Tuesday, 2015-07-07

*** ho has joined #openstack-swift00:01
notmynameho: good morning00:01
redbothat's how we fixed pbr installation problems on lucid too00:01
hogood morning!00:02
notmynameredbo: http://cdn.shopify.com/s/files/1/0535/6917/products/procrastinationdemotivator.jpeg?v=141677629800:02
honotmyname: hello :-)00:02
notmynameho: when you get a chance today (if you do), can you take a look at https://bugs.launchpad.net/swift/+bug/1451710 ?00:03
openstackLaunchpad bug 1451710 in OpenStack Object Storage (swift) "http 503 error reported in swift component after changing the local" [Undecided,New]00:03
notmynameredbo: are you at gophercon this week?00:03
redboYeah, after tomorrow00:03
torgomaticnotmyname: I have not looked at that in the slightest; it's not related to anything I'm working on. Sounds like a bug in the ring-builder CLI; by the time you hit the .set_weight() method of the builder, there's no opportunity to make that goof anymore00:03
honotmyname: Ok, I will check it.00:04
notmynametorgomatic: ah ok00:04
notmynameho: thanks00:04
*** chlong has quit IRC00:08
*** lyrrad has quit IRC00:24
*** lyrrad has joined #openstack-swift00:24
*** lyrrad has quit IRC00:35
*** dmorita has joined #openstack-swift00:37
*** jrichli has joined #openstack-swift00:44
*** shlee322 has quit IRC00:47
*** jrichli has joined #openstack-swift00:48
*** zhill has quit IRC00:49
*** tongli has joined #openstack-swift01:09
*** jrichli has quit IRC01:11
*** jasondotstar has quit IRC01:12
*** gyee has quit IRC01:13
*** tongli_ has joined #openstack-swift01:13
*** shlee322 has joined #openstack-swift01:14
*** tongli has quit IRC01:14
*** tongli_ has quit IRC01:18
*** jasondotstar has joined #openstack-swift01:25
*** chlong has joined #openstack-swift01:29
openstackgerritMerged openstack/python-swiftclient: Added check for negative segment-size  https://review.openstack.org/19446401:36
*** tongli has joined #openstack-swift01:43
*** tongli has quit IRC01:47
*** strycore has quit IRC01:49
*** jkugel has joined #openstack-swift01:54
honotmyname: as for bug 1451710, i couldn't reproduce (it worked) it so i put a comment and memo on launchpad.02:07
openstackbug 1451710 in OpenStack Object Storage (swift) "http 503 error reported in swift component after changing the local" [Undecided,New] https://launchpad.net/bugs/145171002:07
*** mutoulbj has joined #openstack-swift02:12
*** jkugel has quit IRC02:15
*** jkugel has joined #openstack-swift02:16
honotmyname: s/ (it worked) it so/it (it worked) so/02:18
*** jkugel has quit IRC02:21
*** jkugel has joined #openstack-swift02:23
*** mutoulbj_ has joined #openstack-swift02:27
*** hrou has joined #openstack-swift02:28
*** mutoulbj has quit IRC02:30
*** kota_ has joined #openstack-swift02:38
*** ChanServ sets mode: +v kota_02:38
*** tongli has joined #openstack-swift02:46
notmynameho: thank you02:46
notmynameho: I marked it as incomplete pending more info from the submitter02:47
*** jkugel has quit IRC02:47
honotmyname: you are welcome :) yeah, we need more info for further investigation.02:49
*** sanchitmalhotra has joined #openstack-swift03:06
mattoliverausigh.. its always the way. Decide to go implement something that should be "easy" and then you find yourself down a rabbit hole of refactoring and wondering why we did things certain ways. I think I need to break for lunch.. that'll help the brain :)03:21
*** shlee322 has quit IRC03:21
mattoliverauAlthough I may need to ask questions later.. but that's future matt's problem.. present matt will go eat  :P And question way be in the form of in-line comments in my own patch later :)03:21
*** jkugel has joined #openstack-swift03:21
*** jkugel1 has joined #openstack-swift03:24
*** jkugel has quit IRC03:26
*** sanchitmalhotra1 has joined #openstack-swift03:29
*** sanchitmalhotra has quit IRC03:31
*** sanchitmalhotra has joined #openstack-swift03:38
*** sanchitmalhotra1 has quit IRC03:40
*** sanchitmalhotra1 has joined #openstack-swift03:49
*** sanchitmalhotra has quit IRC03:51
*** sanchitmalhotra has joined #openstack-swift04:03
*** sanchitmalhotra1 has quit IRC04:05
*** tongli has quit IRC04:08
*** SkyRocknRoll has joined #openstack-swift04:16
*** shlee322 has joined #openstack-swift04:42
*** zhill has joined #openstack-swift04:43
*** vinsh has quit IRC04:51
*** sanchitmalhotra1 has joined #openstack-swift05:07
*** sanchitmalhotra has quit IRC05:09
*** SkyRocknRoll has quit IRC05:27
*** Trozz_ has left #openstack-swift05:45
*** peluse has quit IRC05:50
*** peluse has joined #openstack-swift05:51
*** ig0r_ has joined #openstack-swift05:51
*** _hrou_ has joined #openstack-swift05:54
*** ig0r__ has quit IRC05:55
*** hrou has quit IRC05:57
*** SkyRocknRoll has joined #openstack-swift06:02
*** saltsa has joined #openstack-swift06:03
*** vinsh has joined #openstack-swift06:08
*** _hrou_ has quit IRC06:10
*** zhill has quit IRC06:11
*** vinsh has quit IRC06:13
*** jith_ has joined #openstack-swift06:27
jith_hi all.. what is the difference between user auth token and composite token?06:28
*** sanchitmalhotra has joined #openstack-swift06:36
*** sanchitmalhotra1 has quit IRC06:38
mattoliveraujith_: auth token is what you as a user will use to login. A copisite token is a second token that a service using swift might use.06:43
*** fifieldt has joined #openstack-swift06:43
*** proteusguy has quit IRC06:47
*** adiantum has quit IRC07:09
*** geaaru has joined #openstack-swift07:12
*** shlee322 has quit IRC07:17
*** rledisez has joined #openstack-swift07:20
jith_mattoliverau: Thanks.. all objects corresponding to one project will save in one account in swift.. is it so?07:30
jith_all other services use composite token or only glance and cinder uses the composite token?07:32
mattoliveraujith_: yes, in general.  if setting up services like glace, you can set that up in different ways. But generally speaking a project/tennent maps to a swift account.07:33
mattoliverauYou can set up services to connect as a partiular user, use the projects account, or use a composite key (users key but a services key) depending on your setup and what you want to do.07:34
mattoliverau*users key and a services key == composite key07:35
*** jistr has joined #openstack-swift07:37
*** jordanP has joined #openstack-swift07:41
jith_mattoliverau: thanks a lot.. is it i can either connect the swift using composite key or as a user using some account? i am sorry if i am wrong.. i am beginner07:46
mattoliveraujith_: this might help: http://docs.openstack.org/developer/swift/overview_auth.html#openstack-service-using-composite-tokens and might be more outdated now but this is good also: http://specs.openstack.org/openstack/swift-specs/specs/done/service_token.html07:48
*** JelleB is now known as AbyssOne08:00
*** AbyssOne is now known as JelleB08:01
*** joeljwright has joined #openstack-swift08:03
*** ChanServ sets mode: +v joeljwright08:03
*** adiantum has joined #openstack-swift08:08
*** vinsh has joined #openstack-swift08:10
*** chlong has quit IRC08:13
*** vinsh has quit IRC08:14
jith_mattoliverau: thanks08:16
*** shlee322 has joined #openstack-swift08:21
*** okdas has joined #openstack-swift08:33
*** proteusguy has joined #openstack-swift08:36
*** SkyRocknRoll has quit IRC08:43
hokota_: around?08:45
kota_ho: yup08:46
hokota_: as for the hack-a-thon, I will be arrived there Aug 8th. how about you?08:47
hokota_: my concern is the ibm code can be used from 9th.08:48
kota_ho: I'm planning to arraved at Aug 9th.08:48
kota_ho: maybe 9th around noon.08:48
hokota_: i see. it's good.08:49
hokota_: thanks! i will ask to the hotel.08:49
kota_ho: k, sorry I can not help you08:50
hokota_: np, btw, which flight?08:51
kota_wait, check it out.08:51
kota_NH0174 and UA486408:52
kota_connected at Huston08:53
*** joeljwright has left #openstack-swift08:53
kota_not sure for my spel08:53
kota_spelling08:53
kota_s/Huston/Houston/08:54
kota_in Texas08:54
hokota_: NH0008 is mine. connected at sanfran.08:55
hokota_: anyway thanks a lot!08:56
kota_ho: I'm looking forward to seeing you there :-)08:56
hokota_: yeah, we don't have a chance to meet in japan :-)08:57
*** shlee322 has quit IRC08:57
*** haypo has joined #openstack-swift08:58
kota_ho: hah, exactly.09:00
*** SkyRocknRoll has joined #openstack-swift09:01
openstackgerritVictor Stinner proposed openstack/swift: Use six to fix imports on Python 3  https://review.openstack.org/18545309:03
openstackgerritVictor Stinner proposed openstack/swift: Replace urllib(2), urlparse with six.moves.urllib  https://review.openstack.org/19683509:03
haypohi. would it be possible to get reviews on my python 3 before my patches start to rot? :-) see https://review.openstack.org/#/q/status:open+project:openstack/swift+branch:master+topic:py3,n,z to get the list09:04
*** krykowski has joined #openstack-swift09:05
*** sanchitmalhotra1 has joined #openstack-swift09:09
*** sanchitmalhotra has quit IRC09:10
haypothe problem is that my patches are huge, so it's likely that they become in conflict09:13
haypoit's quite slow to get a review in swift. maybe i should just give up to port swift to python3?09:13
*** kei_yama has quit IRC09:17
*** km has quit IRC09:17
*** ho has quit IRC09:17
*** km has joined #openstack-swift09:17
*** ho has joined #openstack-swift09:18
*** kei_yama has joined #openstack-swift09:18
*** ho has quit IRC09:30
*** ho has joined #openstack-swift09:30
openstackgerritOpenStack Proposal Bot proposed openstack/python-swiftclient: Updated from global requirements  https://review.openstack.org/8925009:38
openstackgerritOpenStack Proposal Bot proposed openstack/swift: Updated from global requirements  https://review.openstack.org/8873609:38
*** dmorita has quit IRC09:42
*** shlee322 has joined #openstack-swift09:46
*** ppai has joined #openstack-swift09:58
openstackgerritVictor Stinner proposed openstack/swift: Fix tox -e py34  https://review.openstack.org/19903410:12
*** mutoulbj_ has quit IRC10:23
openstackgerritKota Tsuyuzaki proposed openstack/swift: Fix ECAppIter close backend stream correctly  https://review.openstack.org/19904310:27
*** shlee322 has quit IRC10:30
*** sanchitmalhotra has joined #openstack-swift10:41
*** sanchitmalhotra1 has quit IRC10:43
*** kota_ has quit IRC11:00
*** mutoulbj has joined #openstack-swift11:08
*** delatte has quit IRC11:11
*** nexusz99 has joined #openstack-swift11:15
*** km has quit IRC11:29
*** delatte has joined #openstack-swift11:31
hohaypo: your work for porting python3 is great and i'm really interested in it because it's a good chance to learn python3. from tomorrow i will start to review your patches but i don't have enough skill & time for review so I will do it one by one.11:33
*** kei_yama has quit IRC11:37
*** tongli has joined #openstack-swift11:45
*** jasondotstar has quit IRC11:57
*** chlong has joined #openstack-swift12:05
*** nexusz99 has quit IRC12:06
*** nadeem has joined #openstack-swift12:19
*** ho has quit IRC12:19
*** tongli has quit IRC12:26
*** tongli has joined #openstack-swift12:27
*** nadeem has quit IRC12:27
*** tongli has quit IRC12:32
*** bi_fa_fu_ has joined #openstack-swift12:34
haypo"from tomorrow i will start to review your patches" cool12:35
*** vinsh has joined #openstack-swift12:39
*** bi_fa_fu_ has quit IRC12:41
*** bi_fa_fu has joined #openstack-swift12:41
*** benny_ has joined #openstack-swift12:42
*** SkyRocknRoll has quit IRC12:43
*** NM has joined #openstack-swift12:49
*** bi_fa_fu has quit IRC12:50
*** proteusguy has quit IRC12:53
*** proteusguy has joined #openstack-swift12:53
*** proteusguy has quit IRC12:56
*** proteusguy has joined #openstack-swift12:57
*** bi_fa_fu has joined #openstack-swift12:58
*** krykowski has quit IRC12:59
*** krykowski_ has joined #openstack-swift12:59
*** SkyRocknRoll has joined #openstack-swift12:59
*** benny_ has quit IRC13:02
openstackgerritCharles Hsu proposed openstack/python-swiftclient: Add ability to download objects to particular folder.  https://review.openstack.org/16028313:04
openstackgerritCharles Hsu proposed openstack/python-swiftclient: Add ability to download objects to particular folder.  https://review.openstack.org/16028313:05
*** nadeem has joined #openstack-swift13:06
*** haomaiw__ has quit IRC13:08
*** jasondotstar has joined #openstack-swift13:09
*** jasondot_ has joined #openstack-swift13:10
*** hrou has joined #openstack-swift13:12
*** _hrou_ has joined #openstack-swift13:12
*** jasondot_ has quit IRC13:13
*** jasondotstar has quit IRC13:14
*** jasondotstar has joined #openstack-swift13:14
*** hrou has quit IRC13:16
*** _hrou_ has quit IRC13:17
*** hrou has joined #openstack-swift13:18
*** haomaiwa_ has joined #openstack-swift13:22
*** ppai has quit IRC13:23
*** petertr7_away is now known as petertr713:29
openstackgerritVictor Stinner proposed openstack/swift: Fix tox -e py34  https://review.openstack.org/19903413:32
*** tongli has joined #openstack-swift13:34
*** tongli has quit IRC13:39
*** jasondotstar has quit IRC13:40
*** annegentle has joined #openstack-swift13:42
*** haomaiwa_ has quit IRC13:42
*** janonymous has joined #openstack-swift13:45
*** haomaiwang has joined #openstack-swift13:48
*** thurloat_isgone is now known as thurloat13:48
*** blmartin has joined #openstack-swift13:49
*** proteusguy has quit IRC13:55
*** annegentle has quit IRC13:57
openstackgerritVictor Stinner proposed openstack/swift: Fix tox -e py34  https://review.openstack.org/19903413:58
*** annegentle has joined #openstack-swift14:02
*** benny_ has joined #openstack-swift14:04
*** annegentle has quit IRC14:07
*** acampbell has joined #openstack-swift14:08
*** proteusguy has joined #openstack-swift14:09
*** krykowski_ has quit IRC14:14
*** SkyRocknRoll has quit IRC14:17
*** nadeem has quit IRC14:18
*** tongli has joined #openstack-swift14:20
*** chlong has quit IRC14:25
*** krykowski has joined #openstack-swift14:27
*** SkyRocknRoll has joined #openstack-swift14:30
*** SkyRocknRoll has quit IRC14:31
*** annegentle has joined #openstack-swift14:32
*** minwoob has joined #openstack-swift14:34
*** nadeem has joined #openstack-swift14:39
*** nadeem has quit IRC14:41
*** vinsh has quit IRC14:41
*** annegentle has quit IRC14:43
*** krykowski has quit IRC14:47
*** krykowski has joined #openstack-swift14:49
*** vinsh has joined #openstack-swift15:01
*** bi_fa_fu has quit IRC15:02
*** annegentle has joined #openstack-swift15:04
*** adiantum has quit IRC15:08
*** bi_fa_fu has joined #openstack-swift15:09
*** annegentle has quit IRC15:13
*** annegentle has joined #openstack-swift15:21
*** zhill has joined #openstack-swift15:40
*** proteusguy has quit IRC15:40
*** bi_fa_fu has quit IRC15:42
*** tongli has quit IRC15:50
*** tongli has joined #openstack-swift15:50
*** tongli_ has joined #openstack-swift15:54
*** zaitcev has joined #openstack-swift15:54
*** ChanServ sets mode: +v zaitcev15:54
*** tongli has quit IRC15:54
*** bi_fa_fu has joined #openstack-swift15:54
*** adiantum has joined #openstack-swift15:56
*** annegentle has quit IRC15:56
*** tongli_ has quit IRC15:58
*** jistr has quit IRC16:00
*** tongli has joined #openstack-swift16:00
*** mgarza_ has joined #openstack-swift16:00
*** mfalatic has joined #openstack-swift16:01
*** annegentle has joined #openstack-swift16:03
*** tongli_ has joined #openstack-swift16:04
*** tongli has quit IRC16:04
*** benny_ has quit IRC16:04
*** thurloat is now known as thurloat_isgone16:07
*** jistr has joined #openstack-swift16:07
*** tongli_ has quit IRC16:09
*** annegentle has quit IRC16:10
*** haomaiw__ has joined #openstack-swift16:16
*** lyrrad has joined #openstack-swift16:16
*** benny_ has joined #openstack-swift16:16
portantenotmyname: curious about the state of python3 and swift, are you folks going to support both 2 and 3 soon?16:17
*** haomaiwang has quit IRC16:18
*** petertr7 is now known as petertr7_away16:23
*** krykowski has quit IRC16:23
*** tongli has joined #openstack-swift16:24
*** jistr has quit IRC16:35
*** jordanP has quit IRC16:38
*** haypo has quit IRC16:39
notmynameportante: hmm...depends. are you coming back to join us on swift dev work? ;-)16:39
*** haypo has joined #openstack-swift16:40
*** annegentle has joined #openstack-swift16:40
*** haypo has left #openstack-swift16:41
*** NM has quit IRC16:44
*** annegentle has quit IRC16:44
portantehmm...depends. ;)16:46
*** acampbell has quit IRC16:46
*** NM has joined #openstack-swift16:47
*** tongli has quit IRC16:50
notmynameportante: :-)16:51
*** tongli has joined #openstack-swift16:51
notmynameso in seriousness, victor has been working on it, but it's big and (like every other patch in swift) it takes a while to get landed16:51
*** jon1111 has joined #openstack-swift16:52
portantenotmyname: yes, I am sure it will be big, and wonder what the performance impact will be, + = -16:52
notmynameportante: I don't think that's a motivator (or at least I haven't heard that before)16:53
portanteor is go going to be swifts' future direction?16:53
notmynameyet to be seen16:53
portanteor maybe swift should use swift? ;)16:53
portanteredbo, have you looked at swift?16:53
notmynamewe'll need to see a lot of benchmark comparisons between python and go, and I haven't seen anything yet16:53
* portante had to ask that question ...16:54
jon1111If I have a 4 replica ring and half of the cluster is down (region with half of the replicas). Would DELETE of objects fail? (network part)16:54
notmynamejon1111: with 4 replicas, you have a quorum of 3 (4 // 2 + 1 = 3)16:55
notmynamejon1111: so, yes. PUTs and DELETEs would fail16:55
notmynamejon1111: except....16:55
clayg404 is the new 20416:56
*** tongli has quit IRC16:56
notmynamejon1111: they will fail if swift can't find *any* 3 nodes to accept the operation. if the primary locations aren't available, swift will choose handoffs. so even if you get 2 primaries and 1 handoff, that's ok16:56
jon1111OK, thanks, that is what I expected.16:56
*** tongli has joined #openstack-swift16:56
jon1111Yep, they fail with a 4XX, I guess. Then later when the link comes up they will get deleted?16:57
*** adiantum has quit IRC16:57
notmynameclayg: I hear that if we just return a 404 really fast, that will be good enough ;-)16:58
*** benny_ has quit IRC16:58
claygnotmyname: it would be *better* because *faster*17:00
*** tongli has quit IRC17:01
*** tongli has joined #openstack-swift17:04
*** lcurtis has joined #openstack-swift17:04
*** aix has quit IRC17:08
*** tongli has quit IRC17:09
*** benny_ has joined #openstack-swift17:10
*** gyee has joined #openstack-swift17:12
*** petertr7_away is now known as petertr717:14
*** tongli has joined #openstack-swift17:17
*** david-lyle has quit IRC17:19
openstackgerritjanonymous proposed openstack/swift: cPickle is deprecated in py3, replacing it from six.moves  https://review.openstack.org/19918817:19
*** tongli has quit IRC17:22
*** proteusguy has joined #openstack-swift17:23
*** petertr7 is now known as petertr7_away17:25
*** annegentle has joined #openstack-swift17:30
*** petertr7_away is now known as petertr717:34
*** annegentle has quit IRC17:35
*** jon1111 has quit IRC17:37
*** bi_fa_fu has quit IRC17:49
openstackgerritBen Martin proposed openstack/swift: Document method to avoid rsync filling root drive  https://review.openstack.org/19920317:52
*** bi_fa_fu has joined #openstack-swift17:53
*** blmartin is now known as blmartin_away17:54
*** blmartin_away has quit IRC17:57
*** blmartin_away has joined #openstack-swift17:58
*** blmartin_away is now known as blmartin17:58
*** mgarza_ has quit IRC18:00
*** marzif_ has joined #openstack-swift18:03
*** tongli has joined #openstack-swift18:04
*** mgarza_ has joined #openstack-swift18:06
*** benny_ has quit IRC18:09
*** tongli has quit IRC18:09
*** tongli has joined #openstack-swift18:11
*** acampbell has joined #openstack-swift18:16
*** acampbel11 has joined #openstack-swift18:20
*** marzif_ has quit IRC18:21
*** acampbell has quit IRC18:24
*** benny_ has joined #openstack-swift18:25
notmynamethis is in py3.5. looks interesting https://www.python.org/dev/peps/pep-0492/18:35
*** CaioBrentano has joined #openstack-swift18:37
*** bi_fa_fu has quit IRC18:37
*** mgarza_ has quit IRC18:38
*** acampbel11 has quit IRC18:39
*** acampbell has joined #openstack-swift18:40
*** mgarza_ has joined #openstack-swift18:45
*** bi_fa_fu has joined #openstack-swift18:50
*** acampbell has quit IRC18:54
*** annegentle has joined #openstack-swift18:56
*** geaaru has quit IRC18:59
*** silor has joined #openstack-swift18:59
*** blmartin has quit IRC19:04
claygnotmyname: meh, polish that turd19:10
*** mgarza_ has quit IRC19:14
portantenotmyname: but coroutines are NOT asynchronous!19:17
portantethis is one of the fundamental problems with cooperative scheduling: folks think they are operating asynchronously but they are not19:17
portanteoh well, beating a dead horse19:18
claygportante: a processor isn't asynchronous either - and yet somehow we managed multi-tasking operating systems even on single core machines?19:18
portanteI am not sure that is the same thing19:19
portantemultiple cores mean async execution19:19
portanteoffloading of I/O is asynchronous, but the non-blocking method is not about performing operations asynchronously, it is about the kernel telling user space that some operation would block, but if it does not say it would block, it is still free to block it19:21
claygI think the cooperative scheduling gives a sufficiently effective "appearence" of asynchronous operation - just like mult-tasking operating systems are effective at making it looks like it's doing multiple things on a single core?19:21
portanteasync I/O is about describing the I/O operation that user spaces wants to happen, queuing to the OS so that it will happen asynchrously to the thread of execution19:21
claygoic - yeah we're talking about different things :)19:21
claygnot surprised - I agree with you :D19:21
portantethis is what I think is the exact problem with this space, the terminology is not used consistently by the core creators, so peps like this continue to keep the waters muddy19:22
claygoh lol - i get it - the module is the worst name ever for something that basically wraps up non-blocking io in an event reactor - yeah too funny19:23
claygyeah rofl19:23
*** mgarza has joined #openstack-swift19:24
portantesometimes I wonder if that web site that generates interesting Ph.D. paper titles is suitable for new documents like this pep19:24
portantehttp://www.besttitlegenerator.com/19:24
*** occupant has joined #openstack-swift19:26
claygtdasilva: you about my friend?19:36
tdasilvaclayg: hello19:39
notmynameclayg: timburke: py3 ideas added to the meeting agenda19:41
notmynamehttps://wiki.openstack.org/wiki/Meetings/Swift19:42
*** bi_fa_fu has quit IRC19:44
claygtorgomatic: OMG - you finally solved overload!?19:45
*** adiantum has joined #openstack-swift19:47
claygtdasilva: i was looking at versioned writes19:49
claygtdasilva: I'm trying to make it through a few reviews today so I didn't dig into the bug - but it seems like it should be fixed - or if it's hard to fix - at least unerstand why we don't care about fixing it19:50
*** david-lyle has joined #openstack-swift19:50
*** bi_fa_fu has joined #openstack-swift19:51
tdasilvaclayg: just saw your comment there, thanks for catching, i will take a look19:52
claygdoes anyone know the syntax in to link to a line number in gerrit?19:53
*** blmartin has joined #openstack-swift19:59
*** wshao has joined #openstack-swift20:00
*** mgarza has quit IRC20:00
*** mgarza_ has joined #openstack-swift20:03
*** chlong has joined #openstack-swift20:05
*** bi_fa_fu has quit IRC20:06
*** marzif_ has joined #openstack-swift20:07
notmynameportante: have you come across SMR drives yet in your perf group?20:15
*** marzif__ has joined #openstack-swift20:16
*** tongli has quit IRC20:18
*** marzif_ has quit IRC20:18
*** petertr7 is now known as petertr7_away20:21
*** petertr7_away is now known as petertr720:21
portanteWe have started talking about them, not sure yet where they fit in the scheme of things20:24
portantethey are kinda wacky20:24
portanteseems like quite a targeted use case20:24
notmynameI'm currently running some tests on them :-)20:25
*** marzif__ has quit IRC20:26
*** hseipp has joined #openstack-swift20:30
*** CaioBrentano has quit IRC20:31
*** jamielennox is now known as jamielennox|away20:41
*** jasondotstar has joined #openstack-swift20:50
*** jamielennox|away is now known as jamielennox20:52
claygmattoliverau: cschwede: what happened with 197395 - I thought we were golden?20:53
*** marzif__ has joined #openstack-swift20:58
*** mutoulbj has quit IRC21:01
*** mutoulbj has joined #openstack-swift21:01
*** mutoulbj has quit IRC21:01
*** wshao has quit IRC21:03
*** silor has quit IRC21:08
*** pgbridge has quit IRC21:10
*** pgbridge has joined #openstack-swift21:11
*** petertr7 is now known as petertr7_away21:11
*** wshao has joined #openstack-swift21:13
*** delatte has quit IRC21:14
*** blmartin has quit IRC21:14
*** jasondotstar has quit IRC21:16
portantenotmyname: and ... ?21:17
notmynameportante: :-(21:17
portanteoh21:17
*** annegentle has quit IRC21:18
notmynameportante: 4k object PUTs with a concurrency of 16: SMR@37/s; PMR@122/s21:18
portanteso little less than 1/3 of the performance21:20
portantebut is there another measure?  Like PUTs per square inch of rack space or something?21:21
portante;)21:21
hrouAh, very interesting, planning on trying with larger objects ?21:21
notmynamehrou: yes. but next I'm going to do reads. I expect them to be worse21:22
hrouYea : )21:22
*** marzif_ has joined #openstack-swift21:26
*** marzif__ has quit IRC21:26
*** CaioBrentano has joined #openstack-swift21:27
*** hseipp has quit IRC21:30
*** marzif_ has quit IRC21:32
notmynameconcurrency of 32, same object sizes: PMR@178/sec; SMR@41/sec21:42
notmynameportante: hrou: ^21:42
hrouJust about the same as your right tests perf wise21:44
hrou*write tests - interesting!21:44
notmynameno, that was writes21:44
notmynamehaven't done reads yet21:44
hrouAh !  Sorry I see you just x2 the concurrency.21:45
notmynameright21:46
notmynamemeanwhile the PMR drives had gone on to also run the 64 and 128 concurrency tests.21:46
notmynameanyway, off to write the read test code21:47
hrouIf those  workloads end up on the same object server -> disk, that's when I'd expect to see the worst performance.  I believe the SMR drives handle the sequential read requirements by maintaining a mapping in memory, but once that's full, it needs to defragment so to speak.21:47
hrouThat's awesome !  Thanks, you're likely the first (or just about) first person to be doing swift -> smr tests, Feel special ? : )21:48
notmynamenot really ;-)21:48
hrouhehe!21:48
hroumaybe not analogous to being the first person on the moon, but lets call it the second best thing ; )21:50
*** wshao has quit IRC21:51
*** tsubic has joined #openstack-swift21:51
*** jkugel1 has quit IRC21:52
*** benny_ has quit IRC21:53
*** mgarza_ has quit IRC21:57
*** annegentle has joined #openstack-swift21:57
*** mgarza has joined #openstack-swift21:59
*** asettle-afk is now known as asettle22:00
*** wshao has joined #openstack-swift22:05
*** benny_ has joined #openstack-swift22:06
mattoliverauMorning22:13
openstackgerritClay Gerrard proposed openstack/swift: Add save command to ring-builder-analyzer  https://review.openstack.org/19932522:14
*** NM has quit IRC22:17
*** wshao has quit IRC22:20
openstackgerritMerged openstack/python-swiftclient: SwiftClient object upload beginning with / or "./"  https://review.openstack.org/18526922:21
openstackgerritMerged openstack/swift: EC Reconstructor: Do not reconstruct existing fragments.  https://review.openstack.org/19327922:24
*** annegentle has quit IRC22:29
openstackgerritTim Burke proposed openstack/python-swiftclient: Properly test raw writes in Python 3  https://review.openstack.org/18414622:40
openstackgerritTim Burke proposed openstack/python-swiftclient: Always decode command-line arguments as UTF-8  https://review.openstack.org/17694522:40
*** lcurtis has quit IRC22:41
*** hrou has quit IRC22:44
*** chlong has quit IRC22:46
*** classicsnail has quit IRC22:48
*** benny_ has quit IRC22:49
*** jkugel has joined #openstack-swift22:53
*** janonymous has quit IRC22:58
*** km has joined #openstack-swift23:02
*** minwoob has quit IRC23:03
*** zaitcev has quit IRC23:05
*** sanchitmalhotra1 has joined #openstack-swift23:08
*** jkugel has left #openstack-swift23:08
*** sanchitmalhotra has quit IRC23:10
*** annegentle has joined #openstack-swift23:12
*** classicsnail has joined #openstack-swift23:16
notmynamelooks like I'm called for jury duty tomorrow23:21
*** zhill has quit IRC23:21
*** kei_yama has joined #openstack-swift23:22
klrmnnotmyname: whee!23:22
notmynamewheee!23:22
mattoliveraunotmyname: does that mean no meeting?23:22
notmynamemattoliverau: unless you're volunteering to run it ;-)23:23
mattoliveraunotmyname: lol, and we'll solve all the things23:23
notmynameI'll probably end up volunteered to do it all ;-)23:24
mattoliveraunotmyname: delgate all the tough things to our amazing PTL23:24
mattoliveraulol ^^ yup ;)23:24
notmyname"who isn't at the meeting"23:24
mattoliverauexactly :P23:24
notmynameok, I'm updating the meeting page now23:24
mattoliverauCool, jury duty tho, that's kinda exciting (saying that cause I've never had to do it).23:25
notmynameit's an interesting experience23:27
notmynameI've done it once before23:27
notmynameon the one hand, it's (at least symbolically) a huge part of our national duty, I think. like voting23:28
notmynameon the other, it's tedious, boring, inconvenient, and most people try really hard to get out of it23:28
*** hrou has joined #openstack-swift23:32
*** annegentle has quit IRC23:34
mattoliverauyeah, it probably is mainly boring, I'm probably just seeing the glamous side TV/movie-ified side which doesn't really exist :)23:35
notmynamewell, you have to take off work, go somewhere else, sit around and wait, listen to a case, make a decision (those last 2 can take days), and get paid about $10 per day for it23:37
notmynamefor people who don't have good employers or work hourly, it can be a really big hardship23:38
claygnotmyname: I think you're getting a bit hit-n-miss on the priority review dashboard?23:38
notmynameclayg: ah? perhaps so :-)23:38
notmynameI'll admit I haven't kept on top of it23:38
notmynameI did spend all day yesterday in LP though ;-)23:38
*** classicsnail has quit IRC23:38
mattoliveraunotmyname: yeah, actually that could really suck for people, $10 a day, and what if its a large trial.23:39
claygnotmyname: We only expect perfection from you at all times23:39
notmynameclayg: the dashboard with the starred patches or the priority reviews wiki page?23:39
claygmattoliverau: well I think there's also some laws that say cooprerations can't doc your salary - so the judge may excuse you if it's going to be a long trial.23:40
claygnotmyname: starred patches > wiki page23:40
notmynameon the dashboard, that huge list of stuff that needs one more +2 to land is a good place to start :-)23:40
claygacctually with the dashboard - I don't think we need to have a *seperate list* of patches?23:40
klrmnone would think jury duty, while being a responsibility, would at least pay minimum wage23:40
claygnotmyname: ok thanks!23:40
*** classicsnail has joined #openstack-swift23:40
notmynameclayg: yeah, 1 list >> 2 lists23:41
claygklrmn: maybe if the defendent had to pay for it?23:41
notmynameclayg: yikes. state should pay. I don't want to be sued and forced to pay for the jury that may or may not find me guilty. also, can I pay more for my jury to get a better outcome? ;-)23:42
mattoliveraulol23:42
klrmnseems to me, if you want to cut down on frivolty suits, the prosecution should pay23:42
claygklrmn: notmyname: looser pays!23:43
klrmnin criminal cases, that'd be the gov't anyway23:43
notmynameklrmn: and then only the rich can seek court approved justice.23:43
klrmnpoint23:43
hrouMost large employers  will pay for up to 3 weeks, but after generally not;  Yea I've had this debate with a lot of folks, one would think there should be a law that forces employeers to pay no longer how long the trial is23:43
claygyou two *sure* you want to go to court?  I hear jurors rates are going up?23:44
notmynameyou can already sue your accuser for court costs in many cases if you are found not guilty. the state should pay at least min wage for juries23:44
notmynamehrou: especially since you go to jail if you don't serve! so the choice can be jail or lose your job23:44
notmynameclayg: I'll update the dashboard ASAP23:45
hrouRight though that's where the art of getting out of it comes into play, that is the various excuses you can use, really depends on the judge, I know someone who simply said he was starting a new job and thought it would reflect badly on him, and apparently that was sufficient enough and he was let go : )23:46
notmynamemaybe tonight. maybe something to do while waiting for whatever happens tomorrow ;-)23:46
notmynamehrou: FWIW, reading the 4KiB objects is pretty much the same for SMR and PMR drives. that surprised me23:47
notmynameI'll be testing with bigger objects next23:47
*** jasondotstar has joined #openstack-swift23:51
*** gyee has quit IRC23:51
*** benny_ has joined #openstack-swift23:58

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