking at this, is if this check fails: https://github.com/openstack/swift/blob/master/swift/obj/server.py#L37619:05
tdasilvathen we get that 503 in the controller19:05
torgomatictdasilva: yeah, I'll go read through that code some more19:06
tdasilvatorgomatic: thanks for taking the time19:09
*** chandankumar has joined #openstack-swift19:12
*** zul has quit IRC19:18
*** kevinc_ has quit IRC19:19
*** annegent_ has quit IRC19:21
*** cdelatte has quit IRC19:25
*** tumf has joined #openstack-swift19:31
*** tsg has quit IRC19:37
*** chandankumar has quit IRC19:38
*** zul has joined #openstack-swift19:42
*** kevinc_ has joined #openstack-swift19:50
nunizacuwhats the correct owners and premissions for recon files?19:51
nunizacuno matter what i set it it changes constantly and deletes19:51
nunizacuyou got idea about this or where this additional 8080 comes from?20:01
*** tumf has quit IRC20:04
*** PsionTheory has quit IRC20:06
*** fifieldt has quit IRC20:11
*** zaitcev has joined #openstack-swift20:17
*** ChanServ sets mode: +v zaitcev20:17
*** tsg has joined #openstack-swift20:34
*** judd7 has quit IRC20:37
torgomaticnunizacu: the additional 8080 is a bug in some older versions of python-swiftclient; it's not actually present in the returned auth URL, I think20:50
torgomaticnunizacu: I could be wrong, though, but tcpdump will tell the truth no matter what20:53
torgomaticnunizacu: my recon files are owned by swift/swift, mode 060020:54
*** shakamunyi has quit IRC20:58
*** tumf has joined #openstack-swift21:00
*** judd7 has joined #openstack-swift21:15
*** pberis has joined #openstack-swift21:17
*** acoles has quit IRC21:18
*** fifieldt has joined #openstack-swift21:25
notmyname /back21:27
notmynamedfg: ya, the swiftstaclu cluster CI is just what peluse said. it's just getting set up. if it's still failing, I need to look in to that.21:28
notmynamedfg: and yes, it's putting logs into cloud files :-)21:28
*** tumf has quit IRC21:34
*** Midnightmyth has quit IRC21:34
*** pberis1 has joined #openstack-swift21:44
*** marzif__ has joined #openstack-swift21:45
*** yuanzz has joined #openstack-swift21:45
*** evanjfraser_ has joined #openstack-swift21:46
*** peluse_ has joined #openstack-swift21:48
*** acorwin_ has joined #openstack-swift21:51
*** notmyname_ has joined #openstack-swift21:51
*** ChanServ sets mode: +v notmyname_21:51
*** jd__` has joined #openstack-swift21:51
*** hugokuo_ has joined #openstack-swift21:51
*** fifieldt has quit IRC21:52
*** pberis has quit IRC21:52
*** peluse has quit IRC21:52
*** jergerber has quit IRC21:52
*** anderstj has quit IRC21:52
*** pconstantine_ has quit IRC21:52
*** hugokuo has quit IRC21:52
*** acorwin has quit IRC21:52
*** marzif_ has quit IRC21:52
*** yuanz has quit IRC21:52
*** annegentle has quit IRC21:52
*** lpabon has quit IRC21:52
*** jd__ has quit IRC21:52
*** notmyname has quit IRC21:52
*** evanjfraser has quit IRC21:52
*** hugokuo_ is now known as hugokuo21:52
*** notmyname_ is now known as notmyname21:52
*** jd__` is now known as jd__21:52
*** acorwin_ is now known as acorwin21:53
*** anderstj has joined #openstack-swift21:53
notmynamehave I missed anything today?22:00
peluse_just got back myself22:01
*** peluse_ is now known as peluse22:01
*** fifieldt has joined #openstack-swift22:02
peluseargh22:05
*** developer has quit IRC22:09
*** ChanServ sets mode: +v peluse22:11
*** peluse is now known as 7YUAAKAK522:11
*** lpabon has joined #openstack-swift22:11
*** annegentle has joined #openstack-swift22:11
*** pberis has joined #openstack-swift22:11
*** peluse has joined #openstack-swift22:11
*** jergerber has joined #openstack-swift22:11
*** 17SAA2F8E has joined #openstack-swift22:11
*** pconstantine_ has joined #openstack-swift22:11
*** 17SAA2BFU has joined #openstack-swift22:11
*** 17SAA2BFI has joined #openstack-swift22:11
*** 17SAA0LVL has joined #openstack-swift22:11
*** evanjfraser has joined #openstack-swift22:11
*** lpabon has quit IRC22:11
*** lpabon has joined #openstack-swift22:11
*** jergerber has quit IRC22:12
*** pberis has quit IRC22:12
*** peluse has quit IRC22:12
*** 17SAA2F8E has quit IRC22:12
*** 17SAA2BFU has quit IRC22:12
*** 17SAA2BFI has quit IRC22:12
*** 17SAA0LVL has quit IRC22:12
*** evanjfraser has quit IRC22:12
openstackgerritJesse J. Cook proposed a change to openstack/python-swiftclient: Python sessions, streaming, and increased pool  https://review.openstack.org/11606522:13
*** dmsimard is now known as dmsimard_away22:14
*** morganfainberg is now known as dolphm322:15
*** dolphm3 is now known as morganfainberg22:15
*** tkay has joined #openstack-swift22:24
*** tumf has joined #openstack-swift22:31
openstackgerritJesse J. Cook proposed a change to openstack/python-swiftclient: Python sessions, streaming, and increased pool  https://review.openstack.org/11606522:34
*** 7YUAAKAK5 is now known as peluse22:38
*** tsg has quit IRC22:39
peluse.22:39
notmynamepeluse: why did your nick change to 7YUAAKAK5?22:40
* notmyname suspects freenode issues22:40
*** judd7 has quit IRC22:40
*** tsg has joined #openstack-swift22:40
pelusenotmyname: no idea whats going on there... maybe this new IRC client did it?22:40
pelusenow it tells me my nick is taken but still let me use it and says I don't know my own pw22:40
*** marcusvrn has quit IRC22:41
notmynamepeluse: no, I saw other nicks change too22:41
* fifieldt wonders if they can be used as summit discount codes22:41
notmynameheh22:41
notmynamefifieldt: I hear they're handing those out like candy22:41
fifieldtdiscounts? perhaps :) I wouldn't know - colleagues mostly do that22:42
fifieldtthough I did hear a rumour that we're selling very fast22:42
fifieldtpast half way to sold out for sure22:42
fifieldtseems a lot of people want to go to paris22:42
notmynamewow22:43
*** peluse has quit IRC22:47
*** peluse has joined #openstack-swift22:48
*** acoles_away has joined #openstack-swift22:50
*** acoles_away is now known as acoles22:51
*** ChanServ sets mode: +v acoles22:51
jokke_nice22:51
*** tumf has quit IRC23:04
*** notmyname has quit IRC23:04
*** ChanServ sets mode: +v peluse23:04
*** notmyname has joined #openstack-swift23:05
*** ChanServ sets mode: +v notmyname23:05
*** Trixboxer has quit IRC23:07
openstackgerritSamuel Merritt proposed a change to openstack/swift: Fix last_modified_date_to_timestamp on non-UTC systems  https://review.openstack.org/11790723:08
torgomatichappy friday; go review some code that involves timezones ^^ :|23:08
*** jasondotstar has quit IRC23:21
*** tkay has quit IRC23:24
*** jasondotstar has joined #openstack-swift23:27
*** kevinc_ has quit IRC23:28
*** tkay has joined #openstack-swift23:29
*** tumf has joined #openstack-swift23:42
*** morganfainberg is now known as SkepticalCore23:44
matt__But it's Saturday23:51
torgomaticstupid timezones :)23:52
*** matt__ is now known as mattoliverau23:52
*** fifieldt has quit IRC23:56

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!