| *** sayali has joined #openstack-swift | 00:00 | |
| *** thumpba_ has quit IRC | 00:09 | |
| *** jrichli has quit IRC | 00:11 | |
| *** agomez has joined #openstack-swift | 00:21 | |
| *** chlong has quit IRC | 00:25 | |
| *** lyrrad has quit IRC | 00:28 | |
| *** hezhiqiang has joined #openstack-swift | 00:32 | |
| *** alejandrito has joined #openstack-swift | 00:32 | |
| * csmart growls at tonyb | 00:34 | |
| csmart | has another coffee with mattoliverau | 00:35 |
|---|---|---|
| *** minwoob has quit IRC | 00:38 | |
| *** peluse has quit IRC | 00:45 | |
| *** peluse has joined #openstack-swift | 00:46 | |
| *** haomaiwang has joined #openstack-swift | 00:49 | |
| tonyb | csmart: Good to know. | 00:54 |
| tonyb | csmart: You know I'm avoiding you right ? | 00:54 |
| tonyb | csmart: Also ask Matt if he wants to have lunch at the duxton with a few locals | 00:54 |
| tonyb | csmart: a lift can be arranged. | 00:54 |
| *** peluse has quit IRC | 00:54 | |
| mattoliverau | tonyb sure sounds awesome :) | 00:57 |
| openstackgerrit | John Dickinson proposed openstack/swift: Fix missing container update https://review.openstack.org/186735 | 00:57 |
| *** haomaiwang has quit IRC | 01:01 | |
| *** david-lyle has quit IRC | 01:01 | |
| *** haomaiwang has joined #openstack-swift | 01:01 | |
| *** alejandrito has quit IRC | 01:01 | |
| hugokuo | notmyname: sloslo : https://www.fuel-infra.org/plugins/catalog.html Fuel doesn't have it so SwiftStack wrote a new plugin in Fuel for this purpose. | 01:02 |
| *** hezhiqiang has quit IRC | 01:03 | |
| hugokuo | This should be only for SwiftStack user in general. But this version is open for any Swift cluster. Next version of this plugin will be only for SwiftStack since it needs SwiftStack's feature to auto deploy Swift cluster. You can have a look at the current version. | 01:03 |
| hugokuo | btw, it's in the buttom | 01:04 |
| *** haomaiwang has quit IRC | 01:08 | |
| *** gyee has quit IRC | 01:24 | |
| *** chlong has joined #openstack-swift | 01:27 | |
| *** agomez has quit IRC | 01:33 | |
| *** peluse has joined #openstack-swift | 01:52 | |
| *** peluse has quit IRC | 01:54 | |
| *** peluse has joined #openstack-swift | 01:54 | |
| *** haomaiwa_ has joined #openstack-swift | 02:00 | |
| *** haomaiwa_ has quit IRC | 02:01 | |
| *** haomaiwa_ has joined #openstack-swift | 02:01 | |
| *** hezhiqiang has joined #openstack-swift | 02:02 | |
| *** agomez has joined #openstack-swift | 02:33 | |
| *** haomaiwa_ has quit IRC | 02:35 | |
| *** haomaiwang has joined #openstack-swift | 02:36 | |
| *** jrichli has joined #openstack-swift | 02:38 | |
| *** sanchitmalhotra has joined #openstack-swift | 02:39 | |
| *** wbhuber_ has quit IRC | 02:47 | |
| *** miurahr has joined #openstack-swift | 02:49 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 02:53 | |
| *** darrenc is now known as darrenc_afk | 02:54 | |
| *** sanchitmalhotra has quit IRC | 02:55 | |
| *** haomaiwang has quit IRC | 03:01 | |
| *** haomaiwang has joined #openstack-swift | 03:01 | |
| *** sloslo has quit IRC | 03:04 | |
| *** darrenc_afk is now known as darrenc | 03:12 | |
| *** kota_ has joined #openstack-swift | 03:39 | |
| *** ChanServ sets mode: +v kota_ | 03:39 | |
| kota_ | hello | 03:39 |
| hugokuo | おはよう | 03:40 |
| kota_ | hugokuo: awesome! | 03:40 |
| *** agomez has quit IRC | 03:49 | |
| notmyname | kota_: good morning | 03:50 |
| kota_ | notmyname: good evening | 03:50 |
| kota_ | notmyname: Did you update my patch ... just rebased? | 03:50 |
| notmyname | kota_: torgomatic and I took a look at https://review.openstack.org/#/c/186735/ today. I pushed up a version over yours | 03:50 |
| kota_ | yup | 03:51 |
| notmyname | more than rebases (it didn't need that). we fixed some tests | 03:51 |
| kota_ | notmyname: great, thanks | 03:51 |
| notmyname | tuens out the tests were only accidentally passing | 03:51 |
| notmyname | *turns | 03:51 |
| notmyname | but now they are better | 03:51 |
| kota_ | sounds good. | 03:52 |
| kota_ | I'll take a look at that today and will make a comment for the change. | 03:53 |
| kota_ | notmyname: thanks for working that. | 03:53 |
| notmyname | great | 03:53 |
| notmyname | kota_: most of the changes we did are in https://review.openstack.org/#/c/186735/12..13/test/unit/proxy/controllers/test_obj.py | 03:53 |
| kota_ | notmyname: yeah, that's what I am seeing. | 03:54 |
| notmyname | kota_: the root of the changes are with line 933 in the new patch on that file | 03:55 |
| notmyname | basically, we had to make sure that each test gets a *new* ring instead of just using the same one over and over for each test | 03:55 |
| notmyname | some of the tests change the ring inside the test, unfortunately | 03:55 |
| notmyname | so that was happening and setting the replica count to 4 in one test and then the tests you wrote accidentally passed. 4 replicas worked, 1 replica didn't | 03:56 |
| notmyname | most of the other changes are to get the tests working with a 1 replica ring | 03:56 |
| kota_ | I see. | 03:57 |
| kota_ | that makes me sense. | 03:57 |
| kota_ | thanks, the information makes me easy to review. | 03:58 |
| notmyname | I think there could be some long-term improvements in that test file. eg to make sure more tests loop over every policy. and adding different policies to be tested (partial replicas, ec, etc). definitely some improvements in how FakeRing works | 03:58 |
| kota_ | yup | 03:59 |
| kota_ | oh, yeah, that affected my tests. | 03:59 |
| notmyname | please feel free to push over any improvements to what we did | 04:00 |
| *** haomaiwang has quit IRC | 04:01 | |
| *** haomaiwang has joined #openstack-swift | 04:01 | |
| kota_ | notmyname: ok, I may do so if needed ;) | 04:01 |
| * notmyname calls it a night | 04:04 | |
| kota_ | notmyname: have a good night ;) | 04:05 |
| *** hrou has quit IRC | 04:07 | |
| mattoliverau | notmyname: night o/ | 04:08 |
| *** sanchitmalhotra has joined #openstack-swift | 04:13 | |
| *** sanchitmalhotra1 has quit IRC | 04:15 | |
| *** peluse has quit IRC | 04:25 | |
| *** peluse has joined #openstack-swift | 04:26 | |
| *** kei_yama has quit IRC | 04:31 | |
| *** kei_yama has joined #openstack-swift | 04:32 | |
| *** garthb has joined #openstack-swift | 04:34 | |
| *** peluse has quit IRC | 04:35 | |
| *** bill_az has quit IRC | 04:40 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 04:54 | |
| *** sanchitmalhotra has quit IRC | 04:56 | |
| *** haomaiwang has quit IRC | 05:01 | |
| *** haomaiwang has joined #openstack-swift | 05:01 | |
| *** sanchitmalhotra has joined #openstack-swift | 05:12 | |
| *** sanchitmalhotra1 has quit IRC | 05:14 | |
| *** SkyRocknRoll has joined #openstack-swift | 05:17 | |
| *** trifon has joined #openstack-swift | 05:18 | |
| *** peluse has joined #openstack-swift | 05:32 | |
| *** proteusguy__ has quit IRC | 05:33 | |
| *** peluse has quit IRC | 05:39 | |
| *** mfalatic has quit IRC | 05:47 | |
| *** garthb has quit IRC | 05:48 | |
| jst_ | hugokuo: thx for info. are you there? | 06:00 |
| hugokuo | jst_: yup | 06:00 |
| jst_ | regarding fuel plugin for swift | 06:00 |
| *** haomaiwang has quit IRC | 06:01 | |
| jst_ | does fuel plugin than enable all the configuration for swift (replicas, weights etc), but just with swiftstack distribution of swift? | 06:01 |
| *** openstackgerrit has quit IRC | 06:01 | |
| jst_ | or is it general purpose plugin? | 06:01 |
| *** haomaiwang has joined #openstack-swift | 06:01 | |
| *** openstackgerrit has joined #openstack-swift | 06:01 | |
| hugokuo | jst_: The plugin requires a pre-deployed Swift cluster. It simple handles configurations for Glance and Keystone in Fuel | 06:02 |
| hugokuo | jst_: So it's not provide entry for managing your Swift cluster | 06:02 |
| hugokuo | You need to manage your swift cluster by your own. For SwiftStack customer, they manage the cluster via SwiftStack software as an out-of-box of Fuel. | 06:03 |
| jst_ | so let's say i install openstack (nova, identity, neutron) with Mirantis Fuel. And than after the installation i install by hand Swift with configuration i want.. after swift cluster deployes, i have some managing options than within Fuel with this plugin | 06:04 |
| jst_ | so storage policies and ring configuration should be deployed in advance, without fuel. | 06:06 |
| jst_ | ok. i found document: http://plugins.mirantis.com/docs/s/w/swiftstack/swiftstack-0.1.0.pdf | 06:07 |
| hugokuo | jst_: base on the purpose of this plugin it's opposite. And I do recommend you to deploy Swift before constructing Fuel OpenStack environment. b/c Fuel runs check tests after deploy the entirement. It uploads a test image via Glance for testing by NOVA | 06:07 |
| hugokuo | jst_: yes. all Swift wise setting should be separated from fuel | 06:08 |
| hugokuo | Fuel doesn't handle too much Swift stuff tho | 06:08 |
| hugokuo | And ... Fuel restrict user from changing many settings once the environment been deployed. | 06:09 |
| *** mac_ified has quit IRC | 06:11 | |
| *** jamielennox is now known as jamielennox|away | 06:12 | |
| jst_ | hugokuo: ok thank you for the information provided | 06:13 |
| hugokuo | For the case to associate a Swift cluster to existing Fuel environment must be done manually since the limit in Fuel. You need to 1) add new endpoint for object-store in Keystone 2) change the setting of Swift section in glance-api which includes several options 3) restart glance 4) upload VM images | 06:13 |
| hugokuo | jst_: my pleasure | 06:13 |
| *** haomaiwang has quit IRC | 06:18 | |
| *** haomaiwa_ has joined #openstack-swift | 06:18 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 06:32 | |
| *** sanchitmalhotra has quit IRC | 06:34 | |
| *** peluse has joined #openstack-swift | 06:36 | |
| *** peluse has quit IRC | 06:45 | |
| *** sayali has quit IRC | 07:00 | |
| *** briancline has quit IRC | 07:00 | |
| *** geaaru has joined #openstack-swift | 07:00 | |
| *** haomaiwa_ has quit IRC | 07:01 | |
| *** haomaiwang has joined #openstack-swift | 07:01 | |
| *** rledisez has joined #openstack-swift | 07:09 | |
| *** sayali has joined #openstack-swift | 07:14 | |
| *** ntt has quit IRC | 07:26 | |
| *** peluse has joined #openstack-swift | 07:42 | |
| *** peluse has quit IRC | 07:49 | |
| *** haomaiwang has quit IRC | 08:01 | |
| *** haomaiwang has joined #openstack-swift | 08:01 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift: Fix missing container update https://review.openstack.org/186735 | 08:02 |
| *** jordanP has joined #openstack-swift | 08:03 | |
| *** acoles_ is now known as acoles | 08:05 | |
| acoles | good morning | 08:09 |
| kota_ | good morning :) | 08:09 |
| acoles | hi kota_ | 08:10 |
| *** mahatic has joined #openstack-swift | 08:10 | |
| *** akle has joined #openstack-swift | 08:11 | |
| *** haomaiwang has quit IRC | 08:12 | |
| acoles | ho: hi! i tried to add you to a review on gerrit and got an error saying you were not registered. | 08:13 |
| acoles | ho: i have no idea why that happened | 08:13 |
| acoles | ho: got me worried that you had gone away :/ but then i saw your review comments today, phew! ho is still with us :) | 08:15 |
| *** openstackgerrit has quit IRC | 08:16 | |
| *** openstackgerrit has joined #openstack-swift | 08:16 | |
| acoles | kota_: its still 4 weeks away but i am excited about coming to tokyo :) | 08:17 |
| *** ktsuyuzaki has joined #openstack-swift | 08:17 | |
| *** KLuka_ has joined #openstack-swift | 08:18 | |
| *** sayali has quit IRC | 08:19 | |
| *** kota_ has quit IRC | 08:19 | |
| *** sayali has joined #openstack-swift | 08:19 | |
| *** haomaiwang has joined #openstack-swift | 08:22 | |
| *** haomaiwang has quit IRC | 08:23 | |
| *** haomaiwang has joined #openstack-swift | 08:23 | |
| *** kei_yama has quit IRC | 08:25 | |
| *** kei_yama has joined #openstack-swift | 08:25 | |
| cschwede | acoles: Good Morning! I have a question regarding http://paste.openstack.org/raw/473886/ | 08:31 |
| cschwede | acoles: i might need more coffee, but what logic is changed there? at the end it returns with req.environ['swift_owner'] set to True before and after applying that patch | 08:32 |
| acoles | cschwede: hello! | 08:32 |
| cschwede | acoles: argh, nevermind, got it. it’s a difference if req.environ['swift_owner'] was set before | 08:33 |
| ho | acoles: hahaha, i'm very busy now and restricted access to irc but i can read email so i could see donagh's patch. :-) | 08:33 |
| acoles | cschwede: before it would return None if swift_owner was already True in the env | 08:33 |
| acoles | cschwede: yes, that!) | 08:34 |
| cschwede | acoles: yep, just realized this :) | 08:34 |
| acoles | cschwede: now, what i did not do in that patch is clear swift_owner from env if allowed=False | 08:34 |
| acoles | cschwede: i dodged that decision ;) | 08:35 |
| *** Alvin_ has joined #openstack-swift | 08:36 | |
| acoles | ho: great, hope all is wel | 08:36 |
| acoles | well | 08:36 |
| *** Alvin_ has left #openstack-swift | 08:36 | |
| acoles | cschwede: so, i'm sure you figured this, but the root of the problem was the original COPY having account AUTH_xxx and that being authorised so swift_owner was True when the actual PUT request called back | 08:39 |
| ho | acoles: thanks! | 08:39 |
| cschwede | acoles: yep, saw that. but then i was also assuming acls would be required too, thus my patch. but i’m removing that, only adding your patch, a small doc update to tempauth (because tempauth seems to be fine), and adding you, Donagh and Hisashi as Co-author | 08:40 |
| cschwede | just need to finish my tests with it | 08:40 |
| *** Alvin_ has joined #openstack-swift | 08:41 | |
| acoles | cschwede: great work, thanks for taking that on | 08:41 |
| acoles | cschwede: are you planning to squash donagh's test into your review? | 08:44 |
| acoles | that would make sense | 08:44 |
| cschwede | acoles: well, if Donagh is ok with that - sure | 08:44 |
| acoles | donagh: are you here? ^^ | 08:44 |
| donagh_ | acoles: yes | 08:45 |
| acoles | donagh_:u ok with cschwede squashing your test into his review? | 08:45 |
| *** peluse has joined #openstack-swift | 08:47 | |
| *** donagh_ has quit IRC | 08:49 | |
| *** donagh_ has joined #openstack-swift | 08:50 | |
| *** SkyRocknRoll_ has joined #openstack-swift | 08:52 | |
| *** KLuka_ has quit IRC | 08:52 | |
| *** peluse has quit IRC | 08:54 | |
| *** haomaiwang has quit IRC | 09:01 | |
| *** haomaiwang has joined #openstack-swift | 09:01 | |
| openstackgerrit | Christian Schwede proposed openstack/swift: Fix copy requests to service accounts in Keystone https://review.openstack.org/227133 | 09:07 |
| donagh_ | cschwede: re-reviewing my functional test path, I see something odd. I put a comment in that patch (just in case you hit the problem I thought might occur). I'll debug it a bit more | 09:11 |
| *** Alvin_ has quit IRC | 09:13 | |
| *** Alvin_ has joined #openstack-swift | 09:13 | |
| openstackgerrit | Christian Schwede proposed openstack/swift: Fix copy requests to service accounts in Keystone https://review.openstack.org/227133 | 09:16 |
| *** portante has quit IRC | 09:20 | |
| *** Alvin_ has quit IRC | 09:22 | |
| *** Alvin_ has joined #openstack-swift | 09:22 | |
| cschwede | donagh_: alright; meanwhile i just added a very minor change to the existing unit tests, and this change fails on master and passes with the patch: https://review.openstack.org/#/c/227133/6/test/unit/common/middleware/test_keystoneauth.py | 09:24 |
| cschwede | doing this there should be no regression in the future (that’s what i hope for ;)) | 09:25 |
| *** Alvin_ has quit IRC | 09:25 | |
| *** alvin_nathan has joined #openstack-swift | 09:25 | |
| *** portante has joined #openstack-swift | 09:34 | |
| *** ktsuyuzaki has quit IRC | 09:37 | |
| *** kota_ has joined #openstack-swift | 09:37 | |
| *** ChanServ sets mode: +v kota_ | 09:37 | |
| *** proteusguy has joined #openstack-swift | 09:45 | |
| ho | acoles: cschwede: donagh: i tested patch 227133 and passed so your code works. now i have one concern about swit_owner in env. i think it is better to remove in case of delay_denial to prevent further troubles. what do you think? | 09:47 |
| patchbot | ho: https://review.openstack.org/#/c/227133/ - Fix copy requests to service accounts in Keystone | 09:47 |
| cschwede | ho: thx for testing! yeah, Alistair mentioned that and i was thinking about it too | 09:48 |
| *** silor has joined #openstack-swift | 09:50 | |
| *** peluse has joined #openstack-swift | 09:51 | |
| ho | cschwede: sounds great! | 09:53 |
| *** peluse has quit IRC | 10:00 | |
| *** haomaiwang has quit IRC | 10:01 | |
| *** haomaiwang has joined #openstack-swift | 10:01 | |
| *** SkyRocknRoll_ has quit IRC | 10:05 | |
| *** hezhiqiang has quit IRC | 10:06 | |
| donagh_ | ho: schwede: thinking about clearing swift_owner in the delay_denial ... I think we *must* do so. It's not just a nice to have. | 10:06 |
| donagh_ | Consider: | 10:06 |
| donagh_ | if ACL grants write access in destination-account, you also have swift_owner at the point...you should only have write access | 10:08 |
| donagh_ | Since copy is for objects that does not matter right now (I think), but if there is ever some object metadata that can only be set by swift_owner, than we've allowed a write ACL to set that metadata | 10:09 |
| *** briancline has joined #openstack-swift | 10:15 | |
| *** akle|2 has joined #openstack-swift | 10:18 | |
| *** SkyRocknRoll_ has joined #openstack-swift | 10:20 | |
| *** briancline has quit IRC | 10:22 | |
| *** akle has quit IRC | 10:22 | |
| openstackgerrit | Christian Schwede proposed openstack/swift: Fix copy requests to service accounts in Keystone https://review.openstack.org/227133 | 10:33 |
| openstackgerrit | Donagh McCabe proposed openstack/swift: Add func tests for COPY and service tokens https://review.openstack.org/227421 | 10:44 |
| *** silor has quit IRC | 10:44 | |
| *** silor has joined #openstack-swift | 10:50 | |
| *** briancline has joined #openstack-swift | 10:54 | |
| *** peluse has joined #openstack-swift | 10:57 | |
| *** silor has quit IRC | 10:59 | |
| *** silor has joined #openstack-swift | 11:00 | |
| *** briancline has quit IRC | 11:00 | |
| *** haomaiwang has quit IRC | 11:01 | |
| *** haomaiwang has joined #openstack-swift | 11:01 | |
| *** peluse has quit IRC | 11:04 | |
| *** silor1 has joined #openstack-swift | 11:04 | |
| *** silor has quit IRC | 11:05 | |
| *** silor1 is now known as silor | 11:05 | |
| *** cdelatte has joined #openstack-swift | 11:28 | |
| *** delattec has joined #openstack-swift | 11:28 | |
| *** sanchitmalhotra1 has quit IRC | 11:37 | |
| *** haomaiwang has quit IRC | 12:01 | |
| *** haomaiwang has joined #openstack-swift | 12:01 | |
| *** peluse has joined #openstack-swift | 12:01 | |
| *** SkyRocknRoll_ has quit IRC | 12:02 | |
| *** peluse has quit IRC | 12:09 | |
| *** hezhiqiang has joined #openstack-swift | 12:10 | |
| *** miurahr has quit IRC | 12:11 | |
| *** alejandrito has joined #openstack-swift | 12:13 | |
| *** annegentle has joined #openstack-swift | 12:22 | |
| *** NM has joined #openstack-swift | 12:31 | |
| *** alejandrito has quit IRC | 12:36 | |
| *** CaioBrentano has quit IRC | 12:37 | |
| *** alejandrito has joined #openstack-swift | 12:38 | |
| *** CaioBrentano has joined #openstack-swift | 12:38 | |
| *** haomaiwang has quit IRC | 12:41 | |
| *** chlong has quit IRC | 12:57 | |
| *** kota_ has quit IRC | 12:58 | |
| *** km has quit IRC | 13:02 | |
| *** hrou has joined #openstack-swift | 13:04 | |
| *** peluse has joined #openstack-swift | 13:06 | |
| *** janonymous_ has joined #openstack-swift | 13:07 | |
| *** bill_az has joined #openstack-swift | 13:09 | |
| *** jkugel has joined #openstack-swift | 13:11 | |
| *** dustins has joined #openstack-swift | 13:11 | |
| *** peluse has quit IRC | 13:14 | |
| *** peluse has joined #openstack-swift | 13:14 | |
| *** km has joined #openstack-swift | 13:15 | |
| *** km has quit IRC | 13:15 | |
| *** wbhuber has joined #openstack-swift | 13:28 | |
| *** esker has quit IRC | 13:30 | |
| *** kei_yama has quit IRC | 13:34 | |
| *** pberis has joined #openstack-swift | 13:42 | |
| openstackgerrit | janonymous proposed openstack/swift: Eventlet green threads not released back to pool https://review.openstack.org/227855 | 13:48 |
| janonymous_ | please review +2 ones if possible to avoid rebasing all the patches : https://review.openstack.org/#/q/owner:%22janonymous+%253Cjaivish.kothari%2540nectechnologies.in%253E%22+status:open,n,z | 13:49 |
| *** annegentle has quit IRC | 13:53 | |
| *** annegentle has joined #openstack-swift | 13:54 | |
| *** marzif has joined #openstack-swift | 14:04 | |
| *** chsc has joined #openstack-swift | 14:09 | |
| *** esker has joined #openstack-swift | 14:11 | |
| *** esker has quit IRC | 14:11 | |
| *** janonymous_ has quit IRC | 14:12 | |
| *** esker has joined #openstack-swift | 14:12 | |
| *** peluse has quit IRC | 14:14 | |
| *** peluse has joined #openstack-swift | 14:15 | |
| *** delatte has joined #openstack-swift | 14:19 | |
| *** delattec has quit IRC | 14:21 | |
| *** marzif has quit IRC | 14:22 | |
| *** cdelatte has quit IRC | 14:22 | |
| *** marzif has joined #openstack-swift | 14:22 | |
| *** david-lyle has joined #openstack-swift | 14:23 | |
| *** jlhinson has joined #openstack-swift | 14:24 | |
| *** peluse has quit IRC | 14:24 | |
| *** akle|2 has quit IRC | 14:32 | |
| *** david-lyle has quit IRC | 14:35 | |
| *** peluse has joined #openstack-swift | 14:38 | |
| *** annegentle has quit IRC | 14:46 | |
| *** thumpba has joined #openstack-swift | 14:47 | |
| *** esker has quit IRC | 14:48 | |
| *** trifon has quit IRC | 14:50 | |
| *** ccavanna has joined #openstack-swift | 14:52 | |
| *** hezhiqiang has quit IRC | 14:53 | |
| *** garthb has joined #openstack-swift | 14:55 | |
| *** peluse has quit IRC | 14:58 | |
| *** marzif has quit IRC | 14:58 | |
| *** peluse has joined #openstack-swift | 14:59 | |
| *** david-lyle has joined #openstack-swift | 15:02 | |
| *** agomez has joined #openstack-swift | 15:07 | |
| *** peluse has quit IRC | 15:07 | |
| *** agomez has quit IRC | 15:07 | |
| *** jkugel1 has joined #openstack-swift | 15:08 | |
| *** jkugel has quit IRC | 15:08 | |
| *** marzif has joined #openstack-swift | 15:09 | |
| *** peluse has joined #openstack-swift | 15:10 | |
| *** david-lyle has quit IRC | 15:18 | |
| *** david-lyle has joined #openstack-swift | 15:20 | |
| *** annegentle has joined #openstack-swift | 15:29 | |
| *** david-lyle has quit IRC | 15:34 | |
| *** peluse_ has joined #openstack-swift | 15:35 | |
| *** peluse has quit IRC | 15:36 | |
| *** peluse has joined #openstack-swift | 15:37 | |
| *** peluse has quit IRC | 15:37 | |
| *** peluse has joined #openstack-swift | 15:43 | |
| *** peluse___ has joined #openstack-swift | 15:43 | |
| *** peluse__ has joined #openstack-swift | 15:43 | |
| *** peluse__ has quit IRC | 15:43 | |
| *** peluse has quit IRC | 15:43 | |
| *** david-lyle has joined #openstack-swift | 15:44 | |
| *** peluse has joined #openstack-swift | 15:44 | |
| *** peluse__ has joined #openstack-swift | 15:44 | |
| *** peluse has quit IRC | 15:45 | |
| *** peluse__ has quit IRC | 15:45 | |
| peluse_ | hello me | 15:47 |
| *** delatte has quit IRC | 15:47 | |
| *** delatte has joined #openstack-swift | 15:47 | |
| *** peluse2 has quit IRC | 15:47 | |
| *** Alvin_nathan_ has joined #openstack-swift | 15:48 | |
| *** david-lyle has quit IRC | 15:48 | |
| *** peluse_ is now known as peluse | 15:48 | |
| *** ChanServ sets mode: +v peluse | 15:49 | |
| *** alvin_nathan has quit IRC | 15:51 | |
| *** gyee has joined #openstack-swift | 15:51 | |
| *** janonymous_ has joined #openstack-swift | 15:53 | |
| *** SkyRocknRoll has quit IRC | 15:54 | |
| *** dustins is now known as dustins|afk | 15:54 | |
| *** minwoob has joined #openstack-swift | 15:56 | |
| *** annegentle has quit IRC | 16:03 | |
| *** lyrrad has joined #openstack-swift | 16:05 | |
| *** bill_az has quit IRC | 16:07 | |
| *** marzif has quit IRC | 16:10 | |
| *** marzif has joined #openstack-swift | 16:10 | |
| *** annegentle has joined #openstack-swift | 16:11 | |
| *** haomaiwang has joined #openstack-swift | 16:20 | |
| *** rledisez has quit IRC | 16:29 | |
| *** mfalatic has joined #openstack-swift | 16:31 | |
| *** haomaiwang has quit IRC | 16:32 | |
| *** haomaiwa_ has joined #openstack-swift | 16:33 | |
| *** bill_az has joined #openstack-swift | 16:39 | |
| *** cdelatte has joined #openstack-swift | 16:39 | |
| *** NM has quit IRC | 16:39 | |
| *** delatte has quit IRC | 16:41 | |
| *** thumpba has quit IRC | 16:41 | |
| *** delattec has joined #openstack-swift | 16:44 | |
| *** cdelatte has quit IRC | 16:45 | |
| *** NM has joined #openstack-swift | 16:48 | |
| *** esker has joined #openstack-swift | 16:49 | |
| *** janonymous_ has quit IRC | 16:49 | |
| *** esker has quit IRC | 16:49 | |
| *** esker has joined #openstack-swift | 16:50 | |
| *** vinsh has joined #openstack-swift | 16:53 | |
| notmyname | good morning | 16:57 |
| mahatic | wow it's so quiet! | 16:57 |
| *** garthb has quit IRC | 16:57 | |
| mahatic | notmyname: morning! | 16:57 |
| peluse | must be friday :) | 16:58 |
| notmyname | have of the world is already on the weekend | 16:58 |
| notmyname | (including you mahatic ;-) | 16:58 |
| peluse | at intel we don't get saturday's off... | 16:58 |
| peluse | only every other sunday :) | 16:59 |
| mahatic | notmyname: yeah, I'm kinda on the weekend - weekend begins with responding to pending emails :P | 16:59 |
| notmyname | heh | 16:59 |
| mahatic | actually it's a holiday here today | 16:59 |
| peluse | mahatic: you know i'm kidding right.... you should go do something funner than email! | 16:59 |
| * peluse on the commute... | 17:00 | |
| mahatic | peluse: I *just* got to emails. Outreachy stuff. Will be off again | 17:00 |
| *** thumpba has joined #openstack-swift | 17:00 | |
| *** haomaiwa_ has quit IRC | 17:01 | |
| *** thumpba has quit IRC | 17:01 | |
| *** thumpba has joined #openstack-swift | 17:01 | |
| *** marzif has quit IRC | 17:01 | |
| *** esker has quit IRC | 17:01 | |
| *** haomaiwa_ has joined #openstack-swift | 17:01 | |
| acoles | oh yeah its the weekend | 17:05 |
| *** marzif has joined #openstack-swift | 17:07 | |
| *** nickchase has joined #openstack-swift | 17:11 | |
| *** zhill has joined #openstack-swift | 17:14 | |
| *** klrmn1 has joined #openstack-swift | 17:16 | |
| *** haomaiwa_ has quit IRC | 17:16 | |
| notmyname | we've got a release coming up. you didn't expect to stop working this weekend, did you? ;-) | 17:21 |
| acoles | no sir | 17:23 |
| jrichli | lol | 17:24 |
| notmyname | I'd just like to point out that the truly important issues are being discussed on the mailing list: what's the allowed limit for git commit messages | 17:25 |
| acoles | notmyname: actually, let me know the answer because i might be about to violate that | 17:26 |
| notmyname | acoles: someone was complaining that 72 chars is the limit. so I'm guessing that's what it is. also, if someone doesn't like your commit message, they can easily edit it in the gerrit UI. there's even a dedicated button just for that | 17:27 |
| acoles | notmyname: oh line length. i thought you meant number o flines | 17:27 |
| notmyname | of course line length. what could possibly be more important than that?! ;-) | 17:28 |
| *** silor1 has joined #openstack-swift | 17:30 | |
| *** silor has quit IRC | 17:33 | |
| *** silor1 is now known as silor | 17:33 | |
| *** dustins|afk is now known as dustins | 17:41 | |
| *** mahatic has quit IRC | 17:53 | |
| *** proteusguy_ has joined #openstack-swift | 17:58 | |
| *** proteusguy has quit IRC | 18:00 | |
| *** agomez has joined #openstack-swift | 18:00 | |
| *** annegentle has quit IRC | 18:08 | |
| *** geaaru has quit IRC | 18:09 | |
| *** pberis has quit IRC | 18:10 | |
| *** trifon has joined #openstack-swift | 18:18 | |
| *** annegentle has joined #openstack-swift | 18:19 | |
| CaioBrentano | Hi! | 18:21 |
| *** silor1 has joined #openstack-swift | 18:21 | |
| CaioBrentano | "The erasure code support in Swift is considered “beta” at this point. Most major functionality is included, but it has not been tested or validated at large scale." | 18:22 |
| CaioBrentano | Is there any release date for production? :-) | 18:23 |
| *** marzif has quit IRC | 18:24 | |
| *** silor has quit IRC | 18:24 | |
| torgomatic | "when it's ready" :) | 18:25 |
| *** silor1 has quit IRC | 18:25 | |
| CaioBrentano | so there is no "official goal" for this feature? | 18:30 |
| torgomatic | in terms of dates? not that I've heard of | 18:32 |
| torgomatic | I mean, the alternative is that we wait for some OpenStack release and then say "it's production-ready because time has passed!" and people use it and it eats their data | 18:32 |
| notmyname | CaioBrentano: the "production readiness" of EC in swift, in my mind, is the point at which we are treating it like replication: there's always improvements, but there aren't bit outstanding known issues that have data corruption or loss as side-effects | 18:33 |
| CaioBrentano | thanks torgomatic and notmyname | 18:34 |
| notmyname | as of right now, we're working towards finishing the known big EC issues, but certainly not all EC issues, by the end of next week for the liberty openstack release | 18:34 |
| *** marcusvrn_ has quit IRC | 18:35 | |
| notmyname | we've got a lot to do in the next 7 days :-/ | 18:35 |
| CaioBrentano | I'm asking because we are having some "vendors invites" to test their Object Storage solution, and most of them mentions EC as an advantage compared to Swift | 18:38 |
| notmyname | yeah, I've seen some of that too | 18:40 |
| *** zhill has quit IRC | 18:40 | |
| *** marzif has joined #openstack-swift | 18:40 | |
| *** zhill has joined #openstack-swift | 18:40 | |
| * notmyname thinks some storage vendors will be in a tight place if their only competitive advantage goes away when swift adds a feature | 18:41 | |
| minwoob | acoles: are you there? | 18:42 |
| *** alejandrito has quit IRC | 18:42 | |
| acoles | minwoob: yes | 18:42 |
| minwoob | acoles: So first, I think 'cp -a' will copy over the xattrs -- writing the helper method is easy for this, but I haven't thought of how to write the tests for it yet. | 18:43 |
| minwoob | acoles: And also, regarding test_GET_wih_missing_and_range_unsatisfiable | 18:43 |
| minwoob | acoles: Could you explain why you've defined get_response inside that test? It seems that the get_response method that you want is the one associated with req, the swob.Request object. | 18:44 |
| *** esker has joined #openstack-swift | 18:44 | |
| minwoob | acoles: When I did it that way, I started getting 503's. | 18:45 |
| minwoob | Which seems to be going in the right direction. | 18:46 |
| *** tsubic has quit IRC | 18:47 | |
| *** dmsimard is now known as dmsimard|afk | 18:47 | |
| *** esker has quit IRC | 18:47 | |
| acoles | minwoob: the get_response i=method is passed to capture_http_requests, which uses it to fake out responses to requests. The same pattern is used in a lot of the tests in that class. There is no real object server for those test Requests to talk to. | 18:48 |
| *** esker has joined #openstack-swift | 18:48 | |
| acoles | re testing the copy function, there are a lot of tests in test_diskfile that create a diskfile, if you look there you'll find some clues to make two valid diskfiles in tow dirs, then try to copy the data file between the dirs with your copy function, and check you can open the diskfile | 18:49 |
| *** ctennis has quit IRC | 18:49 | |
| *** ctennis has joined #openstack-swift | 18:50 | |
| minwoob | acoles: Such a copy_xattrs_too() function would also need to verify that the xattrs have been preserved though, right? | 18:50 |
| minwoob | The test* for it, I mean. | 18:51 |
| *** jordanP has quit IRC | 18:51 | |
| acoles | minwoob: the diskfile will not open if the xattrs have gone missing, as in using DiskFile.open(), because the xattrs have all the metadata that DiskFile looks for when opening a diskfile | 18:52 |
| minwoob | acoles: Okay, just verifying. | 18:52 |
| acoles | minwoob: You could also try it with the kind of tests where I'd like to use it, similar to test.unit.proxy.test_server.TestECMismatchedFA#test_GET_with_missing_durables, but with a nodes_to_copy parameter. | 18:52 |
| *** _hrou_ has joined #openstack-swift | 18:53 | |
| *** ujjain has quit IRC | 18:53 | |
| *** ujjain- has joined #openstack-swift | 18:53 | |
| *** ujjain has joined #openstack-swift | 18:53 | |
| *** ujjain has joined #openstack-swift | 18:53 | |
| *** hrou has quit IRC | 18:55 | |
| *** annegentle has quit IRC | 18:55 | |
| acoles | minwoob: i gotta leave - good luck and thanks for working on it! | 19:03 |
| minwoob | acoles: Sure, np. Thanks for the info. | 19:04 |
| openstackgerrit | Alistair Coles proposed openstack/swift: Enable object server to return non-durable data https://review.openstack.org/215276 | 19:06 |
| acoles | peluse: kota ^^ with claygs new idea for frag preferences. peluse could you check i have the right bug numbers in commit message | 19:07 |
| acoles | whuber: minwoob fyi ^^ could be worth reviewing this version, although i'm sure there's plenty of scope to improve | 19:08 |
| acoles | wbhuber: ^^ | 19:08 |
| acoles | clayg: if you have little time please just read the commit message and tell if i am totally out of whack | 19:08 |
| acoles | good night | 19:09 |
| *** acoles is now known as acoles_ | 19:09 | |
| *** annegentle has joined #openstack-swift | 19:18 | |
| wbhuber | acoles: thx for the pointer. will have the time to review your version. | 19:24 |
| wbhuber | peluse: based on your ppt sent earlier this week, is setting the segment size to 640kb instead of 1mb the sweet spot for EC to handle objects larger than the size? | 19:24 |
| wbhuber | peluse: does specifying a particular EC scheme (i.e. 10+4) weigh the decision in tuning the segment size to 640kb or 1mb? | 19:25 |
| *** pberis has joined #openstack-swift | 19:33 | |
| *** esker has quit IRC | 19:36 | |
| *** esker has joined #openstack-swift | 19:37 | |
| *** esker has quit IRC | 19:37 | |
| *** thumpba has quit IRC | 19:44 | |
| *** jlhinson_ has joined #openstack-swift | 19:49 | |
| *** jlhinson has quit IRC | 19:52 | |
| *** trifon has quit IRC | 19:52 | |
| *** nickchase has left #openstack-swift | 20:00 | |
| peluse | wbhuber: the 640K was based on 10:14 and a proxy chunk size of 64K (64K * 10). I would presume but haven't tested that yeah if you used a different EC ratio and/or different proxy chunk size you could do better by adjusting segment size | 20:00 |
| peluse | wbhuber: well, not totally true. I did try with 1MB segment and 10:14 and showed it was slightly less performance than 640K | 20:01 |
| peluse | acoles_: yeah, will check | 20:01 |
| *** trifon has joined #openstack-swift | 20:07 | |
| wbhuber | peluse: does that also indicate when tuning EC ratio to say 4+2 - we'd use proxy chunk size of 160k (160k *4)? that would be an interesting experiment test | 20:07 |
| *** albertom has quit IRC | 20:07 | |
| peluse | wbhuber: I don't know that I'd go changing proxy chunk size as a first step, I'd probably adjust segment size (the new parm) and see how that works. So if default chunk is 64K and you have 4+2 I'd set segment to 64K * 4 and see how that works | 20:09 |
| *** marzif has quit IRC | 20:09 | |
| *** alejandrito has joined #openstack-swift | 20:11 | |
| minwoob | wbhuber: You wanted to know how the segment size is related to the threshold at which we should tier between EC and replication based on object size, correct? | 20:11 |
| *** breitz has quit IRC | 20:11 | |
| *** breitz has joined #openstack-swift | 20:12 | |
| minwoob | peluse: Do you have some input on that? | 20:12 |
| wbhuber | minwoob: yeah, i was thinking if we are talking segment size = threshold and 640k is the sweet spot for at least EC scheme of 10:14 - we'd tier all objects whose sizes are less than the threshold (segment size) to replication and the larger ones to EC if this makes any sense | 20:13 |
| wbhuber | minwoob: peluse: this has to be evidently internally handled under the covers as long as both policies exist in swift.conf | 20:15 |
| *** albertom has joined #openstack-swift | 20:15 | |
| *** dmsimard|afk is now known as dmsimard | 20:15 | |
| peluse | wbhuber: we're not trying to solve that problem in swift right now - anything related to auto-anything wrt tiering between repl and ec policies. we have the tier'ing bp's which are about the framework in general and then we need tomanually establish the right "rules" for when based on manual testing, etc | 20:18 |
| peluse | wbhuber: right now we're just looking to identify some perf guidelines for EC as a standalone policy | 20:18 |
| peluse | wbhuber: and comparing it to a baseline of repl is a good way to do that... | 20:19 |
| wbhuber | peluse: absolutely you're right - this auto-tiering would be introduced in the "Swift EC 2.0" some kind of version. | 20:20 |
| wbhuber | peluse: tiering bp's? | 20:20 |
| * torgomatic isn't smart enough for auto-tiering | 20:21 | |
| wbhuber | torgomatic, if it would ever exist, you've to be smart to be able to review it! =] | 20:22 |
| torgomatic | wbhuber: let me know how that works out for you :p | 20:22 |
| peluse | so there are bp's up there.... | 20:22 |
| peluse | shiva from netapp I think had the last update | 20:23 |
| peluse | i'll ping him and see if he's coming to tokyo | 20:23 |
| *** rohit_ has joined #openstack-swift | 20:23 | |
| *** ccavanna has quit IRC | 20:26 | |
| *** cutforth has quit IRC | 20:35 | |
| wbhuber | peluse: minwoob: i believe bp you were pointing to is from the third bullet in this old trello post: https://trello.com/c/QWaYXWNf/120-small-file-optimizations | 20:35 |
| wbhuber | peluse: clearly, the third bullet states the m ring for objects < segment size is directing to the replication path | 20:36 |
| *** agomez has quit IRC | 20:37 | |
| *** alejandrito has quit IRC | 20:41 | |
| *** garthb has joined #openstack-swift | 20:44 | |
| *** alejandrito has joined #openstack-swift | 20:46 | |
| jrichli | torgomatic: I am working on the support for footers in the Replication path. I wanted to check my understanding of the conversation on the wire. | 20:51 |
| jrichli | togomatic: I saw the EC conversation specified on the patch that added it | 20:51 |
| jrichli | togomatic: and so I have modified that to be what I think I am striving for with this solution. could you check out http://paste.openstack.org/show/474069/? | 20:52 |
| jrichli | torgomatic: sorry about the handle misspell ... | 20:53 |
| torgomatic | jrichli: uh, let me go read some code... I think there should be a header in the proxy -> obj PUT request that says "hey, here comes some MIME" | 20:54 |
| jrichli | great, thx! | 20:55 |
| torgomatic | yeah,X-Backend-Obj-Metadata-Footer | 20:55 |
| jrichli | oh, right. i didn't add the new ones on there. i just took out what i didn't need ... but its good to know exactly what i need to add! thx | 20:56 |
| torgomatic | np | 20:56 |
| *** breitz has quit IRC | 21:01 | |
| *** hezhiqiang has joined #openstack-swift | 21:02 | |
| *** zhill has quit IRC | 21:02 | |
| *** zhill has joined #openstack-swift | 21:02 | |
| *** dustins has quit IRC | 21:03 | |
| notmyname | python. how does it work | 21:06 |
| jrichli | it's sneaky | 21:16 |
| *** openstackgerrit has quit IRC | 21:16 | |
| *** openstackgerrit has joined #openstack-swift | 21:16 | |
| *** trifon has quit IRC | 21:17 | |
| *** jrichli has quit IRC | 21:22 | |
| *** pberis has quit IRC | 21:31 | |
| *** jlhinson_ has quit IRC | 21:32 | |
| *** NM has quit IRC | 21:35 | |
| notmyname | the python lesson I just learned is that you should never yield inside a finally clause | 21:41 |
| notmyname | ...because it causes an error (Exception RuntimeError: 'generator ignored GeneratorExit' in <generator object foo at 0x10d2a5aa0> ignored): https://gist.github.com/notmyname/bed03c9796bc9dcada1b | 21:47 |
| *** annegentle has quit IRC | 22:01 | |
| *** openstackgerrit has quit IRC | 22:01 | |
| *** openstackgerrit has joined #openstack-swift | 22:01 | |
| *** Kennan_Vacation2 has joined #openstack-swift | 22:02 | |
| *** Kennan_Vacation has quit IRC | 22:02 | |
| peluse | wbhuber: yeah that was a brainstorming conversation... | 22:02 |
| *** hezhiqiang has quit IRC | 22:03 | |
| *** wbhuber has quit IRC | 22:07 | |
| *** pberis has joined #openstack-swift | 22:08 | |
| *** CaioBrentano has quit IRC | 22:08 | |
| *** vinsh_ has joined #openstack-swift | 22:15 | |
| *** _hrou_ has quit IRC | 22:15 | |
| *** alejandrito has quit IRC | 22:16 | |
| *** vinsh has quit IRC | 22:19 | |
| *** trifon has joined #openstack-swift | 22:19 | |
| *** vinsh_ has quit IRC | 22:20 | |
| openstackgerrit | John Dickinson proposed openstack/swift: Fix missing container update https://review.openstack.org/186735 | 22:23 |
| *** jkugel1 has quit IRC | 22:25 | |
| *** mac_ified has joined #openstack-swift | 22:30 | |
| *** openstack has joined #openstack-swift | 22:34 | |
| *** chsc has quit IRC | 22:39 | |
| *** albertom is now known as albertom[afk] | 22:43 | |
| *** openstackstatus has quit IRC | 22:47 | |
| *** openstack has joined #openstack-swift | 22:49 | |
| *** hezhiqiang has joined #openstack-swift | 23:04 | |
| *** hezhiqiang has quit IRC | 23:08 | |
| *** vinsh has joined #openstack-swift | 23:17 | |
| *** trifon has quit IRC | 23:22 | |
| *** zhill has quit IRC | 23:32 | |
| *** marzif has joined #openstack-swift | 23:34 | |
| *** gyee has quit IRC | 23:35 | |
| *** hrou has joined #openstack-swift | 23:39 | |
| openstackgerrit | Merged openstack/swift: Only yield the pending segment on error if it's a SegmentListingError https://review.openstack.org/227569 | 23:40 |
| *** garthb has quit IRC | 23:46 | |
| *** minwoob has quit IRC | 23:48 | |
| *** vinsh has quit IRC | 23:57 | |
| *** NM has joined #openstack-swift | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!