Thursday, 2020-01-23

*** diablo_rojo has joined #openstack-swift00:03
*** paladox is now known as paladox_UK_IN_EU00:20
timburkeseongsoocho, \o/ glad to hear it00:32
openstackgerritTim Burke proposed openstack/swift master: py38: cgi lost some names  https://review.opendev.org/68859400:51
timburkerledisez, if you get a chance, i'd love to get your opinion on the state of the etag_quoter patch -- in particular, the conversation around https://review.opendev.org/#/c/700056/6/swift/common/middleware/etagquoter.py@2100:56
patchbotpatch 700056 - swift - Middleware that allows a user to have quoted Etags - 7 patch sets00:56
*** paladox_UK_IN_EU is now known as paladox01:24
*** gyee has quit IRC01:52
*** psachin has joined #openstack-swift03:36
*** evrardjp has quit IRC05:34
*** evrardjp has joined #openstack-swift05:34
openstackgerritTim Burke proposed openstack/swift master: container-sync: Sync static links similar to how we sync SLOs  https://review.opendev.org/69813905:36
openstackgerritTim Burke proposed openstack/swift master: New Object Versioning mode  https://review.opendev.org/68238205:36
openstackgerritTim Burke proposed openstack/swift master: s3api: Implement object versioning API  https://review.opendev.org/67368205:36
timburke^^^ my edits on the versioning chain05:36
timburkei kept the three patches we had stacked before, but moved the versioning-specific parts of p 698139 to the versioning patch (keeping just the static link stuff, some middleware re-org, and adding sync-to to get_container_info)05:38
patchbothttps://review.opendev.org/#/c/698139/ - swift - container-sync: Sync static links similar to how w... - 15 patch sets05:38
timburkei'm still not entirely happy with the 412 stuff, but when i picked at it, things started to unravel :-/05:38
timburkei'll see what i can do with it tomorrow05:39
*** manuvakery has quit IRC05:42
*** manuvakery has joined #openstack-swift05:42
*** tkajinam has quit IRC06:01
*** tkajinam has joined #openstack-swift06:03
*** ccamacho has quit IRC06:26
*** tkajinam_ has joined #openstack-swift07:01
*** tkajinam has quit IRC07:03
*** heso has joined #openstack-swift07:12
*** tkajinam__ has joined #openstack-swift07:47
*** tkajinam_ has quit IRC07:50
*** ccamacho has joined #openstack-swift07:50
*** manuvakery has quit IRC07:54
*** tesseract has joined #openstack-swift07:56
*** f0o has quit IRC07:58
*** cjloader has quit IRC07:59
*** fungi has quit IRC07:59
*** ormandj has quit IRC07:59
*** toanster has quit IRC07:59
*** rledisez has quit IRC07:59
*** Jeffrey4l has quit IRC07:59
*** godog has quit IRC07:59
*** csmart has quit IRC07:59
*** Anticimex has quit IRC07:59
*** baffle has quit IRC08:02
*** Jeffrey4l has joined #openstack-swift08:02
*** godog has joined #openstack-swift08:02
*** Anticimex has joined #openstack-swift08:02
*** irclogbot_2 has quit IRC08:04
*** irclogbot_2 has joined #openstack-swift08:05
*** rdejoux has joined #openstack-swift08:12
openstackgerritThiago da Silva proposed openstack/swift master: Return correct etag for raw manifest  https://review.opendev.org/70393308:17
*** tkajinam__ has quit IRC08:29
*** tkajinam has joined #openstack-swift08:29
*** tkajinam has quit IRC08:35
*** rpittau|afk is now known as rpittau08:43
*** cjloader has joined #openstack-swift08:45
*** fungi has joined #openstack-swift08:45
*** f0o has joined #openstack-swift08:45
*** ormandj has joined #openstack-swift08:45
*** toanster has joined #openstack-swift08:45
*** rledisez has joined #openstack-swift08:45
*** orwell.freenode.net sets mode: +v rledisez08:45
*** openstackstatus has quit IRC08:48
*** pcaruana has joined #openstack-swift09:30
*** mkrai_ has joined #openstack-swift09:36
mkrai_Hello Everyone, I am trying to understand Swift worflow for PUT request. Is there any document available for the same?09:36
tdasilvamkrai_: not sure if there's any official docs, but found this online: https://openstack.nimeyo.com/104076/openstack-swift-internal-working-of-put-command09:51
mkrai_tdasilva, Thanks I will take a look.09:53
*** jistr has quit IRC10:06
*** jistr has joined #openstack-swift10:06
openstackgerritMerged openstack/swift master: Allow bulk to fwd some headers at tar extraction  https://review.opendev.org/70173611:03
*** rpittau is now known as rpittau|bbl11:21
*** rcernin has quit IRC12:15
*** jvisser has joined #openstack-swift12:16
*** rcernin has joined #openstack-swift12:20
*** rcernin has quit IRC12:26
*** jvisser has quit IRC12:28
*** jvisser has joined #openstack-swift12:28
*** baffle has joined #openstack-swift12:38
*** csmart has joined #openstack-swift12:38
*** mkrai_ has quit IRC12:48
*** pawan-gupta has joined #openstack-swift12:57
*** rpittau|bbl is now known as rpittau13:00
*** jvisser has quit IRC13:23
*** mkrai_ has joined #openstack-swift13:24
*** mkrai_ has quit IRC13:50
openstackgerritMerged openstack/swift stable/train: py3: Fix formpost unicode filename issues  https://review.opendev.org/70365413:55
*** jvisser has joined #openstack-swift14:02
*** jvisser has quit IRC14:07
*** jvisser has joined #openstack-swift14:24
*** mkrai_ has joined #openstack-swift14:28
*** pcaruana has quit IRC15:04
*** psachin has quit IRC15:13
*** gyee has joined #openstack-swift16:14
*** mkrai_ has quit IRC16:18
DHEAnyone seen object servers hanging and ignoring connections until killed off? Stack trace of one of them: https://paste.centos.org/view/8b147894 I guess there's a lock deadlock somehow when syslogging based on the stack?16:23
DHEI assume some package needs updating...16:24
*** ccamacho has quit IRC16:26
ormandjanybody having issues w/ SLO + s3 api using the aws s3 cli (https://github.com/wbingli/awscli-plugin-endpoint) to specify custom endpoint, of course? we're seeing: "The Content-MD5 you specified did not match what we received." (400) errors on each part16:38
*** tesseract has quit IRC16:41
ormandjactually, seems to impact all usage of aws cli16:58
ormandjnot just SLOs16:58
*** mkrai has joined #openstack-swift17:02
timburkeormandj, sounds familiar -- https://bugs.launchpad.net/swift/+bug/1810397 or maybe what got fixed with https://review.opendev.org/#/c/660196/17:08
openstackLaunchpad bug 1810397 in OpenStack Object Storage (swift) "Remove ETag header from request in CompleteMultipartUpload" [Medium,Fix released]17:08
patchbotpatch 660196 - swift - s3api: Allow clients to upload with UNSIGNED-PAYLOAD (MERGED) - 1 patch set17:08
timburkeormandj, what version of swift is this?17:08
*** rdejoux has quit IRC17:09
*** diablo_rojo has quit IRC17:15
*** rpittau is now known as rpittau|afk17:21
*** diablo_rojo has joined #openstack-swift17:22
timburkeDHE, sounds kinda like https://bugs.launchpad.net/swift/+bug/1710328 ... but every py3-ready swift should have that fix :-/17:25
openstackLaunchpad bug 1710328 in OpenStack Object Storage (swift) "object server deadlocks when a worker thread logs something" [High,Fix released] - Assigned to Samuel Merritt (torgomatic)17:25
timburkemight be worth installing gdb and running https://github.com/swiftstack/python-stack-xray/blob/master/python-stack-xray on a hung server to see all of the greenthreads' stacks, too17:26
*** mkrai has quit IRC17:31
*** evrardjp has quit IRC17:34
*** evrardjp has joined #openstack-swift17:34
ormandjtimburke: stein17:36
ormandjeven with a tiny 1 byte file it's happening using aws cli, not using that plugin, just specifying endpoint manually17:37
ormandji verified the right region is being set in aws cli, same for endpoint17:37
ormandjgetting back 400s from swift proxy17:37
ormandj(2.21.0 to be clear)17:38
ormandjlooks like .1 from your post fixes something w/ multipart, but it's failing even on small single part files (1810397), https://review.opendev.org/#/c/660196/ is definitely it though17:39
patchbotpatch 660196 - swift - s3api: Allow clients to upload with UNSIGNED-PAYLOAD (MERGED) - 1 patch set17:39
ormandjx-amz-content-sha256:UNSIGNED-PAYLOAD17:40
timburke*nods* i backported it to stein (p 675437) but it was released in 2.21.117:41
timburkehttps://github.com/openstack/swift/blob/2.21.1/CHANGELOG#L1717:42
timburkei think there's also a aws-cli config option to work around it... let me see...17:43
DHEtimburke: wow that's old... I'm running 2.23.1 so it should be good17:45
timburkethis guy! https://docs.aws.amazon.com/cli/latest/topic/s3-config.html#payload-signing-enabled17:45
ormandji'd rather us upgrade to .1 and fix both of these :)17:46
ormandjthanks for the quick assist on this, and the fix itself17:46
timburkeDHE, agree -- it was just the most-similar thing i could think of :-/17:46
DHEhmmm... maybe disabling threading would help then?17:48
DHEif it's indeed a related issue...17:48
*** psachin has joined #openstack-swift17:59
DHEwas considering UDP based syslogging, but it still uses the same fundamental code so, no...18:21
*** psachin has quit IRC18:43
timburkeclayg, tdasilva i think http://paste.openstack.org/show/788724/ should make recovery from a primary with no archive a good bit easier18:48
*** openstackgerrit has quit IRC18:51
*** heso has quit IRC19:13
ormandjtimburke: so fun thing, ubuntu cloud archives for bionic has 2.21.1 in proposed19:15
ormandjany chance it could be promoted?19:16
timburke¯\_(ツ)_/¯19:17
timburkemaybe ask coreycb?19:17
ormandjcoreycb: hello my new friend ;)19:17
*** gmann is now known as gmann_lunch19:26
*** gmann_lunch is now known as gmann19:52
* DHE is downgrading to python 3.6 in a "throw science at the wall and see what sticks" strategy...20:08
*** openstackstatus has joined #openstack-swift20:09
*** ChanServ sets mode: +v openstackstatus20:09
*** jvisser has quit IRC20:12
coreycbormandj: we're aiming to get that released monday - you can it track at https://bugs.launchpad.net/cloud-archive/+bug/185893420:50
openstackLaunchpad bug 1858934 in Ubuntu Cloud Archive stein "[SRU] stein stable releases" [High,Fix committed]20:50
timburkeDHE, i'm a great fan of experimentalism and empiricism ;-)21:00
timburkelet me know how it goes21:00
timburkeclayg, tdasilva i just noticed something a little funny with versioning...21:01
timburkeyou do an SLO PUT to a versioned container, you get a version id in your headers. so far, so good21:01
timburkeyou turn on heart-beating... no more version id :-/21:02
DHEtimburke: well I'm not a python guy so I don't have many other options if the experts (you) don't know what's up...21:02
*** rcernin has joined #openstack-swift21:04
timburkeDHE, i still think the stack xray script may be helpful -- it should give us a much better picture of what's going on in that object server, hopefully giving us some ideas about why we're waiting on a lock that never releases...21:04
DHEstack xray?21:09
DHEgoogle found it21:09
DHEshould work, i have debug symbols in my python build...21:10
zaitcevthere's also ghis: https://review.opendev.org/7051321:16
patchbotpatch 70513 - swift - Add Guru Meditation - 4 patch sets21:16
zaitcevIt is one of those always bridesmaid, never a bride kind of patches.21:16
*** diablo_rojo has quit IRC21:16
*** irclogbot_2 has quit IRC21:18
*** tonyb has quit IRC21:19
*** aluria has quit IRC21:19
*** cwright has quit IRC21:19
*** openstackstatus has quit IRC21:20
*** cwright has joined #openstack-swift21:26
*** tonyb has joined #openstack-swift21:26
*** irclogbot_1 has joined #openstack-swift21:26
*** diablo_rojo has joined #openstack-swift21:34
*** diablo_rojo has quit IRC21:37
tdasilvatimburke: interesting, going to start going through your patches in a bit21:39
*** diablo_rojo has joined #openstack-swift21:39
*** jvisser has joined #openstack-swift21:45
*** aluria has joined #openstack-swift22:01
*** diablo_rojo__ has joined #openstack-swift22:09
*** diablo_rojo has quit IRC22:11
*** jvisser has quit IRC22:17
mattoliverau morning22:21
*** diablo_rojo__ is now known as diablo_rojo22:46
*** tkajinam has joined #openstack-swift23:07
*** diablo_rojo has quit IRC23:42

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