| *** lpabon has joined #openstack-swift | 00:03 | |
| tnewsome | Hmm... bulk delete is barely faster than one request per delete. | 00:16 |
|---|---|---|
| *** mwstorer has quit IRC | 00:16 | |
| clayg | tnewsome: yeah i mean I don't think bulk delete does any sort of concurrency... | 00:17 |
| tnewsome | clayg: But I had assumed that just the per-request round-trip time between client and server would be causing issues. | 00:18 |
| clayg | tnewsome: are all of the objects in the same container? cause that sqlitedb is just only gunna support so many updates a second... the .pending helps - but eventually that log has to get commited and it just takes as long as it takes | 00:18 |
| tnewsome | clayg: Yes, all in the same container. | 00:18 |
| clayg | well... maybe? but if it's just round trip time you can beat that with more concurrency on the client | 00:18 |
| tnewsome | clayg: Are there any performance numbers on what I might expect out of that? | 00:18 |
| clayg | are your db's on sata (maybe 150/s) or ssds (300-400/s maybe?) | 00:19 |
| tnewsome | clayg: Right, but now I can delete a batch of 1k objects in about 4 minutes. The round trip time is definitely negligible at that point. | 00:19 |
| tnewsome | clayg: They're on the cheapest big disk I could find. :-) | 00:19 |
| tnewsome | This speed will be fast enough for my application, it's just not as fast as I'd expected. | 00:19 |
| clayg | wait so you're only getting 4 deletes a second? | 00:20 |
| tnewsome | No, about 270 deletes/second. | 00:20 |
| tnewsome | Or is it per minute. I forget again. | 00:21 |
| tnewsome | 4.6/second. | 00:21 |
| clayg | well if 1k takes 4 mins, that's about 250 minute - which is quite low by most measures - but maybe if not if you're waiting... | 00:21 |
| clayg | yeah, so I mean you could do quite a bit better than that with some concurrency | 00:21 |
| tnewsome | I'll give it a shot. Thanks. | 00:22 |
| *** nmap911 has quit IRC | 00:37 | |
| *** serverascode has quit IRC | 00:42 | |
| *** serverascode has joined #openstack-swift | 00:43 | |
| *** shri has quit IRC | 00:53 | |
| *** csd has quit IRC | 00:58 | |
| *** bill_az has quit IRC | 01:05 | |
| *** shakamunyi has joined #openstack-swift | 01:05 | |
| *** zhiyan_ is now known as zhiyan | 01:15 | |
| *** byeager has joined #openstack-swift | 01:15 | |
| *** shakayumi has joined #openstack-swift | 01:15 | |
| clayg | rebase! | 01:17 |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Keep status_changed_at up-to-date with status changes. https://review.openstack.org/86714 | 01:17 |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Extend direct_client https://review.openstack.org/88144 | 01:17 |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Add reconciler probetest outline https://review.openstack.org/90934 | 01:17 |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Add Storage Policies to container.backend https://review.openstack.org/90935 | 01:18 |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Update FakeLogger with better log level handling https://review.openstack.org/90932 | 01:18 |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Add write_fake_ring helper to test.unit https://review.openstack.org/88151 | 01:18 |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Add container-reconciler daemon https://review.openstack.org/85908 | 01:18 |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Make FakeRing a use more real code https://review.openstack.org/90933 | 01:18 |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Add LRUCache to common.utils https://review.openstack.org/90938 | 01:18 |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Extend interface on InternalClient https://review.openstack.org/88152 | 01:18 |
| *** ams0 has joined #openstack-swift | 01:18 | |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Add Storage Policy to container put/delete object https://review.openstack.org/90936 | 01:18 |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Enqueue misplaced objects during container replication https://review.openstack.org/90937 | 01:18 |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Merge container storage_policy_index values. https://review.openstack.org/86720 | 01:18 |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Expose container info on deleted containers. https://review.openstack.org/86708 | 01:18 |
| *** shakamunyi has quit IRC | 01:18 | |
| *** saschpe has quit IRC | 01:21 | |
| *** ams0 has quit IRC | 01:22 | |
| portante | nice! | 01:23 |
| *** saschpe has joined #openstack-swift | 01:23 | |
| clayg | peluse_: fwiw on the account reaper branch, i was leaving those notes for myself as much as anything - i'm going to wait and see if https://review.openstack.org/#/c/87078/ will finally merge and then rebase https://review.openstack.org/#/c/88098 on that | 01:30 |
| openstackgerrit | Yuan Zhou proposed a change to openstack/swift: Adding deprecate policy support https://review.openstack.org/86827 | 02:09 |
| *** ams0 has joined #openstack-swift | 02:18 | |
| *** ams0 has quit IRC | 02:23 | |
| *** bach has quit IRC | 02:23 | |
| *** zhiyan is now known as zhiyan_ | 02:37 | |
| *** yuan has joined #openstack-swift | 02:45 | |
| *** shakayumi has quit IRC | 02:58 | |
| openstackgerrit | A change was merged to openstack/swift: cleanup policy parsing a bit more https://review.openstack.org/87078 | 03:03 |
| *** bach has joined #openstack-swift | 03:05 | |
| *** csd has joined #openstack-swift | 03:07 | |
| *** ams0 has joined #openstack-swift | 03:19 | |
| *** ams0 has quit IRC | 03:23 | |
| *** byeager has quit IRC | 03:34 | |
| *** ppai has joined #openstack-swift | 03:59 | |
| *** nshaikh has joined #openstack-swift | 04:05 | |
| *** tzumainn has quit IRC | 04:14 | |
| *** haomaiwang has joined #openstack-swift | 04:16 | |
| *** csd has quit IRC | 04:17 | |
| *** ams0 has joined #openstack-swift | 04:19 | |
| *** chandan_kumar has joined #openstack-swift | 04:20 | |
| *** haomaiwang has quit IRC | 04:20 | |
| *** ams0 has quit IRC | 04:24 | |
| *** haomaiwa_ has joined #openstack-swift | 04:26 | |
| *** chandan_kumar is now known as chandankumar | 04:26 | |
| *** nshaikh has quit IRC | 04:28 | |
| *** haomaiwa_ has quit IRC | 04:43 | |
| *** ams0 has joined #openstack-swift | 04:47 | |
| *** ams0 has quit IRC | 04:52 | |
| *** ams0 has joined #openstack-swift | 04:54 | |
| *** csd has joined #openstack-swift | 05:01 | |
| *** gyee has quit IRC | 05:05 | |
| *** Trixboxer has joined #openstack-swift | 05:08 | |
| *** shri has joined #openstack-swift | 05:14 | |
| *** cheri has joined #openstack-swift | 05:15 | |
| *** haomaiwa_ has joined #openstack-swift | 05:23 | |
| chmouel | portante: i can look that if you want, | 05:26 |
| chmouel | portante: but I am not sure of where you at currently | 05:26 |
| *** ams0 has quit IRC | 05:28 | |
| *** ams0 has joined #openstack-swift | 05:28 | |
| *** ams0 has quit IRC | 05:34 | |
| *** csd has quit IRC | 05:37 | |
| *** kevinc_ has joined #openstack-swift | 05:43 | |
| *** zaitcev has quit IRC | 05:53 | |
| *** kevinc_ has quit IRC | 05:54 | |
| *** ams0 has joined #openstack-swift | 05:57 | |
| *** psharma has joined #openstack-swift | 06:02 | |
| *** cheri has quit IRC | 06:03 | |
| *** haomaiwa_ has quit IRC | 06:07 | |
| *** haomaiwang has joined #openstack-swift | 06:07 | |
| *** ams0 has quit IRC | 06:08 | |
| *** ams0 has joined #openstack-swift | 06:11 | |
| *** csd has joined #openstack-swift | 06:20 | |
| *** haomai___ has joined #openstack-swift | 06:23 | |
| *** csd has quit IRC | 06:25 | |
| *** haomaiwang has quit IRC | 06:26 | |
| *** bach has quit IRC | 06:28 | |
| *** sileht has quit IRC | 06:30 | |
| *** ams0 has quit IRC | 06:30 | |
| *** sileht has joined #openstack-swift | 06:34 | |
| openstackgerrit | Yuan Zhou proposed a change to openstack/swift: Allow to deprecate policy https://review.openstack.org/86827 | 06:36 |
| *** StevenK has quit IRC | 06:44 | |
| *** StevenK has joined #openstack-swift | 06:48 | |
| *** mmcardle has joined #openstack-swift | 07:10 | |
| *** foexle has joined #openstack-swift | 07:14 | |
| *** mmcardle has quit IRC | 07:20 | |
| *** fbo is now known as fbo_away | 07:20 | |
| *** mmcardle has joined #openstack-swift | 07:41 | |
| *** nacim has joined #openstack-swift | 07:45 | |
| *** cheri has joined #openstack-swift | 07:53 | |
| *** acoles_away is now known as acoles | 07:57 | |
| *** zanc has joined #openstack-swift | 08:03 | |
| *** ams0 has joined #openstack-swift | 08:15 | |
| *** mkollaro has joined #openstack-swift | 08:22 | |
| *** mkerrin has quit IRC | 08:22 | |
| *** ekarlso has quit IRC | 08:29 | |
| *** ekarlso has joined #openstack-swift | 08:29 | |
| *** ams0 has quit IRC | 08:48 | |
| *** nshaikh has joined #openstack-swift | 08:50 | |
| *** pconstantine_ has joined #openstack-swift | 08:55 | |
| *** ams0 has joined #openstack-swift | 08:57 | |
| *** mkerrin has joined #openstack-swift | 08:59 | |
| *** AbyssOne has joined #openstack-swift | 09:08 | |
| *** ams0 has quit IRC | 09:13 | |
| *** ams0 has joined #openstack-swift | 09:13 | |
| *** ams0 has quit IRC | 09:17 | |
| *** nshaikh has quit IRC | 09:20 | |
| *** mkollaro has quit IRC | 09:23 | |
| *** mkollaro has joined #openstack-swift | 09:24 | |
| portante | chmouel: thanks, clarkb stepped in a helped me out | 09:38 |
| portante | chmouel: arguably, the existing functional tests should broken out into separate jobs, right now v1 tests only run after v2 tests complete (not sure if a failure of the first run prevents a run of the next) | 09:39 |
| chmouel | portante: I did like that at first but the infra folks asked me to put on one job so not to use resources | 09:40 |
| chmouel | portante: great that it works :) | 09:40 |
| portante | chmouel: hmm | 09:42 |
| *** corrigac has joined #openstack-swift | 09:43 | |
| *** tzumainn has joined #openstack-swift | 09:49 | |
| *** ccorrigan has quit IRC | 09:50 | |
| *** AbyssOne has quit IRC | 09:50 | |
| *** AbyssOne has joined #openstack-swift | 09:51 | |
| *** fbo_away is now known as fbo_away_away | 10:10 | |
| portante | chmouel: yes, I just looked at a failing swift functional test job, and failure of the first job prevents failure of the second | 10:30 |
| portante | I think we should at least always run the second regardless of the first, but since functional tests leave some state change with the cluster, it means the second test run is potentially affected by the first | 10:31 |
| *** bvandenh has joined #openstack-swift | 10:46 | |
| *** bvandenh has quit IRC | 10:52 | |
| *** bvandenh has joined #openstack-swift | 10:52 | |
| *** bvandenh has quit IRC | 10:53 | |
| *** bvandenh has joined #openstack-swift | 10:54 | |
| *** bvandenh has quit IRC | 10:54 | |
| *** bvandenh has joined #openstack-swift | 10:55 | |
| *** bvandenh has quit IRC | 10:57 | |
| *** bvandenh has joined #openstack-swift | 10:57 | |
| *** nshaikh has joined #openstack-swift | 11:03 | |
| *** bvandenh has quit IRC | 11:05 | |
| *** psharma has quit IRC | 11:09 | |
| *** acoles is now known as acoles_away | 11:12 | |
| jokke_ | portante: yup, cannot trust that it will be clean run | 11:16 |
| *** bvandenh has joined #openstack-swift | 11:18 | |
| chmouel | portante: yes that was the reasoning when we added that | 11:27 |
| *** pconstantine_ has quit IRC | 11:27 | |
| *** otoolee has joined #openstack-swift | 11:32 | |
| *** mmcardle has quit IRC | 11:38 | |
| *** nacim has quit IRC | 11:39 | |
| *** ppai has quit IRC | 11:46 | |
| *** piousbox__ has joined #openstack-swift | 11:49 | |
| *** wer has quit IRC | 11:49 | |
| *** piousbox_ has quit IRC | 11:49 | |
| *** wer has joined #openstack-swift | 11:51 | |
| *** nacim has joined #openstack-swift | 11:55 | |
| *** mmcardle has joined #openstack-swift | 11:56 | |
| *** mkollaro has quit IRC | 12:01 | |
| *** mkollaro has joined #openstack-swift | 12:03 | |
| *** jamie_h has joined #openstack-swift | 12:17 | |
| *** wer has quit IRC | 12:20 | |
| *** wer has joined #openstack-swift | 12:20 | |
| *** acoles_away is now known as acoles | 12:26 | |
| *** wer has quit IRC | 12:30 | |
| *** wer has joined #openstack-swift | 12:31 | |
| *** mkollaro1 has joined #openstack-swift | 12:33 | |
| *** mkollaro has quit IRC | 12:33 | |
| *** mmcardle1 has joined #openstack-swift | 12:38 | |
| *** Honghui has joined #openstack-swift | 12:40 | |
| Honghui | Hello, Any can recommend hardware for swift. We test the Dell R720xd, but with it's H310 card, the performance is terrible. | 12:42 |
| *** haomaiwa_ has joined #openstack-swift | 12:44 | |
| *** haomai___ has quit IRC | 12:45 | |
| *** mmcardle1 has quit IRC | 12:47 | |
| *** wer has quit IRC | 12:48 | |
| *** mmcardle has quit IRC | 12:48 | |
| *** mmcardle has joined #openstack-swift | 12:48 | |
| portante | can you share more details about your setup? | 12:51 |
| *** wer has joined #openstack-swift | 12:55 | |
| jokke_ | Honghui: really depends of your needs. If you want, you can PM me and I'll connect you with our folks | 12:55 |
| *** haomai___ has joined #openstack-swift | 12:57 | |
| *** haomaiwa_ has quit IRC | 12:57 | |
| *** fbo_away_away is now known as fbo_away_away_aw | 13:05 | |
| *** nacim has quit IRC | 13:05 | |
| *** nacim has joined #openstack-swift | 13:06 | |
| *** zul has joined #openstack-swift | 13:07 | |
| *** pberis has joined #openstack-swift | 13:17 | |
| *** cheri has quit IRC | 13:19 | |
| *** nshaikh has quit IRC | 13:20 | |
| *** changbl has quit IRC | 13:24 | |
| openstackgerrit | A change was merged to openstack/python-swiftclient: Add functional tests for python-swiftclient https://review.openstack.org/76355 | 13:25 |
| *** nacim has quit IRC | 13:33 | |
| *** chandankumar is now known as chandankumar|afk | 13:37 | |
| *** NM1 has joined #openstack-swift | 13:49 | |
| *** zanc has quit IRC | 13:50 | |
| *** nacim has joined #openstack-swift | 13:51 | |
| *** zanc has joined #openstack-swift | 13:51 | |
| *** tdasilva has joined #openstack-swift | 13:54 | |
| *** Trixboxer has quit IRC | 13:54 | |
| *** bill_az has joined #openstack-swift | 14:13 | |
| *** NM1 has quit IRC | 14:16 | |
| *** NM1 has joined #openstack-swift | 14:20 | |
| *** haomai___ has quit IRC | 14:32 | |
| *** haomaiwa_ has joined #openstack-swift | 14:32 | |
| *** bach has joined #openstack-swift | 14:35 | |
| *** kevinc_ has joined #openstack-swift | 14:51 | |
| *** shakayumi has joined #openstack-swift | 14:51 | |
| openstackgerrit | paul luse proposed a change to openstack/swift: Add Storage Policy support to the Account Reaper https://review.openstack.org/88098 | 14:51 |
| openstackgerrit | Peter Portante proposed a change to openstack/swift: Add a env var to use in-memory obj server in func https://review.openstack.org/92381 | 14:56 |
| openstackgerrit | Christian Schwede proposed a change to openstack/python-swiftclient: Allow get_account and get_container to return an iterator https://review.openstack.org/74845 | 14:57 |
| openstackgerrit | Peter Portante proposed a change to openstack/swift: Add a env var to use in-memory obj server in func https://review.openstack.org/92381 | 14:58 |
| *** zhiyan_ is now known as zhiyan | 15:01 | |
| *** jergerber has joined #openstack-swift | 15:04 | |
| notmyname | good morning world | 15:05 |
| portante | notmyname: good morning | 15:06 |
| creiht | notmyname: I was hoping to have some 3d printed swift emlems printed by the summit | 15:08 |
| creiht | but turns out it is a bit more difficult than I expected | 15:08 |
| notmyname | creiht: doing it vertically or laying down? | 15:10 |
| creiht | laying down | 15:10 |
| creiht | well the difficult part is turing the image into something that is 3d printable | 15:11 |
| creiht | :) | 15:11 |
| *** Honghui has quit IRC | 15:16 | |
| luisbg | notmyname, morning :) | 15:16 |
| NM1 | Good morning :) | 15:17 |
| NM1 | Who else is going to the summit? :) | 15:17 |
| notmyname | creiht: I'm working on tracking down versioning stuff for you. did you see the ML thread? seems that version numbers use pip ordering, not debian ordering (which is different). still tracking stuff down... | 15:18 |
| creiht | yeah I saw that | 15:20 |
| creiht | notmyname: we worked around it | 15:20 |
| notmyname | creiht: good. but the bigger process issue I'm trying to fix now | 15:20 |
| creiht | I mainly wanted to bring it up so that others wouldn't have a similar issue | 15:20 |
| creiht | yeah thanks for bringing it up | 15:21 |
| creiht | It also saddens me that pip is the defacto deployment strategy | 15:21 |
| notmyname | yes, me too | 15:21 |
| creiht | at least what everything defaults to | 15:21 |
| notmyname | and people wonder why ops are so frustrated with openstack | 15:22 |
| creiht | and leads to why we have some of this crazyness | 15:22 |
| *** krtaylor has joined #openstack-swift | 15:23 | |
| notmyname | NM1: I'll be at the summit | 15:24 |
| *** bvandenh has quit IRC | 15:24 | |
| *** haomaiwa_ has quit IRC | 15:29 | |
| *** haomaiwa_ has joined #openstack-swift | 15:30 | |
| *** adrian_otto has joined #openstack-swift | 15:30 | |
| portante | NM1: o/ | 15:31 |
| *** shakayumi has quit IRC | 15:31 | |
| *** bach has quit IRC | 15:31 | |
| adrian_otto | hey guys. Does anyone know if a "400 Header Line Too Long" error in swiftclient when using an keystone v3 auth token… if that's a known problem with an existing bug/patch? | 15:32 |
| adrian_otto | example: http://logs.openstack.org/71/92171/4/check/gate-solum-devstack-dsvm/6c9f39b/logs/screen-g-api.txt.gz#_2014-05-06_01_09_46_089 | 15:32 |
| *** shakamunyi has joined #openstack-swift | 15:32 | |
| adrian_otto | from what I can tell, the only header being sent there is the X-Auth-Token header, so I suppose it's safe to conclude that the token length is longer than expected? | 15:33 |
| *** zul has quit IRC | 15:34 | |
| notmyname | adrian_otto: yes, keystone pki headers are _huge_ | 15:34 |
| *** haomaiwa_ has quit IRC | 15:34 | |
| notmyname | adrian_otto: and you can set the header limit in the swift configs | 15:35 |
| notmyname | adrian_otto: but swift defaults to 4k headers | 15:35 |
| notmyname | adrian_otto: which implies that your auth header alone is 4k+ (keep that in mind when thinking about per-request overhead for small objects) | 15:35 |
| adrian_otto | well this is Glance that's using swift on the back end to store an image | 15:36 |
| adrian_otto | we are just using the auth token that keystone gave to us | 15:36 |
| notmyname | adrian_otto: doesn't matter who the client is. it's keystone that is giving you the big token | 15:36 |
| adrian_otto | and based on the content of the log, it does not appear to be 4k to me. | 15:36 |
| adrian_otto | let me check that | 15:37 |
| adrian_otto | it is 8371 characters. Terrific. | 15:38 |
| notmyname | wow | 15:39 |
| adrian_otto | any chance you know which configuration setting to adjust? | 15:39 |
| adrian_otto | hopefully we can drop that into our devstack config | 15:39 |
| adrian_otto | as a local.conf setting or something | 15:40 |
| notmyname | adrian_otto: max_header_size in swift.conf | 15:41 |
| notmyname | adrian_otto: oh, and it defaults to 8k, not 4k | 15:41 |
| adrian_otto | awesome, thanks notmyname | 15:41 |
| notmyname | adrian_otto: also, it's exposed in a request to /info, so you can always query the target swift cluster to see what's supported | 15:41 |
| adrian_otto | I'll submit a swift bug to suggest adjusting the default value to accommodate keystone v3 | 15:41 |
| notmyname | adrian_otto: no. that would be rejected | 15:41 |
| adrian_otto | why? | 15:42 |
| * notmyname goes looking for the last patch to do that... | 15:42 | |
| adrian_otto | we can't have a keystone and a swift that don't work together | 15:42 |
| acoles | notmyname: https://review.openstack.org/#/c/66615/ <- that it? | 15:43 |
| adrian_otto | if we plan to continue to use X-Auth-Token headers, then we either need keystone to make shorter auth tokens, or swift to accept longer ones | 15:43 |
| adrian_otto | ok, that is merged. | 15:44 |
| notmyname | adrian_otto: yes that one. but note it didn't change things | 15:44 |
| notmyname | adrian_otto: thanks | 15:44 |
| notmyname | acoles: thanks | 15:44 |
| acoles | notmyname: adrian_otto: default is 8192 https://github.com/openstack/swift/blob/master/etc/swift.conf-sample#L51 | 15:45 |
| notmyname | adrian_otto: the keystone token is unbounded in length. which means that we actually don't have any number we could set as the default that would make it "just work" | 15:45 |
| creiht | heh | 15:45 |
| creiht | adrian_otto: it is really a design flaw of keystone | 15:45 |
| adrian_otto | oh, dear. | 15:45 |
| creiht | the discussion has gone round and round forever | 15:45 |
| adrian_otto | sorry to open an old wound guys | 15:46 |
| creiht | lol | 15:46 |
| creiht | not a wond | 15:46 |
| creiht | wound | 15:46 |
| notmyname | adrian_otto: no, it's not painful | 15:46 |
| creiht | sending around 8K headers for every request is not my idea of good design :) | 15:47 |
| *** zhiyan is now known as zhiyan_ | 15:47 | |
| adrian_otto | definitely not | 15:47 |
| creiht | adrian_otto: btw, there is a way to set keystone to use the old style uuid tokens | 15:48 |
| creiht | well at least there was | 15:48 |
| creiht | I hope they didn't get rid of that for v3 | 15:48 |
| notmyname | creiht: I'm not sure if there is in v3 | 15:48 |
| notmyname | I don't know | 15:48 |
| adrian_otto | oh? | 15:48 |
| notmyname | adrian_otto: set keystone to use uuid tokens | 15:48 |
| adrian_otto | I will check with that team to see what advice they can offer | 15:48 |
| adrian_otto | ok, thanks for the pointers notmyname creiht and acoles. This was extremely helpful! | 15:49 |
| creiht | adrian_otto: yeah, it requires swift to make validation calls | 15:49 |
| notmyname | adrian_otto: small tokens, but then you have to check with token for every uncached uuid token. so higher load on the keystone service | 15:49 |
| creiht | to keystone though | 15:49 |
| adrian_otto | makes sense | 15:49 |
| Alex_Gaynor | Elliptic curves :-( | 15:50 |
| notmyname | creiht: adrian_otto: also, you should know that keystone v3 changes something that was globally unique pre-v3 into something not globally unique. | 15:50 |
| creiht | adrian_otto: the reason the default tokens are so big is because the token is basically an encrypted/signed version of all of your endpoints | 15:50 |
| *** bach has joined #openstack-swift | 15:50 | |
| notmyname | creiht: adrian_otto: and these things are used in ACLs, potentially exposing a lot of data | 15:50 |
| notmyname | acoles is the one who found this | 15:50 |
| creiht | notmyname: yeah this discussion was had a long time ago :( | 15:50 |
| creiht | of course they didn't listen :( | 15:50 |
| acoles | adrian_otto: https://github.com/openstack/keystone/blob/master/etc/keystone.conf.sample#L1266 where you can config keystone to use UUID tokens, but I ran into problems getting it to work today with v3 (serendipty!). YMMV I hope | 15:51 |
| notmyname | adrian_otto: suffice to say that swift+keystone v3 is somewhat scary for deployers. or at least you have to know a lot about the internals of the system (catalog size, identity schemas) before doing it | 15:51 |
| *** kevinc_ has quit IRC | 15:52 | |
| notmyname | acoles: do you have a link to the non-unique account name thing? | 15:52 |
| acoles | notmyname: sec | 15:52 |
| *** bach has quit IRC | 15:53 | |
| acoles | notmyname: https://blueprints.launchpad.net/swift/+spec/keystone-v3-support | 15:53 |
| *** mwstorer has joined #openstack-swift | 15:53 | |
| notmyname | adrian_otto: but TBH, if you are just using swift as a glance target and only putting multi-GB objects, an 8k+ header isn't that big of a deal | 15:53 |
| acoles | ^^ just so happens i wrote a blueprint to gather v3 related issues, the acl thing bug report is linked off here | 15:53 |
| *** bach has joined #openstack-swift | 15:53 | |
| notmyname | see it | 15:54 |
| openstackgerrit | paul luse proposed a change to openstack/swift: Add Storage Policy support to the Account Reaper https://review.openstack.org/88098 | 15:54 |
| *** changbl has joined #openstack-swift | 15:54 | |
| notmyname | adrian_otto: you should read over that blueprint and also the linked bug report https://bugs.launchpad.net/swift/+bug/1299146 | 15:54 |
| acoles | notmyname: i need to push my patches swiftclient and keystoneauth patches again to get them linked from the blueprint | 15:55 |
| *** kevinc_ has joined #openstack-swift | 15:55 | |
| adrian_otto | acoles: tx!! | 15:56 |
| creiht | acoles: thanks for diging into all of that | 15:56 |
| creiht | I've been scared to look at the v3 support stuff | 15:57 |
| acoles | creiht: it has been a pleasure | 15:57 |
| acoles | creiht: er, not | 15:57 |
| creiht | lol | 15:57 |
| *** zul has joined #openstack-swift | 15:58 | |
| notmyname | lol | 15:58 |
| notmyname | adrian_otto: anything else we can help answer about keystone or swift here? | 15:59 |
| adrian_otto | notmyname: Nope, I'm all set for now. Looks like I have a short term workaround, and a long term fix to come back to. | 16:00 |
| notmyname | adrian_otto: great | 16:00 |
| acoles | adrian_otto: if you get UUID tokens working with keystone v3 could you ping me back. would motivate me to try again. i probably have misconfig somewhere. thx | 16:00 |
| *** acoles is now known as acoles_away | 16:03 | |
| adrian_otto | we jsut merged this as a workaround: https://review.openstack.org/92354 | 16:04 |
| adrian_otto | I will file a tech debt bug on our end to revisit UUID tokens so we will revisit that at a later time | 16:04 |
| adrian_otto | I will make a note in that bug to ping you | 16:04 |
| notmyname | ok | 16:05 |
| notmyname | pbr version numbers use the http://legacy.python.org/dev/peps/pep-0440/ format | 16:07 |
| *** zul has quit IRC | 16:07 | |
| notmyname | bah, somthing else coming up first before I get into that (but there will be some changes in the way the swift version is tracked, i think) | 16:08 |
| *** zul has joined #openstack-swift | 16:13 | |
| *** NM1 has quit IRC | 16:20 | |
| *** ams0 has joined #openstack-swift | 16:22 | |
| *** gyee has joined #openstack-swift | 16:30 | |
| *** ams0 has quit IRC | 16:32 | |
| peluse_ | portante: are the gerrit 'gate-swift-tox-func' supposed to be passing now? | 16:39 |
| *** zhiyan_ is now known as zhiyan | 16:42 | |
| *** shakamunyi has quit IRC | 16:42 | |
| *** ams0 has joined #openstack-swift | 16:43 | |
| *** elambert has joined #openstack-swift | 16:46 | |
| portante | ah, umm, err, oh, hmm | 16:48 |
| portante | the feature/ec branch needs the tox.ini file changes with the "func" environment | 16:49 |
| notmyname | peluse_: that phrase "supposed to be" makes it easy :-) | 16:50 |
| notmyname | portante: peluse_: merge from master should clear it up, right? | 16:50 |
| peluse_ | portante: gotcha, I assume we'll get that on our next rebase | 16:50 |
| peluse_ | :) | 16:50 |
| portante | great | 16:50 |
| notmyname | merge, not rebase. important difference :-) | 16:50 |
| peluse_ | notmyname: you wanna fire that up? We're due about now anyway | 16:50 |
| notmyname | peluse_: I don't have time right now. I'm happy to, but I probably won't get to it today | 16:51 |
| peluse_ | notmyname: cool. I only have about 30 min of "free" time myself today, I'll try to take a look and see if there are conflicts I'll clean them up for you and if not will let you know and you can pull the trigger whenever convenient | 16:53 |
| *** zhiyan is now known as zhiyan_ | 16:53 | |
| *** nacim has quit IRC | 16:56 | |
| openstackgerrit | David Goetz proposed a change to openstack/swift: xLO bug with auth tokens expiring during download. https://review.openstack.org/92165 | 16:58 |
| *** NM1 has joined #openstack-swift | 16:58 | |
| *** gyee has quit IRC | 16:58 | |
| *** ams0 has quit IRC | 17:06 | |
| *** bach has quit IRC | 17:08 | |
| *** Midnightmyth has joined #openstack-swift | 17:14 | |
| notmyname | reminder that tomorrow's team meeting is for summit prep https://wiki.openstack.org/wiki/Meetings/Swift | 17:20 |
| *** byeager has joined #openstack-swift | 17:21 | |
| *** chandankumar|afk has quit IRC | 17:24 | |
| *** zul has quit IRC | 17:25 | |
| *** mmcardle has quit IRC | 17:32 | |
| *** cheri has joined #openstack-swift | 17:34 | |
| *** shri has joined #openstack-swift | 17:35 | |
| *** jamie_h has quit IRC | 17:44 | |
| portante | fwiw: the gate/check job for the in-process functional tests appears to be working | 17:52 |
| portante | notmyname: once we get master merged onto feature/ec, then we'll be able to make that gate/check job a proper voting one | 17:53 |
| notmyname | portante: ack | 17:53 |
| *** chandan_kumar has joined #openstack-swift | 17:58 | |
| *** byeager has quit IRC | 17:59 | |
| *** byeager has joined #openstack-swift | 18:02 | |
| *** cheri has quit IRC | 18:03 | |
| *** NM1 has quit IRC | 18:07 | |
| *** kevinc_ has quit IRC | 18:11 | |
| *** byeager has quit IRC | 18:15 | |
| *** bach has joined #openstack-swift | 18:16 | |
| *** mkollaro1 has quit IRC | 18:16 | |
| *** bach has quit IRC | 18:19 | |
| *** ams0 has joined #openstack-swift | 18:20 | |
| *** bach has joined #openstack-swift | 18:20 | |
| *** chandan_kumar has quit IRC | 18:22 | |
| *** bach_ has joined #openstack-swift | 18:23 | |
| *** bach has quit IRC | 18:23 | |
| *** zaitcev has joined #openstack-swift | 18:31 | |
| *** ChanServ sets mode: +v zaitcev | 18:31 | |
| *** ams0 has quit IRC | 18:42 | |
| *** byeager has joined #openstack-swift | 18:43 | |
| *** ams0 has joined #openstack-swift | 18:44 | |
| *** zul has joined #openstack-swift | 18:44 | |
| *** byeager has quit IRC | 18:51 | |
| *** byeager has joined #openstack-swift | 18:57 | |
| *** zul has quit IRC | 19:02 | |
| *** byeager has quit IRC | 19:04 | |
| *** zul has joined #openstack-swift | 19:06 | |
| *** ams0 has quit IRC | 19:27 | |
| *** ams0_ has joined #openstack-swift | 19:27 | |
| *** pberis has quit IRC | 19:30 | |
| openstackgerrit | A change was merged to openstack/swift: Add Storage Policy support to the Account Reaper https://review.openstack.org/88098 | 19:47 |
| notmyname | https://twitter.com/RunAbovecom/status/463760077728788480 | 20:02 |
| notmyname | CrackerJackMack: ^ | 20:02 |
| notmyname | creiht: ^ | 20:02 |
| notmyname | sorry CrackerJackMack | 20:02 |
| notmyname | anyone in here from RunAbove? | 20:04 |
| *** sunglasses has quit IRC | 20:07 | |
| *** ams0_ has quit IRC | 20:07 | |
| *** byeager has joined #openstack-swift | 20:08 | |
| *** ams0 has joined #openstack-swift | 20:17 | |
| *** NM1 has joined #openstack-swift | 20:21 | |
| *** gyee has joined #openstack-swift | 20:24 | |
| *** byeager_ has joined #openstack-swift | 20:26 | |
| *** byeager has quit IRC | 20:26 | |
| *** ams0 has quit IRC | 20:58 | |
| *** bach_ has quit IRC | 21:00 | |
| *** ams0 has joined #openstack-swift | 21:03 | |
| *** bach has joined #openstack-swift | 21:03 | |
| *** bach has quit IRC | 21:04 | |
| *** tdasilva has left #openstack-swift | 21:04 | |
| *** ams0 has quit IRC | 21:12 | |
| *** shri has quit IRC | 21:15 | |
| *** shri has joined #openstack-swift | 21:15 | |
| *** Midnightmyth has quit IRC | 21:19 | |
| *** ams0 has joined #openstack-swift | 21:19 | |
| *** foexle has quit IRC | 21:20 | |
| chmouel | notmyname: i have a twitter handle if you like | 21:33 |
| chmouel | notmyname: he should be involve with the swift install https://twitter.com/gierschv | 21:34 |
| chmouel | they have just launched, i would doubt they have 1PB (except if they include their swift install from their parent company product ovh called hubic http://www.ovhtelecom.fr/hubiC/) | 21:35 |
| notmyname | chmouel: I'm a little familiar with hubic (ie I've heard it exists) | 21:40 |
| notmyname | chmouel: looks like I'm going to talk to him on the phone tomorrow | 21:40 |
| chmouel | cool :) | 21:40 |
| *** ams0 has quit IRC | 21:42 | |
| peluse_ | notmyname: FYI there are some conflicts on merge to ec branch. Addressing now... will shoot a note when resolved/ready | 21:44 |
| notmyname | peluse_: thanks | 21:44 |
| *** kevinc_ has joined #openstack-swift | 21:45 | |
| *** ams0 has joined #openstack-swift | 21:46 | |
| *** bach has joined #openstack-swift | 21:47 | |
| notmyname | chmouel: answer is 1.7 billion https://twitter.com/RunAbovecom/status/463797346338238464 | 22:01 |
| chmouel | notmyname: i didn't think they were so popular | 22:01 |
| chmouel | notmyname: i mean i think they released their service like last week or two weeks ago | 22:01 |
| notmyname | april 8, according to twitter. so just about one month | 22:02 |
| chmouel | oh okay maybe i just heard about it | 22:02 |
| notmyname | 650 obj/sec for 30 days is pretty close to 1.7 billion | 22:03 |
| * notmyname just heard of it today | 22:03 | |
| clayg | neato | 22:03 |
| chmouel | well i guess they have a niche of some very heavy users cause even in france i haven't heard it before two weeks ago | 22:04 |
| *** bach has quit IRC | 22:06 | |
| *** zul has quit IRC | 22:07 | |
| *** zul has joined #openstack-swift | 22:09 | |
| *** NM1 has quit IRC | 22:10 | |
| *** NM1 has joined #openstack-swift | 22:14 | |
| *** NM1 has quit IRC | 22:19 | |
| *** ams0 has quit IRC | 22:20 | |
| *** zul has quit IRC | 22:23 | |
| *** ams0 has joined #openstack-swift | 22:26 | |
| *** byeager_ has quit IRC | 22:29 | |
| creiht | nice | 22:30 |
| *** bach has joined #openstack-swift | 22:30 | |
| *** changbl has quit IRC | 22:49 | |
| *** bach has quit IRC | 22:56 | |
| *** bach has joined #openstack-swift | 22:59 | |
| *** byeager has joined #openstack-swift | 23:00 | |
| *** bach has quit IRC | 23:00 | |
| *** byeager has quit IRC | 23:08 | |
| *** byeager has joined #openstack-swift | 23:12 | |
| *** jergerber has quit IRC | 23:14 | |
| *** bach has joined #openstack-swift | 23:27 | |
| *** bach has quit IRC | 23:36 | |
| *** kevinc_ has quit IRC | 23:42 | |
| *** elambert has quit IRC | 23:44 | |
| *** praveenkumar has quit IRC | 23:51 | |
| *** Edward-Zhang has joined #openstack-swift | 23:55 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!