Tuesday, 2015-04-07

*** ho has joined #openstack-swift00:02
*** Gue______ has joined #openstack-swift00:10
*** thumpba_ has quit IRC00:11
*** zul has joined #openstack-swift00:23
*** zhill has joined #openstack-swift00:29
*** zhill has quit IRC00:33
*** dmorita has joined #openstack-swift00:36
*** jkugel has joined #openstack-swift00:57
*** mjfork has joined #openstack-swift01:04
*** occup4nt is now known as occupant01:04
*** annegentle has joined #openstack-swift01:06
hogood morning guys!01:13
*** mjfork has quit IRC01:20
*** annegentle has quit IRC01:22
*** kota_ has joined #openstack-swift01:22
*** annegentle has joined #openstack-swift01:24
*** annegentle has quit IRC01:29
*** annegent_ has joined #openstack-swift01:30
mattoliverauho: morning01:37
*** annegent_ has quit IRC01:39
*** sluo_wfh has joined #openstack-swift01:39
*** sluo_wfh has quit IRC01:40
homattoliverau: morning! what is your local time? 12:30?01:40
mattoliverau11:40 now. Change in time over the weekend :)01:40
homattoliverau: I see. thanks :-)01:43
*** km has joined #openstack-swift02:03
*** Gue______ has quit IRC02:12
*** chlong has quit IRC02:42
*** chlong has joined #openstack-swift02:44
*** nshaikh has joined #openstack-swift03:31
*** jamielennox is now known as jamielennox|away03:47
*** gyee has quit IRC03:50
*** nshaikh has quit IRC04:30
howhen I ran functest for 166576, I sometime met "DiskFileError: Bad fragment index: None: invalid literal for int() with base 10: 'None'" error.04:33
hohttp://paste.openstack.org/show/199094/04:33
hoDid you see it before?04:33
hos/sometime/sometimes/04:35
*** vinsh has joined #openstack-swift04:40
*** vinsh has quit IRC04:43
kota_ho: seems to relate to patch set 16759504:46
kota_ho: I did not test yet04:47
hokota_: thanks. yeah, I believe it is not related the patch.04:47
kota_ho: have you already setup ec policy? current some functional test pick up a policy randomly.04:48
kota_ho: Bad fragment index: None seems to be caused by a lack of fragment index in the backend request so...04:50
hokota_: yes, I wrote policy_type = erasure_coding in swift.conf.04:52
hokota_: but I didn't care about functest for it.04:52
hos/about functest for it/about it for functest/ sorry...04:53
kota_ho: no worries.04:53
kota_ho: I'm now thinking the reason it would happen...04:54
*** nshaikh has joined #openstack-swift04:56
kota_ho: object versioning would hit around copy issue from ec policy to another but I saw the fix at the pach 17000304:57
hokota_: Did you mention about SWIFT_TEST_POLICY? I checked https://review.openstack.org/#/c/167595/04:58
*** Akshat has joined #openstack-swift04:58
hokota_: thanks I will check the patch https://review.openstack.org/#/c/167595/04:58
kota_ho: yes, we could to set concrete policy to be tested with 16759504:59
hokota_: s%https://review.openstack.org/#/c/167595/%1777304:59
kota_ho: but it didn't be merged yet.04:59
kota_ho: not sure but the copy request might have issues, so if you find something wrong, please feel free to raise that.05:00
*** nshaikh has left #openstack-swift05:00
kota_ho: I am going to try to figure out, too.05:01
hokota_: thanks for your help. first, I will specify SWIFT_TEST_POLICY in swift.conf then run functest several times.05:02
kota_perhaps, versions_middleware is still in reviwing?05:02
kota_s/reviwing/reviewing/05:02
hokota_: yes. https://review.openstack.org/#/c/134347/05:04
kota_ho: thanks, not yet confirmed but maybe you might hit a bug to copy to another policy container (i.e from old version (e.g. replicated) to versioning container (e.g ec) without rerouting controller)05:06
hokota_: thanks :-)05:08
kota_ho: normal copied object has been already fixed at 170003 but  object-versioning looks still to use GETorHEAD directly05:09
kota_ho: anyways, SWIFT_TEST_POLICY would trim the issue because if you specified it, whole containers in the functests will be a same policy.05:10
hokta_: I just started the test with SWIFT_TEST_POLICY. I will let you know the test result after I make several trials05:12
hos/kta_/kota_/ really sorry...05:13
kota_ho: no problem :P05:13
*** ppai has joined #openstack-swift05:36
*** SkyRocknRoll has joined #openstack-swift05:42
*** zhill has joined #openstack-swift05:47
*** zhill has quit IRC05:47
*** SkyRocknRoll has quit IRC06:05
*** SkyRocknRoll has joined #openstack-swift06:07
*** Bsony has joined #openstack-swift06:14
*** silor has joined #openstack-swift06:15
*** Bsony has quit IRC06:22
cschwedeGood Morning!06:43
hokota_: I executed the functest after SWIFT_TEST_POLICY set in os.env.  I could not see the problem anymore (0/5 times).06:46
hocschwede: good morning!06:46
*** silor has quit IRC06:47
*** jistr has joined #openstack-swift06:47
cschwedeho: hello!06:48
hokota_: btw I saw same traceback in both logs, proxy-server and object-server. from log size point of view, i think we don't need to have it in proxy-server. what do you think? (maybe useful for trouble shooting?)06:49
kota_ho: sorry I'm back just now and sounds good not to appear the problem and then...07:06
*** Bsony has joined #openstack-swift07:07
kota_ho: about traceback, if we have really duplicated traceback, yes we can suppress it, imo.07:08
kota_ho: however I have to look at the reason deeply before making agree.07:09
kota_ho: the reason -> why we dump the duplicated log to proxy and object-sever07:10
kota_anyways, good morning cschwede :)07:10
cschwedehello kota_ !07:12
hokota_: thanks! after ec merge finish I will refactor it if there is no reason.07:14
*** silor has joined #openstack-swift07:14
*** joeljwright has joined #openstack-swift07:16
*** Bsony has quit IRC07:18
cschwedenotmyname: small question on https://etherpad.openstack.org/p/ec_merge_plan - the second reviewer who is adding a +2 also adds a +A? Or the third reviewer? Or later?07:19
mattoliveraucschwede: morning, I think the second.. But could be wrong, but either case it can't merge until the -2 is unplugged so should be OK07:20
*** mmcardle has joined #openstack-swift07:21
cschwedemattoliverau: Good Evening! Yes, I agree - but then I thought better ask one more time…07:22
mattoliveraulol07:22
cschwedeAnyways, I can +A my +2 later on as well07:22
mattoliverautrue :)07:22
*** nshaikh has joined #openstack-swift07:34
*** chlong has quit IRC07:34
*** geaaru has joined #openstack-swift07:48
*** jistr has quit IRC07:50
*** km_ has joined #openstack-swift07:54
*** km has quit IRC07:56
*** jistr has joined #openstack-swift08:09
*** jordanP has joined #openstack-swift08:11
*** dmorita has quit IRC08:50
*** kota_ has quit IRC08:54
*** nshaikh has quit IRC08:59
*** Akshat has quit IRC09:11
*** km_ has quit IRC09:13
*** jordanP has quit IRC09:20
*** Akshat has joined #openstack-swift09:30
*** Akshat has quit IRC09:32
*** ujjain has joined #openstack-swift09:39
*** jordanP has joined #openstack-swift09:44
*** aix has joined #openstack-swift09:45
*** foexle has joined #openstack-swift09:52
*** chlong has joined #openstack-swift10:06
*** jistr has quit IRC10:30
*** tobberydberg has joined #openstack-swift10:45
*** jistr has joined #openstack-swift10:51
*** jistr is now known as jistr|biab10:52
*** ujjain has quit IRC10:56
*** ujjain has joined #openstack-swift10:56
*** kei_yama has quit IRC11:06
*** mmcardle has quit IRC11:26
*** AbyssOne__ has quit IRC11:27
*** tab___ has joined #openstack-swift11:28
*** Akshat has joined #openstack-swift11:29
*** AbyssOne has joined #openstack-swift11:31
*** ujjain has quit IRC11:33
*** jkugel has quit IRC11:36
*** SkyRocknRoll has quit IRC11:42
*** mahatic has joined #openstack-swift11:44
*** dencaval has joined #openstack-swift11:46
*** chlong has quit IRC11:49
*** jistr|biab is now known as jistr11:49
*** chlong has joined #openstack-swift11:51
*** ho has quit IRC11:52
openstackgerritOpenStack Proposal Bot proposed openstack/swift: Updated from global requirements  https://review.openstack.org/8873612:06
*** mmcardle has joined #openstack-swift12:19
*** Akshat has quit IRC12:25
*** Akshat has joined #openstack-swift12:26
*** jkugel has joined #openstack-swift12:39
*** jkugel1 has joined #openstack-swift12:40
*** jkugel has quit IRC12:43
openstackgerritKamil Rykowski proposed openstack/swift: Replaced setting run_pause with standard interval  https://review.openstack.org/17117012:53
*** ppai has quit IRC12:55
*** annegentle has joined #openstack-swift12:58
*** fifieldt has joined #openstack-swift13:00
*** krykowski has joined #openstack-swift13:01
*** Akshat has quit IRC13:05
*** annegentle has quit IRC13:06
*** annegentle has joined #openstack-swift13:06
*** openstackgerrit has quit IRC13:07
*** openstackgerrit has joined #openstack-swift13:07
*** ppai has joined #openstack-swift13:09
*** ppai has quit IRC13:23
openstackgerritKamil Rykowski proposed openstack/swift: Replaced setting run_pause with standard interval  https://review.openstack.org/17117013:26
*** chlong has quit IRC13:51
*** skret has joined #openstack-swift14:10
*** annegentle has quit IRC14:16
*** petertr7 has joined #openstack-swift14:23
*** panbalag has joined #openstack-swift14:28
*** panbalag has left #openstack-swift14:28
*** tsg_ has joined #openstack-swift14:33
*** erlon has joined #openstack-swift14:34
*** devlaps has quit IRC14:38
*** lpabon has joined #openstack-swift14:41
*** annegentle has joined #openstack-swift14:47
*** panbalag has joined #openstack-swift14:49
*** annegentle has quit IRC14:58
*** annegentle has joined #openstack-swift15:01
*** annegent_ has joined #openstack-swift15:06
*** annegentle has quit IRC15:07
*** vinsh has joined #openstack-swift15:08
*** panbalag has left #openstack-swift15:15
*** jrichli has joined #openstack-swift15:15
*** annegentle has joined #openstack-swift15:17
*** annegent_ has quit IRC15:17
*** tsg_ has quit IRC15:20
*** panbalag has joined #openstack-swift15:28
panbalagHi, does anyone know how to retrieve the metadata keys stored on an account, container or object?15:29
glangeHEAD requests and look at the headers returned?15:30
joeljwrightthe metadata keys are returned in the headers of a HEAD request, they start 'x-object-meta-', 'x-container-meta-' and 'x-account-meta-' I think15:32
glangeswift has extensive documentation :)15:33
*** foexle has quit IRC15:33
panbalagglange, yeah...I tried posting the metadata "-H 'X-Object-Meta-subject1:science'  -H 'X-Object-Meta-subject2:literature'..but they dont seem to come up in the HEAD information..15:37
panbalagglange, my bad..i posted object meta data on a container... (a mistake a copy paste previous command :)15:39
panbalagglange, it works now15:39
glangecool15:39
*** devlaps has joined #openstack-swift15:40
*** gyee has joined #openstack-swift15:54
*** bkopilov has quit IRC16:01
panbalagwho should we ask for if we need a new one?16:04
panbalagsorry wrong window16:04
*** panbalag has left #openstack-swift16:04
*** krykowski has quit IRC16:06
*** tsg_ has joined #openstack-swift16:07
*** bkopilov has joined #openstack-swift16:14
*** welldannit has joined #openstack-swift16:14
*** aerwin has joined #openstack-swift16:18
notmynamegood morning16:23
*** jistr has quit IRC16:27
pelusemorning16:33
peluseanyone wishing to help the EC errort:  a good place to start is the user docs as they attempt (need eyes please!) to explain everyting :) https://review.openstack.org/#/c/169990/16:36
glangeerrort?  that's a bad typo :)16:38
*** devlaps has quit IRC16:39
*** thumpba has joined #openstack-swift16:44
peluseyeah, that's why I need doc reviewers :)16:47
pelusebut in my defense, the f and the r keys are pretty damned close togather...16:47
*** thumpba_ has joined #openstack-swift16:47
tdasilvaor togetha as they say around here16:49
*** thumpba has quit IRC16:50
pelusetdasilva, hey thanks for the last doc push man - good stuff!16:51
tdasilvapeluse: I hope it was helpful...16:51
pelusetdasilva, for sure - so seeing it all together did they make good sense?16:51
tdasilvabut yeah, I agree we still need more eyes on it...16:51
tdasilvapeluse: I think overall it is pretty good. There were some areas that I wanted to make bigger changes, but wasn't sure if they were correct, maybe I'll add comments or questions16:52
pelusetdasilva, ok cool.  now go review the code patches :)16:53
tdasilvahaha...half-way through 16998916:53
pelusesweet16:53
*** aix has quit IRC16:55
*** jordanP has quit IRC16:58
peluseglange, any chance of enticing you to into looking at this stuff too?  Would be awesome to get your thoughts on any/all of it!16:58
*** mmcardle has quit IRC17:03
*** tsg_ has quit IRC17:04
*** lpabon has quit IRC17:09
*** geaaru has quit IRC17:16
*** annegentle has quit IRC17:27
*** annegentle has joined #openstack-swift17:27
notmynameswift version history https://wiki.openstack.org/wiki/Swift/version_map17:30
*** annegent_ has joined #openstack-swift17:31
*** annegentle has quit IRC17:32
*** jkugel1 has quit IRC17:32
*** mjfork has joined #openstack-swift17:45
portantenotmyname: nice18:00
notmynameI was asked to put some dates on it, so I'll do that next18:00
notmynamealso, it's portante!!18:00
portantejust a swift wannabe18:00
*** jamielennox|away is now known as jamielennox18:01
*** jkugel has joined #openstack-swift18:05
*** tongli has joined #openstack-swift18:08
cschwedenotmyname: nice mapping (the version history). but then it doesn’t accord to https://launchpad.net/swift/kilo ?18:18
notmynamecschwede: how do you mean?18:18
notmynamethe dates?18:19
cschwedenotmyname: no, 2.2.1 and 2.2.2 are marked as kilo series: https://launchpad.net/swift/kilo/2.2.2 https://launchpad.net/swift/kilo/2.2.118:19
notmynameoh right. so you'd say that on my table, the openstack releases are the /end/ of that series18:20
notmyname2.2.1 and 2.2.2 are only tracked as "kilo" for openstack purposes. ie they happened during the openstack kilo dev cycle calendar time18:21
cschwedethat’s how i understood this until this morning, when a colleague told me that 2.2.1 and 2.2.2 are marked as kilo18:21
notmynamehow is that different?18:22
notmynameso, yeah, to talk abotu what's in the kilo release, we talk about anything that has happened since juno. which includes 2.2.1 and 2.2.218:23
*** khivin has joined #openstack-swift18:24
cschwedeyes, i agree on this. but i think 2.2.1 and 2.2.2 are kind of bugfix for the juno release, ie „are juno“. of course they are also included in kilo18:24
cschwedeanyways, i think the map is much better than the launchpad view18:25
*** zhill has joined #openstack-swift18:27
*** sweeper has left #openstack-swift18:27
notmynamecschwede: heh, ok :-)18:30
notmynamewell definitely feel free to update it to be more clear, if needed18:31
*** lpabon has joined #openstack-swift18:37
*** cdelatte has joined #openstack-swift18:42
*** dencaval has quit IRC18:54
*** annegent_ has quit IRC18:56
*** annegentle has joined #openstack-swift18:57
*** jistr has joined #openstack-swift19:07
*** annegent_ has joined #openstack-swift19:12
*** annegentle has quit IRC19:13
*** annegent_ has quit IRC19:17
*** annegentle has joined #openstack-swift19:18
*** mahatic has quit IRC19:22
*** openstackgerrit has quit IRC19:22
*** openstackgerrit has joined #openstack-swift19:22
tdasilvapeluse: around?19:33
peluseyup19:33
tdasilvathere were some new comments on the doc which I could take a stab at fixing, but wanted to check with you first19:33
tdasilvanot sure if you are already working on it19:33
tdasilvaor clayg ??19:33
pelusenope, feel free my friend19:34
tdasilvacool..one more thing...are the docs supposed to also be limited to 80 chars per line?19:34
tdasilvaI was wondering that yesterday and noticed that jrichli noticed the same issue towards the end of the ec doc...19:35
pelusethat's a good question, I dont know but have been doping so except for bullteted lists which seem to piss off RST19:35
klrmnpeluse: bulleted lists are okay with wrapping too, so long as the first letter of each line lines up19:37
klrmn(at least that's how it's been working with SwiftStack's .rst files)19:37
peluseklrmn, ahh, cool.  thanks!19:38
klrmnpeluse: happy to have a ready answer for a swift-related problem =)19:38
*** vinsh has quit IRC19:39
*** tsg has joined #openstack-swift19:42
*** silor has quit IRC19:48
*** jistr has quit IRC19:55
*** annegent_ has joined #openstack-swift19:59
*** tsg has quit IRC20:01
*** annegentle has quit IRC20:02
*** tsg has joined #openstack-swift20:05
*** lpabon has quit IRC20:15
*** vinsh has joined #openstack-swift20:24
*** tongli has quit IRC20:28
*** tsg_ has joined #openstack-swift20:29
*** tsg has quit IRC20:30
*** tgohad has joined #openstack-swift20:31
*** tsg_ has quit IRC20:33
*** yuan has quit IRC20:35
*** aerwin has quit IRC20:57
*** fifieldt has quit IRC21:08
*** annegent_ has quit IRC21:13
tdasilvapeluse, torgomatic, clayg: why did we choose "2" success messages from storage node for the .durable file? why not 1 or 3 or quorum?21:15
*** tobberydberg has quit IRC21:17
*** fifieldt has joined #openstack-swift21:21
torgomatictdasilva: I'm not sure. It matches how many replicas have to be successfully stored in a 3-replica case, but that's just a guess.21:21
*** chlong has joined #openstack-swift21:25
*** jrichli_ has joined #openstack-swift21:26
*** jrichli has quit IRC21:26
*** annegentle has joined #openstack-swift21:28
*** jrichli_ has quit IRC21:31
pelusewe chose 2 because we figured that was the minimum amout to know it would get propogated21:32
peluse> than that would mean potentially more latency and we figured it wasn't neccessary21:32
*** openstackgerrit has quit IRC21:37
*** openstackgerrit has joined #openstack-swift21:37
pelusetdasilva, if you're updating the doc patch right now go ahead and mark it WIP just so anyone else thinking of pushing over knows that an active update is happening so they should wait21:39
tdasilvapeluse: just added a comment there, I think since it is not my patch I can't mark -1 for Workflow21:42
tdasilvapeluse: just trying to figure out what changes are needed in overview_policies.rst, I think we should at least mention EC there21:44
pelusetdasilva, OK cool.  Sounds good, when I wrote that one we purposely left EC out since it didn't exist yet but yeah it would be good to add mention of it now :)21:47
pelusein fact, I used to have a bullet point in the opening paragraph of Storage Policies "* Different Durability Characteristics" that would be good to add now21:52
*** mjfork has quit IRC21:52
tdasilvawhere can I find it?21:52
peluseoh, its long gone I'm sure, you'll have to make some shit up :)21:53
pelusebut let me look...21:53
peluseand then maybe a small blub in the 'configuring policies' section that has a link to how to configure for EC21:53
tdasilvajust noticed we removed feature/sp and only kept feature/sp-review :(21:54
peluseahh crap, the diskFile routing stuff should probably be mentioned in the under the hood section as well21:54
tdasilvayeah, we also don't talk a lot about the changes for PUT in the proxy21:55
peluseyeah in both cases we dont' want a ton of detail but mentioning there are provisions for organzing the code in various modules based on policy is probably a good thing to do21:57
* peluse has to run off to meetings for a bit...21:58
openstackgerritOpenStack Proposal Bot proposed openstack/swift: Updated from global requirements  https://review.openstack.org/8873621:59
tdasilvapeluse: found this https://review.openstack.org/#/c/8582422:01
*** skret has quit IRC22:05
brianclinequestion... I know container sync was getting some love at some point a few months ago - has that still marched onwards or is it on hold at the moment?22:05
*** remix_tj has quit IRC22:11
*** remix_tj has joined #openstack-swift22:11
*** gyee has quit IRC22:11
*** jkugel has quit IRC22:13
*** gyee has joined #openstack-swift22:13
notmynameFWIW, my PTL email for the upcoming elections http://lists.openstack.org/pipermail/openstack-dev/2015-April/060870.html22:16
notmynamebriancline: some. let me look in the logs22:17
notmynamebriancline: well, ok. no a lot ;-)22:19
notmyname*not22:19
*** annegentle has quit IRC22:29
*** annegentle has joined #openstack-swift22:30
mattoliverauMorning22:37
tdasilvamattoliverau: morning :)22:42
notmynameyo22:43
*** annegentle has quit IRC22:43
*** joeljwright has quit IRC22:48
brianclinethanks - no biggie, just crossed my mind23:15
*** petertr7 has quit IRC23:31
*** ho has joined #openstack-swift23:44
*** annegentle has joined #openstack-swift23:44
*** annegentle has quit IRC23:50

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