| *** rcernin has joined #openstack-swift | 00:08 | |
| *** zaitcev has quit IRC | 00:22 | |
| viks___ | clayg: Ok.. thanks | 00:24 |
|---|---|---|
| *** zaitcev has joined #openstack-swift | 00:35 | |
| *** ChanServ sets mode: +v zaitcev | 00:35 | |
| *** gyee has quit IRC | 00:48 | |
| *** rcernin has quit IRC | 00:58 | |
| *** spsurya has joined #openstack-swift | 01:05 | |
| *** rcernin has joined #openstack-swift | 01:14 | |
| *** baojg has joined #openstack-swift | 01:17 | |
| *** BjoernT_ has joined #openstack-swift | 01:44 | |
| *** BjoernT_ has quit IRC | 01:46 | |
| *** BjoernT has quit IRC | 01:47 | |
| *** zaitcev has quit IRC | 02:07 | |
| *** zaitcev has joined #openstack-swift | 02:21 | |
| *** ChanServ sets mode: +v zaitcev | 02:21 | |
| *** BjoernT has joined #openstack-swift | 02:22 | |
| *** zaitcev has quit IRC | 02:35 | |
| *** BjoernT has quit IRC | 02:46 | |
| *** zaitcev has joined #openstack-swift | 02:49 | |
| *** ChanServ sets mode: +v zaitcev | 02:49 | |
| kota_ | hello world | 02:56 |
| mattoliverau | o/ | 03:02 |
| *** ccamacho has joined #openstack-swift | 03:06 | |
| kota_ | mattoliverau: o/ | 03:25 |
| *** psachin has joined #openstack-swift | 03:33 | |
| *** zaitcev has quit IRC | 03:34 | |
| *** gkadam has joined #openstack-swift | 03:41 | |
| *** gkadam has quit IRC | 03:41 | |
| *** gkadam has joined #openstack-swift | 03:43 | |
| *** zaitcev has joined #openstack-swift | 03:45 | |
| *** ChanServ sets mode: +v zaitcev | 03:45 | |
| *** gkadam has quit IRC | 04:00 | |
| *** manuvakery has joined #openstack-swift | 04:46 | |
| *** zaitcev has quit IRC | 06:02 | |
| *** zaitcev has joined #openstack-swift | 06:16 | |
| *** ChanServ sets mode: +v zaitcev | 06:16 | |
| *** rcernin has quit IRC | 06:16 | |
| *** rcernin has joined #openstack-swift | 06:31 | |
| *** ccamacho has quit IRC | 07:10 | |
| *** tesseract has joined #openstack-swift | 07:10 | |
| *** e0ne has joined #openstack-swift | 07:22 | |
| *** rcernin has quit IRC | 07:56 | |
| *** tkajinam has quit IRC | 08:11 | |
| *** rcernin has joined #openstack-swift | 08:23 | |
| *** rcernin has quit IRC | 08:29 | |
| *** rcernin has joined #openstack-swift | 08:38 | |
| *** zaitcev has quit IRC | 09:01 | |
| *** rcernin has quit IRC | 09:04 | |
| *** zaitcev has joined #openstack-swift | 09:14 | |
| *** ChanServ sets mode: +v zaitcev | 09:14 | |
| *** tdasilva has joined #openstack-swift | 09:25 | |
| *** ChanServ sets mode: +v tdasilva | 09:25 | |
| *** baojg has quit IRC | 10:15 | |
| *** baojg has joined #openstack-swift | 10:16 | |
| *** zaitcev has quit IRC | 11:04 | |
| *** zaitcev has joined #openstack-swift | 11:17 | |
| *** ChanServ sets mode: +v zaitcev | 11:17 | |
| *** manuvakery has quit IRC | 11:39 | |
| *** rcernin has joined #openstack-swift | 12:35 | |
| *** zaitcev_ has joined #openstack-swift | 12:39 | |
| *** ChanServ sets mode: +v zaitcev_ | 12:39 | |
| *** rcernin has quit IRC | 12:40 | |
| *** zaitcev has quit IRC | 12:43 | |
| *** BjoernT has joined #openstack-swift | 13:19 | |
| *** BjoernT has quit IRC | 13:23 | |
| *** BjoernT has joined #openstack-swift | 13:23 | |
| *** dsariel has joined #openstack-swift | 13:31 | |
| clayg | tdasilva: nice find on the account_acls prior art! | 13:39 |
| clayg | tdasilva: in fact, I just realized the next patch in the chain does puts use_symlinks in /info because i'm working on making func tests bifricate | 13:40 |
| clayg | I'm not sure we need to have func tests pass/skip w/o static_link=True? | 13:41 |
| tdasilva | clayg: I think we might in the future | 13:41 |
| clayg | 🤔 ??? | 13:41 |
| tdasilva | clayg: if I were for example to run the master func tests against an older swift | 13:41 |
| tdasilva | not sure how likely that is...in the case of the multinode-rolling-upgrade gate tests, it's the opposite | 13:42 |
| tdasilva | where we run an "older" previous release func tests against "master" swift | 13:43 |
| tdasilva | so maybe not... | 13:43 |
| tdasilva | clayg: I had that idea of expanding the multinode-rolling-upgrade tests to run like a setup and verify set of tests and in that case we always use newer tests. So in that case it would be nice to be able to test for static_link=true | 13:45 |
| tdasilva | clayg: see p https://review.opendev.org/#/c/626663/ | 13:46 |
| patchbot | patch 626663 - swift - Add rolling upgrade tests - 7 patch sets | 13:46 |
| clayg | tdasilva: that is brilliant and amazing! 😍 | 13:52 |
| tdasilva | clayg: heh, my hope is that it will be useful for testing py2 to py3 upgrade | 13:55 |
| openstackgerrit | Thiago da Silva proposed openstack/swift master: Add params to get_*_metadata internal client methods https://review.opendev.org/676414 | 14:15 |
| ormandj | heya folks, what are people doing to cleanup orphan segments in +segments folders when clients die? | 14:20 |
| ormandj | we've unfortunately seen a lot of bad clients/connections/etc that lead to partial chunked data sitting in a +segments (s3 compat layer access) without metadata associated with it | 14:21 |
| *** zaitcev_ is now known as zaitcev | 14:24 | |
| *** BjoernT_ has joined #openstack-swift | 14:35 | |
| *** BjoernT has quit IRC | 14:36 | |
| *** ianychoi has joined #openstack-swift | 15:20 | |
| *** dsariel has quit IRC | 15:32 | |
| *** zaitcev_ has joined #openstack-swift | 15:36 | |
| *** ChanServ sets mode: +v zaitcev_ | 15:36 | |
| *** zaitcev has quit IRC | 15:40 | |
| *** diablo_rojo has joined #openstack-swift | 15:48 | |
| *** tesseract has quit IRC | 15:56 | |
| openstackgerrit | Thiago da Silva proposed openstack/swift master: fix ec overview doc formatting https://review.opendev.org/676446 | 16:01 |
| *** fungi has quit IRC | 16:42 | |
| *** fungi has joined #openstack-swift | 16:43 | |
| *** e0ne has quit IRC | 16:51 | |
| *** diablo_rojo has quit IRC | 16:52 | |
| *** mrjk has quit IRC | 16:54 | |
| *** psachin has quit IRC | 16:57 | |
| *** tdasilva has quit IRC | 16:58 | |
| *** diablo_rojo has joined #openstack-swift | 17:05 | |
| *** mrjk has joined #openstack-swift | 17:08 | |
| *** diablo_rojo has quit IRC | 17:10 | |
| *** diablo_rojo has joined #openstack-swift | 17:12 | |
| *** spsurya has quit IRC | 17:14 | |
| *** tdasilva has joined #openstack-swift | 17:23 | |
| *** ChanServ sets mode: +v tdasilva | 17:23 | |
| *** diablo_rojo has quit IRC | 17:24 | |
| *** diablo_rojo has joined #openstack-swift | 17:24 | |
| *** diablo_rojo has quit IRC | 17:51 | |
| *** tdasilva has quit IRC | 17:53 | |
| *** zaitcev__ has joined #openstack-swift | 17:54 | |
| *** ChanServ sets mode: +v zaitcev__ | 17:54 | |
| *** zaitcev_ has quit IRC | 17:57 | |
| *** gyee has joined #openstack-swift | 18:06 | |
| *** e0ne has joined #openstack-swift | 18:08 | |
| *** diablo_rojo has joined #openstack-swift | 18:08 | |
| *** e0ne has quit IRC | 18:10 | |
| *** diablo_rojo has quit IRC | 18:19 | |
| openstackgerrit | Merged openstack/swift master: s3api: Translate 503 to S3-style 503s https://review.opendev.org/674901 | 18:26 |
| ormandj | what ec_type are most people settling on nowadays? | 18:34 |
| *** zaitcev__ has quit IRC | 18:46 | |
| clayg | we've been using isa-l forever | 18:53 |
| clayg | in fact I think all swiftstack clusters do - never had any isa-l specific issue | 18:55 |
| ormandj | that's great to know, thanks clayg. what data/parity mix did you end up determining worked best for you? i'm in the middle of trying to figure out what direction i want to go | 18:58 |
| ormandj | i didn't really see any best practices for that/segment size as well in the docs, and figured i'd get a good starting point before i start trying to benchmark/calculate durability/etc | 18:59 |
| *** zaitcev__ has joined #openstack-swift | 19:00 | |
| *** ChanServ sets mode: +v zaitcev__ | 19:00 | |
| DHE | depends on your needs. more parity is more resilient but increases capacity requirements. that can be offset with more fragments but CPU usage goes up and more network connections are required | 19:01 |
| ormandj | sure, and gotta balance recovery times vs. failure rates/etc to make sure you get the durability you need | 19:02 |
| DHE | I'm leaning towards 10-20 data fragments and 3-5 parity fragments... | 19:02 |
| DHE | and probably isa-l with cauchy (sp?) due to the rs_vand issue with 5+ parity fragments | 19:02 |
| ormandj | i just figure there's at least a wet-finger-in-the-wind range people are finding works well on most typical deployments, which are gonna be spinny disks on some number of servers, some local some wan | 19:02 |
| clayg | one thing you have to watch out in small or mid-size clusters is the write amplification under heavy concurrency with clients - 100 concurrent PUTs on a 20+5 can make 2500 backend connections, you need to think about your per-disk write queue depth | 19:08 |
| ormandj | that's a good callout | 19:10 |
| *** tdasilva has joined #openstack-swift | 19:13 | |
| *** ChanServ sets mode: +v tdasilva | 19:13 | |
| ormandj | clayg: thanks for the protips | 19:30 |
| *** zaitcev_ has joined #openstack-swift | 19:31 | |
| *** ChanServ sets mode: +v zaitcev_ | 19:31 | |
| *** zaitcev__ has quit IRC | 19:34 | |
| openstackgerrit | Merged openstack/swift master: Add params to get_*_metadata internal client methods https://review.opendev.org/676414 | 20:14 |
| *** e0ne has joined #openstack-swift | 20:14 | |
| *** diablo_rojo has joined #openstack-swift | 20:20 | |
| *** diablo_rojo has quit IRC | 20:37 | |
| *** diablo_rojo has joined #openstack-swift | 20:37 | |
| *** diablo_rojo has quit IRC | 20:41 | |
| *** diablo_rojo has joined #openstack-swift | 20:48 | |
| kota_ | morning | 20:56 |
| mattoliverau | Morning | 20:59 |
| timburke | meeting time! | 21:00 |
| *** dsariel has joined #openstack-swift | 21:00 | |
| timburke | clayg, zaitcev_, rledisez, alecuyer ^^^ | 21:01 |
| *** diablo_rojo has quit IRC | 21:04 | |
| *** diablo_rojo has joined #openstack-swift | 21:09 | |
| *** e0ne has quit IRC | 21:11 | |
| *** e0ne has joined #openstack-swift | 21:13 | |
| *** e0ne has quit IRC | 21:13 | |
| kota_ | it looks like the original error is `No such file or directory` | 21:36 |
| rledisez | kota_: yeah, that's right | 21:36 |
| tdasilva | right, it looks like it is trying to rename from 0dbdb93b1ef1360902c56454414f0d221565777702.34279#0.data to 0dbdb93b1ef1360902c56454414f0d221565777702.34279#0#d.data | 21:37 |
| kota_ | so imo we should dig the point to touch real file... | 21:37 |
| kota_ | maybe rename_vfile method | 21:38 |
| rledisez | dig you check that the original file exists in the KV? | 21:38 |
| rledisez | *did | 21:38 |
| rledisez | eg: by making a fuse mount? | 21:39 |
| tdasilva | rledisez: no | 21:40 |
| rledisez | tdasilva: swift-mount-losf /some/mount/path | 21:41 |
| tdasilva | ah! nice | 21:42 |
| rledisez | you'll be able to navigate (readonly) in the KV | 21:42 |
| rledisez | and running some commande like swift-object-info | 21:42 |
| rledisez | there is a flag to enable to delete objects, but it's a bad idea to use it ;) | 21:42 |
| tdasilva | hehe, ok | 21:43 |
| mattoliverau | oh wow, cool! Nice tool | 21:44 |
| kota_ | hmm, rledisez suggest a way to debug so I'm going to look for my breakfast to get fresh brain | 21:44 |
| rledisez | take your time kota_, i'm about to leave | 21:45 |
| tdasilva | and i'm about to go to sleep :) | 21:45 |
| rledisez | take your time too tdasilva, a good night provides better debug skills ;) | 21:46 |
| tdasilva | rledisez, kota_ hopefully we'll have more feedback next week | 21:46 |
| tdasilva | rledisez: Darrel is running the tests internally i'm just relaying info.. | 21:47 |
| tdasilva | I have my losf saio, but haven't run into this yet, will keep trying to reproduce | 21:47 |
| clayg | what'd i miss!? | 21:47 |
| zaitcev_ | a meeting | 21:56 |
| timburke | basically, you guys are doing great work :-) | 22:01 |
| timburke | keep it up! | 22:01 |
| *** diablo_rojo has quit IRC | 22:05 | |
| *** diablo_rojo has joined #openstack-swift | 22:05 | |
| openstackgerrit | Merged openstack/swift master: fix ec overview doc formatting https://review.opendev.org/676446 | 22:17 |
| openstackgerrit | Merged openstack/swift master: py3: port test/functional/test_versioned_writes.py https://review.opendev.org/675710 | 22:17 |
| openstackgerrit | Merged openstack/swift stable/stein: s3api: Allow clients to upload with UNSIGNED-PAYLOAD https://review.opendev.org/675437 | 22:17 |
| openstackgerrit | Merged openstack/swift stable/stein: Fix quarantine when object path is not a directory https://review.opendev.org/675435 | 22:17 |
| zaitcev_ | timburke: do you happen to know as PTL when Train branches? | 22:21 |
| timburke | zaitcev_, https://releases.openstack.org/train/schedule.html indicates that we should have a final release around the first week of october or so | 22:23 |
| timburke | that's somewhat separate from the actual branch getting created; i forget exactly when that would happen -- probably another week or two after? | 22:23 |
| timburke | out of curiosity, why? | 22:24 |
| zaitcev_ | Our automated systems will start building py3 based release then. | 22:40 |
| *** BjoernT_ has quit IRC | 22:40 | |
| timburke | 👍 | 22:41 |
| *** tkajinam has joined #openstack-swift | 22:50 | |
| *** zaitcev_ has quit IRC | 22:59 | |
| *** tyreymer has joined #openstack-swift | 23:05 | |
| *** zaitcev_ has joined #openstack-swift | 23:13 | |
| *** ChanServ sets mode: +v zaitcev_ | 23:13 | |
| openstackgerrit | Merged openstack/swift stable/stein: reconciler: Enqueue right work for shard containers https://review.opendev.org/675433 | 23:14 |
| *** diablo_rojo has quit IRC | 23:18 | |
| *** diablo_rojo has joined #openstack-swift | 23:18 | |
| *** diablo_rojo has quit IRC | 23:31 | |
| *** tyreymer has quit IRC | 23:41 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: py3: fix up listings on sharded containers https://review.opendev.org/671167 | 23:47 |
| openstackgerrit | Tim Burke proposed openstack/swift master: py3: fix up swift-orphans https://review.opendev.org/670932 | 23:52 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!