| *** dmsimard is now known as dmsimard_away | 00:00 | |
| *** david-lyle is now known as david-lyle_afk | 00:06 | |
| *** rmcall has joined #openstack-swift | 00:06 | |
| *** tdasilva has quit IRC | 00:17 | |
| *** dmorita has joined #openstack-swift | 00:29 | |
| *** sungju has joined #openstack-swift | 00:39 | |
| *** oomichi has joined #openstack-swift | 00:43 | |
| *** rmcall has quit IRC | 00:47 | |
| *** rmcall has joined #openstack-swift | 00:49 | |
| *** rmcall has quit IRC | 00:49 | |
| openstackgerrit | Tushar Gohad proposed openstack/swift: EC: Allow tuning ec_object_segment_size per policy https://review.openstack.org/132389 | 01:01 |
|---|---|---|
| *** gyee has quit IRC | 01:40 | |
| *** haomaiwang has quit IRC | 01:47 | |
| * charz Testing stackforge/swift3 and fujita/swift3 with s3test | 01:51 | |
| *** nomad411 has joined #openstack-swift | 01:58 | |
| *** tamizh_geek has joined #openstack-swift | 01:58 | |
| *** nosnos has joined #openstack-swift | 02:02 | |
| *** tamizh_geek has quit IRC | 02:03 | |
| *** haomaiwang has joined #openstack-swift | 02:06 | |
| openstackgerrit | Tushar Gohad proposed openstack/swift: EC: Allow tuning ec_object_segment_size per policy https://review.openstack.org/132389 | 02:07 |
| openstackgerrit | Tushar Gohad proposed openstack/swift: WIP EC: HTTP PUT/HEAD changes for object size/etag reporting https://review.openstack.org/135488 | 02:11 |
| *** tsg has quit IRC | 02:13 | |
| *** NM has joined #openstack-swift | 02:24 | |
| *** X019 has quit IRC | 02:32 | |
| *** tamizh_geek has joined #openstack-swift | 02:38 | |
| *** X019 has joined #openstack-swift | 02:41 | |
| *** km has joined #openstack-swift | 02:46 | |
| *** km is now known as Guest30892 | 02:47 | |
| *** exploreshaifali has joined #openstack-swift | 02:58 | |
| *** nomad411 has quit IRC | 03:02 | |
| *** nomad411 has joined #openstack-swift | 03:06 | |
| *** tamizh_geek has quit IRC | 03:06 | |
| *** miqui has joined #openstack-swift | 03:07 | |
| *** nomad411 has quit IRC | 03:10 | |
| *** tkay has quit IRC | 03:14 | |
| *** tkay has joined #openstack-swift | 03:14 | |
| *** tkay has quit IRC | 03:17 | |
| *** nosnos has quit IRC | 03:20 | |
| *** tsg has joined #openstack-swift | 03:21 | |
| *** exploreshaifali has quit IRC | 03:22 | |
| *** tsg has quit IRC | 03:30 | |
| *** tsg has joined #openstack-swift | 03:31 | |
| *** NM has quit IRC | 03:31 | |
| *** tsg_ has joined #openstack-swift | 03:40 | |
| *** tsg has quit IRC | 03:40 | |
| *** X019 has quit IRC | 03:47 | |
| *** X019 has joined #openstack-swift | 03:51 | |
| *** exploreshaifali has joined #openstack-swift | 03:56 | |
| *** exploreshaifali has quit IRC | 03:59 | |
| *** rmcall has joined #openstack-swift | 04:03 | |
| *** X019 has quit IRC | 04:05 | |
| *** X019 has joined #openstack-swift | 04:08 | |
| *** nosnos has joined #openstack-swift | 04:12 | |
| *** flwang has quit IRC | 04:21 | |
| *** aswadr has joined #openstack-swift | 04:22 | |
| *** exploreshaifali has joined #openstack-swift | 04:24 | |
| *** flwang has joined #openstack-swift | 04:24 | |
| *** panbalag has quit IRC | 04:41 | |
| *** tellesnobrega has quit IRC | 04:44 | |
| *** zaitcev has quit IRC | 05:06 | |
| *** rmcall has quit IRC | 05:07 | |
| *** ppai has joined #openstack-swift | 05:34 | |
| *** tsg_ has quit IRC | 05:37 | |
| *** miqui has quit IRC | 05:44 | |
| *** mahatic_ has quit IRC | 05:44 | |
| *** k4n0 has joined #openstack-swift | 05:52 | |
| *** rmcall has joined #openstack-swift | 05:55 | |
| *** david-lyle_afk has quit IRC | 06:12 | |
| *** nshaikh has joined #openstack-swift | 06:34 | |
| *** k4n0 has quit IRC | 06:39 | |
| *** judd7 has quit IRC | 06:43 | |
| *** k4n0 has joined #openstack-swift | 06:53 | |
| *** rmcall has quit IRC | 07:13 | |
| *** kevinbenton has quit IRC | 07:22 | |
| *** k4n0 has quit IRC | 07:31 | |
| *** mmcardle has joined #openstack-swift | 07:35 | |
| *** k4n0 has joined #openstack-swift | 07:44 | |
| *** exploreshaifali has quit IRC | 07:46 | |
| *** nellysmitt has joined #openstack-swift | 07:53 | |
| *** SkyRocknRoll has joined #openstack-swift | 07:54 | |
| *** sungju has quit IRC | 08:02 | |
| *** exploreshaifali has joined #openstack-swift | 08:05 | |
| *** exploreshaifali has quit IRC | 08:06 | |
| *** exploreshaifali has joined #openstack-swift | 08:08 | |
| *** geaaru has joined #openstack-swift | 08:34 | |
| *** Guest30892 has quit IRC | 08:41 | |
| *** jordanP has joined #openstack-swift | 08:49 | |
| *** NM has joined #openstack-swift | 08:50 | |
| *** rawat_vedams has joined #openstack-swift | 08:53 | |
| *** NM has quit IRC | 08:54 | |
| *** dmorita has quit IRC | 09:03 | |
| *** nosnos has quit IRC | 09:07 | |
| *** rawat_vedams has quit IRC | 09:08 | |
| *** nosnos has joined #openstack-swift | 09:08 | |
| *** oomichi has quit IRC | 09:16 | |
| *** acoles_away is now known as acoles | 09:29 | |
| fifieldt_ | notmyname, looks like the translation proposal is failing for swift at the moment -- https://jenkins.openstack.org/job/swift-propose-translation-update/63/console | 09:31 |
| *** ppai has quit IRC | 09:32 | |
| fifieldt_ | it's looking for a locale directory that doesn't exist | 09:32 |
| *** nshaikh has quit IRC | 09:33 | |
| fifieldt_ | probably because it moved from local to swift/locale | 09:33 |
| fifieldt_ | so, that's a bug in the i18n script | 09:33 |
| fifieldt_ | I'm lodging that now | 09:33 |
| *** tamizh_geek has joined #openstack-swift | 09:33 | |
| *** joeljwright has joined #openstack-swift | 09:35 | |
| *** nshaikh has joined #openstack-swift | 09:36 | |
| fifieldt_ | https://bugs.launchpad.net/openstack-i18n/+bug/1394120 there's the bug link. also posted on the i18n ML | 09:36 |
| *** brnelson has quit IRC | 09:37 | |
| *** brnelson has joined #openstack-swift | 09:37 | |
| *** ppai has joined #openstack-swift | 09:45 | |
| *** exploreshaifali has quit IRC | 09:51 | |
| *** geaaru has quit IRC | 09:55 | |
| *** tamizh_geek has quit IRC | 10:00 | |
| *** geaaru has joined #openstack-swift | 10:08 | |
| *** Anticimex has quit IRC | 10:12 | |
| *** dmsimard_away is now known as dmsimard | 10:17 | |
| *** Anticimex has joined #openstack-swift | 10:18 | |
| *** aix has joined #openstack-swift | 10:23 | |
| *** X019 has quit IRC | 10:28 | |
| *** tellesnobrega has joined #openstack-swift | 10:38 | |
| *** tamizh_geek has joined #openstack-swift | 10:41 | |
| *** flwang has quit IRC | 10:44 | |
| *** tellesnobrega has quit IRC | 10:56 | |
| *** xianghui_ is now known as xianghui | 10:58 | |
| *** X019 has joined #openstack-swift | 11:04 | |
| *** erlon has joined #openstack-swift | 11:07 | |
| *** haomaiwang has quit IRC | 11:19 | |
| *** tamizh_geek has quit IRC | 11:32 | |
| *** haomaiwang has joined #openstack-swift | 11:37 | |
| *** tellesnobrega has joined #openstack-swift | 11:40 | |
| *** dmsimard is now known as dmsimard_away | 11:43 | |
| *** ekarlso- has quit IRC | 11:44 | |
| *** ekarlso- has joined #openstack-swift | 11:49 | |
| *** aswadr has quit IRC | 11:50 | |
| *** aix has quit IRC | 11:52 | |
| *** nellysmitt has quit IRC | 12:05 | |
| *** Anticimex has quit IRC | 12:13 | |
| *** panbalag has joined #openstack-swift | 12:15 | |
| *** tamizh_geek has joined #openstack-swift | 12:17 | |
| *** dmsimard_away is now known as dmsimard | 12:18 | |
| *** Anticimex has joined #openstack-swift | 12:19 | |
| *** ppai has quit IRC | 12:35 | |
| *** tdasilva has joined #openstack-swift | 12:35 | |
| *** tdasilva has quit IRC | 12:39 | |
| *** a1|away is now known as AbyssOne | 12:39 | |
| *** mahatic has joined #openstack-swift | 12:41 | |
| *** tellesnobrega has quit IRC | 12:42 | |
| *** aix has joined #openstack-swift | 12:42 | |
| *** ekarlso- has quit IRC | 12:50 | |
| *** ekarlso- has joined #openstack-swift | 12:50 | |
| *** flwang has joined #openstack-swift | 12:52 | |
| *** mmcardle has quit IRC | 13:05 | |
| *** tamizh_geek has quit IRC | 13:11 | |
| *** SkyRocknRoll has quit IRC | 13:14 | |
| *** NM has joined #openstack-swift | 13:15 | |
| *** exploreshaifali has joined #openstack-swift | 13:20 | |
| *** nellysmitt has joined #openstack-swift | 13:34 | |
| *** mmcardle has joined #openstack-swift | 13:37 | |
| *** tellesnobrega has joined #openstack-swift | 13:40 | |
| *** tellesnobrega has quit IRC | 13:40 | |
| *** miqui has joined #openstack-swift | 13:51 | |
| *** jistr has joined #openstack-swift | 13:53 | |
| *** jistr is now known as jistr|mtg | 13:53 | |
| *** exploreshaifali has quit IRC | 14:02 | |
| openstackgerrit | Daniel Wakefield proposed openstack/python-swiftclient: Fix misplaced check for None in SwiftUploadObject. https://review.openstack.org/133107 | 14:06 |
| *** panbalag has quit IRC | 14:12 | |
| openstackgerrit | Andreas Jaeger proposed openstack/swift: Fix translation setup https://review.openstack.org/135606 | 14:13 |
| fifieldt_ | ^^ that's a simple fix for the translation import problem we were discussing yeserday notmyname | 14:16 |
| *** mahatic has quit IRC | 14:20 | |
| *** mahatic has joined #openstack-swift | 14:30 | |
| *** openstackgerrit has quit IRC | 14:33 | |
| *** openstackgerrit has joined #openstack-swift | 14:34 | |
| *** zul has joined #openstack-swift | 14:35 | |
| *** SkyRocknRoll__ has joined #openstack-swift | 14:37 | |
| *** nellysmitt has quit IRC | 14:41 | |
| *** pberis1 has quit IRC | 14:46 | |
| *** openstackgerrit has quit IRC | 14:49 | |
| *** openstackgerrit has joined #openstack-swift | 14:50 | |
| *** nosnos has quit IRC | 14:52 | |
| *** geaaru has quit IRC | 14:52 | |
| *** exploreshaifali has joined #openstack-swift | 14:54 | |
| *** zul has quit IRC | 14:55 | |
| *** nellysmitt has joined #openstack-swift | 15:00 | |
| *** tamizh_geek has joined #openstack-swift | 15:02 | |
| *** fbo has quit IRC | 15:04 | |
| *** bill_az has joined #openstack-swift | 15:05 | |
| *** geaaru has joined #openstack-swift | 15:05 | |
| *** fbo has joined #openstack-swift | 15:10 | |
| *** zul has joined #openstack-swift | 15:16 | |
| *** pberis1 has joined #openstack-swift | 15:20 | |
| *** zaitcev has joined #openstack-swift | 15:29 | |
| *** ChanServ sets mode: +v zaitcev | 15:29 | |
| *** judd7 has joined #openstack-swift | 15:36 | |
| *** tdasilva has joined #openstack-swift | 15:49 | |
| *** lpabon has joined #openstack-swift | 16:01 | |
| zaitcev | Have we found a sponsor and location for the next Hackathon? | 16:04 |
| *** tamizh_geek has quit IRC | 16:14 | |
| *** rmcall has joined #openstack-swift | 16:15 | |
| *** tamizh_geek has joined #openstack-swift | 16:26 | |
| *** nshaikh has quit IRC | 16:30 | |
| *** nellysmitt has quit IRC | 16:39 | |
| *** SkyRocknRoll__ has quit IRC | 16:42 | |
| *** gyee has joined #openstack-swift | 16:45 | |
| *** tamizh_geek has quit IRC | 16:49 | |
| openstackgerrit | Alistair Coles proposed openstack/swift-specs: Spec for at-rest encryption. https://review.openstack.org/123220 | 16:50 |
| acoles | torgomatic: ^^ just a rebase to new dir structure | 16:51 |
| *** tamizh_geek has joined #openstack-swift | 16:53 | |
| *** k4n0 has quit IRC | 16:56 | |
| notmyname | good morning | 16:57 |
| notmyname | fifieldt_: thanks. I'll look at it this morning | 16:57 |
| mahatic | good morning | 16:57 |
| notmyname | zaitcev: not yet. I was waiting until I got back from the summit. it's something I need to do soon | 16:57 |
| notmyname | reminder to all that the meeting today is in 2 hours | 16:58 |
| zaitcev | notmyname: thanks. As usual there are some budgetary issues and I'm being poked about it. | 16:58 |
| notmyname | zaitcev: of course | 16:58 |
| zaitcev | I heard the M summit is slated for Tokyo | 16:58 |
| zaitcev | I'll be there even if I have to pay my way | 16:59 |
| peluse | morning | 16:59 |
| notmyname | zaitcev: next 2 are vancouver and tokyo | 16:59 |
| notmyname | how do I find the gerrit patch from the jenkins output? I want to look at the translation patch | 17:05 |
| notmyname | fifieldt_: ^ | 17:05 |
| *** tkay has joined #openstack-swift | 17:08 | |
| *** IRTermite has quit IRC | 17:09 | |
| *** IRTermite has joined #openstack-swift | 17:09 | |
| *** NM has quit IRC | 17:10 | |
| *** NM has joined #openstack-swift | 17:10 | |
| NM | Next summits are going to be on Vancouver and Tokyo? | 17:13 |
| eikke | yes | 17:18 |
| *** gvernik has joined #openstack-swift | 17:19 | |
| eikke | swifterdarrell: should I put that bindings code in swift.common.utils as it is now, or create a new module? (it's > 3200 lines now, that module) | 17:22 |
| *** atan8 has joined #openstack-swift | 17:35 | |
| *** tamizh_geek has quit IRC | 17:37 | |
| *** geaaru has quit IRC | 17:38 | |
| *** tamizh_geek has joined #openstack-swift | 17:41 | |
| swifterdarrell | eikke: I'd put it in utils, but that's open for debate | 17:42 |
| swifterdarrell | eikke: if you did pull it out, I'd probably expect all the ctypes-wrappers-for-syscall/glibc-functions to be in that new file | 17:43 |
| swifterdarrell | eikke: with some name that reflects that | 17:43 |
| eikke | hmh, k. I called it 'swift.common.splice' for now, and will push it like that, but that's open for debate, I won't bikeshed over this (but I do think having a proper module is the way to go) | 17:44 |
| * swifterdarrell shrugs | 17:45 | |
| swifterdarrell | I think one file per callable function (e.g. splice()) is a bit much, but I agree it's not worth bikeshedding over ;) I'll +2 it as you did it | 17:46 |
| *** jordanP has quit IRC | 17:56 | |
| *** aix has quit IRC | 17:57 | |
| *** erlon has quit IRC | 17:59 | |
| *** rmcall has quit IRC | 18:01 | |
| *** rmcall has joined #openstack-swift | 18:01 | |
| *** nshaikh has joined #openstack-swift | 18:03 | |
| notmyname | reminder that the swift meeting is in 50 minutes | 18:10 |
| notmyname | (I keep saying things because this is the first meeting since the US time change) | 18:10 |
| *** tamizh_geek has quit IRC | 18:11 | |
| *** mmcardle has quit IRC | 18:11 | |
| *** Guest99387 has joined #openstack-swift | 18:22 | |
| *** Guest99387 is now known as linuxaddicts | 18:22 | |
| *** exploreshaifali has quit IRC | 18:23 | |
| *** tdasilva has quit IRC | 18:31 | |
| *** zul has quit IRC | 18:31 | |
| linuxaddicts | hi, i want to scale test swift as a glance back end. | 18:36 |
| linuxaddicts | will uploading few 100 images, creating boot volume and booting off these volumes be a good start? how would this really test swift? | 18:37 |
| clayg | linuxaddicts: well that would test a lot of things - do you have another glance backend you'd be able to repeat the test and compare the results with? | 18:40 |
| linuxaddicts | sure, the other one would be the default | 18:41 |
| zaitcev | I'm pretty sure Swift will lose. | 18:43 |
| notmyname | zaitcev: so full of cheer and hope | 18:43 |
| clayg | linuxaddicts: the default is what... a local file system? | 18:44 |
| zaitcev | 100 of 10G images is merely 1TB, so for this test all you need to do is buy a $110 2T hybrid drive and plop it into Glance node, format it with XFS, done. | 18:44 |
| linuxaddicts | clayg, yes the default is the local file system. | 18:45 |
| zaitcev | To make a Swift this small, you need a ton of processes started, integrate it with your auth. I know beginners keep blowing stuff up doing that. | 18:45 |
| zaitcev | But most importantly, Swift will keep 3 copies and will send all the traffic through gazillion processes on the same CPU and will lose in performance. | 18:46 |
| *** tdasilva has joined #openstack-swift | 18:46 | |
| zaitcev | It basically is one of those silly micro-benchmarks at which we're very bad | 18:46 |
| clayg | stupid ec has too many words | 18:47 |
| clayg | so an object is a fragment archive, so that means it holds a number of fragments, one for each segment | 18:47 |
| zaitcev | Now if linuxaddicts had 1000 clients all uploading and launching VMs on 100 compute nodes, then we would crush Glance's default backend into a bloody pancake due to scalability | 18:48 |
| notmyname | linuxaddicts: zaitcev is right and you mentioned "scale". in the swift world, "scale" is generally though of in billions of objects and PB of data. not 100. | 18:48 |
| notmyname | linuxaddicts: what is it you're trying to test of find out? do you have a specific use case you're building for? | 18:48 |
| clayg | but the metadata name is "x-object-sysmeta-ec-segment-size" - isn't that the size of the fragment? or do I have to back calculate from the segment-size how much I need to read off the object to get one fragment from this object that is needed to rebuild the segment? | 18:49 |
| zaitcev | If we're interested winning comparos like the one linuxaddicts was proposing, we ought to accelerate the "in-process" Swift that portante was proposing. IMHO though. | 18:51 |
| notmyname | zaitcev: you mean tdasilva? | 18:52 |
| swifterdarrell | clayg: at least one of peluse's summit talks had a whole slide (IIRC) for a terminology glossary... | 18:52 |
| clayg | swifterdarrell: yeah well, i'm not sure we're internally consistent | 18:52 |
| zaitcev | I didn't think if tdasilva was going to write it himself | 18:52 |
| zaitcev | But yeah, those two :-) | 18:52 |
| swifterdarrell | clayg: *nod* if so, surely everyone would agree that's undesirable | 18:53 |
| notmyname | zaitcev: tdasilva and ignacio were talking about it. the single-process swift idea has some disadvantages. like mixing net io with disk io and eventlet not really working well there | 18:54 |
| clayg | swifterdarrell: well it may be that there's more than one way to look at it, or it may be that the ec-metadata *is* the segment-size and I need to use math to get the fragment-size - in which case it seems less convienent to store the segment-size, but at least it's consistent use of the words | 18:54 |
| swifterdarrell | zaitcev: notmyname: optimizing Swift to serve 1T of data from one server well, just doesn't seem interesting to me... | 18:54 |
| notmyname | swifterdarrell: :-) | 18:55 |
| zaitcev | swifterdarrell: as long as you guys keep winning deals that start with 2 racks and go up, I can see why | 18:55 |
| swifterdarrell | clayg: *nod* I have no idea, myself, but a single glossary page *somewhere* for Swift EC stuff would be helpful if it doesn't already exist, to keep everything in line | 18:55 |
| linuxaddicts | notmyname, the use case i have in mind is to see how swift will behave in a scale environment. and since you clarified the data set will be huge and 100's may not really be worthwhile. | 18:55 |
| notmyname | swifterdarrell: http://specs.openstack.org/openstack/swift-specs/specs/in_progress/erasure_coding.html#terminology | 18:56 |
| swifterdarrell | zaitcev: there's a lot of room between "2 racks" and "1 2T disk"... | 18:56 |
| swifterdarrell | zaitcev: but I can understand the allure of hyperbole ;) | 18:56 |
| notmyname | swifterdarrell: are you saying that hyperbole is the best thing ever? | 18:57 |
| linuxaddicts | the next question is what will the above test i mentioned really infer if i were to run it? | 18:57 |
| swifterdarrell | notmyname: oh, gosh, I wouldn't go *that* far | 18:57 |
| notmyname | linuxaddicts: most of the time people try out swift by building asmaller cluster (eg 10-20% of what they want) and then putting a similar percentage of their target workload to it | 18:57 |
| zaitcev | yeah | 18:58 |
| *** nshaikh has left #openstack-swift | 18:59 | |
| notmyname | linuxaddicts: there's a lot of differences in a local disk and swift (beyond perf differences for a give workload). swift is designed to provide durable storage that's always available and scales really well. in your proposed test, as described, it's more of a works vs doesn't work. then you have to make a choice based on cost, requirements, etc | 19:00 |
| notmyname | linuxaddicts: but as zaitcev pointed out, you might in fact just be testing glance's ability to handle your use case | 19:00 |
| notmyname | swift meeting time in #openstack-meeting | 19:00 |
| linuxaddicts | notmyname, thanks for confirming. i believe the same it will be more of a glance test rather than swift. | 19:01 |
| peluse | so the EC design spec is that single glossary you guys are speaking of | 19:01 |
| peluse | will send link after meeting | 19:01 |
| clayg | peluse: notmyname already sent http://specs.openstack.org/openstack/swift-specs/specs/in_progress/erasure_coding.html#terminology | 19:05 |
| clayg | peluse: but it didn't really say what x-object-sysmeta-ec-segment-size was exactly | 19:05 |
| clayg | although I think interestingly enough it is the segment-size, and I need to use get_segment_info to figure out the fragment_size :\ | 19:07 |
| peluse | clayg, we already read the whole object when doing a reconstruction and buffer up fragments to get a segment size to feed to pyeclib | 19:08 |
| peluse | why do you need to know fragment size? | 19:09 |
| clayg | yup that did it | 19:11 |
| peluse | BTW it does look like the metadata section in the design spec needs to be updated in that area... | 19:11 |
| clayg | peluse: i'm looking at the decode path, I don't think we want the proxy to read the whole object when doing a GET ;) | 19:12 |
| clayg | peluse: it seems like you need to figure out the fragment_size when building up your fragment_payloads for decode | 19:13 |
| clayg | peluse: i.e. how much to i read from each node before handing it off the list of fragments to pyeclib to generate the segment which I can send out to the client | 19:13 |
| peluse | a segment_szie worth | 19:14 |
| clayg | peluse: nope, a fragment_size worth | 19:14 |
| clayg | i started reading a segment size and I kept getting garbage back from decode | 19:14 |
| *** tkay has quit IRC | 19:14 | |
| peluse | oh yeah, sorry was thinking encode | 19:15 |
| clayg | yeah I think we're pretty solid on encode | 19:15 |
| clayg | but since sam is on vacation this week I thought I'd look at decode a bit more so hopefully I'll be able to contribute to a working GET path | 19:16 |
| *** tkay has joined #openstack-swift | 19:16 | |
| clayg | also on my list is that concurrent GET review you posted on priority reviews | 19:16 |
| peluse | sweet | 19:17 |
| peluse | think I'm gonna head home and work from there after the meeting, if you have time for a quick phone call would like to chat a little about current reconstructor updated_deleted() stuff | 19:18 |
| clayg | peluse: link me some context? | 19:19 |
| peluse | yup https://review.openstack.org/#/c/131872/ and am currently working on update() | 19:20 |
| clayg | peluse: but yeah I can chat, now or soon (w/i ~2-3 hours) would good | 19:20 |
| *** jordanP has joined #openstack-swift | 19:20 | |
| peluse | OK, it will take me about 30 min to get home after the swift meeting, will ping you when I'm back online | 19:20 |
| clayg | k | 19:22 |
| *** bkopilov has quit IRC | 19:25 | |
| *** zul has joined #openstack-swift | 19:26 | |
| *** rcrit has quit IRC | 19:29 | |
| *** rcrit has joined #openstack-swift | 19:30 | |
| *** exploreshaifali has joined #openstack-swift | 19:38 | |
| mattoliverau | Cools well I'm going back to my holiday, have fun peoples | 19:53 |
| * portante comes out with the sound of Lurch from the Adams Family TV show ... | 19:53 | |
| portante | zaitcev: you rang? | 19:53 |
| acoles | mattoliverau: take it easy | 19:53 |
| zaitcev | portante: mentioned you in conjunction with in-process something. | 19:53 |
| notmyname | portante: just missed the meeting | 19:54 |
| portante | story of my life | 19:54 |
| acoles | portante: its not a bad way to live :) | 19:54 |
| portante | :) | 19:55 |
| acoles | notmyname: here's another brown bag talk for your youtube index https://www.youtube.com/watch?v=EMLK5XsU0_8 | 19:57 |
| notmyname | thanks | 19:58 |
| *** zul has quit IRC | 19:59 | |
| *** acoles is now known as acoles_away | 19:59 | |
| *** cutforth has joined #openstack-swift | 19:59 | |
| *** tdasilva has quit IRC | 20:00 | |
| *** gvernik has quit IRC | 20:00 | |
| *** zul has joined #openstack-swift | 20:10 | |
| *** tdasilva has joined #openstack-swift | 20:13 | |
| *** dmsimard is now known as dmsimard_away | 20:16 | |
| cebruns | Hi - new to Swift, and I am looking for a way to determine the amount of capacity a cluster has. Free/Used/etc - is there a command that provides that? | 20:25 |
| notmyname | cebruns: do you already have a swift cluster, or are you asking in general? | 20:26 |
| clayg | doesn't recon have something for doing a a distributed df - like for doing drive distribution histograms? | 20:26 |
| cebruns | I'm hacking around with a SAIO installation. | 20:26 |
| notmyname | ya, recon has a good way to see capacity used/available on individual drives | 20:27 |
| cebruns | Yeah - I tried swift-recon on the SAIO and it reports "Error. No drive info available" - I imagine it's because they're loopback devices, but wasn't sure. | 20:28 |
| notmyname | cebruns: ya | 20:29 |
| peluse | clayg, ready when you are. call me at 480 554 3688 or I can call you - either way | 20:29 |
| clayg | lemme find a phone - you don't skype right? | 20:29 |
| peluse | clayg, I can probably make that work if you can't find a phone. I use Lync w/softphone | 20:29 |
| notmyname | cebruns: so in that case, swift usage, what with all the complexities of replication etc, is really 2 sums (depending on what you're looking for) | 20:29 |
| notmyname | cebruns: the first is the sum of `df` on all of the drives. that gives you the total capacity used in your cluster: all replicas, all metadata, all filesystem overhead | 20:30 |
| notmyname | cebruns: the second is the sum of the "bytes-used" in every account. that's the total of the usable capacity used in the cluster (ie what you probably do billing on) | 20:31 |
| *** linuxaddicts has quit IRC | 20:31 | |
| clayg | peluse: facebook:clay.gerrard <- me skype | 20:31 |
| cebruns | notmyname: That makes sense. Thanks! Guess I need to install some "real drives" to check it out. | 20:32 |
| *** nomad411 has joined #openstack-swift | 20:33 | |
| *** NM has quit IRC | 20:34 | |
| clayg | thanks paul! | 20:47 |
| peluse | no, thank you! | 20:47 |
| *** rcrit has quit IRC | 20:50 | |
| *** rcrit has joined #openstack-swift | 20:55 | |
| tdasilva | hi, i have a question about versioned writes: lwhat should be the behavior if during a PUT the COPY of the previous obj to the 'versions' container was successful, but the actual PUT failed | 20:57 |
| tdasilva | we don't remove the copy that was made right? | 20:58 |
| notmyname | tdasilva: what's the current behavior? | 20:58 |
| tdasilva | notmyname: sorry, i believe that is the current behavior, just wanted to confirm that i wasn't missing something | 20:59 |
| tdasilva | notmyname: I guess my follow up question would be was the the intended purpose or just something that was overlooked, do we care in other words? | 21:01 |
| notmyname | tdasilva: IIRC that was the intended purpose (but that was years ago) | 21:01 |
| tdasilva | notmyname: ok, I'm not going to be concerned with that for now then | 21:03 |
| peluse | clayg so wrt the hashes.pkl notion, actually the implementation I was pursuing was using the fact that I, as the local reconstructor, know what I'm interested in from my partners, the FI that matches my node index, so that I can | 21:05 |
| peluse | now pass that on via get)hashes and the hashes.pkl will only be caluclated for relevant .data files | 21:05 |
| peluse | meaning on hashes.pkl format and set of related functions... but requires the stamp file thing still. | 21:05 |
| peluse | not a ton of code to do it this way so will complete it and then we can look closer | 21:05 |
| peluse | on=one there one post ago.... | 21:06 |
| peluse | "one hashes.pkj format" | 21:06 |
| *** shakamunyi_ has joined #openstack-swift | 21:11 | |
| *** jasondotstar has joined #openstack-swift | 21:12 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 21:12 |
| *** jasondotstar has quit IRC | 21:13 | |
| *** cutforth has quit IRC | 21:13 | |
| *** jasondotstar has joined #openstack-swift | 21:14 | |
| clayg | peluse: maybe if hashes.pkl is {fragment_index: {suffix: hash}} | 21:15 |
| clayg | peluse: then you're still only decode_hashes_pkl.keys() away from your list of stamps | 21:16 |
| clayg | peluse: and maybe also there's an oppertunity to optomize the number of dirs to traverse if you have 10K si[3]'s and only 1 si[2] | 21:17 |
| clayg | peluse: but it's fine, start with something that works | 21:17 |
| peluse | yeah, I think I might have to go that route anyway. think I'm about to run into an issue with how I was trying to do it.... will keep you posted | 21:21 |
| *** gyee has quit IRC | 21:26 | |
| *** bill_az has quit IRC | 21:26 | |
| openstackgerrit | paul luse proposed openstack/swift: SAIO updates to support EC development environment. https://review.openstack.org/128154 | 21:27 |
| *** gyee has joined #openstack-swift | 21:28 | |
| *** gyee has quit IRC | 21:28 | |
| *** gyee has joined #openstack-swift | 21:29 | |
| *** gyee has quit IRC | 21:29 | |
| *** gyee has joined #openstack-swift | 21:30 | |
| *** gyee has quit IRC | 21:31 | |
| clayg | man i am having terrible luck with gerrit today :\ | 21:38 |
| *** exploreshaifali has quit IRC | 21:39 | |
| *** shakamunyi_ has quit IRC | 21:59 | |
| *** rmcall has quit IRC | 22:05 | |
| *** rmcall has joined #openstack-swift | 22:07 | |
| *** zul has quit IRC | 22:07 | |
| *** miqui has quit IRC | 22:09 | |
| *** openstackgerrit has quit IRC | 22:10 | |
| *** openstackgerrit has joined #openstack-swift | 22:10 | |
| *** zul has joined #openstack-swift | 22:11 | |
| *** jasondotstar has quit IRC | 22:24 | |
| *** mahatic has quit IRC | 22:33 | |
| *** jistr|mtg has quit IRC | 22:42 | |
| *** zul has quit IRC | 22:43 | |
| *** rmcall has quit IRC | 22:47 | |
| *** rmcall has joined #openstack-swift | 22:49 | |
| *** gyee has joined #openstack-swift | 22:55 | |
| *** gyee has quit IRC | 22:55 | |
| *** jasondotstar has joined #openstack-swift | 23:08 | |
| *** nomad411 has quit IRC | 23:14 | |
| *** jordanP has quit IRC | 23:17 | |
| *** gyee has joined #openstack-swift | 23:18 | |
| *** jasondotstar has quit IRC | 23:33 | |
| *** jasondotstar has joined #openstack-swift | 23:36 | |
| *** judd7 has quit IRC | 23:39 | |
| openstackgerrit | Merged openstack/swift: Adds v1 API documentation to doc/source/api https://review.openstack.org/135102 | 23:39 |
| openstackgerrit | Merged openstack/swift: Fix translation setup https://review.openstack.org/135606 | 23:41 |
| openstackgerrit | Nicolas Trangez proposed openstack/swift: Rework `splice` and `tee` https://review.openstack.org/135319 | 23:43 |
| eikke | swifterdarrell: ^^ (you asked me to ping when I made the changes) | 23:45 |
| *** zul has joined #openstack-swift | 23:47 | |
| eikke | 94% test coverage on the module ^_^ (and the remaining 6 lines would need some really old glibc) | 23:48 |
| notmyname | eikke: nice :-) | 23:53 |
| peluse | anyone seen/heard from gholt latelt? Was looking for an old blog series he did on consistent hashing but can't seem to find it (last known location http://greg.brim.net/page/building_a_consistent_hashing_ring.html) | 23:53 |
| eikke | notmyname: ty | 23:55 |
| eikke | time for a series episode & sleep now, 1am here | 23:55 |
| *** joeljwright has quit IRC | 23:55 | |
| *** sungju has joined #openstack-swift | 23:55 | |
| notmyname | peluse: seems all the domain names I've known him to use no longer respond. | 23:57 |
| notmyname | ahale: dfg: do you know where gholt's blog went? | 23:57 |
| ahale | nope no idea sorry | 23:58 |
| *** tdasilva has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!