| *** jamielennox|away is now known as jamielennox | 00:02 | |
| mattoliverau | Morning swift land, fyi, I'm going away for a few nights with the family.. trying to actually enjoy being unemployed. So if you don't hear much from me, that's why.. but I'll still be contactable, just might be a little delayed in responding :) | 00:27 |
|---|---|---|
| *** alenavolk has joined #openstack-swift | 00:40 | |
| *** links has joined #openstack-swift | 00:54 | |
| *** alenavolk has quit IRC | 01:04 | |
| *** tovin07_ has joined #openstack-swift | 01:04 | |
| *** alenavolk has joined #openstack-swift | 01:05 | |
| *** alenavolk has joined #openstack-swift | 01:08 | |
| *** alenavolk has joined #openstack-swift | 01:08 | |
| *** alenavolk has quit IRC | 01:09 | |
| *** alenavolk has joined #openstack-swift | 01:09 | |
| *** alenavolk has joined #openstack-swift | 01:10 | |
| *** alenavolk has quit IRC | 01:13 | |
| *** alenavolk has joined #openstack-swift | 01:13 | |
| kota_ | good morning | 01:28 |
| *** Sukhdev has joined #openstack-swift | 01:47 | |
| *** lan2 has joined #openstack-swift | 01:51 | |
| *** jamielennox is now known as jamielennox|away | 02:00 | |
| *** alenavolk has joined #openstack-swift | 02:09 | |
| *** alenavolk has quit IRC | 02:11 | |
| *** alenavolk has joined #openstack-swift | 02:11 | |
| *** winggundamth has joined #openstack-swift | 02:18 | |
| *** alenavolk has joined #openstack-swift | 02:28 | |
| *** klrmn has quit IRC | 02:36 | |
| *** janonymous has joined #openstack-swift | 02:52 | |
| mattoliverau | kota_: morning | 03:03 |
| *** two_tired has joined #openstack-swift | 03:03 | |
| kota_ | mattoliverau: o/ | 03:19 |
| *** psachin has joined #openstack-swift | 03:23 | |
| *** lan1 has joined #openstack-swift | 03:28 | |
| *** lan2 has quit IRC | 03:31 | |
| *** alenavolk has quit IRC | 03:53 | |
| *** jamielennox|away is now known as jamielennox | 03:53 | |
| *** jaosorior has joined #openstack-swift | 04:52 | |
| *** chsc has joined #openstack-swift | 04:53 | |
| *** chsc has quit IRC | 05:08 | |
| *** winggundamth has quit IRC | 05:28 | |
| *** ChubYann has quit IRC | 05:49 | |
| *** geaaru_ has quit IRC | 05:50 | |
| *** gyee has joined #openstack-swift | 06:08 | |
| *** chsc has joined #openstack-swift | 06:18 | |
| *** chsc has joined #openstack-swift | 06:18 | |
| *** rcernin has joined #openstack-swift | 06:19 | |
| *** chsc has quit IRC | 06:30 | |
| *** gyee has quit IRC | 06:31 | |
| *** pcaruana has joined #openstack-swift | 07:11 | |
| *** pcaruana has quit IRC | 07:12 | |
| *** pcaruana has joined #openstack-swift | 07:13 | |
| *** rcernin has quit IRC | 07:14 | |
| *** rcernin has joined #openstack-swift | 07:14 | |
| *** skudlik has joined #openstack-swift | 07:18 | |
| *** cbartz has joined #openstack-swift | 07:39 | |
| *** hoonetorg has quit IRC | 07:39 | |
| *** Sukhdev has quit IRC | 07:50 | |
| *** geaaru has joined #openstack-swift | 07:52 | |
| *** hoonetorg has joined #openstack-swift | 07:53 | |
| *** SkyRocknRoll has joined #openstack-swift | 08:04 | |
| *** adriant has quit IRC | 08:28 | |
| *** SkyRocknRoll has quit IRC | 09:06 | |
| *** oshritf has joined #openstack-swift | 09:26 | |
| *** mvk has quit IRC | 09:36 | |
| *** hseipp has joined #openstack-swift | 09:38 | |
| *** oshritf has quit IRC | 09:48 | |
| *** lan2 has joined #openstack-swift | 09:49 | |
| *** skudlik has quit IRC | 09:50 | |
| *** lan1 has quit IRC | 09:53 | |
| *** tovin07_ has quit IRC | 09:55 | |
| *** mvk has joined #openstack-swift | 10:06 | |
| *** kei_yama has quit IRC | 10:27 | |
| *** skudlik has joined #openstack-swift | 11:02 | |
| *** oshritf has joined #openstack-swift | 11:31 | |
| *** NM has joined #openstack-swift | 11:34 | |
| *** oshritf has quit IRC | 11:48 | |
| *** NM has quit IRC | 12:04 | |
| *** mvk has quit IRC | 12:07 | |
| *** chlong has joined #openstack-swift | 12:29 | |
| *** mvk has joined #openstack-swift | 12:32 | |
| *** foutatoro has joined #openstack-swift | 12:37 | |
| *** catintheroof has joined #openstack-swift | 12:41 | |
| *** chlong has quit IRC | 12:55 | |
| foutatoro | hi all, i would like to know how to deploy backup server with openstack-swift in order to retrieve data of containers in case of node failure ? | 12:58 |
| *** lan2 has quit IRC | 13:00 | |
| *** klamath has joined #openstack-swift | 13:06 | |
| *** klamath has quit IRC | 13:06 | |
| *** klamath has joined #openstack-swift | 13:07 | |
| *** NM has joined #openstack-swift | 13:31 | |
| *** MVenesio has joined #openstack-swift | 13:38 | |
| foutatoro | any suggestion ? | 13:45 |
| *** catinthe_ has joined #openstack-swift | 13:53 | |
| *** catinth__ has joined #openstack-swift | 13:53 | |
| *** catintheroof has quit IRC | 13:54 | |
| *** links has quit IRC | 13:57 | |
| *** catinthe_ has quit IRC | 13:57 | |
| *** oshritf has joined #openstack-swift | 14:06 | |
| *** oshritf has quit IRC | 14:12 | |
| *** two_tired has quit IRC | 14:14 | |
| *** two_tired has joined #openstack-swift | 14:29 | |
| *** psachin has quit IRC | 14:38 | |
| *** lucasxu has joined #openstack-swift | 14:56 | |
| *** mvk_ has joined #openstack-swift | 14:57 | |
| *** d0ugal has quit IRC | 14:57 | |
| *** chsc has joined #openstack-swift | 15:06 | |
| *** chsc has joined #openstack-swift | 15:06 | |
| *** abhitechie has joined #openstack-swift | 15:06 | |
| *** d0ugal has joined #openstack-swift | 15:07 | |
| *** _JZ_ has joined #openstack-swift | 15:23 | |
| *** alenavolk has joined #openstack-swift | 15:29 | |
| *** two_tired has quit IRC | 15:33 | |
| *** chlong has joined #openstack-swift | 15:38 | |
| *** klrmn has joined #openstack-swift | 15:41 | |
| *** gyee has joined #openstack-swift | 15:42 | |
| *** chsc has quit IRC | 15:47 | |
| notmyname | good morning | 15:48 |
| notmyname | foutatoro: I'm a little confused by your question (or what you're looking for). swift automatically stores multiple copies of everything in the system and works around any failures to ensure both high durability of the data and high availability of the system | 15:48 |
| *** skudlik has quit IRC | 15:50 | |
| *** pcaruana has quit IRC | 15:59 | |
| *** mvk_ has quit IRC | 16:03 | |
| *** lucasxu has quit IRC | 16:04 | |
| *** mvk has quit IRC | 16:07 | |
| *** cbartz has quit IRC | 16:08 | |
| *** klrmn has quit IRC | 16:13 | |
| *** rcernin has quit IRC | 16:14 | |
| *** lucasxu has joined #openstack-swift | 16:18 | |
| clayg | @notmyname: maybe "data of containers" means *containers* not buckets? | 16:23 |
| clayg | In which case. Swift is great for that! | 16:23 |
| *** hseipp has quit IRC | 16:25 | |
| *** JimCheung has joined #openstack-swift | 16:25 | |
| clayg | So. Container sharding. If one wanted to "rebalance" a container ring by hand because say one part was HUGE - can anyone think of a reasonable interface for that beyond direct manipulation of the replica2part2dev table? | 16:27 |
| clayg | Maybe an input file that "pins" parts to disks then you reduce the devices weight? Or a part->weight input and careful plumbing for "weight of one part" so that "parts wanted" gets decremented differently for *some* parts!? | 16:28 |
| *** rcernin has joined #openstack-swift | 16:31 | |
| *** abhitechie has quit IRC | 16:38 | |
| *** abhitechie has joined #openstack-swift | 16:39 | |
| timburke | good morning | 16:39 |
| *** chsc has joined #openstack-swift | 16:42 | |
| *** chsc has joined #openstack-swift | 16:42 | |
| *** abhitechie has quit IRC | 16:42 | |
| *** abhitechie has joined #openstack-swift | 16:43 | |
| *** jaosorior is now known as jaosorior_away | 16:45 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: fixup! Add container sharding to Swift containers https://review.openstack.org/450491 | 16:51 |
| *** saint_ has joined #openstack-swift | 16:54 | |
| *** chlong has quit IRC | 16:59 | |
| *** mvk has joined #openstack-swift | 17:05 | |
| *** _JZ_ has quit IRC | 17:08 | |
| *** chsc has quit IRC | 17:09 | |
| *** klrmn has joined #openstack-swift | 17:13 | |
| *** lucasxu has quit IRC | 17:14 | |
| *** chlong has joined #openstack-swift | 17:14 | |
| *** mvk_ has joined #openstack-swift | 17:15 | |
| *** chlong has quit IRC | 17:15 | |
| *** saint_ has quit IRC | 17:17 | |
| *** ChubYann has joined #openstack-swift | 17:18 | |
| *** abhitechie has quit IRC | 17:23 | |
| timburke | looking at patch 448240 | 17:23 |
| patchbot | https://review.openstack.org/#/c/448240/ - swift - Enable per policy proxy config options | 17:23 |
| *** vint_bra has joined #openstack-swift | 17:25 | |
| clayg | notmyname: who exactly is going to explain to the ML thread about dropping eventlet Swift's situation? | 17:29 |
| clayg | notmyname: http://lists.openstack.org/pipermail/openstack-dev/2017-May/117161.html | 17:31 |
| notmyname | yeah. I saw that. | 17:31 |
| clayg | like I don't even - where did this idea of "if we make it a community goal suddenly everyone gets a whole bag of round tuits in the mail" come from? | 17:32 |
| clayg | wanna get rid of eventlet - it's easy - rewrite everything (in maybe not-python) - done! | 17:32 |
| *** Sukhdev has joined #openstack-swift | 17:44 | |
| clayg | OMG! when did sam/torgomatic leave this channel? | 17:46 |
| *** tonanhngo has joined #openstack-swift | 17:48 | |
| clayg | neway - for other ring peeps (cschwede_ timburke kota_ ?) I think https://github.com/openstack/swift/blob/master/swift/common/ring/builder.py#L1294 is the only place we decrement parts_wanted on devices | 17:49 |
| clayg | so it seems at least *possible* we could consult a map like `dev['parts_wanted'] -= self.part2weight[part]` ??? | 17:50 |
| *** chsc has joined #openstack-swift | 17:51 | |
| *** chsc has joined #openstack-swift | 17:51 | |
| timburke | ah... for lumpy partitions? i kinda like that... | 17:52 |
| clayg | it's a shoddy solution - we'd have a += parts_wanted during gather - and also do something to move the extra weighted parts to the front of the reassign_parts list | 17:52 |
| timburke | how would we get that feedback to the builder, though? | 17:52 |
| clayg | timburke: yeah the ux would be crap - you'd have to keep a file that basically has a list of parts and their "extraness" | 17:52 |
| clayg | lines and lines of "<part#> <per-part-weight>" | 17:53 |
| clayg | upfront you take the replica*2**part-power count and + replica*sum_of_extra_weights so that the total "weight of one part" that gets smeared across the device weight it slightly smaller overall - as you add capacity (total available weight) and some parts get more and more out of whack weight wise it's a huge maintenance overhead :'( | 17:54 |
| *** mvk_ has quit IRC | 17:54 | |
| *** chlong has joined #openstack-swift | 17:55 | |
| *** skudlik has joined #openstack-swift | 17:55 | |
| timburke | roughly how many parts do we expect people to need to weight at a time? if it's something you only really need to do occasionally to put out a fire, we might be able to get away with having it just be a ring-builder subcommmand... | 17:55 |
| timburke | if we ever want weights for any significant fraction of all parts, though, we'll need to find something better | 17:56 |
| clayg | timburke: oh... just annotate it in the builder... `swift-ring-builder container.builder weight-part 95067 10` | 17:56 |
| timburke | yeah, i don't want yet another file | 17:57 |
| clayg | `swift-ring-builder container.builder weight-part` could list... yeah that's a better idea | 17:57 |
| clayg | I feel like even if you get upto a dozen or more it's already pretty bad - if it was "a significant percentage" like >1% that's already a lot :\ | 17:59 |
| *** Sukhdev has quit IRC | 18:01 | |
| timburke | yeah, so what we'd want to do is have somebody (auditors, maybe?) tally up an actual-bytes-used per partition, then aggregate that across the cluster, and feed *that* back into the builder. might could work | 18:02 |
| clayg | i definitely want the auditors to tally up useful shit! | 18:03 |
| clayg | wasn't someone at the operator feedback asking about holding connections draining bytes backend connections or something? I think lp bug #1572719 is still a thing - but I guess "high priority" might be lie :'( | 18:06 |
| openstack | Launchpad bug 1572719 in OpenStack Object Storage (swift) "Client may hold socket open after ChunkWriteTimeout" [High,Confirmed] https://launchpad.net/bugs/1572719 | 18:06 |
| *** geaaru has quit IRC | 18:09 | |
| *** lucasxu has joined #openstack-swift | 18:12 | |
| *** itlinux has joined #openstack-swift | 18:21 | |
| *** catinth__ has quit IRC | 18:28 | |
| *** catintheroof has joined #openstack-swift | 18:30 | |
| *** lucasxu has quit IRC | 18:41 | |
| *** itlinux has quit IRC | 18:41 | |
| *** itlinux has joined #openstack-swift | 18:42 | |
| *** lucasxu has joined #openstack-swift | 18:43 | |
| *** itlinux has quit IRC | 18:48 | |
| *** lucasxu has quit IRC | 18:58 | |
| *** lucasxu has joined #openstack-swift | 19:03 | |
| *** alenavolk has quit IRC | 19:19 | |
| *** skudlik has quit IRC | 19:33 | |
| *** mathiasb has quit IRC | 19:42 | |
| *** mathiasb has joined #openstack-swift | 19:44 | |
| *** skudlik has joined #openstack-swift | 19:47 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: fixup! Add container sharding to Swift containers https://review.openstack.org/450491 | 19:56 |
| *** alenavolk has joined #openstack-swift | 19:58 | |
| clayg | yay container sharding! | 19:58 |
| timburke | i'm just trying to get green balls on that so we can squash it in -- there's a lot of good stuff there, but i'm sure there's still a ways to go | 19:59 |
| *** JimCheung has quit IRC | 20:00 | |
| timburke | really not sure how PUT o1, o2, o3 resulted in ROWIDs 3, 1, 2 but w/e http://logs.openstack.org/91/450491/12/check/gate-swift-python27-ubuntu-xenial/4104479/console.html | 20:00 |
| *** JimCheung has joined #openstack-swift | 20:00 | |
| *** alenavolk has quit IRC | 20:04 | |
| *** JimCheung has quit IRC | 20:05 | |
| *** chsc has quit IRC | 20:06 | |
| *** alenavolk has joined #openstack-swift | 20:07 | |
| *** chlong has quit IRC | 20:07 | |
| *** chsc has joined #openstack-swift | 20:11 | |
| *** chsc has joined #openstack-swift | 20:11 | |
| *** chlong has joined #openstack-swift | 20:30 | |
| *** JimCheung has joined #openstack-swift | 20:31 | |
| *** JimCheung has quit IRC | 20:36 | |
| *** JimCheung has joined #openstack-swift | 20:39 | |
| *** chlong has quit IRC | 20:41 | |
| *** skudlik has quit IRC | 20:46 | |
| *** zaitcev has joined #openstack-swift | 21:02 | |
| *** ChanServ sets mode: +v zaitcev | 21:02 | |
| *** skudlik has joined #openstack-swift | 21:02 | |
| zaitcev | May 22 16:58:42 rhev-a24c-01 kernel: traps: swift-proxy-ser[1668] trap invalid opcode ip:7f28726574bd sp:7fff9881d6f0 error:0 | 21:02 |
| zaitcev | May 22 16:58:42 rhev-a24c-01 kernel: in liberasurecode.so.1.4.0[7f2872652000+d000] | 21:02 |
| zaitcev | eye dundu nufin | 21:02 |
| clayg | heh | 21:02 |
| clayg | can you tell if it's something in the proxy or reconstructor thats generating that? is it killing the process (segfault)? which backend? | 21:04 |
| zaitcev | proxy | 21:04 |
| timburke | zaitcev: what EC backend? isa-l? | 21:05 |
| zaitcev | how do I find out? I only have {"name": "ec0603", "aliases": "ec0603"} | 21:06 |
| zaitcev | well, the configured part is | 21:06 |
| zaitcev | [storage-policy:1] | 21:06 |
| zaitcev | name = ec0603 | 21:06 |
| zaitcev | policy_type = erasure_coding | 21:06 |
| zaitcev | ec_type = liberasurecode_rs_vand | 21:06 |
| notmyname | oh, interesting. so it's the software-only liberasurecode one | 21:07 |
| *** _JZ_ has joined #openstack-swift | 21:07 | |
| timburke | huh. nope, libec's RS vand... | 21:07 |
| zaitcev | I'm just suspicious if it actually obeys | 21:07 |
| zaitcev | I swear it used to work fine. But I didn't touch it in a while. Now I went to retrieve an object and it causes proxy to segfault immediately. | 21:08 |
| zaitcev | sounds like gdb time | 21:08 |
| timburke | yeah, that'd be best i think | 21:09 |
| timburke | at least it's reliable! | 21:09 |
| zaitcev | heh! | 21:10 |
| *** alenavolk has quit IRC | 21:14 | |
| *** catintheroof has quit IRC | 21:22 | |
| *** MVenesio has quit IRC | 21:23 | |
| *** MVenesio has joined #openstack-swift | 21:28 | |
| *** catintheroof has joined #openstack-swift | 21:29 | |
| *** NM has quit IRC | 21:34 | |
| *** _JZ_ has quit IRC | 21:35 | |
| *** vint_bra1 has joined #openstack-swift | 21:54 | |
| *** vint_bra2 has joined #openstack-swift | 21:56 | |
| *** vint_bra has quit IRC | 21:57 | |
| *** lucasxu has quit IRC | 21:58 | |
| *** MVenesio has quit IRC | 21:59 | |
| *** vint_bra1 has quit IRC | 21:59 | |
| *** tdasilva has quit IRC | 22:01 | |
| *** itlinux has joined #openstack-swift | 22:14 | |
| *** adriant has joined #openstack-swift | 22:19 | |
| *** klamath has quit IRC | 22:25 | |
| *** itlinux has quit IRC | 22:42 | |
| *** itlinux has joined #openstack-swift | 23:01 | |
| *** itlinux has quit IRC | 23:05 | |
| *** vint_bra2 has quit IRC | 23:16 | |
| *** NM has joined #openstack-swift | 23:19 | |
| *** skudlik has quit IRC | 23:27 | |
| *** chsc has quit IRC | 23:34 | |
| *** kei_yama has joined #openstack-swift | 23:36 | |
| *** catintheroof has quit IRC | 23:40 | |
| *** catintheroof has joined #openstack-swift | 23:48 | |
| *** tdasilva has joined #openstack-swift | 23:54 | |
| notmyname | kota_: good morning. are you around yet? | 23:55 |
| *** catintheroof has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!