| *** sams-gleb has quit IRC | 00:01 | |
| *** jistr has quit IRC | 00:03 | |
| *** jistr has joined #openstack-swift | 00:07 | |
| *** dmorita has quit IRC | 00:32 | |
| *** mingyu has joined #openstack-swift | 00:45 | |
| *** mingyu has quit IRC | 00:48 | |
| *** dmorita has joined #openstack-swift | 01:13 | |
| *** dmorita has quit IRC | 01:18 | |
| jrichli | notmyname, mattoliverau: have a great time at LCA | 01:19 |
|---|---|---|
| *** janonymous has left #openstack-swift | 01:19 | |
| kota_ | good morning | 01:35 |
| kota_ | timburke, clayg, acoles_: oh, that reclaim_age test patch has got merged? great! i was thinking of starting from that point for this week's my review work but it's unnecessary. | 01:37 |
| *** sams-gleb has joined #openstack-swift | 01:58 | |
| *** mrda is now known as mrda-lca | 02:00 | |
| *** mrda-lca is now known as mrda | 02:00 | |
| *** sams-gleb has quit IRC | 02:02 | |
| *** geaaru has quit IRC | 02:05 | |
| *** mingyu has joined #openstack-swift | 02:11 | |
| *** mingyu_ has joined #openstack-swift | 02:14 | |
| *** mingyu has quit IRC | 02:15 | |
| *** stevemar has quit IRC | 02:21 | |
| *** stevemar has joined #openstack-swift | 02:21 | |
| *** links has joined #openstack-swift | 02:25 | |
| *** mingyu_ has quit IRC | 02:34 | |
| *** links has quit IRC | 02:38 | |
| *** mingyu has joined #openstack-swift | 02:39 | |
| *** links has joined #openstack-swift | 02:46 | |
| *** winggundamth has joined #openstack-swift | 03:04 | |
| *** mingyu has quit IRC | 03:06 | |
| *** Jeffrey4l_ has quit IRC | 03:21 | |
| *** Jeffrey4l has joined #openstack-swift | 03:21 | |
| *** jrichli has quit IRC | 03:23 | |
| *** tqtran has joined #openstack-swift | 03:49 | |
| *** tqtran has quit IRC | 03:50 | |
| *** sams-gleb has joined #openstack-swift | 04:00 | |
| *** psachin has joined #openstack-swift | 04:01 | |
| *** sams-gleb has quit IRC | 04:04 | |
| *** SkyRocknRoll has joined #openstack-swift | 04:16 | |
| *** SkyRocknRoll has quit IRC | 04:19 | |
| *** geaaru has joined #openstack-swift | 04:31 | |
| *** ppai has joined #openstack-swift | 04:56 | |
| *** chosafine has joined #openstack-swift | 04:59 | |
| *** SkyRocknRoll has joined #openstack-swift | 05:01 | |
| *** klrmn has quit IRC | 05:12 | |
| *** SkyRocknRoll has quit IRC | 05:59 | |
| *** sams-gleb has joined #openstack-swift | 06:02 | |
| *** chosafine has quit IRC | 06:02 | |
| *** sams-gleb has quit IRC | 06:07 | |
| *** SkyRocknRoll has joined #openstack-swift | 06:13 | |
| *** bkopilov_ has joined #openstack-swift | 06:14 | |
| *** silor has joined #openstack-swift | 06:35 | |
| *** silor1 has joined #openstack-swift | 06:40 | |
| *** silor has quit IRC | 06:43 | |
| *** silor1 is now known as silor | 06:43 | |
| *** tqtran has joined #openstack-swift | 06:54 | |
| *** kei_yama_ has joined #openstack-swift | 07:05 | |
| *** kei_yama has quit IRC | 07:06 | |
| *** SkyRocknRoll has quit IRC | 07:14 | |
| *** tqtran has quit IRC | 07:15 | |
| *** tesseract has joined #openstack-swift | 07:21 | |
| *** hogepodge_ has joined #openstack-swift | 07:24 | |
| *** SkyRocknRoll has joined #openstack-swift | 07:30 | |
| *** pcaruana has joined #openstack-swift | 07:34 | |
| *** hogepodge_ has quit IRC | 07:37 | |
| *** sams-gleb has joined #openstack-swift | 07:56 | |
| *** sams-gleb has quit IRC | 08:06 | |
| *** sams-gleb has joined #openstack-swift | 08:06 | |
| *** oshritf has joined #openstack-swift | 08:09 | |
| *** kei_yama_ has quit IRC | 08:10 | |
| openstackgerrit | Anh Tran proposed openstack/swift-specs: Removed redundant words: 'in', 'this' https://review.openstack.org/420557 | 08:11 |
| *** kei_yama has joined #openstack-swift | 08:13 | |
| openstackgerrit | Anh Tran proposed openstack/swift: Removed redundant 'is' https://review.openstack.org/420561 | 08:17 |
| *** rledisez has joined #openstack-swift | 08:19 | |
| *** sams-gleb has quit IRC | 08:21 | |
| *** sams-gleb has joined #openstack-swift | 08:22 | |
| *** sams-gleb has quit IRC | 08:26 | |
| *** trananhkma has joined #openstack-swift | 08:30 | |
| *** mingyu has joined #openstack-swift | 08:48 | |
| *** fgemein has joined #openstack-swift | 09:01 | |
| *** hseipp has joined #openstack-swift | 09:04 | |
| *** thurloat has quit IRC | 09:06 | |
| *** thurloat has joined #openstack-swift | 09:08 | |
| *** fgemein has left #openstack-swift | 09:13 | |
| *** jordanP has joined #openstack-swift | 09:15 | |
| *** cbartz has joined #openstack-swift | 09:18 | |
| *** winggundamth has quit IRC | 09:33 | |
| *** dmorita has joined #openstack-swift | 09:34 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift: Fix unnecessary for-loop and mis docs https://review.openstack.org/420607 | 09:34 |
| *** dmorita has quit IRC | 09:38 | |
| *** asettle has joined #openstack-swift | 09:46 | |
| *** asettle is now known as Guest58506 | 09:46 | |
| *** winggundamth has joined #openstack-swift | 09:46 | |
| *** mvk has quit IRC | 09:47 | |
| *** Guest58506 has quit IRC | 09:53 | |
| *** hseipp has quit IRC | 09:56 | |
| *** hseipp has joined #openstack-swift | 10:02 | |
| *** sams-gleb has joined #openstack-swift | 10:07 | |
| *** mvk has joined #openstack-swift | 10:20 | |
| *** hseipp has quit IRC | 10:22 | |
| openstackgerrit | Andreas Jaeger proposed openstack/swift: Remove broken links https://review.openstack.org/420625 | 10:23 |
| *** hseipp has joined #openstack-swift | 10:27 | |
| *** jcaron has joined #openstack-swift | 10:28 | |
| *** hseipp has quit IRC | 10:29 | |
| *** SkyRocknRoll has quit IRC | 10:30 | |
| *** Shashikant86 has joined #openstack-swift | 10:32 | |
| *** kei_yama has quit IRC | 10:33 | |
| openstackgerrit | Merged openstack/swift: Tighten the move-one-replica test https://review.openstack.org/419107 | 10:40 |
| jcaron | Hi everyone, I am Jean from OVH.com, | 10:42 |
| jcaron | Some of our customers have the need to use a TXT resource record instead of the CNAME (middleware CNAME lookup) | 10:42 |
| jcaron | How can this feature be properly implemented ? add the TXT lookup directly in middleware/cname_lookup.py ? rename the middleware (dns_lookup.py) ? | 10:42 |
| jcaron | add a new middleware dedicated for the TXT lookup (code replication .. ) ? | 10:42 |
| jcaron | thanks for your incoming advices | 10:42 |
| *** SkyRocknRoll has joined #openstack-swift | 10:45 | |
| *** mingyu has quit IRC | 10:47 | |
| *** asettle has joined #openstack-swift | 10:53 | |
| *** asettle has quit IRC | 10:54 | |
| *** asettle has joined #openstack-swift | 10:59 | |
| *** asettle__ has joined #openstack-swift | 11:01 | |
| *** asettle has quit IRC | 11:01 | |
| *** jordanP has quit IRC | 11:01 | |
| *** asettle__ has quit IRC | 11:10 | |
| *** asettle has joined #openstack-swift | 11:11 | |
| *** asettle is now known as Guest74297 | 11:11 | |
| *** Shashikant86 has quit IRC | 11:15 | |
| *** Guest74297 has quit IRC | 11:17 | |
| *** AmosT has joined #openstack-swift | 11:29 | |
| *** AmosT has left #openstack-swift | 11:30 | |
| *** silor has quit IRC | 11:35 | |
| *** silor has joined #openstack-swift | 11:36 | |
| *** ganders has joined #openstack-swift | 11:40 | |
| *** asettle_ has joined #openstack-swift | 11:50 | |
| *** asettle_ is now known as asettle | 11:52 | |
| *** bkopilov_ has quit IRC | 11:57 | |
| *** bkopilov_ has joined #openstack-swift | 11:58 | |
| *** Shashikant86 has joined #openstack-swift | 12:02 | |
| *** SkyRocknRoll has quit IRC | 12:24 | |
| *** McMurlock1 has joined #openstack-swift | 12:25 | |
| *** winggundamth has quit IRC | 12:34 | |
| *** SkyRocknRoll has joined #openstack-swift | 12:35 | |
| openstackgerrit | Merged openstack/swift: Removed redundant 'is' https://review.openstack.org/420561 | 12:36 |
| openstackgerrit | Merged openstack/swift: Cleanup tests from empty suffix quarantined db fix https://review.openstack.org/405134 | 12:36 |
| openstackgerrit | Merged openstack/swift: Remove broken links https://review.openstack.org/420625 | 12:36 |
| *** McMurlock1 has quit IRC | 12:38 | |
| *** McMurlock1 has joined #openstack-swift | 12:41 | |
| *** McMurlock1 has quit IRC | 12:42 | |
| *** McMurlock1 has joined #openstack-swift | 12:42 | |
| *** klamath has joined #openstack-swift | 12:44 | |
| *** klamath has quit IRC | 12:44 | |
| *** klamath has joined #openstack-swift | 12:45 | |
| *** Shashikant86 has quit IRC | 13:00 | |
| *** oshritf_ has joined #openstack-swift | 13:12 | |
| *** oshritf has quit IRC | 13:15 | |
| *** ppai has quit IRC | 13:32 | |
| *** hseipp has joined #openstack-swift | 13:34 | |
| *** mingyu has joined #openstack-swift | 13:34 | |
| *** psachin has quit IRC | 13:34 | |
| *** mingyu has quit IRC | 13:51 | |
| *** silor has quit IRC | 14:02 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift: Fix unnecessary for-loop and mis docs https://review.openstack.org/420607 | 14:07 |
| *** Shashikant86 has joined #openstack-swift | 14:08 | |
| kota_ | sorry, cschwede i fixed pep8 error in the patch | 14:09 |
| cschwede | kota_: thx, still LGTM! | 14:10 |
| kota_ | cschwede: thanks for your quick response! | 14:10 |
| kota_ | oops? | 14:11 |
| kota_ | still misindent??? | 14:11 |
| kota_ | sorry, I may push wrong patch | 14:11 |
| cschwede | kota_: but the diff from 1 to 2 looks like it fixes the pep8 error? https://review.openstack.org/#/c/420607/1..2/swift/common/ring/builder.py | 14:12 |
| patchbot | patch 420607 - swift - Fix unnecessary for-loop and mis docs | 14:12 |
| *** Shashikant86 has quit IRC | 14:12 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift: Fix unnecessary for-loop and mis docs https://review.openstack.org/420607 | 14:12 |
| cschwede | http://logs.openstack.org/07/420607/2/check/gate-swift-pep8-ubuntu-xenial/ce4d0a7/console.html#_2017-01-16_14_10_45_007670 | 14:13 |
| kota_ | cschwede: I thought but i got mistake again | 14:13 |
| kota_ | cschwede: patch 3 should pass pep8 | 14:13 |
| cschwede | ah yes, patch #1 fixed one of the two lines. makes sense | 14:13 |
| kota_ | cschwede: yeah, sorry. my mistake, again. | 14:15 |
| kota_ | my head seems to go sleep :/ | 14:15 |
| cschwede | kota_: well, all good things come in threes - pep8 passed :) | 14:18 |
| kota_ | yey | 14:18 |
| kota_ | cschwede: thanks! | 14:19 |
| * kota_ is going to offline | 14:20 | |
| *** acoles_ is now known as acoles | 14:24 | |
| *** links has quit IRC | 14:30 | |
| *** Shashikant86 has joined #openstack-swift | 14:33 | |
| acoles | clayg: jrichli: git branch | wc -l == 137 :P | 14:36 |
| *** mingyu has joined #openstack-swift | 14:36 | |
| acoles | kota_: actually the reclaim age patch 374419 didn't land yet, just some associated tests, but looks like cschwede may be on it :) | 14:38 |
| patchbot | https://review.openstack.org/#/c/374419/ - swift - Move documented reclaim_age option to correct loca... | 14:38 |
| *** caiobrentano_ has joined #openstack-swift | 14:40 | |
| kota_ | acoles: oh, really | 14:48 |
| *** mingyu has quit IRC | 14:51 | |
| *** mingyu has joined #openstack-swift | 14:53 | |
| *** Shashikant86 has quit IRC | 14:55 | |
| *** Shashikant86 has joined #openstack-swift | 15:03 | |
| cschwede | acoles: indeed, i was looking at that patch! and with your and mahatic's comments i'm fine merging that patch - it's in the gate now :) | 15:10 |
| mahatic | cschwede: thanks for looking at it! | 15:11 |
| acoles | cschwede: thanks! | 15:23 |
| *** sams-gleb has quit IRC | 15:25 | |
| *** sams-gleb has joined #openstack-swift | 15:25 | |
| *** sams-gleb has quit IRC | 15:30 | |
| *** bkeller` has joined #openstack-swift | 15:32 | |
| *** sams-gleb has joined #openstack-swift | 15:42 | |
| *** mvk has quit IRC | 15:50 | |
| *** Shashikant86 has quit IRC | 16:00 | |
| *** klrmn has joined #openstack-swift | 16:01 | |
| *** bkeller` has quit IRC | 16:07 | |
| *** bkeller`_ has joined #openstack-swift | 16:07 | |
| *** bkeller`_ is now known as bkeller` | 16:08 | |
| *** Shashikant86 has joined #openstack-swift | 16:12 | |
| *** bkeller`_ has joined #openstack-swift | 16:13 | |
| *** bkeller` has quit IRC | 16:14 | |
| *** bkeller`_ has quit IRC | 16:14 | |
| *** jrichli has joined #openstack-swift | 16:15 | |
| *** bkeller` has joined #openstack-swift | 16:15 | |
| timburke | good morning | 16:16 |
| *** mingyu has quit IRC | 16:17 | |
| *** hseipp has quit IRC | 16:21 | |
| mahatic | timburke: good morning. Isn't it a holiday in the US? | 16:21 |
| *** hseipp has joined #openstack-swift | 16:21 | |
| timburke | i've got something i need to fix for a release this week :-) | 16:22 |
| mahatic | oic :) | 16:22 |
| *** mingyu has joined #openstack-swift | 16:23 | |
| *** cbartz has quit IRC | 16:23 | |
| *** mingyu has quit IRC | 16:27 | |
| *** mingyu has joined #openstack-swift | 16:27 | |
| openstackgerrit | Merged openstack/swift: Move documented reclaim_age option to correct location https://review.openstack.org/374419 | 16:35 |
| openstackgerrit | Merged openstack/swift: Fix unnecessary for-loop and mis docs https://review.openstack.org/420607 | 16:36 |
| *** McMurlock1 has quit IRC | 16:44 | |
| *** oshritf_ has quit IRC | 16:50 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 16:53 |
| *** ganders has quit IRC | 16:54 | |
| *** tqtran has joined #openstack-swift | 16:57 | |
| *** ganders has joined #openstack-swift | 17:02 | |
| *** sams-gleb has quit IRC | 17:10 | |
| *** sams-gleb has joined #openstack-swift | 17:11 | |
| *** mingyu has quit IRC | 17:12 | |
| clayg | morning | 17:13 |
| *** sams-gleb has quit IRC | 17:15 | |
| *** mingyu has joined #openstack-swift | 17:19 | |
| *** chsc has joined #openstack-swift | 17:21 | |
| *** klrmn has quit IRC | 17:28 | |
| *** tesseract has quit IRC | 17:29 | |
| *** rledisez has quit IRC | 17:29 | |
| *** nottrobin has quit IRC | 17:30 | |
| *** hseipp has quit IRC | 17:31 | |
| *** nottrobin has joined #openstack-swift | 17:31 | |
| *** mingyu has quit IRC | 17:37 | |
| *** McMurlock1 has joined #openstack-swift | 17:37 | |
| *** JimCheung has joined #openstack-swift | 17:40 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift: Pretend *some* parts min_part_hours_passed https://review.openstack.org/311226 | 17:45 |
| *** Shashikant86 has quit IRC | 17:51 | |
| *** McMurlock1 has quit IRC | 17:53 | |
| *** Shashikant86 has joined #openstack-swift | 17:53 | |
| *** Shashikant86 has quit IRC | 17:56 | |
| clayg | kota_: part of the need for the mockig was because on dev it's pretty easy to get a mtime that == on modified files :\ | 17:57 |
| *** dmorita has joined #openstack-swift | 17:57 | |
| clayg | kota_: in patch 419787 I found using the files inode number to be a bit more reliable and works the same as python stdlib os.path.samefile | 17:58 |
| patchbot | https://review.openstack.org/#/c/419787/ - swift - Better optimistic lock in get_hashes | 17:58 |
| clayg | kota_: in particular I guess PavelK had a particularlly low resolution mtime on a COW filesystem (containers) and found the test to be unreliable w/o the mocks | 17:59 |
| *** mvk has joined #openstack-swift | 18:06 | |
| acoles | clayg: I am going to try to loop back to those ^^ suffix hashing patches this week - we're still preferring your 4 (or 5?) separated patches right? I saw your +2 come then go on patch 402043. | 18:11 |
| patchbot | https://review.openstack.org/#/c/402043/ - swift - Optimize hash calculation when suffix hash invalid... | 18:11 |
| clayg | acoles: idk :'( | 18:11 |
| acoles | clayg: heh you got a +2 on one of them patch 418692 | 18:16 |
| patchbot | https://review.openstack.org/#/c/418692/ - swift - Optimize noop case for suffix rehash | 18:16 |
| *** huo has joined #openstack-swift | 18:16 | |
| *** huo has quit IRC | 18:16 | |
| acoles | and +1 from Pavel | 18:16 |
| *** klrmn has joined #openstack-swift | 18:16 | |
| acoles | so maybe I'll start there | 18:17 |
| clayg | jesus | 18:20 |
| clayg | one of the main reasons I was sure we should pull these apart was because we had multiple bug fixes squashed in with a *minor performance enhancement* | 18:20 |
| clayg | so naturally everyone starts with that ;) | 18:21 |
| timburke | may as well pick off the easy one, right? :P | 18:21 |
| acoles | oh gerrit! | 18:28 |
| clayg | oh look! now acoles can review ring patches -> p 420607 | 18:30 |
| patchbot | https://review.openstack.org/#/c/420607/ - swift - Fix unnecessary for-loop and mis docs (MERGED) | 18:30 |
| * acoles attempts to infer useful info from "Related Changes" | 18:30 | |
| acoles | clayg: I knew it's a public holiday over there, I didn't know it was national take the p day :) | 18:31 |
| clayg | acoles: I had to google | 18:33 |
| acoles | really? TIL | 18:33 |
| clayg | yeah i'm not sure there's an exactly equivilent 'merican expression https://en.wikipedia.org/wiki/Taking_the_piss | 18:34 |
| acoles | clayg: huh. toodle-pip :P | 18:37 |
| *** acoles is now known as acoles_ | 18:38 | |
| *** sams-gleb has joined #openstack-swift | 19:12 | |
| *** sams-gleb has quit IRC | 19:17 | |
| *** lespaul has joined #openstack-swift | 19:23 | |
| clayg | you | 19:27 |
| clayg | ` | 19:27 |
| clayg | ~ | 19:27 |
| clayg | whoa! | 19:27 |
| clayg | i've learned recently that `<return>~.` is some magic when you ssh connection goes to sleep - but while trying to remember the corect sequence apparently my ssh connection was re-established - computers are funny | 19:28 |
| clayg | cschwede: did you ever figure out the thing with openstack-client slowness with swift download? | 19:29 |
| openstackgerrit | Clay Gerrard proposed openstack/swift: Better optimistic lock in get_hashes https://review.openstack.org/419787 | 19:30 |
| clayg | ^ just a spelling error | 19:31 |
| clayg | aside from patch 418691 - I think patch 419787 is the most imporant in the sequence | 19:31 |
| patchbot | https://review.openstack.org/#/c/418691/ - swift - Fix performance regression with hash invalidations | 19:31 |
| patchbot | https://review.openstack.org/#/c/419787/ - swift - Better optimistic lock in get_hashes | 19:31 |
| clayg | ... if you're prioritizing | 19:31 |
| clayg | of course they both depend on patch 418689 | 19:31 |
| patchbot | https://review.openstack.org/#/c/418689/ - swift - Extract test pattern to helper | 19:31 |
| lespaul | Hello. I'm trying to create a container using a storage policy however, I'm getting a 400 Bad Request Invalid X-Storage-Policy-Index 2. I can create a container with the default policy fine. Any ideas? | 19:33 |
| clayg | lespaul: client api uses the names in /info | 19:34 |
| clayg | x-storage-policy: <name> | 19:34 |
| clayg | or alias | 19:34 |
| lespaul | the name matches fine though. i didn't put any alias, it's commented out. | 19:35 |
| *** xlucas has joined #openstack-swift | 19:35 | |
| lespaul | ok let me double check.. | 19:35 |
| clayg | oh - maybe the backend is rejecting it - sorry - restart object servers? | 19:35 |
| clayg | ... or reload | 19:36 |
| *** PavelK has joined #openstack-swift | 19:38 | |
| lespaul | once i put an object into a container, is there a command to show which nodes/drives the replicas are contained? | 19:44 |
| *** ganders has quit IRC | 19:50 | |
| clayg | lespaul: swift-get-nodes is pretty good for that - it'll list ring primaries and handoffs | 19:53 |
| clayg | lespaul: you can then query them to see which nodes it landed on (generally the primaries unless you're multi-region write_affinity or something went wrong) | 19:53 |
| *** SkyRocknRoll has quit IRC | 19:54 | |
| *** jerrygb has joined #openstack-swift | 19:57 | |
| clayg | debating if I should try again to consolidate all the suffix hashing fixes or give up and review ec fragment duplication... | 20:06 |
| PavelK | clayg: hi. I'm not sure tha I understand what means +A patch - should I abandon patch 402043? | 20:13 |
| patchbot | https://review.openstack.org/#/c/402043/ - swift - Optimize hash calculation when suffix hash invalid... | 20:13 |
| openstackgerrit | Merged openstack/swift: Confirm receipt of SLO PUT with etag https://review.openstack.org/390901 | 20:37 |
| clayg | PavelK: idk :'( | 20:37 |
| clayg | PavelK: what do you think about the getmtime vs. stat/inode/samefile check? | 20:38 |
| clayg | oh, you +1'd it | 20:39 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 20:40 |
| PavelK | clayg: yes, thats fine - if you merge idea of patch 418691 on it | 20:41 |
| patchbot | https://review.openstack.org/#/c/418691/ - swift - Fix performance regression with hash invalidations | 20:41 |
| clayg | yeah, i haven't really come up with a diff I like to merge them :'( | 20:41 |
| clayg | you'd think the safe_get_inode and read_hashes helpers would make it easier - but idk - there's still this weird schizophrenic thing were we want safe_get_inode under the directory lock but we also want to flag do_listdir/set-hashes-{} and safe_get_inode of any existing file when something goes wrong in consolidate hashes | 20:44 |
| clayg | I *think* I want like "with dir_lock: try: hashes = _do_consolidate; except Exception: reset-the-business; inode = safe_get_inode()" | 20:45 |
| clayg | ... maybe that's it... | 20:45 |
| *** zaitcev has quit IRC | 20:48 | |
| *** zaitcev has joined #openstack-swift | 20:50 | |
| *** ChanServ sets mode: +v zaitcev | 20:50 | |
| PavelK | I'm not sure that dir_lock should hold the caller - once can someone forget to lock it. So I preffer function "give me hashes and something how I can compare that it was not changed when I lock the dir later". | 20:50 |
| lespaul | clayg: it worked, thanks. where would i find documentation for the more esoteric commands in swift? | 20:50 |
| clayg | except if something goes wrong in conslidate hashes you still need the inode | 20:50 |
| PavelK | may be | 20:51 |
| clayg | you could say "call conslidate hashes and have it return the thing - then if that goes wrong go get it anyway" | 20:51 |
| clayg | I think you can still make a function that does both "safe_conslidate_hashes_and_get_inode" - but you can method extract that to just do what I had above (lock, _private_unsafe_conslidate, safe_get_inode) - and have it return all the state you need (maybe empty hashes, state-indicating-do_listdir, inode) | 20:52 |
| clayg | anyway - i'm less worried about cleanliness than correctness - but I need something understandable to reason about correctness | 20:53 |
| clayg | at a minimum patch 402043 is missing the test(s) form https://review.openstack.org/#/c/419787/4/test/unit/obj/test_diskfile.py | 20:54 |
| patchbot | https://review.openstack.org/#/c/402043/ - swift - Optimize hash calculation when suffix hash invalid... | 20:54 |
| patchbot | patch 419787 - swift - Better optimistic lock in get_hashes | 20:54 |
| clayg | ... but I didn't even notice those tests were needed until you pointed it out to me in the split up changes! | 20:54 |
| PavelK | I'm scared of picle.loads(.pkl) CPU utization, but I'm sure that there are more important thinks where you can spend time | 20:54 |
| clayg | CPU? those files are *tiny* | 20:55 |
| PavelK | You know it better than me :-) | 20:56 |
| clayg | i mean they have to be? suffixes are fixed width - 0xfff - how long can it take to pickle deserialize that? | 20:57 |
| * clayg wishes he was timburke could cook up one of those cute timeit lines in 2 seconds | 20:57 | |
| *** jrichli_ has joined #openstack-swift | 20:58 | |
| timburke | i dunno, but it looks like we've had trouble with pickle being slow before -- https://bugs.launchpad.net/swift/+bug/1031954 | 20:59 |
| openstack | Launchpad bug 1031954 in OpenStack Object Storage (swift) "Slow Ring Loading in 2.7 due to Ring Unpickling" [High,Fix released] - Assigned to Darrell Bishop (darrellb) | 20:59 |
| *** jamielennox|away is now known as jamielennox | 20:59 | |
| clayg | https://gist.github.com/clayg/c6a5c41b12c7f0c4075386d3f23d6d81 | 21:02 |
| clayg | cPickle is definately faster | 21:03 |
| PavelK | clayg: yes, it should be small. I don't know swift well, I just little know diskfile.py :-). I'm leaving a comment near to patch that read_pickle can be placed next to write_pickle helper - and it seems good for me | 21:03 |
| clayg | PavelK: I thought about putting it in utils - but didn't want to try and go fix the updater to use it | 21:04 |
| PavelK | thanks for hard argument! | 21:04 |
| timburke | clayg: yeah, cschwede et al. totally figured it out. the follow-up was https://review.openstack.org/#/c/416249/ | 21:04 |
| patchbot | patch 416249 - keystoneauth - Prevent MemoryError when logging response bodies (MERGED) | 21:04 |
| clayg | timburke: thanks - i found some reference to that on the bug | 21:04 |
| clayg | timburke ... and I saw that one also merged ... | 21:04 |
| clayg | PavelK: thanks for pointing out all these bugs! | 21:06 |
| clayg | and all your help with the solution and review | 21:06 |
| PavelK | "It's my work" :-) | 21:07 |
| clayg | ... and for not stopping to push on finishing the rest of the cleanup! | 21:07 |
| clayg | lol @ "it's a damn good thing I'm paid to put up with these %$$holes" | 21:08 |
| PavelK | clayg: I have not answer for my first question - what to do with old messed patch? Close it as abandoned or what you mean with "+A it" | 21:08 |
| clayg | PavelK: I apparently can't bring myself to +A patch 402043 as is - I could try to come up with a list of things that would need to be addressed - but ultimately it's just "what is the diff applied ontop of patch 402043 that makes me comfortable with the end state" | 21:11 |
| patchbot | https://review.openstack.org/#/c/402043/ - swift - Optimize hash calculation when suffix hash invalid... | 21:11 |
| patchbot | https://review.openstack.org/#/c/402043/ - swift - Optimize hash calculation when suffix hash invalid... | 21:11 |
| clayg | if I had that diff I'd probably just push it up as a follow up and tell gerrit to merge your patch | 21:11 |
| clayg | all this crap and all these extra patches and non-sense was me struggling to come up with the followup diff | 21:12 |
| clayg | if the split up patches got some traction - because they were easier to review - it would have validated that splitting them up was the right way to go - and maybe it'd be easier to iterate and merge the fixes seperately | 21:13 |
| clayg | but... we're ... 3-4 working days in on that? | 21:13 |
| clayg | and I've learned it's hard to merge some of the changes together (we were just talking about it - my split patches conflict with each other - we can't just merge them all w/o a rebase) | 21:13 |
| clayg | so it's a big $%^&*ing mess | 21:13 |
| clayg | so I'm not going to tell you the right thing to do - because I have no idea | 21:14 |
| *** sams-gleb has joined #openstack-swift | 21:15 | |
| *** sams-gleb has quit IRC | 21:19 | |
| clayg | PavelK: notmyname can attest this whole thing has been a bit of an existential crisis for me | 21:20 |
| *** chosafine has joined #openstack-swift | 21:20 | |
| clayg | but no one seems to want to tell me how we should proceed either | 21:20 |
| ahale | well, i have mostly no idea what you're doing since i havent paid any attention to swift for ages, but i think its cool you're looking at hashes stuff clayg, it used to give me nightmares sometimes. and i also think the coolest parts of swift are the hard bits when noone knows cos its not been done before | 21:24 |
| clayg | ahale: not me, PavelK | 21:28 |
| clayg | +1 PavelK is awesome for looking at the hard stuff | 21:28 |
| ahale | thats how little attention i have paid! , yeah +1 to that :) | 21:28 |
| *** Jeffrey4l has quit IRC | 21:34 | |
| *** Jeffrey4l has joined #openstack-swift | 21:35 | |
| *** lespaul has quit IRC | 21:37 | |
| *** chosafine has quit IRC | 21:41 | |
| *** dfflanders has joined #openstack-swift | 21:52 | |
| PavelK | clayg: conflicts was the reason why I made one patch. But I think that your patches can be merged independently and the fourth - "inode" can be rebased on them and merged later. | 21:52 |
| PavelK | clayg: So I leave my patch as is and abandon it after your patches were merged, right? | 21:54 |
| clayg | heh, idk == "I don't know" ;) | 21:55 |
| *** sams-gleb has joined #openstack-swift | 21:56 | |
| clayg | maybe folks are scared to review the other suffix fixes because they think I get frustrated things are moving to slowly and go +A you patch anyway (I totaly might!) | 21:56 |
| *** chosafine has joined #openstack-swift | 21:57 | |
| clayg | for the moment I'm trying to review patch 219165 (look at the size of the diff to proxy test_server !?) | 21:59 |
| patchbot | https://review.openstack.org/#/c/219165/ - swift - EC Fragment Duplication - Foundational Global EC C... | 21:59 |
| *** jrichli_ has quit IRC | 22:09 | |
| *** ChanServ sets mode: +v jrichli | 22:13 | |
| *** geaaru has quit IRC | 22:14 | |
| *** sams-gleb has quit IRC | 22:21 | |
| *** sams-gleb has joined #openstack-swift | 22:22 | |
| *** sams-gleb has quit IRC | 22:26 | |
| jrichli | yay, bouncer restored :-) | 22:28 |
| *** jerrygb has quit IRC | 22:34 | |
| clayg | wtf is frags_by_byte_order? | 22:37 |
| clayg | like ... endianness !? | 22:37 |
| *** lespaul has joined #openstack-swift | 22:38 | |
| lespaul | Hello. I uploaded an object to my cluster (6 nodes w/ 2 drives each). I ran swift-get-nodes and found the replicas on the target nodes. I then powered of 2 of the nodes w/ replicas. I checked swift-get-nodes again and it still lists the same. Should the replica be copied to the handoff nodes? | 22:55 |
| lespaul | *Shouldn't | 22:56 |
| *** Durgesh has joined #openstack-swift | 23:10 | |
| *** xlucas has quit IRC | 23:12 | |
| lespaul | I don't think swift-get-nodes shows accurate information. If I run swift-get-nodes -a /etc/swift/ <ring.gz> <account> then make up a non-existent container and object, a result is still displayed. Thoughts? | 23:16 |
| *** jerrygb has joined #openstack-swift | 23:22 | |
| *** Durgesh has quit IRC | 23:22 | |
| *** klamath has quit IRC | 23:25 | |
| *** jerrygb has quit IRC | 23:26 | |
| *** jerrygb has joined #openstack-swift | 23:34 | |
| torgomatic | lespaul: swift-get-nodes just performs a ring lookup and shows you the results; it might be where a thing that exists is, it might be where a thing that exists will be after replication, or it might be where a thing that does not exist would go if you created it | 23:35 |
| *** chsc has quit IRC | 23:35 | |
| *** kei_yama has joined #openstack-swift | 23:43 | |
| notmyname | hello, world | 23:46 |
| lespaul | torgomatic: that makes sense. is there a command that will give a real-time location of the replicas or fragments? | 23:50 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!