Friday, 2017-07-14

*** itlinux has quit IRC00:00
*** OrionX has quit IRC00:01
torgomatictimburke: to be fair, 0 parity frags is like storing a single replica, but with even worse durability00:16
timburketorgomatic: yup. not saying it's a good idea, but it *was* something that came up once we got duplicated frags00:17
timburke(in fact, upon investigation, i determined it was kinda a *terrible* idea as i recall)00:18
*** tonanhngo has quit IRC00:19
*** itlinux has joined #openstack-swift00:19
*** tonanhngo has joined #openstack-swift00:20
*** JimCheung has quit IRC00:21
*** JimCheung has joined #openstack-swift00:22
*** JimCheung has quit IRC00:27
*** tonanhngo_ has joined #openstack-swift00:30
*** JimCheung has joined #openstack-swift00:30
*** tonanhngo has quit IRC00:31
*** tonanhngo_ has quit IRC00:34
openstackgerritClay Gerrard proposed openstack/swift master: Add multiple worker processes strategy to reconstructor  https://review.openstack.org/47841600:34
clayg^ i'm just throwing up my progress today, but I think it's getting close - i need to go over it again00:35
claygi found lp bug #1626290 in a weird state that made it hard to find in search :\00:35
openstackLaunchpad bug 1626290 in OpenStack Object Storage (swift) "[object-reconstructor] section is not documented" [Wishlist,Confirmed] https://launchpad.net/bugs/162629000:35
claygtimburke: you mis-understood the use-case00:36
timburkealmost certainly00:36
claygone thing you hit sometimes with uploads to swift of large objects is speed of a single spindle00:37
claygif you do a 7+0x3 in a single region, and compare it to a 3-replica single region - durability should be similar-ish - but your aggregate iops/disk-write-bw available to single client PUT stream should be higher00:38
clayg... that's all00:38
claygi may still be a terrible idea - and maybe even for durability reasons - but I remember thinking when you added to the ec duplication comparison you didn't seem to have some context of a previous out-of-band discussion on the thought experiment00:39
*** JimCheung has quit IRC00:40
claygor maybe you were reading the shorthand of a duplicated policy 7+0x3 like a math equation (i.e. 7 + (0 * 3) instead of the intended  (k + m) * duplication factor which is always the way to read swift ec duplication schemas since there is no such thing as just duplicating the parity frags)00:41
claygi don't recall00:41
claygg'night!00:41
mattoliveraumorning00:48
timburkekota_: on https://github.com/openstack/liberasurecode/commit/74eaa37 -- should we have added something to the error codes at https://github.com/openstack/liberasurecode/blob/1.5.0/include/erasurecode/erasurecode.h#L390-L401 ? like EBADFRAGSET or something?00:48
mattoliverauand g'night clayg00:48
timburkegood morning mattoliverau!00:49
*** JimCheung has joined #openstack-swift00:53
*** tovin07_ has joined #openstack-swift00:55
*** kiennt has joined #openstack-swift01:07
*** Renich has quit IRC01:10
*** tovin07_ has quit IRC01:17
*** tovin07_ has joined #openstack-swift01:18
*** itlinux has quit IRC01:37
*** OrionX_ has quit IRC01:37
*** itlinux has joined #openstack-swift01:50
*** itlinux has quit IRC01:51
*** JimCheung has quit IRC01:54
*** RayLei has joined #openstack-swift01:58
RayLeiI encountered a swauth bug, the certification takes two times to pass https://user-images.githubusercontent.com/23159763/28195049-95232d8a-687a-11e7-9e33-ac6a645c59a9.png01:58
kota_good morning01:59
kota_timburke: ack01:59
RayLeigood morning01:59
kota_timburke: it seems good idea to add the error case, currently liberasurecode raise UNKNOWN error with syslog message like "Encountered error in backend decode function!"02:05
kota_clayg, timburke: IIRC, current pyeclib doesn't support striping (i.e. m=0)02:07
kota_but... i think, i had seen the striping driver in pyeclib... but not in use so need to maintain if you want to use.02:08
*** JimCheung has joined #openstack-swift02:09
kota_clayg, timburke: https://github.com/openstack/pyeclib/blob/master/pyeclib/core.py#L17802:10
*** JimCheung has quit IRC02:14
RayLeiI guess clayg is asleep02:16
claygclayg never sleeps - he only defrags02:17
RayLeimay be you can help me look at the above certification02:20
*** frank_young has joined #openstack-swift02:20
*** JimCheung has joined #openstack-swift02:26
*** JimCheung has quit IRC02:31
*** frank_young has quit IRC02:34
*** frank_young has joined #openstack-swift02:34
*** frank_young is now known as frank-young02:35
*** frank_young has joined #openstack-swift02:36
openstackgerritTuan Luong-Anh proposed openstack/swift master: Update URL home-page in documents according to document migration  https://review.openstack.org/48375502:39
*** ukaynar has quit IRC02:55
*** ukaynar has joined #openstack-swift02:55
*** ukaynar has quit IRC02:59
*** JimCheung has joined #openstack-swift03:03
*** JimCheung has quit IRC03:08
*** silor has joined #openstack-swift03:27
*** itlinux has joined #openstack-swift03:28
*** psachin has joined #openstack-swift03:31
openstackgerritTuan Luong-Anh proposed openstack/swift master: Update URL home-page in documents according to document migration  https://review.openstack.org/48375503:33
*** gkadam has joined #openstack-swift03:36
*** OrionX has joined #openstack-swift03:47
*** OrionX_ has joined #openstack-swift03:48
*** OrionX has quit IRC03:52
openstackgerritOpenStack Proposal Bot proposed openstack/swift master: Updated from global requirements  https://review.openstack.org/8873603:53
*** kei_yama has quit IRC03:55
*** tovin07_ has quit IRC03:59
*** tovin07_ has joined #openstack-swift04:00
*** kei_yama has joined #openstack-swift04:03
RayLeiWhy swift certification, the first failure, the second successful?04:04
*** itlinux has quit IRC04:14
*** links has joined #openstack-swift04:18
frank_youngYou mean, the first time swift authentication fail, but the second time success ??04:27
*** frank_young has quit IRC04:44
*** frank_young has joined #openstack-swift04:45
*** RayLei has quit IRC04:50
*** JimCheung has joined #openstack-swift05:04
*** frank_young has quit IRC05:07
*** frank_young has joined #openstack-swift05:07
*** cbartz has joined #openstack-swift05:09
*** JimCheung has quit IRC05:09
*** kiennt has quit IRC05:23
*** abhitechie has joined #openstack-swift05:26
*** xrb has joined #openstack-swift05:26
*** abhitechie has quit IRC05:31
*** abhitechie has joined #openstack-swift05:33
*** abhinavtechie has joined #openstack-swift05:35
*** abhitechie has quit IRC05:35
*** kiennt has joined #openstack-swift05:36
*** gyee has quit IRC05:37
*** abhinavtechie has quit IRC05:43
*** abhitechie has joined #openstack-swift05:48
*** rcernin has joined #openstack-swift05:52
*** kiennt has quit IRC05:58
*** OrionX_ has quit IRC05:58
*** xrb has quit IRC06:01
*** silor has quit IRC06:02
*** kiennt has joined #openstack-swift06:03
*** remix_tj has quit IRC06:13
*** cbartz has quit IRC06:17
*** pcaruana has joined #openstack-swift06:25
*** silor has joined #openstack-swift06:27
*** tovin07_ has quit IRC06:30
*** klrmn has quit IRC06:34
*** tovin07_ has joined #openstack-swift06:36
*** remix_tj has joined #openstack-swift06:37
*** silor1 has joined #openstack-swift06:54
*** xrb has joined #openstack-swift06:56
*** silor has quit IRC06:56
*** silor1 is now known as silor06:56
*** RayLei has joined #openstack-swift06:56
RayLeihello06:56
*** RayLei has quit IRC06:57
*** JimCheung has joined #openstack-swift07:01
*** JimCheung has quit IRC07:05
*** cschwede_ has joined #openstack-swift07:10
*** cschwede has quit IRC07:13
*** tesseract has joined #openstack-swift07:14
*** cbartz has joined #openstack-swift07:15
*** kei_yama has quit IRC07:19
*** tovin07__ has joined #openstack-swift07:28
*** tovin07_ has quit IRC07:30
*** cbartz has quit IRC07:33
*** abhitechie has quit IRC07:35
*** RayLei has joined #openstack-swift07:49
*** OrionX has joined #openstack-swift07:58
*** JimCheung has joined #openstack-swift08:02
*** OrionX has quit IRC08:02
*** JimCheung has quit IRC08:07
*** tovin07 has joined #openstack-swift08:07
*** tovin07__ has quit IRC08:08
*** abhitechie has joined #openstack-swift08:09
*** joeljwright has joined #openstack-swift08:11
*** ChanServ sets mode: +v joeljwright08:11
*** abhitechie has quit IRC08:27
*** kei_yama has joined #openstack-swift08:51
*** kei_yama has quit IRC08:54
*** mvk has quit IRC09:32
*** joeljwright has quit IRC09:51
*** pcaruana has quit IRC09:56
*** abhitechie has joined #openstack-swift10:00
*** mvk has joined #openstack-swift10:01
*** JimCheung has joined #openstack-swift10:03
*** JimCheung has quit IRC10:08
*** pcaruana has joined #openstack-swift10:09
*** xrb has quit IRC10:15
*** xrb has joined #openstack-swift10:16
*** tovin07 has quit IRC10:20
*** frank_young has quit IRC10:22
*** cbartz has joined #openstack-swift10:22
*** abhitechie has quit IRC10:44
*** kiennt has quit IRC10:47
*** frank_young has joined #openstack-swift10:58
*** OrionX has joined #openstack-swift10:59
*** OrionX has quit IRC11:03
*** RayLei has quit IRC11:10
*** pcaruana has quit IRC11:25
*** d0ugal has joined #openstack-swift11:37
*** d0ugal has quit IRC11:39
*** d0ugal has joined #openstack-swift11:40
*** d0ugal has quit IRC11:40
*** d0ugal has joined #openstack-swift11:40
*** d0ugal has quit IRC11:44
*** dja has quit IRC11:46
*** frank_young has quit IRC11:47
*** dja has joined #openstack-swift11:50
*** frank_young has joined #openstack-swift11:57
*** frank_young has quit IRC12:01
*** JimCheung has joined #openstack-swift12:05
*** JimCheung has quit IRC12:09
*** MVenesio has joined #openstack-swift12:12
*** frank_young has joined #openstack-swift12:15
*** ukaynar has joined #openstack-swift13:05
*** JimCheung has joined #openstack-swift13:06
*** JimCheung has quit IRC13:11
*** szaher has quit IRC13:18
*** mat128 has joined #openstack-swift13:22
*** ukaynar has quit IRC13:25
*** Dinesh_Bhor has quit IRC13:26
*** frank_young has quit IRC13:26
cschwede_sorry for the delay folks - here is the updated bug triage list: https://etherpad.openstack.org/p/swift-bug-triage-list13:37
cschwede_some lines include a Gerrit link; I combined two scripts to find unlinked patches. Thus bugs in state NEW with a gerrit patch are likely at least in progress13:38
*** gkadam has quit IRC13:38
cschwede_acoles, clayg, tdasilva, notmyname ^13:38
acolescschwede_: thanks, great work13:39
cschwede_acoles: bummer, i messed up the launchpad urls. one second13:39
acolescschwede_: hehe I hadn't noticed until you said13:39
*** catintheroof has joined #openstack-swift13:40
cschwede_fixed13:42
acolescschwede_: tdasilva: so what is the best way to update the etherpad - add your nick to start of line when you take on bugs to triage, strikethrough when closed, make it bold is in doubt and want consensus?13:43
acoless/is/if/13:43
cschwede_acoles: sounds like a good plan to me13:44
cschwede_acoles: maybe add the keyword triage to the bug as well? to ensure we can filter on that in the future within launchpad13:45
cschwede_especially if the bug stays in state NEW13:45
tdasilvacschwede_: so ideally all bugs that are not in the initial state of 'undecided' should be 'triaged', no?13:47
acolescschwede_: there is a status 'Triaged'13:47
tdasilvashould we move bugs from New to Triaged? and what does triaged mean? confirmed??13:48
tdasilvaheh13:48
tdasilvatoo many status options13:49
cschwede_acoles: hmm, after thinking about it, a bug should either be Confirmed, or Invalid, or not?13:49
acolesstatus 'New' is defined as 'not looked at' so ideally we'd like to have zero New (at least for some nano-second! )13:49
cschwede_there should be no old bugs in New after the triage?13:49
cschwede_right13:49
tdasilvasounds good13:50
cschwede_so ignore my triage comment above :=13:50
cschwede_:)13:50
acolesso we need 'Looked at but still don't know what to do with this'13:50
tdasilvaacoles: don't know for what reason? missing info?13:51
tdasilvaor can't confirm it's a bug or not?13:51
acolestdasilva: need another opinion, so I guess it just stays New until we get consensus13:51
acolesbut I liked cschwede_ idea that we somehow tag to differentiate New that we have looked at from newer News (!)13:52
tdasilvaacoles: trying to write all that down in the etherpad...13:52
acolescschwede_: can the scripts filter bugs with no activity, and since a date - that might make my thought moot i.e. it owuld be easy to find stuff that arrrived since we started this process?13:53
cschwede_acoles: yes, already did that. the age is the amount of days without an update13:54
cschwede_so the bugs triaged in the last couple days are not included13:54
acolesright13:54
*** pcaruana has joined #openstack-swift13:56
*** RayLei has joined #openstack-swift13:59
acolestdasilva: I just hijacked your how to use tags section :)13:59
tdasilvalol, i kind of meant it to mean like what tags to use for what purpose :P13:59
tdasilvashould we detail that?13:59
acolesyeah I suspected you did, just bump my stuff down :P14:00
tdasilvaheheh14:00
tdasilvaacoles: i was hoping you would come up with the tags ;)14:00
*** frank_young has joined #openstack-swift14:02
acolestdasilva: this is the editor I have always wanted ... I think changes and see them happen before my eyes :)14:03
tdasilvarofl14:03
tdasilvaacoles, cschwede_: should we also add tags for components?14:03
tdasilvawe have some like docs, translation, ec14:04
tdasilvawondering if we should things like, middleware, proxy, object-server, container-server, account-server???14:04
tdasilvais that helpful at all?14:04
acolesthe server-types I can imagine so - people may tag middlewares separately14:06
*** szaher has joined #openstack-swift14:06
acolesprobably not owrth creating per-middleware tags up front, they can't all have bugs!14:07
acolesworth*14:07
*** joelinux_ has quit IRC14:08
cschwede_tdasilva: sounds like a good idea. we should also add the wishlist tag if it's not a bug but sth that would be useful14:10
acolescschwede_: there is a status for wishlist IIRC14:10
*** SkyRocknRoll has joined #openstack-swift14:10
tdasilvaImportance14:10
* cschwede_ needs to look into launchpad and not only the api results14:11
tdasilvaheh14:11
cschwede_tdasilva: right, let's use the importance for it!14:11
acolestdasilva: sorry, IDNRC - Importance14:12
tdasilvawhat do you think of my break down of the status on the etherpad? does it make sense to say not to use Triaged and just use Confirmed?14:12
cschwede_that said, there is "Undecided, needs more discussion" in the prio as well14:12
RayLeii meet a bug about swift v1 auth https://user-images.githubusercontent.com/23159763/28195049-95232d8a-687a-11e7-9e33-ac6a645c59a9.png14:12
cschwede_tdasilva: sounds good to me14:12
cschwede_RayLei: you should make sure that your auth token gets invalidated asap14:13
RayLeiIf you do not restart memcached, I use the old token to be able to access swift twice in succession14:14
cschwede_RayLei: and delete that image asap14:14
cschwede_RayLei: expiring the token does not fix this, but please do yourself a favor and remove that image and expire your token!14:14
tdasilvacschwede_, acoles: so I made a suggestion that a bug should be left Undecided only when it has not been triaged. As part of the triaging, we should always set the Importance14:16
tdasilvaImportance can change as more people look at the bug14:16
cschwede_+114:16
RayLei+114:17
acolestdasilva: looks good14:18
*** mat128 has quit IRC14:22
*** mat128 has joined #openstack-swift14:23
*** RayLei has quit IRC14:26
*** ukaynar has joined #openstack-swift14:31
*** chlong has joined #openstack-swift14:48
*** links has quit IRC14:49
*** psachin has quit IRC14:56
*** lucasxu has joined #openstack-swift14:59
*** OrionX has joined #openstack-swift15:01
*** JimCheung has joined #openstack-swift15:02
*** OrionX has quit IRC15:04
*** rcernin has quit IRC15:05
*** ukaynar_ has joined #openstack-swift15:06
*** JimCheung has quit IRC15:06
*** ukaynar has quit IRC15:07
*** MVenesio has quit IRC15:25
*** klrmn has joined #openstack-swift15:26
openstackgerritMerged openstack/swift master: Add X-Copy-From-Account to api-ref  https://review.openstack.org/48311415:32
notmynamegood morning15:32
notmynamecschwede_: thanks for the bug etherpad15:33
*** mvk has quit IRC15:35
notmynameFYI in case you see it, new setuptools release seems to be causing problems http://lists.openstack.org/pipermail/openstack-dev/2017-July/119694.html15:38
*** SkyRocknRoll has quit IRC15:47
openstackgerritOpenStack Proposal Bot proposed openstack/swift master: Updated from global requirements  https://review.openstack.org/8873615:51
*** itlinux has joined #openstack-swift15:52
notmynamethere's a new cloud out there now. alibaba. their object storage looks like s3 https://www.alibabacloud.com/product/oss?spm=a3c0i.o31978en.675768.dnavproductsb1.9ebe31Rkh9B515:59
*** klrmn has quit IRC16:03
*** jeblair has quit IRC16:09
*** jeblair has joined #openstack-swift16:11
*** oshritf has joined #openstack-swift16:12
claygso I'm thinking I might go ahead and square away lp bug #1626290 today real quick (after I commute and work on this other thing I need to do)16:13
openstackLaunchpad bug 1626290 in OpenStack Object Storage (swift) "[object-reconstructor] section is not documented" [Wishlist,Confirmed] https://launchpad.net/bugs/162629016:13
claygI guess I'll make it orthogonal to patch 478416 - but maybe if one of them merges we can rebase/fixup which ever is left still in review to catch up?16:14
patchbothttps://review.openstack.org/#/c/478416/ - swift - Add multiple worker processes strategy to reconstr...16:14
*** ukaynar_ has quit IRC16:14
*** frank_young has quit IRC16:14
*** tonanhngo has joined #openstack-swift16:15
acolesclayg: I had made a note to at least do that as a follow up but yeah if you get there then we'll rebase as necessary16:24
timburkegood morning16:24
*** oshritf has quit IRC16:25
timburkekota_: huh... i'd entirely missed the striping driver. looks like it doesn't touch libec at all? the more i play around with this, the more i want as much logic as possible in libec, with pyeclib being a minimal shim to get us python bindings for it16:27
*** xrb has quit IRC16:28
timburkei'd also *really* like to see us get rid of the C code in pyeclib. we've got ctypes; why can't we just use python?16:28
*** chsc has joined #openstack-swift16:29
*** cbartz has left #openstack-swift16:30
*** abhitechie has joined #openstack-swift16:34
tdasilvatimburke: the more we push logic to ec, the easier it will be to port pyeclib to a golang binding16:35
*** dosaboy_ is now known as dosaboy16:35
timburketdasilva: bingo!16:35
*** JimCheung has joined #openstack-swift16:41
*** klrmn has joined #openstack-swift16:53
*** psachin has joined #openstack-swift16:53
*** mvk has joined #openstack-swift16:58
*** psachin has quit IRC17:15
*** gyee has joined #openstack-swift17:20
*** pcaruana has quit IRC17:41
*** cschwede_ has quit IRC17:46
*** xrb has joined #openstack-swift17:51
*** saint_ has quit IRC18:10
*** frank_young has joined #openstack-swift18:13
*** frank_young has quit IRC18:18
*** spotz has quit IRC18:33
*** silor has quit IRC18:39
*** spotz has joined #openstack-swift18:45
*** lucasxu has quit IRC18:47
openstackgerritThiago da Silva proposed openstack/swift master: Small updates to saio doc  https://review.openstack.org/48401618:57
*** OrionX has joined #openstack-swift19:03
*** OrionX has left #openstack-swift19:03
*** MVenesio has joined #openstack-swift19:16
*** mat128 has quit IRC19:38
*** catintheroof has quit IRC20:01
*** JimCheung has quit IRC20:09
*** JimCheung has joined #openstack-swift20:09
*** frank_young has joined #openstack-swift20:37
*** frank_young has quit IRC20:41
*** ukaynar has joined #openstack-swift21:10
*** chlong has quit IRC21:19
*** MVenesio has quit IRC21:38
*** OrionX_ has joined #openstack-swift21:59
*** frank_young has joined #openstack-swift22:02
*** OrionX_ has quit IRC22:06
*** frank_young has quit IRC22:06
*** itlinux has quit IRC22:13
*** vint_bra has quit IRC22:18
*** xrb has quit IRC22:19
*** xrb has joined #openstack-swift22:19
*** xrb has quit IRC22:19
*** xrb has joined #openstack-swift22:20
*** xrb has quit IRC22:20
*** xrb has joined #openstack-swift22:21
*** xrb has quit IRC22:21
*** xrb has joined #openstack-swift22:21
*** xrb has quit IRC22:22
*** xrb has joined #openstack-swift22:22
*** xrb has quit IRC22:23
*** d0ugal has joined #openstack-swift22:29
openstackgerritTim Burke proposed openstack/liberasurecode master: flat_xor_hd: better validate args  https://review.openstack.org/48361022:53
*** OrionX has joined #openstack-swift23:02
*** OrionX has quit IRC23:06
*** JimCheung has quit IRC23:07
*** JimCheung has joined #openstack-swift23:07
openstackgerritTim Burke proposed openstack/liberasurecode master: flat_xor_hd: better validate args  https://review.openstack.org/48361023:24
*** itlinux has joined #openstack-swift23:27
*** frank_young has joined #openstack-swift23:40
*** itlinux_ has joined #openstack-swift23:43
*** frank_young has quit IRC23:45
*** itlinux has quit IRC23:49
*** ukaynar has quit IRC23:50
*** chsc has quit IRC23:55

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!