| *** haomaiwang has quit IRC | 00:01 | |
| *** haomaiwang has joined #openstack-swift | 00:01 | |
| *** kota_ has joined #openstack-swift | 00:17 | |
| *** ChanServ sets mode: +v kota_ | 00:17 | |
| kota_ | good morning | 00:29 |
|---|---|---|
| mattoliverau | kota_: morning | 00:30 |
| *** m_kazuhiro has joined #openstack-swift | 00:30 | |
| *** sanchitmalhotra has joined #openstack-swift | 00:35 | |
| *** haomaiwang has quit IRC | 00:36 | |
| *** sanchitmalhotra1 has quit IRC | 00:36 | |
| *** esker has quit IRC | 00:42 | |
| *** ho has joined #openstack-swift | 00:45 | |
| ho | good morning! | 00:49 |
| *** esker has joined #openstack-swift | 00:50 | |
| mattoliverau | ho: morning | 00:50 |
| kota_ | ho, mattoliverau: morning :) | 00:54 |
| *** parker7788 has left #openstack-swift | 00:54 | |
| *** kei_yama has quit IRC | 01:03 | |
| *** kei_yama has joined #openstack-swift | 01:04 | |
| *** baojg has joined #openstack-swift | 01:07 | |
| *** patchbot has quit IRC | 01:11 | |
| *** patchbot` has joined #openstack-swift | 01:11 | |
| *** ujjain has quit IRC | 01:12 | |
| *** patchbot` is now known as patchbot | 01:12 | |
| *** ujjain- has joined #openstack-swift | 01:12 | |
| *** esker has quit IRC | 01:32 | |
| *** esker has joined #openstack-swift | 01:33 | |
| *** haomaiwang has joined #openstack-swift | 01:35 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 01:39 | |
| *** sanchitmalhotra has quit IRC | 01:42 | |
| *** david-ly_ has joined #openstack-swift | 01:51 | |
| *** david-lyle has quit IRC | 01:54 | |
| *** esker has quit IRC | 01:57 | |
| *** haomaiwang has quit IRC | 02:01 | |
| *** sanchitmalhotra has joined #openstack-swift | 02:01 | |
| *** haomaiwang has joined #openstack-swift | 02:01 | |
| *** sanchitmalhotra1 has quit IRC | 02:03 | |
| *** esker has joined #openstack-swift | 02:06 | |
| ho | mattolverau, kota_: morning! | 02:09 |
| *** SkyRocknRoll has joined #openstack-swift | 02:16 | |
| mattoliverau | Looks like I'll be flying into Tokyo on the Monday... at least I don't have to worry about jetlag. :) | 02:21 |
| kota_ | mattoliverau: definitely. when are you planning to land to Japan? | 02:27 |
| kota_ | mattoliverau: strictly time...I mean. | 02:27 |
| *** esker has quit IRC | 02:31 | |
| mattoliverau | I land at 19:30 in Narita, so I suspect I wont get into Tokyo until quite late :( | 02:31 |
| kota_ | exactly, it's late time :( In addition, everyone have to spend about an hour to get into downtown... | 02:33 |
| kota_ | but that's good info, maybe I have to re-plan the date of something like party(?) I'd like to host. | 02:35 |
| kota_ | Manday might not be good :/ | 02:35 |
| kota_ | s/Manday/Monday/ | 02:35 |
| *** esker has joined #openstack-swift | 02:38 | |
| mattoliverau | kota_: well I'd love to come, but if you can only do Monday that's fine, I don't want to ruin anyone else's fun :) | 02:43 |
| kota_ | mattoliverau: not yet dicieded anything. I'm now surveying which date is suitable for anyone. | 02:44 |
| *** kei_yama has quit IRC | 02:47 | |
| mattoliverau | kota_: oh ok :) | 02:47 |
| *** kei_yama has joined #openstack-swift | 02:47 | |
| *** kei_yama has quit IRC | 02:48 | |
| *** kei_yama has joined #openstack-swift | 02:48 | |
| *** wbhuber has quit IRC | 02:49 | |
| mattoliverau | kota_: also my wife is flying up on the Saturday (after summit) and we'll spend 2 weeks holidaying in your lovely country :) | 02:52 |
| *** silor has joined #openstack-swift | 02:56 | |
| *** haomaiwang has quit IRC | 03:01 | |
| *** haomaiwang has joined #openstack-swift | 03:01 | |
| *** esker has quit IRC | 03:05 | |
| *** esker has joined #openstack-swift | 03:10 | |
| *** SkyRocknRoll has quit IRC | 03:25 | |
| *** esker has quit IRC | 03:26 | |
| *** baojg has quit IRC | 03:31 | |
| *** esker has joined #openstack-swift | 03:35 | |
| *** ppai has joined #openstack-swift | 03:42 | |
| ho | mattoliverau: +10000 | 03:42 |
| kota_ | mattoliverau: really nice!!! | 03:53 |
| *** esker has quit IRC | 03:54 | |
| *** jamielennox is now known as jamielennox|away | 03:56 | |
| *** haomaiwang has quit IRC | 04:01 | |
| *** esker has joined #openstack-swift | 04:01 | |
| *** haomaiwang has joined #openstack-swift | 04:01 | |
| *** baojg has joined #openstack-swift | 04:04 | |
| *** mahatic has joined #openstack-swift | 04:05 | |
| openstackgerrit | Alexandra Settle proposed openstack/swift: Moving DLO functionality doc to middleware https://review.openstack.org/219991 | 04:15 |
| *** nadeem has joined #openstack-swift | 04:21 | |
| openstackgerrit | Alexandra Settle proposed openstack/swift: Moving DLO functionality doc to middleware https://review.openstack.org/219991 | 04:22 |
| *** resker has joined #openstack-swift | 04:23 | |
| *** esker has quit IRC | 04:24 | |
| *** hrou has quit IRC | 04:35 | |
| *** nadeem has quit IRC | 04:35 | |
| *** resker has quit IRC | 04:44 | |
| *** esker has joined #openstack-swift | 04:48 | |
| *** SkyRocknRoll has joined #openstack-swift | 04:57 | |
| *** haomaiwang has quit IRC | 05:01 | |
| *** haomaiwang has joined #openstack-swift | 05:01 | |
| *** flwang1 has quit IRC | 05:07 | |
| *** mahatic has quit IRC | 05:13 | |
| *** esker has quit IRC | 05:16 | |
| *** eranrom has joined #openstack-swift | 05:20 | |
| *** esker has joined #openstack-swift | 05:22 | |
| *** nadeem has joined #openstack-swift | 05:33 | |
| *** mahatic has joined #openstack-swift | 05:40 | |
| *** nadeem has quit IRC | 05:41 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift: Don't send commits for quorum *BAD* requests on EC https://review.openstack.org/220059 | 05:44 |
| *** esker has quit IRC | 05:45 | |
| *** bkopilov has joined #openstack-swift | 05:49 | |
| *** esker has joined #openstack-swift | 05:51 | |
| *** csmart has quit IRC | 05:59 | |
| *** haomaiwang has quit IRC | 06:01 | |
| *** csmart has joined #openstack-swift | 06:01 | |
| *** haomaiwang has joined #openstack-swift | 06:01 | |
| *** pksingh has quit IRC | 06:05 | |
| *** portante_ has quit IRC | 06:09 | |
| *** ndk has quit IRC | 06:09 | |
| *** portante has joined #openstack-swift | 06:09 | |
| *** trifon has quit IRC | 06:10 | |
| *** ndk has joined #openstack-swift | 06:10 | |
| *** mattoliverau has quit IRC | 06:13 | |
| *** mattoliverau has joined #openstack-swift | 06:14 | |
| *** ChanServ sets mode: +v mattoliverau | 06:14 | |
| mahatic | mattoliverau: good evening! | 06:17 |
| mattoliverau | mahatic: o/ | 06:21 |
| *** trifon has joined #openstack-swift | 06:24 | |
| *** esker has quit IRC | 06:25 | |
| *** esker has joined #openstack-swift | 06:27 | |
| *** baojg has quit IRC | 06:36 | |
| *** mahatic has quit IRC | 06:38 | |
| *** mahatic has joined #openstack-swift | 06:39 | |
| openstackgerrit | Matthew Oliver proposed openstack/swift-specs: Update sharding spec, notably CountingTrie and shrinking https://review.openstack.org/201000 | 06:42 |
| *** esker has quit IRC | 06:45 | |
| *** esker has joined #openstack-swift | 06:50 | |
| *** haomaiwang has quit IRC | 07:01 | |
| *** haomaiwang has joined #openstack-swift | 07:01 | |
| *** rledisez has joined #openstack-swift | 07:05 | |
| *** fifieldt has joined #openstack-swift | 07:13 | |
| *** esker has quit IRC | 07:17 | |
| *** esker has joined #openstack-swift | 07:23 | |
| openstackgerrit | Mahati Chamarthy proposed openstack/swift: Probe tests for object encryption https://review.openstack.org/220897 | 07:24 |
| openstackgerrit | Mahati Chamarthy proposed openstack/swift: Modified tests to include real crypto https://review.openstack.org/211451 | 07:24 |
| openstackgerrit | Mahati Chamarthy proposed openstack/swift: Encrypt flag for probe tests https://review.openstack.org/216974 | 07:24 |
| openstackgerrit | Matthew Oliver proposed openstack/swift: Add container and account reverse listings https://review.openstack.org/120709 | 07:24 |
| openstackgerrit | Christian Schwede proposed openstack/swift: Add notification middleware https://review.openstack.org/196755 | 07:31 |
| *** joeljwright has joined #openstack-swift | 07:31 | |
| *** ChanServ sets mode: +v joeljwright | 07:31 | |
| *** trifon has quit IRC | 07:35 | |
| *** geaaru has joined #openstack-swift | 07:35 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 07:43 | |
| *** sanchitmalhotra has quit IRC | 07:45 | |
| *** trifon has joined #openstack-swift | 07:45 | |
| *** esker has quit IRC | 07:52 | |
| *** chlong has quit IRC | 07:56 | |
| *** esker has joined #openstack-swift | 07:59 | |
| *** haomaiwang has quit IRC | 08:01 | |
| *** haomaiwang has joined #openstack-swift | 08:01 | |
| *** baojg has joined #openstack-swift | 08:01 | |
| *** baojg has quit IRC | 08:18 | |
| *** acoles_ is now known as acoles | 08:20 | |
| acoles | good morning | 08:22 |
| *** esker has quit IRC | 08:24 | |
| mahatic | good morning | 08:26 |
| *** esker has joined #openstack-swift | 08:29 | |
| *** jistr has joined #openstack-swift | 08:30 | |
| *** itlinux has joined #openstack-swift | 08:40 | |
| *** itlinux has quit IRC | 08:40 | |
| *** itlinux has joined #openstack-swift | 08:41 | |
| kota_ | acoles, mahatic: good morning | 08:42 |
| mahatic | kota_: o/ | 08:45 |
| *** baojg has joined #openstack-swift | 08:46 | |
| itlinux | System Information: Model: MacBook Pro (13-inch, Retina, Early 2015) • CPU: Intel Core i7-5557U (4 Threads, 2 Cores) @ 3.10 GHz • Memory: 16.00 GB • Uptime: 11 days • Disk Space: 999.70 GB • Graphics: Intel Iris Graphics 6100 • Screen Resolution: 1680 x 1050 (HiDPI Mode) • OS: OS X Yosemite (Version 10.10.5, Build 14F27) | 08:47 |
| openstackgerrit | Christian Schwede proposed openstack/swift: Add containeralias middleware https://review.openstack.org/62494 | 08:49 |
| *** itlinux has quit IRC | 08:52 | |
| *** itlinux has joined #openstack-swift | 08:53 | |
| *** aix has quit IRC | 08:53 | |
| *** esker has quit IRC | 08:54 | |
| openstackgerrit | Mahati Chamarthy proposed openstack/swift: Probe tests for object encryption https://review.openstack.org/220897 | 08:56 |
| openstackgerrit | Mahati Chamarthy proposed openstack/swift: Encrypt flag for probe tests https://review.openstack.org/216974 | 08:56 |
| *** sanchitmalhotra has joined #openstack-swift | 09:00 | |
| *** haomaiwang has quit IRC | 09:01 | |
| *** haomaiwa_ has joined #openstack-swift | 09:01 | |
| *** jordanP has joined #openstack-swift | 09:01 | |
| *** sanchitmalhotra1 has quit IRC | 09:02 | |
| *** esker has joined #openstack-swift | 09:05 | |
| *** bsdkurt has quit IRC | 09:10 | |
| *** wuhg has quit IRC | 09:11 | |
| *** aix has joined #openstack-swift | 09:18 | |
| *** trifon has quit IRC | 09:19 | |
| *** trifon has joined #openstack-swift | 09:19 | |
| *** trifon has quit IRC | 09:22 | |
| *** trifon has joined #openstack-swift | 09:23 | |
| *** esker has quit IRC | 09:25 | |
| *** esker has joined #openstack-swift | 09:33 | |
| openstackgerrit | Alistair Coles proposed openstack/python-swiftclient: Initialise delete_object mock before it is called https://review.openstack.org/221219 | 09:39 |
| acoles | notmyname: ^^ an attempt at fixing that bug! its a stab in the dark but does it make any sense? | 09:40 |
| joeljwright | acoles: that bug looks like it might be a timing issue | 09:45 |
| joeljwright | those deletes are processed on a thread pool, so the test might well be happening before the calls have been completed in some cases | 09:46 |
| *** wer has quit IRC | 09:47 | |
| *** wer has joined #openstack-swift | 09:47 | |
| joeljwright | I I would have hoped they would be guaranteed to complete if the call to upload was completed though… | 09:48 |
| joeljwright | but looking at the test, there doesn't seem to be any consumption of the results of calling swiftclient.shell.main(argv) | 09:49 |
| joeljwright | ok, ignore me | 09:51 |
| joeljwright | the call to the shell should make sure it's done | 09:51 |
| joeljwright | back to the drawing board... | 09:51 |
| acoles | joeljwright: hi! yep i had same thought process but I can't see how the delete_segment threads would not complete before the assertion is made. | 09:53 |
| joeljwright | acoles: yeah, the call to upload certainly waits for the results of the deletes | 09:55 |
| joeljwright | very odd | 09:55 |
| *** baojg has quit IRC | 09:56 | |
| *** baojg has joined #openstack-swift | 09:56 | |
| acoles | joeljwright: so my guess is that maybe the dynamic creation of mock return_value for delete_object is racing on the two delete_segment threads, with only one result being set of course, and that only then getting on call counted. | 09:56 |
| * acoles waves hands | 09:56 | |
| joeljwright | acoles: sounds like a reasonable guess… | 09:57 |
| acoles | do we have any mocking experts here? | 09:57 |
| *** baojg has quit IRC | 10:00 | |
| *** haomaiwa_ has quit IRC | 10:01 | |
| *** haomaiwang has joined #openstack-swift | 10:01 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 10:05 | |
| *** esker has quit IRC | 10:05 | |
| *** sanchitmalhotra has quit IRC | 10:06 | |
| *** esker has joined #openstack-swift | 10:11 | |
| openstackgerrit | Jiri Suchomel proposed openstack/swift: Let object-info find files in a given directory https://review.openstack.org/189258 | 10:20 |
| *** silor has quit IRC | 10:20 | |
| *** trifon has quit IRC | 10:28 | |
| *** esker has quit IRC | 10:33 | |
| *** aix has quit IRC | 10:35 | |
| *** aix has joined #openstack-swift | 10:36 | |
| *** thurloat_isgone is now known as thurloat | 10:37 | |
| *** esker has joined #openstack-swift | 10:39 | |
| *** sanchitmalhotra has joined #openstack-swift | 10:58 | |
| *** thurloat is now known as thurloat_isgone | 10:59 | |
| *** thurloat_isgone is now known as thurloat | 11:00 | |
| *** sanchitmalhotra1 has quit IRC | 11:00 | |
| *** haomaiwang has quit IRC | 11:01 | |
| *** 92AAB1O7Y has joined #openstack-swift | 11:01 | |
| *** trifon has joined #openstack-swift | 11:07 | |
| *** esker has quit IRC | 11:10 | |
| *** 92AAB1O7Y has quit IRC | 11:12 | |
| *** silor has joined #openstack-swift | 11:16 | |
| *** silor has quit IRC | 11:18 | |
| *** SkyRocknRoll has quit IRC | 11:18 | |
| *** esker has joined #openstack-swift | 11:18 | |
| *** flwang1 has joined #openstack-swift | 11:18 | |
| *** SkyRocknRoll has joined #openstack-swift | 11:20 | |
| *** silor has joined #openstack-swift | 11:21 | |
| *** mahatic has quit IRC | 11:23 | |
| *** marcusvrn_ has joined #openstack-swift | 11:25 | |
| *** mahatic has joined #openstack-swift | 11:26 | |
| *** flwang1 has quit IRC | 11:26 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift: WIP: PUT/GET path for global ec cluster https://review.openstack.org/219165 | 11:34 |
| *** esker has quit IRC | 11:36 | |
| *** kota_ has quit IRC | 11:39 | |
| *** early has quit IRC | 11:41 | |
| *** esker has joined #openstack-swift | 11:45 | |
| *** flwang1 has joined #openstack-swift | 11:48 | |
| *** early has joined #openstack-swift | 11:49 | |
| *** SkyRocknRoll has quit IRC | 11:51 | |
| *** bkopilov has quit IRC | 12:00 | |
| openstackgerrit | Hisashi Osanai proposed openstack/swift: Add functional test for access control (PUT/DELETE) with Keystone https://review.openstack.org/202411 | 12:01 |
| *** esker has quit IRC | 12:07 | |
| *** eandersson has joined #openstack-swift | 12:11 | |
| *** esker has joined #openstack-swift | 12:17 | |
| *** m_kazuhiro has quit IRC | 12:19 | |
| *** chlong has joined #openstack-swift | 12:22 | |
| *** daemontool_ has quit IRC | 12:23 | |
| *** flwang1 has quit IRC | 12:26 | |
| *** flwang1 has joined #openstack-swift | 12:26 | |
| *** km has quit IRC | 12:27 | |
| *** mahatic has quit IRC | 12:29 | |
| *** esker has quit IRC | 12:38 | |
| *** kei_yama has quit IRC | 12:40 | |
| *** flwang1 has quit IRC | 12:42 | |
| *** daemontool_ has joined #openstack-swift | 12:42 | |
| *** esker has joined #openstack-swift | 12:44 | |
| *** mahatic has joined #openstack-swift | 12:46 | |
| *** ppai has quit IRC | 12:47 | |
| *** ho has quit IRC | 12:50 | |
| *** esker has quit IRC | 12:52 | |
| *** esker has joined #openstack-swift | 12:52 | |
| *** esker has quit IRC | 12:54 | |
| *** proteusguy__ has quit IRC | 12:55 | |
| *** esker has joined #openstack-swift | 12:56 | |
| *** esker has quit IRC | 12:56 | |
| *** clyps_ has joined #openstack-swift | 12:56 | |
| *** jordanP has quit IRC | 12:58 | |
| *** mahatic has quit IRC | 12:59 | |
| *** marzif has joined #openstack-swift | 13:09 | |
| *** tongli has joined #openstack-swift | 13:10 | |
| *** marzif has quit IRC | 13:10 | |
| *** haomaiwang has joined #openstack-swift | 13:10 | |
| *** changbl has quit IRC | 13:11 | |
| *** marzif has joined #openstack-swift | 13:11 | |
| *** proteusguy__ has joined #openstack-swift | 13:12 | |
| *** hrou has joined #openstack-swift | 13:16 | |
| *** dustins has joined #openstack-swift | 13:17 | |
| *** daemontool_ has quit IRC | 13:22 | |
| *** breitz has joined #openstack-swift | 13:23 | |
| *** marzif has quit IRC | 13:23 | |
| *** marzif_ has joined #openstack-swift | 13:23 | |
| *** marzif_ has quit IRC | 13:26 | |
| *** marzif_ has joined #openstack-swift | 13:26 | |
| *** marzif has joined #openstack-swift | 13:29 | |
| *** marzif_ has quit IRC | 13:29 | |
| *** itlinux has quit IRC | 13:46 | |
| *** ctrath has joined #openstack-swift | 13:51 | |
| *** nadeem has joined #openstack-swift | 13:56 | |
| *** haomaiwang has quit IRC | 14:01 | |
| *** haomaiwang has joined #openstack-swift | 14:01 | |
| *** jistr is now known as jistr|call | 14:01 | |
| *** ctrath has quit IRC | 14:02 | |
| *** yland has joined #openstack-swift | 14:02 | |
| *** marzif_ has joined #openstack-swift | 14:03 | |
| *** marzif has quit IRC | 14:03 | |
| openstackgerrit | Merged openstack/python-swiftclient: Initialise delete_object mock before it is called https://review.openstack.org/221219 | 14:06 |
| *** sanchitmalhotra has quit IRC | 14:06 | |
| *** ctrath has joined #openstack-swift | 14:07 | |
| *** nadeem has quit IRC | 14:08 | |
| *** nadeem has joined #openstack-swift | 14:08 | |
| *** marzif_ has quit IRC | 14:11 | |
| *** marzif has joined #openstack-swift | 14:11 | |
| yland | Hello my name is Yolande Amate, I am new to open-source and would like to contribute to openstack-swift. I have successfull installed openstack enabling swift on ubuntu14.04 through devstack. I would like to test some functionalities of openstack like uploading an ubuntu vm image with swift and test access to the os. When I try swift upload I am requested to give " ST_AUTH, ST_USER, and ST_KEY" , so please I would like to know how create a swift accou | 14:15 |
| yland | nt, or what are the default for newly installed swift. | 14:15 |
| *** wer_ has joined #openstack-swift | 14:15 | |
| *** wer has quit IRC | 14:16 | |
| yland | I also disabled nova and glance while installing openstack, since I got an error while they were both enabled alongside swift. | 14:17 |
| *** jistr|call is now known as jistr | 14:20 | |
| *** wbhuber has joined #openstack-swift | 14:21 | |
| *** prometheanfire has joined #openstack-swift | 14:23 | |
| prometheanfire | does 2.4.0 support any py3 variants? | 14:23 |
| prometheanfire | 3.3 or 3.4 for instance? | 14:23 |
| prometheanfire | think all the deps support it | 14:23 |
| *** dustins has quit IRC | 14:25 | |
| *** bkopilov has joined #openstack-swift | 14:27 | |
| *** jrichli has joined #openstack-swift | 14:32 | |
| prometheanfire | gonna assume no for now, tox doesn't show it | 14:41 |
| *** jlhinson has joined #openstack-swift | 14:42 | |
| openstackgerrit | Doug Hellmann proposed openstack/python-swiftclient: Add links for release notes tool https://review.openstack.org/221329 | 14:42 |
| acoles | yland: if you are using devstack then you will be using keystone identity service so you will need to use the OS options with swift CLI | 14:43 |
| acoles | yland: type swift --os-help | 14:43 |
| MooingLemur | prometheanfire: pretty sure py3 support is only "in progress" | 14:43 |
| acoles | prometheanfire: MooingLemur is correct | 14:43 |
| prometheanfire | MooingLemur: ya, that was my last update on the subject | 14:43 |
| cschwede | yland: Hello! in your devstack VM, cd to the devstack directory; there should be a file called openrc. just do a „source openrc“ and you should be set | 14:43 |
| acoles | yland: heh - cschwede has better answers than me :) | 14:44 |
| jrichli | mahatic, acoles: great work on feature/crypto! It was nice to see so much activity while I have been out :-) | 14:44 |
| acoles | jrichli: welcome back | 14:45 |
| hrou | jrichli, acoles +1 ! I'm reviewing the changes as we speak ! Looks great so far. | 14:45 |
| jrichli | acoles: I feel really great. I didn't work all weekend - but maybe 2 hours | 14:45 |
| acoles | hrou: hold fire on keymaster review - i have another change i'd like to push to gerrit | 14:46 |
| *** wer_ has quit IRC | 14:46 | |
| hrou | acoles, sure thing ! That sounds great, I'm just reviewing the structure which I imagine won't change too much, will hold off on leaving comments though. | 14:47 |
| acoles | jrichli: hrou : i'd like to make the keymaster not have to persist key id's for container and account, but rather always calculate them. | 14:47 |
| jrichli | acoles: I noticed that. so that means keeping the key_id at a header, right? | 14:47 |
| jrichli | I had only added that for a temporary workaround, but I am fine with keeping it there | 14:48 |
| acoles | jrichli: no. i didnt push that change yet. at least i dont think i did | 14:48 |
| jrichli | I think that is what your comments led me to | 14:48 |
| *** wer_ has joined #openstack-swift | 14:48 | |
| acoles | jrichli: i'll push the change - that will be clearer | 14:49 |
| hrou | acoles, jrichli - I recall at the hackathon jrichli had a concern from a debugging point of view in terms of just dynamically calculating them ? | 14:49 |
| jrichli | hrou - in acoles comments, he had mentioned that the key_id was still saved, so you could go from that. | 14:49 |
| jrichli | hrou - but maybe acoles came up with something even better. we shall see | 14:49 |
| acoles | jrichli: yeah on the object. here i am just talking about the container and account keys | 14:50 |
| acoles | and no i dont have anything better! | 14:50 |
| *** baojg has joined #openstack-swift | 14:50 | |
| acoles | I just spent a lot of time figuring out why I had 144 func tests errors before realising that resetswift is needed if the proxy is restarted, because otherwise it is impossible to list existing containers | 14:52 |
| *** david-ly_ is now known as david-lyle | 14:53 | |
| jrichli | acoles: I run resetswift all the time. maybe I shouldn't do it so much | 14:53 |
| acoles | jrichli: we avoid it in production :D :D | 14:54 |
| hrou | After sticking crypto in the pipeline ? yea : ) | 14:54 |
| *** ctrath has quit IRC | 14:54 | |
| *** dustins has joined #openstack-swift | 14:54 | |
| hrou | I've gotten to used to it that on our semi production systems (used for testing larger clusters) I've changed resetswift to just be a stop / start swift, to avoid unintentional pain : ) | 14:55 |
| jrichli | acoles: I think the newline issue you mentioned in email is the same issue I have been debugging. the new line is probably there because the etag was not properly decrypted: the crypto-meta it gets is old/incorrect sometimes | 14:55 |
| *** ctrath has joined #openstack-swift | 14:55 | |
| acoles | jrichli: yeah i'd guess its same cause. my only contribution may be to explain the intermittency ;) | 14:56 |
| jrichli | acoles: if you change to always use the same iv for all etags, issues go away | 14:57 |
| *** ctrath has quit IRC | 14:57 | |
| *** SkyRocknRoll has joined #openstack-swift | 14:57 | |
| jrichli | the intermittency is that the etag crypto-meta only sometimes gets mixed on the wrong response | 14:57 |
| *** dustins has quit IRC | 14:57 | |
| jrichli | so, here is what can happen: | 14:57 |
| jrichli | fake_footers sets the correct etag in the response headers it passes along | 14:58 |
| acoles | jrichli: hrou so, i'd really like to get some stuff merged because the patch chain is just impeding progress and rebase mistakes are just causing unnecessary churn on gerrit | 14:58 |
| *** minwoob has joined #openstack-swift | 14:58 | |
| jrichli | then the decrypter gets an old etag in its response headers. what the heck happened in between? i donno | 14:58 |
| jrichli | acoles: agreed. let me know what I can do to help with that. | 14:59 |
| acoles | jrichli: we have a -1 comment on the crypto module patch due to not checking size of key | 14:59 |
| hrou | acoles, yea that sounds like a good plan to me, we can always propose additional changes, so maybe don't worry too much about the reviews and just push them in ? | 15:00 |
| yland | cschwede, acoles thanks for the help. Now I get "Endpoint for object-store not found - have you specified a region?" with swift list and swift upload commands | 15:00 |
| acoles | jrichli: I was intending to fix that next | 15:00 |
| jrichli | acoles: oh, I thought you said mahatic fixed that, but on a different patch. is that right? | 15:00 |
| *** fifieldt has quit IRC | 15:00 | |
| *** ctrath has joined #openstack-swift | 15:00 | |
| *** haomaiwang has quit IRC | 15:01 | |
| *** haomaiwang has joined #openstack-swift | 15:01 | |
| *** haomaiwang has quit IRC | 15:01 | |
| acoles | jrichli: err, did i say that ? not aware of it. | 15:01 |
| hrou | jrichli, I was going through all the changes this morning, I didn't see that, unless it wasn't on the crypto feature branch : / | 15:02 |
| hrou | https://review.openstack.org/#/q/status:open+project:openstack/swift+branch:feature/crypto,n,z | 15:02 |
| jrichli | acoles: I must have interpreted something incorrectly | 15:02 |
| *** dustins has joined #openstack-swift | 15:02 | |
| jrichli | hrou: the -1 was cleared, but no changes were made for it | 15:03 |
| jrichli | because of rebasing | 15:03 |
| hrou | jrichli, ah gotcha, the rebase I guess | 15:03 |
| hrou | Right | 15:03 |
| hrou | jrichli, btw re: the etag - well the solution we're proposing for conditional get may just fix that issue for you ; ) Given the fixed iv .. : ) | 15:04 |
| *** kevinc_ has joined #openstack-swift | 15:05 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: Trivial Key Master for encryption https://review.openstack.org/193749 | 15:05 |
| acoles | hrou: jrichli ^^ | 15:05 |
| acoles | can you try that in your patch chains | 15:06 |
| cschwede | yland: hmm, strange - that should be set too. anyways, try with: export OS_REGION_NAME="RegionOne“ | 15:06 |
| hrou | acoles, perfect that's great I'm going to take a look at that now. | 15:06 |
| jrichli | hrou: true, but I want to find out how the response headers can get mixed like that. it reveals a problem that needs to be fixed | 15:06 |
| jrichli | hrou: of course, it could be a problem that would go away when fake_footers go away - but i haven't proved that yet | 15:07 |
| acoles | jrichli: how many patches need to be applied to get that error? does it need the container listing patch? does it only happen with the COPY middleware? | 15:07 |
| hrou | jrichli, yea I think that's prudent i.e. at least to identify the route case, who knows it may lead to other things as well. | 15:07 |
| acoles | hrou: thanks. its getting smaller each time :) | 15:08 |
| hrou | acoles, which makes reviewing that much easier : ) | 15:08 |
| *** baojg has quit IRC | 15:09 | |
| jrichli | acoles: I haven't tried outside of COPY yet. I haven't looked yet, but i will most likely need to add a similar test case to use - COPY is making a lot more requests happen at once, which I think is what brings it out | 15:09 |
| *** janonymous_ has joined #openstack-swift | 15:09 | |
| hrou | acoles, really quick silly question .. for trivial not important (get to it later) changes I think you mentioned you use "NIT ... followed by comment" ? | 15:10 |
| acoles | jrichli: yep. btw I needed to resolve a few conflicts to apply the copy middleware patch - same for you? | 15:10 |
| *** trifon has quit IRC | 15:11 | |
| jrichli | acoles: that was a few weeks ago :-) what type of conflicts? I might have made some obvious/explainable resolutions, but I don't remember anything significant. | 15:11 |
| acoles | hrou: yes use "nit: " for trivial stuff like a comment typo, or just fix it and push a new version to gerrit. | 15:12 |
| acoles | jrichli: there was a method in obj.py i think that got deleted and there was a change in that method on the crypto branch | 15:12 |
| hrou | acoles, thanks! | 15:12 |
| jrichli | acoles: i missed that. dont suppose you remember the name of the method? i can see if its in my repo | 15:13 |
| acoles | think it was obj.py handle_copy_request | 15:14 |
| *** changbl has joined #openstack-swift | 15:14 | |
| acoles | also, i rebased the copy middleware patch on master branch, so grab the latest from patch 156923 | 15:15 |
| patchbot | acoles: https://review.openstack.org/#/c/156923/ | 15:15 |
| *** mahatic has joined #openstack-swift | 15:16 | |
| jrichli | acoles: yes, I see that the method is no longer there. there were changes made which were only on feature/crypto? What was the nature of the change? | 15:19 |
| hrou | jrichli, acoles I pulled it down a while ago too but it was fairly trivial for me, that or I just forgot. | 15:19 |
| *** wbhuber has quit IRC | 15:19 | |
| janonymous_ | Do someone have trouble when services are run through openstack command , but when checking through swift-init command nothing is shown, instead of services running | 15:22 |
| jrichli | acoles: btw, the total errors/failures I have with my COPY+crypto is (SKIP=17, errors=5, failures=10) | 15:23 |
| jrichli | functests | 15:23 |
| *** wbhuber has joined #openstack-swift | 15:23 | |
| hrou | Hey janonymous_ what are "openstack" command | 15:25 |
| hrou | janonymous_, do you mean the openstack CLI for swift, I'm not having any issues with that on my system (i.e. using the openstack cli on a running cluster and swift-init showing the processes) | 15:26 |
| acoles | jrichli: https://review.openstack.org/#/c/203310/11/swift/proxy/controllers/obj.py | 15:26 |
| janonymous_ | hrou: Hey, yes. | 15:26 |
| janonymous_ | hrou: Is it working fine everytime. i mean starting services with openstack commands first .. because when i was doing that pid files were not created but still processes were running | 15:28 |
| jrichli | acoles: that diff only shows the part where we remove the "remove_items" call. I see that the copy_headers_into method has changed since feature/crypto so that it copies the sysmeta as well. | 15:34 |
| hrou | janonymous_, so in my case I still started the swift process (well the proxy on a given host) through swift-init and I'm using the openstack swift cli to upload/download, so not sure if that matches what you're describing | 15:34 |
| janonymous_ | hrou: I'll try to check running it several times using fresh installation i guess for better results. | 15:34 |
| *** SkyRocknRoll has quit IRC | 15:34 | |
| yland | cschwede: I tried the export command and I also ran "env | grep OS_" which gave me some info including a line containing "OS_REGION_NAME=RegionOne" but I still get the same error even after running ./stack.sh. | 15:35 |
| janonymous_ | hrou: Ohh.. not that case , slight different one. | 15:35 |
| jrichli | acoles: the way that I added crypto into the COPY as middleware is I added the items that I know we have added by hand. maybe I should have made use of git merge, but it sounds like this resulted in fewer errors/failures | 15:36 |
| cschwede | yland: how does your devstack config look like? can you share it using http://paste.openstack.org/ ? | 15:37 |
| hrou | janonymous_, did you actually start swift through the openstack cli ? I didn't think that was possible, only listing services, let me know if it is I can try it : ) | 15:38 |
| mahatic | acoles, jrichli hello, I just through the long scroll back :) I didn fix up the Crypto key length, because was not sure if that'd be the only review | 15:39 |
| mahatic | went through* | 15:39 |
| acoles | mahatic: ok np i'm on it | 15:40 |
| yland | cschwede: do you mean my local.conf file? | 15:40 |
| janonymous_ | hrou: I prepared a fresh vm and saw status of services through openstack cli it showed running, but when i saw it using swift-init no process were running | 15:40 |
| cschwede | yland: yep | 15:40 |
| mahatic | acoles: thanks | 15:41 |
| janonymous_ | hrou: I'll try again with new setup i think | 15:41 |
| *** joeljwright has quit IRC | 15:41 | |
| janonymous_ | hrou: maybe i did something stupid.. | 15:41 |
| hrou | janonymous_, how are you listing the status ? 'openstack service show swift' ? | 15:42 |
| acoles | jrichli: k. it was just a thought re the crypto meta mixup, that maybe something went awry, but maybe not. | 15:42 |
| janonymous_ | openstack service show | 15:42 |
| *** devlaps has joined #openstack-swift | 15:43 | |
| hrou | janonymous_, yep, so the "enabled" field ? Interesting I didn't think that implied the service was up and running, not sure what it really implied. | 15:43 |
| janonymous_ | hrou: yes | 15:44 |
| hrou | janonymous_, right if I stop swift it still lists it as enabled, I don't think it implies the service is running let me check one thing. | 15:44 |
| yland | cschwede: local.conf (http://paste.openstack.org/show/450422/) and stackrc(http://paste.openstack.org/show/450423/). This are the only file I modified. | 15:46 |
| *** barker has joined #openstack-swift | 15:53 | |
| *** barker has quit IRC | 15:57 | |
| cschwede | yland: you don’t need to modify stackrc to disable other services. have a look at http://docs.openstack.org/developer/devstack/configuration.html#swift | 16:02 |
| cschwede | yland: also, a minimal local.conf for swift with keystone only might look like this: https://gist.github.com/cschwede/a54173fdfc745865233b | 16:02 |
| *** gyee has joined #openstack-swift | 16:03 | |
| hrou | acoles, jrichli, mahatic - I guess at some point we'll end up changing the trivial keymaster to not touch account / container if we decide on the reduced scope (obj/obj meta only)? | 16:03 |
| *** annegentle has joined #openstack-swift | 16:05 | |
| acoles | hrou: right. container key will be needed because we use that to encrypt the object metadata in container listing. | 16:05 |
| acoles | account key would be redundant i think | 16:05 |
| mahatic | yup, container key will be needed | 16:05 |
| hrou | acoles, mahatic - good point re: container keys ! | 16:06 |
| mahatic | acoles: but we probably would still keep that in trivial keymaster, since we need it for future anyway? | 16:06 |
| mahatic | account key I mean | 16:06 |
| hrou | mahatic, maybe ! It depends, actually, assuming we go with a reduced scope, maybe we need to ignore any keys provided in the encryption / decryption middleware, as on paper we're not in control of the keymaster (and a simple implementation may just always provide a key regardless of the path). | 16:09 |
| acoles | hrou: should we drop it now? make patch smaller :P seriously, if its not used then should we have it there in anticipation of future or leave it to future-us to add back in? | 16:11 |
| acoles | jrichli: ^^ | 16:11 |
| *** jistr has quit IRC | 16:11 | |
| *** mahatic has quit IRC | 16:13 | |
| hrou | acoles, jrichli actually I believe the latest and greatest patch sets have what we want on the engine side (patch 214438) so we'll ignore anything account lvl anyway at this point. | 16:16 |
| patchbot | hrou: https://review.openstack.org/#/c/214438/ | 16:16 |
| *** mahatic has joined #openstack-swift | 16:16 | |
| *** esker has joined #openstack-swift | 16:16 | |
| hrou | So even if the trivial keymaster provides a key, just a question of if we want to simplify it even further. | 16:16 |
| mahatic | hrou: thatś true | 16:17 |
| notmyname | good morning | 16:20 |
| jrichli | acoles hrou mahatic: IMO, we could take out the account key in the trivial_keymaster, and it can just be added when / if it is going to be used in the future. | 16:20 |
| yland | cschwede: Thank very much, all seems to work fine now. | 16:20 |
| acoles | jrichli: +1 | 16:21 |
| hrou | janonymous_, I'm going to ask some folks I know on the openstack CLI team what the status field should represent (as I see some talk for other components that it may indeed represent if the service is up and running), or if I'm adventurous I'll look at the code later today : ) | 16:21 |
| hrou | janonymous_, anyway I played around with it and its always listed as enabled for me, regardless of the status of the swift processes | 16:22 |
| acoles | jrichli: method md5hex(s) in test_crypto.py in patch 193826 appears to be unused? does it get used further up the chain? | 16:22 |
| patchbot | acoles: https://review.openstack.org/#/c/193826/ | 16:22 |
| *** mahatic has quit IRC | 16:22 | |
| janonymous_ | hrou : yes, that was my concern :) | 16:27 |
| janonymous_ | hrou: thanks for looking on the same :) | 16:27 |
| hrou | janonymous_, anytime ! I like using the openstack CLI, glad others do as well (maybe like was a strong word ; ) | 16:27 |
| janonymous_ | hrou: I'll troubleshoot also for it :) | 16:28 |
| janonymous_ | also | 16:28 |
| *** jlhinson has quit IRC | 16:29 | |
| *** jlhinson has joined #openstack-swift | 16:29 | |
| yland | Is it possible to upload an operating system vm image with swift and test/run the os with horizon? | 16:29 |
| jrichli | acoles: you are correct. blow that sucker away! :-) | 16:30 |
| janonymous_ | hrou: Thanks. | 16:30 |
| acoles | jrichli: its toast | 16:30 |
| jrichli | thx! | 16:30 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-swiftclient: Updated from global requirements https://review.openstack.org/89250 | 16:31 |
| acoles | notmyname: how was the museum? | 16:33 |
| notmyname | acoles: we had a good time. | 16:35 |
| acoles | jrichli: how come all your first commits are on friday and saturdays? what's going on rest of the week? :P | 16:35 |
| hrou | acoles, lol! | 16:36 |
| acoles | notmyname: great. that changelog patch took some persuasion! | 16:36 |
| jrichli | acoles: I know, right? I have been a little better about being able to work on 'my stuff' at work lately | 16:36 |
| notmyname | yeah, it did. that's for beating it a few times :-) | 16:36 |
| jrichli | acoles notmyname: I was wondering if it would be ok to add the crypto middleware to the pipeline for just the feature/crypto gate tests | 16:38 |
| notmyname | jrichli: yeah, that makes sense | 16:39 |
| acoles | notmyname: np. be interesting to see if patch 221219 fixes it. | 16:39 |
| patchbot | acoles: https://review.openstack.org/#/c/221219/ | 16:39 |
| jrichli | do I have the necessary permissions? | 16:39 |
| notmyname | jrichli: what do you mean? | 16:40 |
| acoles | jrichli: not sure you need any | 16:40 |
| acoles | if its in the sample file it will be used, i think | 16:40 |
| jrichli | I am not sure what is involved. Just figured that there might be some control over who can change gate tests | 16:40 |
| openstackgerrit | John Dickinson proposed openstack/python-swiftclient: Add links for release notes tool https://review.openstack.org/221329 | 16:40 |
| jrichli | acoles: oh, I see. | 16:41 |
| acoles | jrichli: not sure if its proxy-server.conf-sample or the one under doc/saio that gets used | 16:41 |
| openstackgerrit | Alistair Coles proposed openstack/swift: Cryptography module to be used by middleware https://review.openstack.org/193826 | 16:42 |
| acoles | hrou: mahatic: jrichli: ^^ i made a few fixups and added check on key size. | 16:43 |
| *** rledisez has quit IRC | 16:43 | |
| acoles | jrichli: could you ask Christian Cachin to review that one more time? | 16:43 |
| jrichli | acoles: sure, thanks | 16:44 |
| hrou | acoles, looks great will review, only thing that quickly comes to mind is if we should check the iv length while we're at it. (though we do generate this locally) | 16:46 |
| acoles | hrou: the Cipher class will raise an exception if iv is wrong length - there is a test in test_crypto.py | 16:47 |
| hrou | acoles, ah perfect | 16:48 |
| acoles | jrichli: devstack uses whatever pipeline is in /etc/proxy-server.conf-sample (then patches it for correct auth middleware) so if you push a patch with crypto middlewares in the pipeline that should do the trick. obviously it will need the conf sections for those entry points and the lines in setup.cfg so will have to be based on other patches. | 16:51 |
| *** trifon has joined #openstack-swift | 16:55 | |
| jrichli | acoles: sounds good, thx! | 16:58 |
| *** aix has quit IRC | 16:58 | |
| *** barker has joined #openstack-swift | 17:01 | |
| openstackgerrit | John Dickinson proposed openstack/swift: Merge branch 'master' into feature/hummingbird https://review.openstack.org/221410 | 17:02 |
| notmyname | dfg: ^ | 17:02 |
| *** trifon has quit IRC | 17:03 | |
| *** trifon has joined #openstack-swift | 17:04 | |
| dfg | notmyname: cool thx. can i just +2 / approve it? | 17:05 |
| notmyname | dfg: yup | 17:05 |
| dfg | notmyname: thanks a bunch | 17:06 |
| notmyname | dfg: any time :-) | 17:06 |
| notmyname | dfg: probably would be a good idea to try to do that more regularly between you me and redbo. once every week or two | 17:06 |
| *** eandersson has quit IRC | 17:07 | |
| dfg | notmyname: ya. i think its been almost 2 months or something | 17:08 |
| *** AndreiaKumpera has joined #openstack-swift | 17:12 | |
| cschwede | yland: great that it works now - you’re welcome! | 17:12 |
| *** zhill has joined #openstack-swift | 17:15 | |
| *** barker has quit IRC | 17:17 | |
| *** barker has joined #openstack-swift | 17:31 | |
| clayg | morning | 17:31 |
| *** acoles is now known as acoles_ | 17:34 | |
| *** thurloat is now known as thurloat_isgone | 17:42 | |
| *** mvandijk has left #openstack-swift | 17:46 | |
| *** aix has joined #openstack-swift | 18:01 | |
| hrou | jrichli, this version of the keymaster still has the logic for versioned gets, in theory this will go away once we validate the copy middleware and versioned objs? | 18:07 |
| jrichli | hrou: yes. but I figure we could keep it in for now so that any new errors in those related tests would be revealed - and not looked over b/c of an existing failure. But I am flexible on that. | 18:18 |
| *** bil has joined #openstack-swift | 18:19 | |
| bil | Hi | 18:20 |
| bil | Can someone help me with a question? | 18:20 |
| hrou | jrichli, yea keep it for now but eventually remove I think ? I find it complicates the get path as we have use case where we don't find the key on a path but will not look up if the response has the magical id key path stored. | 18:21 |
| jrichli | hrou: I haven't yet been able to look at the latest patch from acoles | 18:22 |
| hrou | jrichli, oh no worries for now ! Basically all that logic hasn't changed its still there, we'll revisit it later. | 18:23 |
| jrichli | bil: everyone on will do their best to help. What is the question? | 18:23 |
| *** geaaru has quit IRC | 18:26 | |
| openstackgerrit | David Goetz proposed openstack/swift: use recon diskusage so you can weed out unmounted drives https://review.openstack.org/221446 | 18:27 |
| bil | I need to retrieve X-Storage-Url and X-Auth-Token | 18:27 |
| *** T0m_ has joined #openstack-swift | 18:30 | |
| *** openstackgerrit has quit IRC | 18:31 | |
| *** openstackgerrit has joined #openstack-swift | 18:31 | |
| jrichli | bil: Do you mean as a client, or in some piece of middleware, or something else? | 18:31 |
| bil | several ways one as a client | 18:32 |
| T0m_ | Is there a way to check the cluster status in Swift? | 18:32 |
| bil | using curl | 18:32 |
| bil | and othe using swiftclient | 18:32 |
| bil | I tried curl -i -H "Content-Type: application/json" -d ' | 18:33 |
| bil | curl -i -H "Content-Type: application/json" -d ' | 18:33 |
| bil | { | 18:33 |
| openstackgerrit | David Goetz proposed openstack/swift: Better error output in TestHealthCheck https://review.openstack.org/212188 | 18:33 |
| wbhuber | bil: does "swift stat" work for you? | 18:34 |
| hrou | T0m_, you can use swift-init status <server type> which will list the processes PIDs | 18:34 |
| T0m_ | would that need to be ran on each server though? I wondered if from one mode you can do a sort of “check status” and it would report back the health of each node | 18:35 |
| hrou | T0m_, for that you need to run it on each, its pretty easy to script something up to accomplish what you want, but others may be aware of something that already exists. | 18:36 |
| jrichli | T0m_ wbhuber: I think that "swift stat" might be what T0m_ is looking for | 18:36 |
| T0m_ | Ok thats fine - new to swift and just setup 2 nodes in a cluster and just wondered if there was a way to check the health | 18:37 |
| notmyname | T0m_: yeah, what you're looking for is swift-recon | 18:37 |
| bil | yes | 18:37 |
| T0m_ | Ah ok, think I did come across that but i’ll have to look at configuring it - thanks | 18:37 |
| notmyname | `swift stat` gives info about your particular account. not cluster healp info | 18:37 |
| hrou | T0m_, jrichli wbhuber - yea swift stat : ) could work too but it really depends what you're looking for, as that'd be for the account your credentials point too. | 18:38 |
| *** silor has quit IRC | 18:38 | |
| notmyname | T0m_: there's a CLI tool you can run from any machine that (1) has connectivity to everything in teh cluster and (2) the ring file(s) | 18:38 |
| notmyname | T0m_: the only "configuring" you need to do is to make sure you have recon in the pipeline in the configs. which will be the case if you used upstream docs for configuring it | 18:39 |
| T0m_ | i wondered if you could see if nodes were in sync, and maybe the count of quarantined objects, or any reported disk failures | 18:39 |
| notmyname | yup | 18:39 |
| wbhuber | bil: great, then do a "swift --verbose stat" | 18:39 |
| notmyname | that's what it's for | 18:39 |
| wbhuber | bil: you'd see "StorageUrl" and "Auth Token" which are what you wanted in the first place. | 18:40 |
| * notmyname needs to go to lunch now | 18:40 | |
| torgomatic | dfg: ping | 18:40 |
| wbhuber | notmyname: you cannot. | 18:40 |
| wbhuber | notmyname: jk | 18:40 |
| notmyname | wbhuber: heh. tell that to my wife who is waiting downstairs for me ;-) | 18:41 |
| wbhuber | notmyname: she wins. ;-) | 18:41 |
| bil | here is what I sent http://pastebin.com/u8w365EJ | 18:45 |
| bil | i changed my real ip | 18:45 |
| wbhuber | bil: you can put that in a file called "openrc" or "adminrc" similar to this: http://docs.openstack.org/cli-reference/content/cli_openrc.html | 18:46 |
| wbhuber | bil: then do a "source openrc" | 18:46 |
| *** jlhinson has quit IRC | 18:46 | |
| wbhuber | bil: then run "swift --verbose stat" directly | 18:46 |
| bil | ok I will try | 18:47 |
| dfg | torgomatic: whats up? | 18:47 |
| bil | thanks | 18:47 |
| wbhuber | bil: http://pastebin.com/JwT9arcz | 18:48 |
| wbhuber | bil: THat should be your openrc | 18:48 |
| wbhuber | bil: change the IP address as well | 18:48 |
| wbhuber | bil: do a "source openrc" | 18:48 |
| bil | ok thanks a lot | 18:49 |
| bil | i will try it now | 18:49 |
| *** T0m_ has left #openstack-swift | 18:49 | |
| *** garthb has joined #openstack-swift | 18:49 | |
| torgomatic | dfg: I was thinking about redoing the ratelimiting on slo since we're doing this range thing now and I wanted to run my ideas past you for a sanity check | 18:52 |
| bil | wbhuber | 18:53 |
| torgomatic | since clients will be able to put a range in a slo manifest, the 1 MiB segment size limit stops helping, because they can just request 1 byte from the 1 MiB segment | 18:54 |
| bil | wbhuber:In the end I need to do this in python so first I want to try it with curl | 18:54 |
| bil | and see what response I get | 18:54 |
| torgomatic | and requiring ranges to be at least 1 MiB isn't that great for the folks who want the range stuff | 18:54 |
| torgomatic | so I wanted to (a) limit slo segment size to 1 byte or more (so no 0-byte segments because that's dumb) and then ratelimit only small segments/ranges, where "small" is "less than 1 MiB" | 18:55 |
| torgomatic | there should probably be a "(b)" in there somewhere, because that's two things, but you see what I mean | 18:56 |
| torgomatic | basically if the client wants big chunks, feed them as fast as they can eat it | 18:57 |
| bil | wbhuber:or maybe via swift command | 18:57 |
| clayg | torgomatic: dfg is all about ratelimiting | 18:57 |
| dfg | thats true. | 18:57 |
| torgomatic | but if the client wants small chunks, feed them slowly | 18:57 |
| clayg | torgomatic: yeah %$#@ 'em | 18:57 |
| dfg | torgomatic: there's still the 1000 max size right? | 18:57 |
| torgomatic | dfg: yeah, I wasn't going to change that | 18:58 |
| torgomatic | there's still SLO-of-SLOs for people who need lots of segments | 18:58 |
| dfg | ya | 18:58 |
| dfg | everybody's favorite feature | 18:59 |
| clayg | ummm... i am totally lost on the rfc822 explosion of patches | 19:00 |
| torgomatic | heh, I love that one... people are like "what if I need more than 1000 segments?" and I'm like "let me tell you about exponents" | 19:01 |
| clayg | dfg: oh thanks for the push on patch 212188 !!! | 19:01 |
| patchbot | clayg: https://review.openstack.org/#/c/212188/ | 19:01 |
| clayg | does it still need a go fmt? | 19:01 |
| dfg | torgomatic: ya thats sounds fine. the only thing i could think of is the PUTs would be expensive with all the HEADs to the 1 byte files. but thats not really a concern i guess. | 19:01 |
| torgomatic | dfg: cool, thanks. I'll go work on that then. | 19:02 |
| dfg | so would they get like the first 10 chunks with no limiting and then start after that? | 19:03 |
| dfg | thats how it does it right? | 19:03 |
| torgomatic | dfg: first 10 small chunks, yeah... so if they've got a bunch of 1+ MiB chunks, those don't count, but then they get 10 small ones free and then ratelimit to 1 small per second | 19:03 |
| dfg | ya cool. | 19:03 |
| dfg | clayg: np. i should have fixed tat gofmt problem. unless i screwed up the commit | 19:04 |
| minwoob | clayg: I see that https://bugs.launchpad.net/swift/+bug/1491908 is necessary in order for container sync to work, but shouldn't try to make it so the proper response is emitted on the 409 in the normal case, and have the container sync case handle the 409s in its own way? | 19:12 |
| openstack | Launchpad bug 1491908 in OpenStack Object Storage (swift) "proxy server reports 202 even though EC backend didn't accept object" [Undecided,New] | 19:12 |
| minwoob | shouldn't* we | 19:12 |
| minwoob | This seems to be a correctness issue if the proxy keeps returning 202 for the 409s. | 19:13 |
| minwoob | Basically, the proposed idea is to have the proxy server return the correct response upon 409, but then in container sync to implement a logic that can flexibly handle the 409s. | 19:15 |
| *** jlhinson has joined #openstack-swift | 19:16 | |
| *** jkugel has joined #openstack-swift | 19:25 | |
| *** aix has quit IRC | 19:32 | |
| *** yland has quit IRC | 19:40 | |
| minwoob | Hey peluse - The commit message for your alternate fragments patch (https://review.openstack.org/#/c/213147/) says there are new probe tests. However, I don't see any attached. | 19:41 |
| minwoob | Were you going to add them in? | 19:42 |
| jrichli | all: If you have some time, please review patch 205803 | 19:49 |
| patchbot | jrichli: https://review.openstack.org/#/c/205803/ | 19:49 |
| jrichli | It would be great to get it in the next release. Thank you! | 19:49 |
| clayg | minwoob: yes I think that's the correct thing (re: teaching container sync to handle 409 means "already synced") - but I don't know how to do it with a smooth upgrade | 19:49 |
| clayg | minwoob: like you can't really upgrade your proxy to start spewing 409's until you're sure all your container-sync clusters (clients) are ready to eat them :'( | 19:49 |
| minwoob | clayg: Right. So it seems that such a patch to fix the proxy server would need to include the logic for handling the 409s in the container sync case. | 19:52 |
| minwoob | Is that going to be particularly difficult? | 19:52 |
| clayg | minwoob: but that patch would need to be applied to *both* clusters *at the same time* | 19:53 |
| minwoob | I see. | 19:53 |
| minwoob | Should we wait for post-Liberty for that, then? | 19:57 |
| clayg | minwoob: i'm not really sure :\ | 19:58 |
| clayg | minwoob: I guess updating container sync to handle 409's would just be "the right thing to do" | 19:58 |
| clayg | minwoob: when the proxy can start returning 409's instead of 202's is maybe more complicated if we're worried about breaking container sync for existing deployments | 19:58 |
| clayg | it's po | 19:58 |
| clayg | minwoob: it's possible we could add a shim in the container sync middleware on the target cluster that's like "support_legacy_container_sync_clients" and then you can turn it on to capture the real 409 response and make it a 202 for container sync requests only - maybe | 19:59 |
| eranrom | clayg, minwoob Incidently, I was spending today working on this: Here is another option: | 20:00 |
| minwoob | clayg: Yeah, I was thinking along those lines. | 20:00 |
| eranrom | If we follow Clay's previous idea of using Expect 100-continue + timestamp for container sync to prevent multiple uploads, we can translate the 409 to 417 only if the request has the expect 100 continue header, thus not breaking older versions. but perhaps I am missing the beginning of this conversation | 20:02 |
| bil | I need to get storage url fron server I need to do it via V2 keystone using python could someone help me or atleast direct me | 20:02 |
| bil | I got some response via curl | 20:02 |
| *** dustins has quit IRC | 20:04 | |
| *** lcurtis has joined #openstack-swift | 20:06 | |
| eranrom | ok sorry for the interruption, you were discussing the 409->202 in general with container sync being only part of the picture. So what I am suggesting is not relevant. | 20:06 |
| clayg | eranrom: I like the general idea of making the container-sync client somehow explicitly demonstrate that it is "new enough" to recieve the 409 and translate to 202 otherwise | 20:07 |
| *** barker has quit IRC | 20:07 | |
| clayg | eranrom: no I think you're on the right track | 20:07 |
| clayg | minwoob: new plan - do what eranrom thinks is best - eranrom is smart | 20:07 |
| minwoob | +1 | 20:08 |
| eranrom | well, thanks :-) but if the translation problem goes beyond container sync then I do not see how to do that. Except perhaps the shim idea :-) thiking... | 20:09 |
| minwoob | eranrom: Where is the 417 coming from, btw | 20:09 |
| eranrom | thinking... | 20:09 |
| minwoob | eranrom: From the above explanation. | 20:09 |
| eranrom | its ExpectationFailure: client says expect 100-continue, server says do not continue by sending this | 20:10 |
| eranrom | +1 for the "support_legacy_container_sync_clients"idea with the following additional detail: the "shim" in the container sync middleware will translate from 409 to 202 only if the request has the container sync auth header (so that not every request gets translated back) and from 409 to 417 if we also have the expect header | 20:19 |
| *** barker has joined #openstack-swift | 20:25 | |
| *** ekarlso has quit IRC | 20:38 | |
| openstackgerrit | Merged openstack/swift: Merge branch 'master' into feature/hummingbird https://review.openstack.org/221410 | 20:43 |
| *** tongli has quit IRC | 20:44 | |
| *** trifon_ has joined #openstack-swift | 20:51 | |
| *** trifon has quit IRC | 20:53 | |
| *** kevinc___ has joined #openstack-swift | 20:55 | |
| *** cdelatte has quit IRC | 20:56 | |
| *** kevinc_ has quit IRC | 20:57 | |
| *** ekarlso has joined #openstack-swift | 20:57 | |
| *** ekarlso has quit IRC | 20:57 | |
| *** ekarlso has joined #openstack-swift | 20:57 | |
| *** aix has joined #openstack-swift | 21:01 | |
| *** changbl has quit IRC | 21:07 | |
| *** marzif has quit IRC | 21:09 | |
| *** flwang1 has joined #openstack-swift | 21:09 | |
| *** marzif has joined #openstack-swift | 21:09 | |
| *** lcurtis has quit IRC | 21:13 | |
| *** bil has quit IRC | 21:14 | |
| clayg | notmyname: yay settings -> preferences -> change view = "New Screen" !!! | 21:14 |
| notmyname | just something to try out and see if it sits well with you | 21:15 |
| openstackgerrit | Bill Huber proposed openstack/swift: Reconstructor GET excludes user_agent in log https://review.openstack.org/221506 | 21:17 |
| *** haypo has quit IRC | 21:20 | |
| *** nadeem has quit IRC | 21:33 | |
| *** eranrom has quit IRC | 21:35 | |
| peluse__ | minwoob, wrt the probe tests, yeah I think they ended up on a different patch when I broke a larger one apart a while back. I'll update the commit msg, thanks | 21:36 |
| openstackgerrit | David Goetz proposed openstack/swift: go: assign startTime to Now at beginning of replication run https://review.openstack.org/221512 | 21:36 |
| *** devlaps has quit IRC | 21:38 | |
| *** nadeem has joined #openstack-swift | 21:39 | |
| *** jrichli has quit IRC | 21:44 | |
| clayg | notmyname: boooo - the "f" shortcut doesn't work in new screen! | 21:44 |
| notmyname | clayg: and scrolling (in my browser) seems wonky. I'm not in love with the new look | 21:45 |
| *** darrenc_ is now known as darrenc | 21:47 | |
| clayg | notmyname: I don't think the "unified diff" view is available anymore either :\ | 21:50 |
| clayg | goodness - why did they have to take out so much good stuff to make the crappy parts less annoying!? | 21:50 |
| *** jamielennox|away is now known as jamielennox | 21:51 | |
| *** jkugel has quit IRC | 21:52 | |
| peluse__ | who is 'goodness'? :) | 21:54 |
| *** marcusvrn_ has quit IRC | 21:55 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift: Re-organise ssync tests https://review.openstack.org/220198 | 21:59 |
| openstackgerrit | Clay Gerrard proposed openstack/swift: Fix invalid frag_index header in ssync_sender when reverting EC tombstones https://review.openstack.org/217830 | 21:59 |
| *** thurloat_isgone is now known as thurloat | 21:59 | |
| *** barker has quit IRC | 22:02 | |
| *** jlhinson has quit IRC | 22:03 | |
| *** annegentle has quit IRC | 22:13 | |
| briancline | are there any known issues around DLOs and publicly-readable container ACLs? | 22:14 |
| briancline | seeing some weird issues on 2.2.2, probably something dumb but figured I'd ask | 22:14 |
| torgomatic | briancline: I'd guess that it's related to container listings; DLOs need container listings to be able to do stuff, so if you don't have .rlistings set on the public container, it won't work | 22:16 |
| torgomatic | other than that, I got nothing | 22:16 |
| briancline | notmyname: ahhh that would make sense. I'll give that a shot. thanks | 22:17 |
| *** proteusguy__ has quit IRC | 22:17 | |
| openstackgerrit | Merged openstack/swift: go: improve replication logging https://review.openstack.org/220333 | 22:18 |
| briancline | yep, that did it. I figured it was something silly | 22:28 |
| *** proteusguy__ has joined #openstack-swift | 22:30 | |
| *** wbhuber is now known as wbhuber_ | 22:40 | |
| *** hrou has quit IRC | 22:48 | |
| *** ctrath has quit IRC | 22:49 | |
| *** janonymous_ has quit IRC | 22:51 | |
| *** asettle has joined #openstack-swift | 22:51 | |
| *** wbhuber_ has quit IRC | 22:52 | |
| *** nadeem has quit IRC | 22:55 | |
| *** kaliya has joined #openstack-swift | 22:59 | |
| kaliya | Hi in Juno I have this recurrent error in swift logs ( http://paste.openstack.org/raw/450899/ ) but I don't know how to troubleshoot further. Glance works fine on swift and object uploads/downloads from swift also is perfect. I think that it might be something old... There are no other hitting requests coming in nor other operations... But this entry is very frequent. Suggestions? | 23:01 |
| *** ho has joined #openstack-swift | 23:01 | |
| kaliya | Seems that it tries to do a listdir on a file | 23:02 |
| ho | good morning! | 23:06 |
| kaliya | hi ho | 23:08 |
| ho | kaliya: hello | 23:09 |
| *** km has joined #openstack-swift | 23:10 | |
| *** asettle is now known as asettle-afk | 23:15 | |
| torgomatic | kaliya: it says right there in the output: "Permission denied: '/var/lib/glance/node/2/objects/85/809/2ac8393f1b22d216fdc9915e9968a809'" | 23:26 |
| torgomatic | go make sure swift can read that directory | 23:26 |
| *** chlong has quit IRC | 23:27 | |
| *** chlong has joined #openstack-swift | 23:28 | |
| *** cdelatte has joined #openstack-swift | 23:34 | |
| *** cdelatte has quit IRC | 23:37 | |
| *** asettle-afk is now known as asettle | 23:38 | |
| *** kei_yama has joined #openstack-swift | 23:39 | |
| *** hrou has joined #openstack-swift | 23:44 | |
| *** asettle has quit IRC | 23:48 | |
| *** minwoob has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!