Thursday, 2015-08-20

*** sanchitmalhotra1 has quit IRC00:00
*** sanchitmalhotra1 has joined #openstack-swift00:06
*** barker has quit IRC00:06
*** sanchitmalhotra has quit IRC00:08
*** nakagawamsa has joined #openstack-swift00:09
*** chsc has quit IRC00:13
openstackgerritVictor Stinner proposed openstack/swift: pep8: Don't override '_' symbol  https://review.openstack.org/21349500:13
openstackgerritVictor Stinner proposed openstack/swift: pep8: Fix usage of the l10n _('...') function  https://review.openstack.org/21349400:13
*** gyee has quit IRC00:33
*** sanchitmalhotra has joined #openstack-swift00:36
*** lyrrad has quit IRC00:36
*** sanchitmalhotra1 has quit IRC00:38
*** sanchitmalhotra1 has joined #openstack-swift00:40
*** sanchitmalhotra has quit IRC00:42
openstackgerritAkihito Takai proposed openstack/swift: Fix the comment in [1].  https://review.openstack.org/21491500:46
*** nakagawamsa has quit IRC00:49
*** nakagawamsa has joined #openstack-swift01:01
*** sanchitmalhotra1 has quit IRC01:03
*** minwoob has quit IRC01:09
*** sanchitmalhotra has joined #openstack-swift01:09
claygpeluse: bah I ment to -1 the ec buckets patch so you'd see it - but it's fine - i'm still working on it anyway01:11
*** haigang has joined #openstack-swift01:16
*** barker has joined #openstack-swift01:22
*** bill_az_ has quit IRC01:25
*** sanchitmalhotra1 has joined #openstack-swift01:25
notmynametorgomatic: I think this is what you were looking for https://wiki.openstack.org/wiki/Swift/version_map ;-)01:25
notmynametorgomatic: and that's what I'll merge into the changelog (why I put a -2 on it)01:25
notmynamehmm...looks like I need to update that01:26
*** sanchitmalhotra has quit IRC01:27
*** pksingh has quit IRC01:27
*** bill_az has joined #openstack-swift01:34
*** bill_az has quit IRC01:35
*** ccavanna has joined #openstack-swift01:40
*** barker has quit IRC01:42
openstackgerritTim Burke proposed openstack/swift: Add delete markers to versioned_writes middleware  https://review.openstack.org/21492201:49
*** haomaiwang has joined #openstack-swift01:53
notmynametimburke: wait. from that patch subject line, it implies you've added "real" versioning ^01:53
notmyname(that's really cool!)01:53
*** barker has joined #openstack-swift02:00
*** barker has quit IRC02:08
*** haomaiwang has quit IRC02:09
*** jrichli has joined #openstack-swift02:09
*** haomaiwang has joined #openstack-swift02:10
*** baojg has joined #openstack-swift02:16
*** sanchitmalhotra has joined #openstack-swift02:26
*** sanchitmalhotra1 has quit IRC02:28
openstackgerritJanie Richling proposed openstack/swift: Encrypting/Decrypting object metadata  https://review.openstack.org/20331002:31
*** dbarman has quit IRC02:31
openstackgerritJohn Dickinson proposed openstack/swift: Add OpenStack release names to changelog  https://review.openstack.org/21472802:34
*** vinsh has quit IRC02:42
*** vinsh has joined #openstack-swift02:42
*** bkopilov has quit IRC02:43
openstackgerritHisashi Osanai proposed openstack/swift: test/(functional/probe):Replace python print operator with print function (pep H233, py33)  https://review.openstack.org/20660302:45
openstackgerritJanie Richling proposed openstack/swift: Decrypting Container Listing  https://review.openstack.org/21443802:45
*** vinsh has quit IRC02:46
*** sanchitmalhotra1 has joined #openstack-swift02:58
*** vinsh has joined #openstack-swift03:00
*** sanchitmalhotra has quit IRC03:01
*** vinsh has quit IRC03:05
*** haomaiwang has quit IRC03:09
*** haomaiwang has joined #openstack-swift03:10
openstackgerritHisashi Osanai proposed openstack/swift: Python3: do not use im_self/im_func/func_closure  https://review.openstack.org/20306603:12
*** hezhiqiang has joined #openstack-swift03:13
*** NM has joined #openstack-swift03:33
*** Kennan2 is now known as Kennan03:36
*** baojg has quit IRC03:47
*** NM has quit IRC03:50
*** baojg has joined #openstack-swift03:54
*** SkyRocknRoll has joined #openstack-swift03:58
*** sanchitmalhotra has joined #openstack-swift03:58
*** sanchitmalhotra1 has quit IRC04:01
*** chenhuayi has quit IRC04:06
*** sanchitmalhotra1 has joined #openstack-swift04:09
*** haomaiwang has quit IRC04:09
*** haomaiwang has joined #openstack-swift04:10
*** sanchitmalhotra has quit IRC04:11
*** hrou has quit IRC04:18
*** jrichli has quit IRC04:21
*** bkopilov has joined #openstack-swift04:30
*** baojg has quit IRC04:37
*** baojg has joined #openstack-swift04:38
*** ppai has joined #openstack-swift04:42
*** darrenc is now known as darrenc_afk04:46
*** mjfork has quit IRC04:56
*** zul has quit IRC05:04
*** darrenc_afk is now known as darrenc05:06
*** haomaiwang has quit IRC05:09
*** haomaiwang has joined #openstack-swift05:10
*** SkyRocknRoll has quit IRC05:12
*** asettle has joined #openstack-swift05:14
*** haigang has quit IRC05:14
*** zaitcev has quit IRC05:16
*** haomaiwang has quit IRC05:19
openstackgerritCharles Hsu proposed openstack/python-swiftclient: Increase httplib._MAXHEADERS to 256.  https://review.openstack.org/21414405:24
*** silor has joined #openstack-swift05:27
*** barker has joined #openstack-swift05:27
*** silor1 has joined #openstack-swift05:28
*** silor has quit IRC05:31
*** silor1 is now known as silor05:31
*** hezhiqiang has quit IRC05:32
*** hezhiqiang has joined #openstack-swift05:32
*** SkyRocknRoll has joined #openstack-swift05:42
*** barker has quit IRC05:51
*** haomaiwa_ has joined #openstack-swift06:00
openstackgerritMerged openstack/swift: pep8: Fix hacking H232 warnings (octal)  https://review.openstack.org/20614906:00
*** akle|2 has quit IRC06:03
*** akle|2 has joined #openstack-swift06:03
*** asettle has quit IRC06:09
*** haomaiwa_ has quit IRC06:09
*** sanchitmalhotra has joined #openstack-swift06:12
*** haomaiwang has joined #openstack-swift06:13
*** sanchitmalhotra1 has quit IRC06:14
*** haomaiwang has quit IRC06:17
*** akle|2 has quit IRC06:21
*** baojg has quit IRC06:22
*** akle|3 has joined #openstack-swift06:23
*** baojg has joined #openstack-swift06:24
*** haomaiwang has joined #openstack-swift06:29
*** hezhiqiang has quit IRC06:35
*** haigang has joined #openstack-swift06:37
*** hezhiqiang has joined #openstack-swift06:37
*** mahatic has joined #openstack-swift06:38
*** haomaiwang has quit IRC06:47
*** sanchitmalhotra1 has joined #openstack-swift06:52
openstackgerritMerged openstack/swift: pep8: replace deprecated calls to assert_()  https://review.openstack.org/21349106:52
*** haomaiwang has joined #openstack-swift06:52
*** sanchitmalhotra has quit IRC06:54
*** rledisez has joined #openstack-swift07:04
*** peluse_ has joined #openstack-swift07:05
*** peluse has quit IRC07:08
*** haomaiwang has quit IRC07:09
*** haomaiwang has joined #openstack-swift07:10
*** mahatic has quit IRC07:17
openstackgerritOpenStack Proposal Bot proposed openstack/swift: Updated from global requirements  https://review.openstack.org/8873607:21
*** SkyRocknRoll has quit IRC07:24
*** hezhiqia_ has joined #openstack-swift07:25
*** hezhiqiang has quit IRC07:26
openstackgerritOpenStack Proposal Bot proposed openstack/swift: Updated from global requirements  https://review.openstack.org/17434307:27
*** mahatic has joined #openstack-swift07:27
*** hezhiqia_ has quit IRC07:31
*** hezhiqiang has joined #openstack-swift07:31
*** mahatic has quit IRC07:35
*** mahatic has joined #openstack-swift07:37
*** mahatic has quit IRC07:38
*** geaaru has joined #openstack-swift07:42
*** haomaiwang has quit IRC08:09
*** SkyRocknRoll has joined #openstack-swift08:10
*** haomaiwa_ has joined #openstack-swift08:10
*** wbhuber has quit IRC08:16
*** asettle has joined #openstack-swift08:19
*** asettle has quit IRC08:24
*** jistr has joined #openstack-swift08:28
*** janonymous has joined #openstack-swift08:28
*** jordanP has joined #openstack-swift08:28
*** hezhiqiang has quit IRC08:30
*** hezhiqiang has joined #openstack-swift08:34
*** silor1 has joined #openstack-swift08:38
*** silor has quit IRC08:39
*** silor1 is now known as silor08:39
*** aix has joined #openstack-swift08:41
*** ho has quit IRC08:57
*** acoles_ is now known as acoles09:05
*** haomaiwa_ has quit IRC09:09
*** mahatic has joined #openstack-swift09:09
*** haomaiwang has joined #openstack-swift09:10
*** NM has joined #openstack-swift09:11
*** tsubic has quit IRC09:13
*** jistr has quit IRC09:19
*** jistr has joined #openstack-swift09:20
openstackgerritKazuhiro MIYAHARA proposed openstack/swift: Fix dispersion-reports error message  https://review.openstack.org/21369709:27
mahaticacoles: Morning! This might be silly, I need to take jrichli's changes from patch 214438 - without taking a checkout of that change, what happens if I rebase feature/crypto onto my local branch (where I'm making my changes)09:29
patchbotmahatic: https://review.openstack.org/#/c/214438/09:29
openstackgerritKazuhiro MIYAHARA proposed openstack/swift: Fix dispersion-reports error message  https://review.openstack.org/21369709:30
*** tsubic has joined #openstack-swift09:45
acolesmahatic: hi! i need to better understand what you want to do :) do you want to apply change from 214438 to your working branch?09:46
mahaticacoles: :) yes, (I'm hung on rebasing!) Last time I submitted it, I got "Merge Conflict" in gerrit.09:47
acolesmahatic: and what is your working branch based on? did you start from feature/ec?09:49
acolesfeature/crypto i mean :)09:49
mahatic:D I was confused for a sec. Yes, I took the tail branch (git checkout --- and get checkout FETcH_HEAD thing)09:50
*** aix has quit IRC09:51
*** hezhiqiang has quit IRC09:51
mahaticacoles: it is based off patch 20345409:52
patchbotmahatic: https://review.openstack.org/#/c/203454/09:52
acolesmahatic: ok, so you started from 203454 and now you want to include the changes from 214438?09:53
mahaticacoles: yeah09:54
acolesmahatic: i am just looking at the patch chains09:57
mahaticsure, thanks09:58
*** mahatic_ has joined #openstack-swift10:00
*** mahatic has quit IRC10:00
acolesso, 203454 is based on 203310 and 214438 is also based on 203310. but 203454 is abandoned, so do you really want to be based on 203454?10:02
*** Kennan2 has joined #openstack-swift10:03
acolesis your review that we are discussing patch 211451?10:03
patchbotacoles: https://review.openstack.org/#/c/211451/10:03
mahatic_yeah, that's the one10:04
*** Kennan has quit IRC10:04
mahatic_I'm looking at those patches you mentioned, just a moment please10:04
acolesso do you want the changes from 203454 (abandoned)?10:05
acolesk, i'm just going to grab coffee, bbiab10:05
mahatic_acoles: nope, I just want changes from 21443810:09
*** haomaiwang has quit IRC10:09
mahatic_I took changes from 203454 when it was not abandones10:09
*** haomaiwang has joined #openstack-swift10:10
acolesmahatic_: ok let me figure out the steps to take - seems like we need to get your changes based onto 21443810:10
*** dbarman has joined #openstack-swift10:11
mahatic_acoles: yeah, thanks. Or I could make my changes (and redo some existing) once I have 214438 on current branch10:12
acolesmahatic_: try this http://paste.openstack.org/show/422617/10:21
acolesthere are some merge conflicts you will have to resolve10:22
mahatic_sure, will try that10:22
*** baojg has quit IRC10:33
*** baojg has joined #openstack-swift10:34
*** mahatic has joined #openstack-swift10:35
*** mahatic_ has quit IRC10:35
*** mahatic has quit IRC10:36
*** mahatic has joined #openstack-swift10:36
*** baojg has quit IRC10:38
*** david__ has joined #openstack-swift10:45
*** haomaiwang has quit IRC10:48
mahaticacoles: so it worked :) Moving forward, how do I pick changes - I upload my changes and then jrichli does her stuff and uploads, then how do i pick those again?10:48
*** dbarman has quit IRC10:48
mahaticI mean how do I get jrchili's changes into my local?10:50
acolesmahatic: great :)10:54
acolesmahatic: each time the review you are based on changes (in this case 214438 now) then you need to rebase your patch10:55
acolesmahatic: so you would do 'git review -d 214438' to get latest version of that10:55
mahaticacoles: okay. "git review -d 214438" and then "git checkout my_branch" ?10:56
acolesthen checkout your branch and 'git rebase -i review/janie_richling/jlr_clisting'10:56
acolesyup10:56
acolesthen fix any merge conflicts, git add, git rebase --continue just like you have just done10:57
acolesmahatic: you're probably going to have to do this quite often so it will become familiar :)10:58
mahaticacoles: okay :) thanks so much! you saved my day! Glad to have you my timezone :D10:58
mahaticacoles: yeah :)10:58
acolesmahatic: no probs10:59
*** SkyRocknRoll has quit IRC11:05
*** kei_yama has quit IRC11:09
*** kairo_ has joined #openstack-swift11:11
*** kairo has quit IRC11:14
*** aix has joined #openstack-swift11:16
*** SkyRocknRoll has joined #openstack-swift11:21
*** kairo_ has quit IRC11:22
mahaticacoles: since it created a new branch "review/mahati/crypto_mahati", does it hurt to rename it to something smaller?11:24
*** NM has quit IRC11:25
acolesmahatic: no you can rename it to whatever you want. gerrit tracks patches based on the change id in the commit message. your branch name may appear as the topic in gerrit but i'm not aware that is has any significance.11:26
mahaticacoles: alright. thanks!11:27
*** sanchitmalhotra1 has quit IRC11:35
*** kairo has joined #openstack-swift11:36
*** bkopilov has quit IRC11:42
*** david__ has quit IRC11:55
*** robefran has quit IRC12:01
*** hezhiqiang has joined #openstack-swift12:04
*** haigang has quit IRC12:18
*** rledisez has quit IRC12:21
*** rledisez has joined #openstack-swift12:21
*** asettle has joined #openstack-swift12:22
*** asettle has quit IRC12:23
*** km has quit IRC12:32
*** vinsh has joined #openstack-swift12:35
*** barker has joined #openstack-swift12:36
*** ccavanna has quit IRC12:36
*** robefran has joined #openstack-swift12:38
*** NM has joined #openstack-swift12:45
*** chlong has joined #openstack-swift12:45
*** SkyRocknRoll has quit IRC12:45
*** annegentle has joined #openstack-swift12:50
*** jistr is now known as jistr|mtg12:53
*** ppai has quit IRC12:54
*** aix has quit IRC12:54
*** aix has joined #openstack-swift13:00
*** chlong has quit IRC13:00
*** bill_az has joined #openstack-swift13:00
*** barker has quit IRC13:03
*** chlong has joined #openstack-swift13:03
*** dustins has joined #openstack-swift13:07
*** hrou has joined #openstack-swift13:14
*** bkopilov has joined #openstack-swift13:15
*** sasik has joined #openstack-swift13:15
*** sasik has left #openstack-swift13:16
*** janonymous_ has joined #openstack-swift13:25
*** dustins has quit IRC13:26
janonymous_Hi..13:26
*** dustins has joined #openstack-swift13:26
janonymous_Could someone suggest, is there need to handle this bug : https://bugs.launchpad.net/swift/+bug/148399713:27
openstackLaunchpad bug 1483997 in OpenStack Object Storage (swift) "Request using Chunked encoding as a header in object GET raises httlib Error" [Undecided,New] - Assigned to janonymous (janonymous)13:27
janonymous_This bug raises an error visible in logs when a http request of GET is made using header Transfer-encoding chunked13:28
*** jamielennox is now known as jamielennox|away13:30
*** zul has joined #openstack-swift13:30
*** ccavanna has joined #openstack-swift13:39
*** SkyRocknRoll has joined #openstack-swift13:41
*** janonymous_ has quit IRC13:42
mattoliverauMorning13:48
*** haomaiwang has joined #openstack-swift13:53
*** lcurtis has joined #openstack-swift13:58
*** blmartin has joined #openstack-swift13:58
tdasilvamattoliverau: good morning! how's going in TX??14:00
hrouMorning mattoliverau / tdasilva !14:00
tdasilvahrou: hi! good morning14:00
mattoliverauhrou, tdasilva: o/14:01
mattoliverauTX is good.. Last day tho. I fly home this evening14:01
hroumattoliverau, seems like it cooled down a bit 90 degrees : )14:02
tdasilvamattoliverau: nice, have a good trip home!14:03
mattoliverauThanks, going to the rackspace office for the day first tho :)14:04
*** aix has quit IRC14:04
*** jistr|mtg is now known as jistr14:08
*** jrichli has joined #openstack-swift14:08
*** aix has joined #openstack-swift14:09
*** haomaiwang has quit IRC14:09
*** bill_az has quit IRC14:10
*** haomaiwang has joined #openstack-swift14:10
*** jlhinson has joined #openstack-swift14:11
kairomorning guys.14:13
*** wbhuber has joined #openstack-swift14:14
*** barker has joined #openstack-swift14:14
*** zul has quit IRC14:17
*** petertr7_away is now known as petertr714:19
*** SkyRocknRoll has quit IRC14:31
openstackgerritAlistair Coles proposed openstack/swift: Make pipeline_property less brittle  https://review.openstack.org/10492614:31
pgbridgegood morning14:41
*** blmartin has quit IRC14:45
*** dustins has quit IRC14:45
*** dustins has joined #openstack-swift14:46
*** blmartin has joined #openstack-swift14:46
*** annegentle has quit IRC14:58
*** annegentle has joined #openstack-swift14:58
*** lpabon has joined #openstack-swift14:59
*** minwoob has joined #openstack-swift14:59
*** zul has joined #openstack-swift15:02
*** barker has quit IRC15:04
*** alejandrito has joined #openstack-swift15:08
*** haomaiwang has quit IRC15:09
*** haomaiwang has joined #openstack-swift15:10
*** aerwin has joined #openstack-swift15:13
*** nadeem has joined #openstack-swift15:19
*** barker has joined #openstack-swift15:20
*** nadeem has quit IRC15:20
*** nadeem has joined #openstack-swift15:21
*** jistr is now known as jistr|mtg15:22
*** mahatic has quit IRC15:24
*** kevinc_ has joined #openstack-swift15:29
*** zul has quit IRC15:33
*** haomaiwang has quit IRC15:35
timburkenotmyname: well, it's enough that i think we can get rid of https://github.com/swiftstack/swift_undelete -- i'm still not happy about the extra data movement, though, so there's still a case for using symlinks15:40
*** petertr7 is now known as petertr7_away15:42
*** petertr7_away is now known as petertr715:48
acolespeluse_: ping15:51
*** hogepodge has quit IRC15:53
*** akle|3 has quit IRC16:02
*** silor has quit IRC16:06
*** esker has quit IRC16:07
*** jkugel has joined #openstack-swift16:07
*** jistr|mtg is now known as jistr16:10
*** david__ has joined #openstack-swift16:14
*** hogepodge has joined #openstack-swift16:23
*** zul has joined #openstack-swift16:24
*** petertr7 is now known as petertr7_away16:25
*** chsc has joined #openstack-swift16:36
*** nadeem has quit IRC16:40
*** bill_az has joined #openstack-swift16:42
*** lyrrad has joined #openstack-swift16:42
*** rledisez has quit IRC16:42
*** jistr has quit IRC16:46
*** barker has quit IRC16:47
openstackgerritMerged openstack/swift: Add OpenStack release names to changelog  https://review.openstack.org/21472816:48
*** barker has joined #openstack-swift16:49
*** Kennan2 has quit IRC16:50
*** bill_az has quit IRC17:02
*** Kennan has joined #openstack-swift17:05
*** jordanP has quit IRC17:06
*** Kennan has quit IRC17:14
*** zhill has joined #openstack-swift17:16
*** kevinc_ has quit IRC17:16
*** Kennan has joined #openstack-swift17:20
*** aix has quit IRC17:22
*** Kennan has quit IRC17:24
*** kevinc_ has joined #openstack-swift17:25
*** Kennan has joined #openstack-swift17:40
*** barker has quit IRC17:41
*** kevinc___ has joined #openstack-swift17:43
*** barker has joined #openstack-swift17:46
*** kevinc_ has quit IRC17:46
*** barker has quit IRC17:51
*** breitz has quit IRC17:52
*** breitz has joined #openstack-swift17:52
*** petertr7_away is now known as petertr717:54
*** barker has joined #openstack-swift17:54
*** dustins has quit IRC17:56
*** chsc has quit IRC17:56
claygmorning17:59
*** gyee has joined #openstack-swift18:00
*** barker has quit IRC18:00
*** gyee has quit IRC18:00
*** gyee has joined #openstack-swift18:03
acolesmorning clayg18:03
*** barker has joined #openstack-swift18:04
kairomorning(3pm) :)18:04
*** barker has quit IRC18:06
acoleskairo: where is it 3pm right now?18:08
*** bill_az has joined #openstack-swift18:09
*** bapalm has quit IRC18:11
*** geaaru has quit IRC18:13
*** bapalm has joined #openstack-swift18:14
acoleshey clayg thanks for taking a look over patch 104926, i just fixed the merge conflict before mattoliverau wielded his abandonment axe over it, has been ages since i looked at it but i will try to follow up your comments18:18
patchbotacoles: https://review.openstack.org/#/c/104926/18:18
kairoacoles: Brazil :)18:18
mattoliverauI like the idea of an abandonment axe, I should get one18:18
acolesclayg: right now i'm hacking diskfile to allow opening of non-durable data files18:19
acolesmattoliverau: i don't think they will let you take it on the plane :)18:19
acoleskairo: oh right18:19
acolesmattoliverau: safe travelling btw18:19
*** zhill has quit IRC18:21
mattoliverauacoles: thanks18:22
mattoliverauProbably not, maybe I need a plastic one18:22
*** zhill has joined #openstack-swift18:22
*** zhill has quit IRC18:29
claygacoles: yeah that's right!  eat it diskfile!18:30
openstackgerritAlistair Coles proposed openstack/swift: Enable object server to return non-durable data  https://review.openstack.org/21527618:30
acolesvoila!18:30
acoles(still WIP)18:30
claygacoles: oh no sorry - that pipleline property thing is totally my fault :'(18:30
claygacoles: I'd love to clean it up for you - but... man - *so* much other stuff I really really want to work on18:31
acolesclayg: +1 "bigger fish" :)18:32
acolesclayg: i decided to make a post it note backlog for myself today but ran out of post it notes :/18:33
claygacoles: oh goodness :'(18:33
claygi love TODO "this was going great until it fell apart and now wtf"18:33
*** nadeem has joined #openstack-swift18:34
claygacoles: so you're thinking that a datafile w/o durable can't clear an older .meta file (I think that's right, and that's how I'm reading the test)18:35
acolesclayg: a non-durable data file at t2 should not cause a meta file at t1 to be deleted. but when we do a non-durable response we should not return the meta file at t1 with the non-durable at t2.18:37
*** dustins has joined #openstack-swift18:37
claygacoles: agreed18:37
claygacoles: nice work18:37
*** zhill has joined #openstack-swift18:37
acolesclayg: i want to sleep on the implementation but thats how i think it should behave. glad we agree :)18:38
wbhuberacoles: isn't this patch related to the "reckless GETs" discussion we had at hackathon?18:39
acolesclayg btw i based this on peluse patch 213147 cos i love how he made get_ondisk_files return a dict.18:39
patchbotacoles: https://review.openstack.org/#/c/213147/18:39
acoleswbhuber: exactly18:39
acoleswbhuber: "optimistic GETs" ;)18:40
wbhuberacoles: politically correct. :)18:40
wbhuberacoles: cannot we merge get_ondisk_files with gather_on_disk_files? both sound redundant18:42
wbhuberbut i understand the intention behind each one18:42
acoleswbhuber: its getting that way, so far i kept them both in case someone else using the diskfile API is calling get_ondisk_files, but actually peluse has changed the return type so maybe we can just combine them. i'll look at it some more.18:45
*** NM has quit IRC18:46
*** NM has joined #openstack-swift18:51
claygacoles: yeah peluse_ is awesome18:51
openstackgerritAlistair Coles proposed openstack/swift: Enable object server to return non-durable data  https://review.openstack.org/21527618:53
acolesclayg: ^^ you made me think how the test could be better. thanks.18:54
claygacoles: now you're just throwing credit around to make other people more interested in your work (that's a *great* strategy btw)18:54
acoleslol18:54
acolesalways spread the blame18:55
claygright!18:55
acolesgoodnight/good evening/enjoy lunch.18:55
*** acoles is now known as acoles_18:57
*** bill_az has quit IRC19:13
*** zul has quit IRC19:23
*** Protux has quit IRC19:27
*** jamielennox|away has quit IRC19:27
*** serverascode has quit IRC19:28
*** nottrobin has quit IRC19:28
*** bill_az has joined #openstack-swift19:29
*** nottrobin has joined #openstack-swift19:32
*** serverascode has joined #openstack-swift19:34
*** Protux has joined #openstack-swift19:35
*** chsc has joined #openstack-swift19:37
*** darrenc has quit IRC19:39
*** lpabon has quit IRC19:40
*** darrenc has joined #openstack-swift19:40
*** breitz has quit IRC19:43
*** breitz has joined #openstack-swift19:43
*** esker has joined #openstack-swift19:46
*** esker has quit IRC19:47
*** alejandrito has quit IRC19:47
*** jamielennox|away has joined #openstack-swift19:47
*** esker has joined #openstack-swift19:47
*** jamielennox|away is now known as jamielennox19:47
*** cppforlife_ has quit IRC19:49
*** alejandrito has joined #openstack-swift19:49
claygtorgomatic: what are the changes we don't have any tests that validate an invalid Range header results in a 400 instead of say a 500 returned when catch_errors gobbles up a ValueError that gets thrown in proxy.controllers.base.GetOrHeadHandler._get_response_parts_iter?19:50
clayg*chances19:50
*** cppforlife_ has joined #openstack-swift19:50
torgomaticclayg: a minute ago I would have said small, but given that you're asking the question, I'm going to go with "nearly 100%"19:50
claygtorgomatic (or anyone): if you happen to have a saio up can you curl with a -H 'Range: fuck' real quick and tell me i'm crazy?19:51
torgomaticone sec19:52
*** resker has joined #openstack-swift19:53
torgomaticclayg: when I do that the Range header seems like it's completely ignored19:53
claygunittests have tacos=3-519:53
claygsweet!  I just have something broken - thank good ness19:53
torgomaticbytes=101-tacos also results in ignored header19:54
claygtorgomatic: well that sounds like the behavior we want!19:54
torgomatichooray!19:54
claygtorgomatic: doh!  it was because my test object is zero-byte19:55
claygso... what the chances ... invalid Range ... on zero byte object ..19:56
*** resker has quit IRC19:56
claygtorgomatic: could still be my broken branch i 'spose19:56
torgomaticclayg: nope, happens over here too19:56
*** resker has joined #openstack-swift19:57
torgomaticRange: bytes=burr-itos on a zero-byte object 500s19:57
torgomaticboo19:57
*** esker has quit IRC19:57
claygtorgomatic: k, thanks - i'm working on the unit test19:57
clayg<hugs>19:57
*** lpabon has joined #openstack-swift19:58
*** zhill has quit IRC19:58
claygtorgomatic: yeah even better - thanks!  I probalby wouldn't have only gotten as far as a failing unitest with a bug report anyway!19:58
*** zhill has joined #openstack-swift20:01
*** ccavanna has quit IRC20:02
*** annegentle has quit IRC20:04
*** zhill has quit IRC20:04
*** resker has quit IRC20:04
*** zhill has joined #openstack-swift20:04
*** annegentle has joined #openstack-swift20:04
*** zhill has quit IRC20:07
*** zhill has joined #openstack-swift20:08
*** alejandrito has quit IRC20:09
*** joeljwright has joined #openstack-swift20:13
*** ChanServ sets mode: +v joeljwright20:13
*** alejandrito has joined #openstack-swift20:13
*** zaitcev has joined #openstack-swift20:17
*** ChanServ sets mode: +v zaitcev20:17
*** robefran has quit IRC20:19
openstackgerritSamuel Merritt proposed openstack/swift: Fix 500 for bogus Range request to 0-byte object.  https://review.openstack.org/21531720:27
torgomaticclayg: ^^20:27
claygtorgomatic: can you come drop some ec range knowledge on me?20:29
*** zul has joined #openstack-swift20:39
claygtorgomatic: I hate those live wire tests tho - they're soooooo flipping slow!20:40
*** albertom has joined #openstack-swift20:43
albertomsomeone using erasure codes?20:43
albertomswift-ring-builder account.builder create 10 3 1 <- which value should I use? since 3 is for 3 replicas20:44
wbhuberalbertom: yes, there are a bunch of us using Swift Erasure Coding.20:44
wbhuberalbertom: Depends on what your EC scheme is20:46
albertomok, I wanto to have 14 storage nodes20:46
wbhuberalbertom: if you have a 3+2 scheme, you can use this: swift-ring-builder ~/builderfile create 18 5 120:46
albertomso for the defaults 10 + 4 would be create 11 14 1 ?20:48
clayghttp://docs.openstack.org/developer/swift/overview_erasure_code.html#using-an-erasure-code-policy would probably be the place to indicate that your rings primary count should be data + parity20:48
albertomalso why 18 ?20:48
wbhuberwell, that's just the # of partitions20:49
claygit's weird because the ec storage polices validate ring has data + parity primary partition placements (replicas) - but I don't see where the docs say it20:49
wbhuberclayg: i am actually looking for it.20:49
wbhubera doc update is much dired.20:49
*** bill_az has quit IRC20:50
claygwbhuber: yeah it looks like a paragraph right after (calculations depending on ec_type) would be a great place to go into the ring20:50
albertomoh i got lost in the doc20:51
albertombut i recall  the number of partitions would be the20:51
albertomnumber of nodes * 100 and then find an exponent of 2 that gets closer to it20:52
albertomso for 14 nodes,20:52
albertom2^1120:52
claygtimburke: so with the delete-marker the object name will still be in the container listing too20:53
timburkeclayg: yes. with a content-type of application/x-delete-marker20:53
*** joeljwright has quit IRC20:57
wbhuberalbertom: i'm not certain about # of nodes * 100 and find the power of 2 to be close to it20:57
wbhuberalbertom: where did you find that?20:57
albertomhe, trying to find the page again20:57
wbhuberi'm looking in swift-ring-builder code (ringbuilder.py)20:58
*** robefran has joined #openstack-swift20:58
claygthere's this one -> but it's a little optomistic -> https://rackerlabs.github.io/swift-ppc/20:59
albertomhttp://docs.openstack.org/developer/swift/deployment_guide.html#ring-preparing20:59
albertomthe part_power part20:59
clayg... that or people don't tend to think very realistically about the max drives - or look very closely at "max part per drive"21:00
claygalbertom: how many devices do you have currently?21:00
wbhuberclayg: we need to make that page EC friendly!21:00
wbhuberi.e. Replicas/EC (data + parity)21:00
albertomclayg: i am only testing erasure codes with ec_type = isa_l_rs_vand                                                                                                         │21:00
claygwbhuber: not my page21:00
albertomI have 14 drives21:00
wbhuber14 per each node?21:01
albertomone node with 14 drives :P21:01
albertomi am just testing the software so I dont care for performance right now21:01
claygno worries, with 14 devices you can go as high as 16 and still fill up the devices before you loose your lunch on inodes21:02
claygbut if you're just testing and not planning on growing the deployment up to more than a few thousand drives you may as well just do like 10 to save some cpu's when you build your rings21:02
wbhuberclayg: before i take off for my kid's school open house, can you quickly open a doc bug for me to include a PP on how to build a ring specifically for EC?21:04
claygPP?21:04
wbhuberparagraph21:04
claygsure!21:04
openstackgerritJanie Richling proposed openstack/swift: Decrypting Container Listing  https://review.openstack.org/21443821:04
wbhuberclayg: ty21:04
albertomSo what happens when you create the ring with a low part_power and then grow beyond the expected ?21:05
claygalbertom: most people can guess w/i two orders of magnitude21:07
claygalbertom: but if you start with 3 drives and end up with >300 you'll get lumpy data distribution and whine at cschwede to finish part power adjustment21:07
*** lpabon has quit IRC21:08
claygmost of the time people start with a few nodes all of which have more than a few devices, and can't realistically grow beyond many thounds of devices in many hundreds of servers without hitting some sort of limit - so it ends up being a mostly non/low-priority issue21:10
albertomso for ec codes i dont need rsync right ?21:13
claygalbertom: container & account still use rsync for rebalance21:14
*** acoles_ has quit IRC21:16
*** zul has quit IRC21:18
*** jrichli has quit IRC21:18
*** acoles_ has joined #openstack-swift21:19
*** ChanServ sets mode: +v acoles_21:19
*** acoles_ is now known as acoles21:19
*** annegentle has quit IRC21:20
*** annegentle has joined #openstack-swift21:20
lcurtisGood Afternoon all...question...if I lose a disk containing async_pendings, am I pretty much hosed?21:24
*** wbhuber_ has joined #openstack-swift21:24
clayglcurtis: not normally21:25
clayglcurtis: most of the time there's other disks that have the same async_pending on them - that or they already got it into the containers and you're fine21:25
lcurtisah...make sense21:26
lcurtissorry--one more hopefully quick question...would sqlite wirtes be considered syncrhonous?21:27
lcurtisguessing writes to the pending files not syncrhonous, but actual merge would be?21:27
clayghrmm... syncrhonous in what sense?  "does an object PUT commit it's existance to the container before returning success"?21:28
lcurtissyncrhonous write to disk21:28
lcurtisso put written to pending file, then merged to sqlite db21:29
claygis there such a thing as an asynchronous write to disk?21:29
*** wbhuber has quit IRC21:29
*** wbhuber_ is now known as wbhuber21:29
lcurtisi believe so, wehre a sync is not explicitly specified, though i could certainly be mistaken21:30
clayglike... page cache without fsync?  or like some sort of crazy libaio... i'm confused I think :D21:30
clayglcurtis: well that's true - because sqlite (and most databases) like get better throughput doing bulk inserts instead of row by row - the object updates get bunched up in the .pending file (that's an append write) and only slammed into sqlite on read21:31
clayg... well not "only" - if you're doing a bunch of writes eventually the update will pause and merge - otherwise the container existence check every 30 seconds will get at least one of the replicas21:32
*** dustins has quit IRC21:32
lcurtisjust curious..trying to figure out ways high container writes21:32
claygah21:33
clayglcurtis: notmyname's thing is good!21:33
claygChange-Id: I737bd0af3f124a4ce3e0862a155e97c1f0ac3e5221:33
lcurtisclayg: awesome! thank you. He had mentioned that before21:34
*** aerwin has quit IRC21:35
lcurtisbookmarked it this time21:36
torgomaticdidn't we merge that thing yet?21:37
claygtorgomatic: no no no - I think it's merged!21:38
torgomaticclayg: oh good21:38
*** annegentle has quit IRC21:42
tdasilvatimburke: interesting patch on versioned writes!21:43
timburketdasilva: i thought you might be :)21:44
tdasilvatimburke: is the marker being added to the "versioned container" or the container holding the past versions?21:45
tdasilvatimburke: the marker being a 0 byte file, right?21:45
*** david__ has quit IRC21:45
hrouMe too : - )  In the sense that it'll make life easier for us crypto folks21:46
tdasilvahrou: how so?21:46
timburketdasilva: versioned container. unless you try to push-marker-delete a delete marker, in which case both. and yes, 0-byte file21:46
hroutdasilva, I guess more accurately phrased the copy middleware will take care of it from our perspective.21:47
hroutdasilva, such that any copy (versioning is just an example) will flow through the crypto middleware as a get and put21:48
*** alejandrito has quit IRC21:49
tdasilvatimburke: was wondering if you would make sense to PUT it in the "X-Version-Location" container??? so that it marks that it was deleted, but doesn't "polute" the source container21:49
tdasilvahrou: got it!21:49
*** bapalm has quit IRC21:49
tdasilvatimburke: just an idea21:50
*** lpabon has joined #openstack-swift21:52
*** bapalm has joined #openstack-swift21:52
*** alejandrito has joined #openstack-swift21:52
timburketdasilva: interesting... clayg did seem worried about that...21:53
timburketdasilva: is it worth breaking pop-to-restore?21:54
tdasilvanot sure it needs to breat21:54
tdasilvabreak21:54
tdasilvatimburke: if i understand what you are trying to say, you are worried about poping a marker?21:55
tdasilvatimburke: in that case, i'd just pop again...21:55
timburketdasilva: yes. what i like about the way i've written it is that a delete is just another version. it can use all the same machinery and has all the same characteristics21:56
*** bapalm has quit IRC21:57
timburkeif we write the delete marker to the versions container, then pop-version deletes to the versioned container would need to look in two places...though i suppose we'd do that anyway...lemme play with it21:57
*** nadeem has quit IRC21:58
*** bapalm has joined #openstack-swift22:03
*** jkugel has quit IRC22:05
*** petertr7 is now known as petertr7_away22:06
*** bapalm has quit IRC22:08
openstackgerritClay Gerrard proposed openstack/swift: WIP: fixups for etag_buckets  https://review.openstack.org/21536022:09
claygpeluse_: ^ please go ahead and squash any of that you think is helpful22:09
*** lpabon has quit IRC22:13
*** NM has quit IRC22:14
*** bapalm has joined #openstack-swift22:14
*** bapalm has quit IRC22:19
*** annegentle has joined #openstack-swift22:20
*** minwoob has quit IRC22:20
*** chlong has quit IRC22:25
*** zul has joined #openstack-swift22:26
*** lcurtis has quit IRC22:28
*** bapalm has joined #openstack-swift22:29
*** annegentle has quit IRC22:33
*** annegentle has joined #openstack-swift22:33
*** bapalm has quit IRC22:34
*** bapalm has joined #openstack-swift22:35
*** annegentle has quit IRC22:38
*** bapalm has quit IRC22:43
*** bapalm has joined #openstack-swift22:43
*** hezhiqiang has quit IRC22:47
*** bapalm has quit IRC22:48
*** bapalm has joined #openstack-swift22:51
*** thurloat_isgone is now known as thurloat22:52
*** NM has joined #openstack-swift22:54
*** hrou has quit IRC23:00
*** chsc has quit IRC23:02
*** thurloat is now known as thurloat_isgone23:03
*** jlhinson has quit IRC23:03
*** alejandrito has quit IRC23:06
*** zhill has quit IRC23:07
*** david__ has joined #openstack-swift23:07
*** zhill has joined #openstack-swift23:08
*** zhill has quit IRC23:12
*** kutija has quit IRC23:15
*** km has joined #openstack-swift23:18
*** kei_yama has joined #openstack-swift23:19
*** ho has joined #openstack-swift23:23
*** blmartin has quit IRC23:30
hogood morning!23:36
*** robefran has quit IRC23:48
openstackgerritTim Burke proposed openstack/swift: Add the ability to specify ranges for SLO segments  https://review.openstack.org/21172623:49
*** hrou has joined #openstack-swift23:55

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