Thursday, 2018-09-13

openstackgerritMerged openstack/swift master: added some bootstrapping info to the top of the SAIO doc  https://review.openstack.org/60220300:35
*** gyee has quit IRC00:59
*** chocolate-elvis has joined #openstack-swift02:16
*** kei_yama has quit IRC03:48
*** kei_yama has joined #openstack-swift04:03
openstackgerritPete Zaitcev proposed openstack/swift master: Replace MIME with PUT+POST for EC and Encryption  https://review.openstack.org/42791105:03
*** hseipp has joined #openstack-swift06:47
*** rcernin has quit IRC07:01
*** tesseract has joined #openstack-swift07:01
*** kei_yama has quit IRC07:22
*** kei_yama has joined #openstack-swift07:27
*** cbartz has joined #openstack-swift07:29
*** geaaru has joined #openstack-swift07:37
openstackgerritTim Burke proposed openstack/swift master: s3api: Include '-' in multipart ETags  https://review.openstack.org/57586007:54
*** d0ugal has joined #openstack-swift08:20
*** cbartz has quit IRC08:25
*** cbartz has joined #openstack-swift08:27
*** d0ugal has quit IRC08:29
*** d0ugal has joined #openstack-swift08:29
*** d0ugal has quit IRC08:35
*** d0ugal has joined #openstack-swift08:36
*** d0ugal has quit IRC08:44
*** d0ugal has joined #openstack-swift08:44
openstackgerritKota Tsuyuzaki proposed openstack/swift master: s3api: Include '-' in multipart ETags  https://review.openstack.org/57586010:29
*** d0ugal has quit IRC11:07
*** d0ugal has joined #openstack-swift11:08
openstackgerritMerged openstack/swift master: s3api: Include '-' in multipart ETags  https://review.openstack.org/57586011:30
*** ianychoi has quit IRC12:04
*** ianychoi has joined #openstack-swift12:11
*** szaher_ has joined #openstack-swift12:24
*** cbartz has quit IRC12:58
*** szaher_ has quit IRC13:02
*** kei_yama has quit IRC13:04
*** szaher has joined #openstack-swift13:05
*** mvenesio has joined #openstack-swift13:50
*** d0ugal has quit IRC13:56
*** d0ugal has joined #openstack-swift13:56
*** zaitcev has joined #openstack-swift13:57
*** ChanServ sets mode: +v zaitcev13:57
*** fatema_ has joined #openstack-swift14:01
*** pcaruana has joined #openstack-swift14:28
*** d0ugal has quit IRC14:59
*** itlinux has quit IRC15:09
*** gyee has joined #openstack-swift15:15
*** fatema has joined #openstack-swift15:24
*** sorrison has joined #openstack-swift15:39
openstackgerritThiago da Silva proposed openstack/swift master: WIP: add multinode func test to swift  https://review.openstack.org/60168615:47
timburkekota_: those new failures fell out of https://review.openstack.org/#/c/577941/ i think :-/15:51
patchbotpatch 577941 - swift - s3_acl: Require swift_owner authz to create buckets (MERGED) - 3 patch sets15:51
*** Sukhdev has joined #openstack-swift15:51
kota_let me look at15:51
timburkethe thing is, it never really made sense that test:tester3 could create buckets via s3 -- they can't create containers via swift!15:52
timburkemaybe we could add a test:tester2 to the sample confs who *is* an account owner?15:52
kota_i just wrote up some assertions to confirm my understanding like https://gist.github.com/bloodeagle40234/b7f700300b4cab1d664abd5ed1de1e8815:53
*** geaaru has quit IRC15:58
*** hseipp has quit IRC15:58
kota_alright16:00
*** itlinux has joined #openstack-swift16:01
kota_https://gist.github.com/bloodeagle40234/b7f700300b4cab1d664abd5ed1de1e8816:01
kota_update the gist16:01
*** sorrison has quit IRC16:11
*** sorrison has joined #openstack-swift16:15
*** itlinux is now known as itlinux-away16:15
kota_timburke: you seems to be right. non-owner, but .admin account can see BucketAlreadyExists when the bucket created by another owner16:15
*** itlinux-away is now known as itlinux16:17
*** itlinux is now known as itlinux-away16:17
openstackgerritClay Gerrard proposed openstack/swift master: Add a chunks_finished to BaseDiskFileWriter  https://review.openstack.org/60238816:24
*** itlinux-away is now known as itlinux16:25
*** itlinux is now known as itlinux-away16:25
*** itlinux-away is now known as itlinux16:25
*** itlinux is now known as itlinux-away16:25
*** itlinux-away is now known as itlinux16:27
timburkekota_: this reminds me that https://review.openstack.org/#/c/575829/ is a thing that hasn't landed yet...16:48
patchbotpatch 575829 - swift - Change PUT bucket conflict error - 4 patch sets16:48
kota_timburke: :P16:49
timburkeit also reminds me that we should probably have func tests both with and without s3_acl enabled16:51
openstackgerritThiago da Silva proposed openstack/swift master: WIP: add multinode func test to swift  https://review.openstack.org/60168616:51
kota_hah ->> https://review.openstack.org/#/c/575829/4/test/functional/s3api/test_bucket.py you already know the AccessDenied case!16:52
patchbotpatch 575829 - swift - Change PUT bucket conflict error - 4 patch sets16:52
kota_exactly, i don't think it can pass with non s3_acl case because it dosn't have the owner info.16:53
timburkethe weird thing to me is why i ever thought i needed AccessDenied in https://review.openstack.org/#/c/577941/3/test/functional/s3api/test_bucket.py16:54
patchbotpatch 577941 - swift - s3_acl: Require swift_owner authz to create buckets (MERGED) - 3 patch sets16:54
sorrisonnotmyname: I have some working code for s3 keystone token cacheing but need some help with how to do cacheing, can I come have a chat with you guys today sometime?16:54
openstackgerritClay Gerrard proposed openstack/swift master: Add a chunks_finished to BaseDiskFileWriter  https://review.openstack.org/60238816:55
openstackgerritClay Gerrard proposed openstack/swift master: Refactor obj.server.ObjectController.PUT  https://review.openstack.org/60220516:55
clayg^ timburke  zaitcev  - ok I think I have the diskfile + object-server ready for the STAGE/UPDATE/COMMIT change now :D16:56
notmynamesorrison: yes! we're in Big Thompson all day today16:56
kota_sorrison: I don't think s3api on top of swift can cache the keystone token because the actual aws signature requires the check with the secret and the sort of request information.16:58
sorrisonkota_: I'll come over in about 10 mins and I can explain my approach16:59
notmynamekota_: maybe. sorrison has some good ideas and his goal is to make it so that s3api requests do not have to hit keystone for every request16:59
kota_sorrison: alright, I'm in the room.16:59
zaitcevnotmyname: it sounds impossible just like Kota just explained, but16:59
*** tesseract has quit IRC17:00
zaitcevIt may be possible to cache hmacs for identical requests, like when some stupid app keeps hitting Swift with the same request again and again, so that only the timestamp changes17:00
notmynamewait until sorrison gets here to explain it. :) IIRC, there is a keystone endpoint where he's able to get some creds info and do what he wants17:02
timburkezaitcev: request timestamp gets baked into the signature to defend against replay attacks :P17:02
timburkeof course, it's still reliant upon client headers (Date and/or x-aws-date iirc?) but there's only a couple-minutes window where you could send *exactly* the same request17:03
timburkepresigned urls are another story, of course17:04
*** sorrison has quit IRC17:05
zaitcevtimburke: yeah... so when replay is desired, then... well we only need to remain in suspense for a few minutes17:05
*** sorrison has joined #openstack-swift17:05
*** e0ne has joined #openstack-swift17:06
e0nenotmyname: hi John. will you have a time tomorrow to discuss swift and horizon integration?17:07
*** sorrison has quit IRC17:07
*** sorrison has joined #openstack-swift17:08
notmynamee0ne: sure. or this afternoon. whatever is good for you17:10
e0nenotmyname: I'm sorry, can't do it today. I'll catch you tomorrow17:11
notmynamee0ne: anything specific you'd like to talk about? if there's somthing I can prepare for....17:12
e0nenotmyname: I would like to fill https://etherpad.openstack.org/p/horizon-feature-gap for the Swift to prepare TODO list to the Horizon team17:13
notmynameok17:13
openstackgerritAlexandre Lécuyer proposed openstack/swift master: WIP - Abstract file and directory operations in replicator/reconstructor  https://review.openstack.org/60240217:16
openstackgerritThiago da Silva proposed openstack/swift master: WIP: add multinode func test to swift  https://review.openstack.org/60168617:21
claygtimburke: right!  and probably some tests too17:22
*** itlinux is now known as itlinux-away17:27
*** itlinux-away is now known as itlinux17:29
*** itlinux is now known as itlinux-away17:29
*** sorrison has quit IRC17:30
*** itlinux-away is now known as itlinux17:32
*** itlinux is now known as itlinux-away17:32
*** sorrison has joined #openstack-swift17:32
*** itlinux-away is now known as itlinux17:34
*** itlinux is now known as itlinux-away17:34
*** itlinux-away is now known as itlinux17:38
*** itlinux is now known as itlinux-away17:38
*** itlinux-away is now known as itlinux17:41
notmynamefatema: http://paste.openstack.org/show/730044/17:46
openstackgerritTim Burke proposed openstack/swift master: Change PUT bucket conflict error  https://review.openstack.org/57582917:46
openstackgerritTim Burke proposed openstack/swift master: Add another user for s3api func tests  https://review.openstack.org/60241117:46
*** e0ne has quit IRC17:48
*** e0ne has joined #openstack-swift17:49
openstackgerritTim Burke proposed openstack/swift master: Change PUT bucket conflict error  https://review.openstack.org/57582917:54
openstackgerritTim Burke proposed openstack/swift master: Add another user for s3api func tests  https://review.openstack.org/60241118:00
openstackgerritTim Burke proposed openstack/swift master: Change PUT bucket conflict error  https://review.openstack.org/57582918:00
kota_nice18:04
*** sorrison has quit IRC18:10
openstackgerritClay Gerrard proposed openstack/swift master: Add a chunks_finished to BaseDiskFileWriter  https://review.openstack.org/60238818:12
openstackgerritClay Gerrard proposed openstack/swift master: Refactor obj.server.ObjectController.PUT  https://review.openstack.org/60220518:12
openstackgerritThiago da Silva proposed openstack/swift master: WIP: add multinode func test to swift  https://review.openstack.org/60168618:12
claygso the proxy.test_mem_server tests are a little janky - we import a half-dozen different TestCases and have at least a couple different ways of making the mem_server versions skip/pass when the test_server versions expect the behavior of the real DiskFile18:14
claygvery rarely the tests expect "go look on the disk" - most of the time it's more subtle like "opening this diskfile in this policy should raise DiskFileDoesNotExist with a timestamp == tombstone"18:15
claygand not only does mem_server not supprot storage policies - it doesn't really support tombstones!?  🤷‍♂️18:15
*** Sukhdev has quit IRC18:23
*** zaitcev has quit IRC18:30
*** e0ne has quit IRC18:31
*** fatema has quit IRC18:32
*** rcernin has joined #openstack-swift19:00
*** m_kazuhiro has joined #openstack-swift19:12
*** itlinux is now known as itlinux-away19:22
*** e0ne has joined #openstack-swift19:23
*** itlinux-away is now known as itlinux19:23
*** itlinux is now known as itlinux-away19:24
*** itlinux-away is now known as itlinux19:24
*** itlinux is now known as itlinux-away19:24
*** itlinux-away is now known as itlinux19:24
*** itlinux is now known as itlinux-away19:24
*** itlinux-away is now known as itlinux19:25
*** itlinux is now known as itlinux-away19:26
*** itlinux-away is now known as itlinux19:33
*** itlinux is now known as itlinux-away19:33
*** itlinux-away is now known as itlinux19:36
*** itlinux is now known as itlinux-away19:36
*** fatema has joined #openstack-swift19:38
*** itlinux-away is now known as itlinux19:49
*** itlinux is now known as itlinux-away19:50
*** itlinux-away is now known as itlinux19:50
*** sorrison has joined #openstack-swift19:55
openstackgerritKota Tsuyuzaki proposed openstack/swift master: Add another user for s3api func tests  https://review.openstack.org/60241119:55
*** zaitcev has joined #openstack-swift19:57
*** ChanServ sets mode: +v zaitcev19:57
openstackgerritKota Tsuyuzaki proposed openstack/swift master: Add another user for s3api func tests  https://review.openstack.org/60241120:00
kota_mem_server always seems to bite us when we want to change the backends.20:03
kota_change something of the backends.20:03
timburkemy impression was that it's *designed* to be that way, to make us think harder about changing the diskfile API and consider poor projects like swift-on-file20:09
zaitcevOr Kinetic20:15
zaitcevalso poor20:15
*** Sukhdev has joined #openstack-swift20:16
*** sorrison has quit IRC20:21
*** sorrison has joined #openstack-swift20:25
openstackgerritThiago da Silva proposed openstack/swift master: WIP: add multinode func test to swift  https://review.openstack.org/60168620:26
openstackgerritKota Tsuyuzaki proposed openstack/swift master: Add ceph-s3 test non-voting job  https://review.openstack.org/56470020:29
*** itlinux is now known as itlinux-away20:33
*** itlinux-away is now known as itlinux20:35
*** pcaruana has quit IRC20:38
*** sorrison has quit IRC20:39
*** sorrison has joined #openstack-swift20:39
*** e0ne has quit IRC20:49
*** e0ne has joined #openstack-swift20:50
*** rcernin has quit IRC20:51
kota_yey, https://review.openstack.org/#/c/602411/ passed with all green20:51
patchbotpatch 602411 - swift - Add another user for s3api func tests - 4 patch sets20:51
*** itlinux has quit IRC20:59
*** e0ne has quit IRC20:59
*** itlinux has joined #openstack-swift21:06
*** e0ne has joined #openstack-swift21:07
*** e0ne has quit IRC21:12
*** sorrison has quit IRC21:13
*** itlinux has quit IRC21:15
*** sorrison has joined #openstack-swift21:15
*** itlinux has joined #openstack-swift21:16
*** e0ne has joined #openstack-swift21:20
kota_http://logs.openstack.org/00/564700/34/check/swift-probetests-centos-7/c367f8b/job-output.txt.gz#_2018-09-13_20_57_09_65402021:22
kota_it looks like we still need to pin eventlet?21:22
*** mvenesio has quit IRC21:26
openstackgerritKazuhiro MIYAHARA proposed openstack/swift master: Add object-expirer new mode to execute tasks from general task queue  https://review.openstack.org/51738921:26
*** rcernin has joined #openstack-swift21:29
*** itlinux has quit IRC21:36
*** itlinux has joined #openstack-swift21:37
openstackgerritClay Gerrard proposed openstack/swift master: Refactor obj.server.ObjectController.PUT  https://review.openstack.org/60220521:45
openstackgerritClay Gerrard proposed openstack/swift master: Allow EC to ignore invalid request ETag  https://review.openstack.org/60246721:45
kota_http://logs.openstack.org/00/564700/34/check/swift-probetests-centos-7/37f9f3f/job-output.txt.gz BrokenPipe again :/21:47
*** itlinux has quit IRC21:50
*** fatema has quit IRC21:57
*** itlinux has joined #openstack-swift22:01
*** itlinux has quit IRC22:02
openstackgerritThiago da Silva proposed openstack/swift master: WIP: add multinode func test to swift  https://review.openstack.org/60168622:02
*** fatema_ has quit IRC22:09
*** zaitcev has quit IRC22:29
*** e0ne has quit IRC22:33
*** zaitcev has joined #openstack-swift22:37
*** ChanServ sets mode: +v zaitcev22:37
*** e0ne has joined #openstack-swift22:41
*** e0ne has quit IRC22:46
*** zaitcev has quit IRC22:51
*** fatema has joined #openstack-swift22:55
*** zaitcev has joined #openstack-swift22:57
*** ChanServ sets mode: +v zaitcev22:57
openstackgerritTim Burke proposed openstack/swift master: Add punch_hole utility function  https://review.openstack.org/54803423:04
*** sorrison has quit IRC23:10
*** kei_yama has joined #openstack-swift23:12
* kota_ is looking at https://review.openstack.org/54803423:16
patchbotpatch 548034 - swift - Add punch_hole utility function - 11 patch sets23:16
*** sorrison has joined #openstack-swift23:22
mattoliveraumorning23:25
StevenKmattoliverau: No it isn't, it's 5:30pm yesterday23:26
*** Sukhdev has quit IRC23:27
mattoliverauStevenK:  wow, is it.. I am sleep deprived :P23:28
StevenKTroll successful :-P23:28
openstackgerritThiago da Silva proposed openstack/swift master: WIP: add multinode func test to swift  https://review.openstack.org/60168623:31
kota_mattoliverau: o/23:34
notmynametdasilva: it's getting fixed23:38
tdasilvait's not because of the white space?23:39
tdasilvamattoliverau: o/23:40
notmynametdasilva: nope23:45
*** fatema has quit IRC23:51
openstackgerritJohn Dickinson proposed openstack/swift master: WIP: add multinode func test to swift  https://review.openstack.org/60168623:58

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