| *** zaitcev has quit IRC | 00:07 | |
| *** gyee has quit IRC | 00:08 | |
| *** kyles_ne_ has quit IRC | 00:15 | |
| *** kyles_ne has joined #openstack-swift | 00:15 | |
| *** kyles_ne has quit IRC | 00:20 | |
| *** rmcall has quit IRC | 00:38 | |
| *** shri has quit IRC | 00:50 | |
| *** judd7 has quit IRC | 01:04 | |
| *** addnull has joined #openstack-swift | 01:24 | |
| *** dmsimard_away is now known as dmsimard | 01:27 | |
| *** dmsimard is now known as dmsimard_away | 01:28 | |
| *** mahatic_ has quit IRC | 01:29 | |
| *** jamiehannaford has joined #openstack-swift | 01:42 | |
| *** tsg has quit IRC | 01:57 | |
| *** mitz_ has quit IRC | 02:11 | |
| *** nosnos has joined #openstack-swift | 02:14 | |
| *** echevemaster has quit IRC | 02:17 | |
| *** mitz_ has joined #openstack-swift | 02:18 | |
| *** hhuang has joined #openstack-swift | 02:27 | |
| *** mrsnivvel has quit IRC | 02:33 | |
| openstackgerrit | A change was merged to openstack/swift: fix a comment https://review.openstack.org/125446 | 02:39 |
|---|---|---|
| *** jamiehannaford has quit IRC | 02:43 | |
| *** X019 has quit IRC | 02:51 | |
| *** mahatic has joined #openstack-swift | 02:51 | |
| *** X019 has joined #openstack-swift | 02:54 | |
| *** Alex_Gaynor has left #openstack-swift | 03:02 | |
| *** openstackgerrit has quit IRC | 03:08 | |
| *** openstackgerrit has joined #openstack-swift | 03:11 | |
| *** mrsnivvel has joined #openstack-swift | 03:13 | |
| *** openstackgerrit has quit IRC | 03:18 | |
| *** openstackgerrit has joined #openstack-swift | 03:19 | |
| *** oomichi has joined #openstack-swift | 03:25 | |
| *** mlauter has joined #openstack-swift | 03:28 | |
| *** nosnos has quit IRC | 03:29 | |
| *** nosnos has joined #openstack-swift | 03:30 | |
| *** DisneyRicky has quit IRC | 03:31 | |
| *** nosnos has quit IRC | 03:34 | |
| *** mlauter has quit IRC | 03:35 | |
| openstackgerrit | Michael Barton proposed a change to openstack/swift: sendfile for GETs https://review.openstack.org/125548 | 04:03 |
| redbo | oh I forgot to run flake8 | 04:07 |
| *** vdreamarkitex has quit IRC | 04:07 | |
| openstackgerrit | Michael Barton proposed a change to openstack/swift: sendfile for GETs https://review.openstack.org/125548 | 04:09 |
| *** nosnos has joined #openstack-swift | 04:19 | |
| *** haomaiwang has joined #openstack-swift | 04:50 | |
| *** DisneyRicky has joined #openstack-swift | 04:59 | |
| *** haomaiwang has quit IRC | 05:02 | |
| *** haomaiwang has joined #openstack-swift | 05:02 | |
| *** SkyRocknRoll__ has joined #openstack-swift | 05:12 | |
| *** haomaiw__ has joined #openstack-swift | 05:15 | |
| *** haomaiw__ has quit IRC | 05:16 | |
| *** haomaiwang has quit IRC | 05:16 | |
| *** haomaiwang has joined #openstack-swift | 05:17 | |
| openstackgerrit | Matthew Oliver proposed a change to openstack/swift: Basic container sharding middleware https://review.openstack.org/125553 | 05:27 |
| *** haomaiwang has quit IRC | 05:35 | |
| *** haomaiwa_ has joined #openstack-swift | 05:35 | |
| *** hhuang has quit IRC | 05:46 | |
| *** ttrumm has joined #openstack-swift | 06:05 | |
| *** ttrumm_ has joined #openstack-swift | 06:08 | |
| *** oomichi__ has joined #openstack-swift | 06:08 | |
| openstackgerrit | Kota Tsuyuzaki proposed a change to openstack/swift: Efficient Replication for Distributed Regions https://review.openstack.org/99824 | 06:10 |
| *** ttrumm has quit IRC | 06:10 | |
| *** oomichi has quit IRC | 06:10 | |
| *** oomichi__ has quit IRC | 06:27 | |
| *** X019 has quit IRC | 06:44 | |
| *** X019 has joined #openstack-swift | 06:44 | |
| openstackgerrit | Matthew Oliver proposed a change to openstack/swift: Basic container sharding middleware https://review.openstack.org/125553 | 06:50 |
| openstackgerrit | Michael Barton proposed a change to openstack/swift: sendfile for GETs https://review.openstack.org/125548 | 06:55 |
| *** SkyRocknRoll__ has quit IRC | 07:23 | |
| *** hhuang has joined #openstack-swift | 07:27 | |
| *** geaaru has joined #openstack-swift | 07:29 | |
| mattoliverau | Well I'm calling it a night, nigth all... if anyone is around.. channel is eerily quiet with everyone at the hackathon | 07:40 |
| *** jistr has joined #openstack-swift | 07:42 | |
| *** mitz_ has quit IRC | 07:53 | |
| *** mitz_ has joined #openstack-swift | 07:55 | |
| *** jamiehannaford has joined #openstack-swift | 07:59 | |
| *** hhuang has quit IRC | 08:25 | |
| *** nellysmitt has joined #openstack-swift | 08:51 | |
| *** kopparam has joined #openstack-swift | 08:55 | |
| *** ttrumm_ has quit IRC | 09:05 | |
| *** ttrumm_ has joined #openstack-swift | 09:07 | |
| *** ttrumm has joined #openstack-swift | 09:11 | |
| *** ttrumm_ has quit IRC | 09:11 | |
| *** ttrumm_ has joined #openstack-swift | 09:12 | |
| *** ttrumm has quit IRC | 09:15 | |
| *** hhuang has joined #openstack-swift | 09:22 | |
| *** ttrumm_ has quit IRC | 09:24 | |
| *** ttrumm has joined #openstack-swift | 09:24 | |
| *** mkollaro has joined #openstack-swift | 09:31 | |
| *** shaifali_ has joined #openstack-swift | 09:38 | |
| *** shaifali_ has quit IRC | 09:38 | |
| *** shaifali_ has joined #openstack-swift | 09:39 | |
| *** shaifali_ has quit IRC | 09:39 | |
| *** shaifali_ has joined #openstack-swift | 09:40 | |
| *** aix has joined #openstack-swift | 09:51 | |
| *** oomichi has joined #openstack-swift | 10:03 | |
| *** nellysmitt has quit IRC | 10:04 | |
| *** vdreamarkitex has joined #openstack-swift | 10:05 | |
| *** hhuang has quit IRC | 10:23 | |
| *** Dafna has quit IRC | 10:31 | |
| *** Dafna has joined #openstack-swift | 10:37 | |
| *** X019 has quit IRC | 10:46 | |
| *** nellysmitt has joined #openstack-swift | 10:46 | |
| *** nosnos has quit IRC | 10:50 | |
| *** nosnos has joined #openstack-swift | 10:51 | |
| *** nosnos has quit IRC | 10:55 | |
| *** mahatic has quit IRC | 10:56 | |
| *** X019 has joined #openstack-swift | 10:58 | |
| *** dmsimard_away is now known as dmsimard | 11:07 | |
| *** mahatic has joined #openstack-swift | 11:21 | |
| *** hhuang has joined #openstack-swift | 11:28 | |
| *** judd7 has joined #openstack-swift | 11:49 | |
| *** NM1 has joined #openstack-swift | 12:08 | |
| *** NM1 has quit IRC | 12:09 | |
| *** addnull has quit IRC | 12:28 | |
| *** shaifali_ has quit IRC | 12:29 | |
| *** X019 has quit IRC | 12:31 | |
| *** X019 has joined #openstack-swift | 12:31 | |
| *** hhuang has quit IRC | 12:32 | |
| *** oomichi has quit IRC | 12:32 | |
| *** fifieldt has joined #openstack-swift | 12:37 | |
| *** elambert has joined #openstack-swift | 12:39 | |
| *** NM1 has joined #openstack-swift | 12:41 | |
| *** elambert has quit IRC | 12:47 | |
| *** miqui has joined #openstack-swift | 12:50 | |
| dmsimard | good morning everyone | 12:59 |
| dmsimard | I'm trying to troubleshoot an issue with object-auditor looping infinitely in a bit of a crazy way. Can't see why. Logs (even in debug) aren't showing any errors or stack traces. | 13:00 |
| dmsimard | I traced the process and perhaps it might talk to you guys because it doesn't ring me a bell: http://paste.openstack.org/show/nAHuIsGoOruZ8yP5g161/ | 13:00 |
| *** NM1 has quit IRC | 13:00 | |
| dmsimard | I've done the usual like checking syslogs, dmesg, checking drives health and permissions - nothing obvious showing up. | 13:00 |
| *** NM1 has joined #openstack-swift | 13:01 | |
| dmsimard | It's only doing this in one of six storage nodes of a cluster split in three zones that is not yet in production. It's really pretty much empty (10GB total used reported by swift-recon) | 13:02 |
| *** mwstorer has joined #openstack-swift | 13:06 | |
| *** ppai has joined #openstack-swift | 13:07 | |
| ahale | the trace kinda looks like it is just doing really fast passes of the mostly empty disk.. does that box have significantly less of the total compared to the rest ? | 13:09 |
| *** elambert has joined #openstack-swift | 13:10 | |
| *** lcurtis has joined #openstack-swift | 13:12 | |
| dmsimard | ahale: Nope, they're all identical as far as I know. Both for hardware and software | 13:16 |
| dmsimard | ahale: It's looping endlessly, I would understand a "fast pass", like, every now and then | 13:16 |
| ahale | ah no i meant less of that total 10GB disk usage | 13:16 |
| dmsimard | ahale: At a glance, I see probably 1GB on that node. | 13:17 |
| dmsimard | No more than 1.5 | 13:17 |
| notmyname | good morning | 13:21 |
| *** zaitcev has joined #openstack-swift | 13:25 | |
| *** ChanServ sets mode: +v zaitcev | 13:25 | |
| *** CaioBrentano has joined #openstack-swift | 13:26 | |
| *** tdasilva has joined #openstack-swift | 13:28 | |
| zaitcev | dmsimard: what exact version are you running? e.g. python -c "import swift; print swift.__version__" | 13:28 |
| dmsimard | zaitcev: 1.13.1, packaged by Ubuntu's cloud archive | 13:28 |
| zaitcev | go go gadget git checkout 1.13.1 | 13:29 |
| *** CaioBrentano has quit IRC | 13:31 | |
| zaitcev | oh, brother... http://www.zaitcev.us/things/dmsimard-auditor.diff | 13:31 |
| zaitcev | Never mind that parallel stuff, but you want that self._sleep() | 13:32 |
| dmsimard | zaitcev: That's the diff from the master branch ? | 13:33 |
| zaitcev | dmsimard: yes | 13:33 |
| dmsimard | zaitcev: git blame's giving me this commit for the patch that introduced the sleep: https://github.com/openstack/swift/commit/d317888a7eae276ae9dddf26a9030d01f6ba00fe | 13:35 |
| *** CaioBrentano has joined #openstack-swift | 13:37 | |
| zaitcev | no, it's actually bcdafa3831111f334caadd30505f051c81abcb1f | 13:37 |
| dmsimard | zaitcev: Oh, and that commit message exactly matches the issue I have :) | 13:38 |
| dmsimard | Do you know what version of Swift is matched with Juno ? Is it 2.0 or 2.1.0 ? | 13:39 |
| zaitcev | I'm pretty certain it's 2.1.0 | 13:39 |
| notmyname | the juno release will include swift 2.2.0 | 13:40 |
| notmyname | we've already released 2.1.0 | 13:40 |
| zaitcev | well yeah... but he needs something to install today | 13:40 |
| notmyname | :-) | 13:41 |
| zaitcev | dmsimard: please continue from here and verify that your problem is actually fixed... I need something urgent to do here, so ttyl | 13:42 |
| dmsimard | 2.2.0 !? :) | 13:42 |
| zaitcev | dmsimard: please continue from here and verify that your problem is actually fixed... I need something urgent to do here, so ttyl | 13:42 |
| zaitcev | er | 13:42 |
| dmsimard | dmsimard: Sure. | 13:43 |
| dmsimard | er :p | 13:43 |
| dmsimard | zaitcev: sure | 13:43 |
| *** nellysmitt has quit IRC | 13:49 | |
| *** CaioBrentano has quit IRC | 13:51 | |
| *** wasmum has joined #openstack-swift | 13:52 | |
| *** NM2 has joined #openstack-swift | 13:54 | |
| *** NM1 has quit IRC | 13:54 | |
| *** CaioBrentano has joined #openstack-swift | 13:54 | |
| *** DisneyRicky has quit IRC | 13:57 | |
| *** tsg has joined #openstack-swift | 13:58 | |
| *** vr1 has joined #openstack-swift | 14:02 | |
| vr1 | hello | 14:02 |
| *** ttrumm has quit IRC | 14:06 | |
| openstackgerrit | Daniel Wakefield proposed a change to openstack/python-swiftclient: Replaces Stacktraces with useful error messages. https://review.openstack.org/125407 | 14:07 |
| *** nellysmitt has joined #openstack-swift | 14:14 | |
| *** zaitcev has quit IRC | 14:16 | |
| *** DisneyRicky has joined #openstack-swift | 14:19 | |
| *** danw_ has joined #openstack-swift | 14:24 | |
| *** tsg has quit IRC | 14:24 | |
| *** kopparam has quit IRC | 14:26 | |
| *** nellysmitt has quit IRC | 14:28 | |
| *** nellysmitt has joined #openstack-swift | 14:29 | |
| *** danw_ has quit IRC | 14:30 | |
| *** nellysmitt has quit IRC | 14:33 | |
| *** vr1 has quit IRC | 14:34 | |
| *** vr1 has joined #openstack-swift | 14:35 | |
| *** nexusz99 has joined #openstack-swift | 14:35 | |
| *** nellysmitt has joined #openstack-swift | 14:38 | |
| NM2 | Does anyone recently migrate from version 1.x to version 2.x ? I'd like to know if there is anything else to worry beside storage policies… | 14:42 |
| *** AnjuT has joined #openstack-swift | 14:49 | |
| lcurtis | all, is it possible to actually run vm's in swift? | 14:51 |
| lcurtis | or used only to store images? | 14:51 |
| jokke_ | lcurtis: swift is for storing "any" objects. Nova would be the compute part | 14:52 |
| lcurtis | thank you jakke | 14:52 |
| lcurtis | so in other words, i could store a kvm img in swift and run from the compute node? | 14:53 |
| jokke_ | lcurtis: you want to store your image in Glance (which may use Swift as a back end) and you run it on Nova ;) | 14:54 |
| lcurtis | lol | 14:54 |
| lcurtis | okay...still more reading to do | 14:54 |
| lcurtis | thank you | 14:54 |
| jokke_ | lcurtis: good place to start from is http://www.openstack.org/software/ | 14:55 |
| lcurtis | but technically i could use either swift or cinder as back end for images? | 14:55 |
| lcurtis | i have been reading there...perhaps just not understanding fully | 14:56 |
| *** nexusz99 has quit IRC | 14:57 | |
| NM2 | lcurtis: You never understand cloud computing on the first reading. I took 6 months to 'really' understand the ring :) | 14:58 |
| jokke_ | lcurtis: Glance supports multiple storage drivers, more info from http://docs.openstack.org/developer/glance/configuring.html#configuring-the-filesystem-storage-backend | 14:58 |
| *** NM2 is now known as NM | 14:59 | |
| lcurtis | would one really need cinder if vm's can be stored in swift? | 15:00 |
| notmyname | lcurtis: you don't run VMs from swift. but swift can store the images. | 15:00 |
| *** nellysmitt has quit IRC | 15:00 | |
| notmyname | lcurtis: cinder is for persistent block devices attached to your Vms | 15:00 |
| lcurtis | okay...so images are ephemeral in swift? | 15:01 |
| notmyname | no | 15:01 |
| notmyname | swift is durable storage | 15:01 |
| lcurtis | hmm | 15:01 |
| notmyname | lcurtis: here's a good overview of swift https://swiftstack.com/openstack-swift/ | 15:01 |
| notmyname | lcurtis: so swift is a place to put data. it can be deployed alongside nova and other openstack components or it can be deployed independently | 15:02 |
| notmyname | lcurtis: one kind of data people put in swift is VM images. and the way to do that using the rest of openstack is to use glance | 15:03 |
| lcurtis | so glance stores the image and runs the image using api calls to swift? | 15:03 |
| lcurtis | ive read quite a few docs and listened to many youtubes on the commute home | 15:04 |
| notmyname | lcurtis: glance tracks where the image is stored. a user sends an API call to glance to store (or load) and image, and glance sends API calls to swift. the VMs don't run from swift, but glance pulls the data to the VM host machine and then runs it there | 15:05 |
| lcurtis | okay...if i wanted to do something like live migration, could i use either cinder or swift | 15:07 |
| lcurtis | Ah...think I get it | 15:14 |
| lcurtis | http://www.gossamer-threads.com/lists/openstack/dev/21956 | 15:14 |
| lcurtis | so images would be static in swift | 15:14 |
| notmyname | lcurtis: correct | 15:15 |
| *** tsg has joined #openstack-swift | 15:15 | |
| lcurtis | and then i could use cinder to achieve live migration | 15:15 |
| lcurtis | coming from vmware sab vsphere bkgd | 15:16 |
| lcurtis | san | 15:16 |
| *** nexusz99 has joined #openstack-swift | 15:16 | |
| *** tgohad has joined #openstack-swift | 15:17 | |
| *** tsg has quit IRC | 15:20 | |
| *** annegent_ has joined #openstack-swift | 15:23 | |
| *** vr1 has quit IRC | 15:24 | |
| lcurtis | notmyname..thanks for your help | 15:30 |
| lcurtis | makes sense now | 15:30 |
| notmyname | great! | 15:31 |
| *** vr1 has joined #openstack-swift | 15:31 | |
| *** vdreamarkitex has quit IRC | 15:32 | |
| *** nexusz99 has quit IRC | 15:38 | |
| swift_fan1 | Hello -- I do not wish to include the "gatekeeper" and "dlo" middlewares into my Swift proxy-server.conf pipeline, but the proxy server error log keeps saying that the "Pipeline was modified" by adding the filters/middleware in. | 15:48 |
| swift_fan1 | Is this a Swift bug? | 15:48 |
| swift_fan1 | notmyname : Or is this intentional by Swift, and if so, then what's the reason ? | 15:49 |
| swifterdarrell | swift_fan1: they're both required | 15:49 |
| swift_fan1 | swifterdarrell : Why/how so? | 15:49 |
| *** rmcall has joined #openstack-swift | 15:49 | |
| swifterdarrell | swift_fan1: gatekeeper ensures system metadata doesn't get out to clients and DLO (dynamic large object) functionality used to be embedded in Swift but was fully extracted into middleware | 15:50 |
| swifterdarrell | swift_fan1: if you don't want to see the messages about them being added, you may explicitly enter them in your pipeline | 15:50 |
| swift_fan1 | swifterdarrell : Thanks. But why is it that they must be included in ? | 15:52 |
| swift_fan1 | swifterdarrell : I can sort of see the reason for gatekeeper, but it seems that dlo can be done without..... | 15:52 |
| swifterdarrell | swift_fan1: because without gatekeeper your Swift will be broken, and without DLO you won't have the same Swift everyone who's ever had swift has had | 15:52 |
| swifterdarrell | swift_fan1: they're simply required | 15:52 |
| swifterdarrell | swift_fan1: DLO is a functionality of Swift and always has been | 15:53 |
| swift_fan1 | swifterdarrell : Is there a list of required Swift middleware vs. non-required Swift middleware ? | 15:54 |
| swifterdarrell | swift_fan1: probably in the code wherever gatekeeper and dlo get added | 15:54 |
| swift_fan1 | swifterdarrell : Or is all the middleware that is initially available with a standard Swift installation, all required to be in the pipeline ? | 15:55 |
| swifterdarrell | swift_fan1: and/or maybe in the docs somewhere? I don't remember | 15:55 |
| swifterdarrell | swift_fan1: they're definitely not all required | 15:55 |
| swift_fan1 | swifterdarrell : For instance, like tempURL or formpost ? | 15:55 |
| swift_fan1 | swifterdarrell : They are in the middleware list that initially just comes with a standard Swift installation, but would they have a good chance of being *non-required* middleware ? | 15:56 |
| openstackgerrit | Simon Lorenz proposed a change to openstack/swift: Added flag to create unique tmp dir per strg node https://review.openstack.org/124088 | 15:58 |
| *** ZBhatti has joined #openstack-swift | 15:59 | |
| *** kyles_ne has joined #openstack-swift | 16:00 | |
| swifterdarrell | swift_fan1: gatekeeper and dlo are probably the only "required" middlewares and they are added automatically in order to not break existing deployments | 16:00 |
| swifterdarrell | swift_fan1: you could consider an auth middleware "required" if you want any permissions enforcement | 16:01 |
| swifterdarrell | swift_fan1: you'd be crazy not to use proxy-logging (twice), but it's not "required" | 16:01 |
| swifterdarrell | swift_fan1: ditto for catch_errors | 16:02 |
| swift_fan1 | swifterdarrell : Why would you care about an existing deployment, if all you're trying to do is set up your own, individual Swift cluster for your own use ? | 16:02 |
| swifterdarrell | swift_fan1: *blink* *bliink* | 16:03 |
| swifterdarrell | swift_fan1: are you asking why the developers of Swift care about deployments of Swift? | 16:03 |
| swift_fan1 | swifterdarrell : I mean in scenarios where there isn't an existing deployment of Swift, and you're just trying to set one up (so there isn't an existing deployment in the first place). | 16:05 |
| swifterdarrell | swift_fan1: fresh deployments should probably be explicit vs. implicit and explicitly add the gatekeeper and dlo middleware | 16:05 |
| swift_fan1 | swifterdarrell : But what if you want your Swift system to not care about dlo functionality ? | 16:06 |
| swift_fan1 | or even gatekeeper | 16:06 |
| swifterdarrell | swift_fan1: you don't really get to do that | 16:06 |
| swifterdarrell | swift_fan1: this is like the bottom of the program where it does GOTO 10 | 16:06 |
| glange | swift_fan1: you can change the code to do what you want to do :) | 16:06 |
| swifterdarrell | swift_fan1: I feel like we already talked about how those are required | 16:06 |
| swift_fan1 | glange swifterdarrell : So, if one wants to disable the dlo functionality for their Swift cluster, the only ways to do so are to (1) change the code itself (2,3,...) anything else that is recommended ? | 16:09 |
| glange | swift_fan1: you maybe could write a middleware that prevents DLOs and use that -- that might be another option | 16:11 |
| swift_fan1 | glange swifterdarrell : Because what if you're setting this up for some other people that you know, and you don't want them to utilize a particular functionality (such as dlo), then how do you make sure they they can't ? | 16:11 |
| swift_fan1 | glange swifterdarrell : Other than just telling them not to use it. | 16:12 |
| swift_fan1 | glange swifterdarrell : There are some reasons for this, but let's not focus on them right now. Thanks | 16:13 |
| glange | another option is to run an old version of swift that doesn't have dlo in it :) | 16:15 |
| swifterdarrell | heh | 16:15 |
| *** CaioBrentano has quit IRC | 16:26 | |
| ahale | if you dont want people to use it set rate limit to a year per segment and kick in after the first segment.. or something | 16:27 |
| *** ttrumm has joined #openstack-swift | 16:28 | |
| *** gyee has joined #openstack-swift | 16:30 | |
| *** vr1 has quit IRC | 16:32 | |
| *** ZBhatti has quit IRC | 16:41 | |
| *** swift_fan1 has quit IRC | 16:45 | |
| *** jistr has quit IRC | 16:45 | |
| *** AnjuT has quit IRC | 16:49 | |
| *** mkollaro has quit IRC | 17:05 | |
| *** shri has joined #openstack-swift | 17:11 | |
| *** tgohad has quit IRC | 17:19 | |
| *** CaioBrentano has joined #openstack-swift | 17:23 | |
| *** rmcall_ has joined #openstack-swift | 17:23 | |
| *** rmcall has quit IRC | 17:24 | |
| *** rmcall_ is now known as rmcall | 17:24 | |
| *** rmcall_ has joined #openstack-swift | 17:31 | |
| *** elambert has quit IRC | 17:31 | |
| *** rmcall has quit IRC | 17:31 | |
| *** mkollaro has joined #openstack-swift | 17:32 | |
| openstackgerrit | Richard Hawkins proposed a change to openstack/swift: Fix metadata overall limits bug https://review.openstack.org/125360 | 17:34 |
| *** ZBhatti has joined #openstack-swift | 17:34 | |
| *** rmcall has joined #openstack-swift | 17:34 | |
| *** rmcall_ has quit IRC | 17:35 | |
| *** mwstorer has quit IRC | 17:37 | |
| *** ttrumm has quit IRC | 17:38 | |
| *** aix has quit IRC | 17:44 | |
| *** kopparam has joined #openstack-swift | 17:48 | |
| *** mkollaro has quit IRC | 17:49 | |
| *** geaaru has quit IRC | 17:50 | |
| *** elambert has joined #openstack-swift | 17:50 | |
| *** elambert has quit IRC | 17:50 | |
| *** kopparam has quit IRC | 17:52 | |
| *** swift_fan has joined #openstack-swift | 17:54 | |
| *** bill_az has joined #openstack-swift | 17:56 | |
| *** annegent_ has quit IRC | 18:00 | |
| *** htrmeira has joined #openstack-swift | 18:01 | |
| *** annegent_ has joined #openstack-swift | 18:08 | |
| swift_fan | ahale : A year per segment ? | 18:11 |
| swift_fan | ahale : Is that a feasible solution ? | 18:12 |
| swift_fan | ahale : Do you literally mean just only 1 segment, per year ? | 18:12 |
| ahale | i guess, its more silly really.. a day would do probably, cant remember if its swift that limits reqs to 24 hours or my lbs | 18:13 |
| swift_fan | glange swifterdarrell (as well). | 18:13 |
| swift_fan | ahale : So, what does setting the rate limit to only 1 segment per day or year, have any relation to restricting others from using dlo ? | 18:14 |
| ahale | its just not really something you'd wanna really hard disable since you want to permit tools to work that might assume its there as it was a core part pf proxy-server in the past | 18:14 |
| swift_fan | Thanks | 18:14 |
| ahale | nah its in the dlo conf only .. check out the samples | 18:14 |
| ahale | it would still accept requests for them, just make it totally impossible to use meaningfully.. discourage rather than disallow | 18:15 |
| *** kopparam has joined #openstack-swift | 18:18 | |
| *** nellysmitt has joined #openstack-swift | 18:23 | |
| NM | Is it possible that swif changed the ring files by itself? In case of a drive failed or something like that. | 18:24 |
| *** kopparam has quit IRC | 18:24 | |
| swift_fan | ahale : That's actually pretty clever! | 18:26 |
| *** jamiehannaford has quit IRC | 18:36 | |
| glange | swift_fan: you can also make your swift cluster have unlimited capacity if you use a write only middleware :) | 18:50 |
| *** tsg has joined #openstack-swift | 18:58 | |
| *** theanalyst has quit IRC | 19:00 | |
| *** kyles_ne has quit IRC | 19:00 | |
| *** kyles_ne has joined #openstack-swift | 19:01 | |
| *** kyles_ne has quit IRC | 19:05 | |
| *** theanalyst has joined #openstack-swift | 19:14 | |
| openstackgerrit | Jay Bryant proposed a change to openstack/swift: Handle os.listdir failures in object-updater https://review.openstack.org/125746 | 19:18 |
| *** mahatic has quit IRC | 19:26 | |
| notmyname | NM: no. swift doesn't change the ring automatically | 19:36 |
| *** ChanServ changes topic to "Swift 2.1.0 https://launchpad.net/swift/+milestone/2.1.0 | Swift Review Dashboard: http://bit.ly/1iVBigF | Priority Reviews: https://wiki.openstack.org/wiki/Swift/PriorityReviews | Ideas: https://wiki.openstack.org/wiki/Swift/ideas | Kilo summit topics: https://etherpad.openstack.org/p/kilo-swift-summit-topics" | 19:45 | |
| notmyname | note the https://etherpad.openstack.org/p/kilo-swift-summit-topics in the topic now. this is the place to add summit session topics | 19:46 |
| *** annegent_ has quit IRC | 19:51 | |
| *** kopparam has joined #openstack-swift | 19:52 | |
| *** bgmccollum has joined #openstack-swift | 19:52 | |
| swift_fan | glange : What do you mean ? | 19:55 |
| swift_fan | ?? | 19:55 |
| swift_fan | glange : unlimited capacity ? | 19:56 |
| notmyname | swift_fan: he's not being serious. or actually, he is, but it's a totally unusable thing | 19:56 |
| notmyname | swift_fan: see http://www.supersimplestorageservice.com (hint: it's a joke) | 19:56 |
| swift_fan | notmyname : I know it was facetious, but I was wondering what this "unusable thing" is, regardless ....... | 19:56 |
| notmyname | write only storage? | 19:57 |
| notmyname | /dev/null as a server? | 19:57 |
| notmyname | *service | 19:57 |
| swift_fan | notmyname : ? | 19:57 |
| glange | swift_fan: if you write data and never read it, why do I even have to store it? | 19:58 |
| glange | I can accept a lot of data if I don't have to store it | 19:59 |
| NM | notmyname: Thanks. I think my ring got corrupted but I found some backups on the /etc/swift/backups dir. I'm thinking about to use then to restore the ring… Any thoughts? | 20:00 |
| *** aerwin has joined #openstack-swift | 20:01 | |
| notmyname | NM: yup. that's what the backups are for :-) | 20:02 |
| *** aerwin has left #openstack-swift | 20:03 | |
| *** aerwin has joined #openstack-swift | 20:03 | |
| NM | notmyname: :) | 20:05 |
| notmyname | NM: so the trick would be to know what's changed in your infra since that backup was made | 20:05 |
| notmyname | NM: ie you don't want to deploy a ring that is going to cause a huge data movement problem in your prod cluster | 20:06 |
| *** occupant has quit IRC | 20:06 | |
| *** occupant has joined #openstack-swift | 20:09 | |
| *** kopparam has quit IRC | 20:09 | |
| *** occupant has quit IRC | 20:09 | |
| *** ZBhatti has quit IRC | 20:09 | |
| openstackgerrit | Alistair Coles proposed a change to openstack/python-swiftclient: Fix cross account upload using --os-storage-url https://review.openstack.org/125759 | 20:09 |
| *** kopparam has joined #openstack-swift | 20:10 | |
| *** occupant has joined #openstack-swift | 20:10 | |
| *** exploreshaifali has joined #openstack-swift | 20:13 | |
| *** kopparam has quit IRC | 20:14 | |
| exploreshaifali | I am trying to install swift in vmware, using vagrant, but getting "vagrant cannot detect virtualbox" error http://pastebin.com/5iGnvJse | 20:16 |
| notmyname | exploreshaifali: does vagrant support vmware these days? | 20:16 |
| notmyname | I know it originally was virtualbox only, but maybe they added vmware fusion support | 20:16 |
| exploreshaifali | so I should try only with virtualbox | 20:17 |
| notmyname | I know that works | 20:17 |
| exploreshaifali | okay | 20:17 |
| exploreshaifali | ; | 20:17 |
| exploreshaifali | :) | 20:17 |
| NM | notmyname: I was checking my ring and found this output object.builder, build version 0 | 20:17 |
| NM | 262144 partitions, 3.000000 replicas, 0 regions, 0 zones, 0 devices, 0.00 balance | 20:17 |
| NM | I cant understand how it has 0 region/zone/devices but still works... | 20:18 |
| NM | Btw, this is a production cluster. | 20:18 |
| *** annegent_ has joined #openstack-swift | 20:18 | |
| *** annegent_ has quit IRC | 20:19 | |
| *** annegent_ has joined #openstack-swift | 20:20 | |
| notmyname | NM: don't push that ring. it doesn't have anything added | 20:21 |
| *** kyles_ne has joined #openstack-swift | 20:21 | |
| openstackgerrit | Richard Hawkins proposed a change to openstack/swift: Fix metadata overall limits bug https://review.openstack.org/125360 | 20:22 |
| NM | notmyname: that is the what is disturbing me: the swift is running with this empty ring. | 20:23 |
| notmyname | NM: a backup is created on create() and on rebalance(). so what you are looking at is the one from the create() call. basically, just setting the replica count and part power | 20:23 |
| notmyname | surely you have a different .ring.gz deployed | 20:23 |
| *** nellysmitt has quit IRC | 20:23 | |
| notmyname | NM: here's the help message that might (but I hope not!) be useful to you: | 20:25 |
| notmyname | swift-ring-builder <ring_file> write_builder [min_part_hours] | 20:25 |
| notmyname | Recreate a builder from a ring file (lossy) if you lost your builder | 20:25 |
| notmyname | backups. (Protip: don't lose your builder backups). | 20:25 |
| notmyname | [min_part_hours] is one of those numbers lost to the builder, | 20:25 |
| notmyname | you can change it with set_min_part_hours. | 20:25 |
| notmyname | NM: if you really have lost your backups and corrupted your ring but you still have a .ring.gz around, you can use that command to (mostly) recreate your builder file | 20:26 |
| notmyname | NM: also, if you're at that point and this is necessary, then you also have to buy clayg a beer for writing it ;-) | 20:26 |
| NM | On 2015 summit I can pay on for him and you for your help :) | 20:27 |
| NM | I think that explains why I can't see the devices on my ring but it still works... | 20:28 |
| notmyname | NM: so if you are at the point where you only have some .ring.gz files but you don't have any builder files, then you need to fix that ASAP and then audit your ring builder backup process | 20:29 |
| NM | The backups files seens to be fine. I ran swift-ring-builder and the output was as expected | 20:31 |
| notmyname | NM: and you have old backup files? eg you should have builder files in the backups dir after the rebalance | 20:31 |
| notmyname | from before when it got whatever corruption you mentioned | 20:31 |
| NM | Yes I have both. | 20:32 |
| notmyname | ah, good | 20:32 |
| notmyname | so you're ok then | 20:32 |
| *** mlauter has joined #openstack-swift | 20:32 | |
| NM | I hope so. I'll restore then and watch my monitor. | 20:32 |
| notmyname | well, wait. don't push deploy yet | 20:33 |
| notmyname | so the builder files don't have to be deployed anywhere | 20:33 |
| notmyname | they're just there to build the ring.gz files | 20:33 |
| NM | That was something that I always wanted to know :) | 20:33 |
| notmyname | but since you're pulling a builder file from backup, then check it out to ensure that everything looks right. like I said, you don't want to push out some ring that's radically different and cause a bunch of data movement (at best) or unavailability (at worse) | 20:34 |
| *** htrmeira has quit IRC | 20:35 | |
| NM | If I just run swift-ring-builder on my ring file, and the output is what I expected, should I deploy? I haven't modify my server/disks. | 20:35 |
| notmyname | NM: well, I don't want to be the guy who tells you to deploy or not ;-) | 20:37 |
| NM | Hahahahaha! In your place I'd say the same thing. | 20:38 |
| notmyname | NM: but yeah, the first level of checking would be so see if the ring-builder output is what you expect. the right zones in the right places, etc | 20:38 |
| NM | Well, it's done. Lets see :D | 20:38 |
| notmyname | "NM, why is our site offline?" "well, some guy on the internet said it probably wouldn't break" | 20:38 |
| *** zul has quit IRC | 20:44 | |
| *** aerwin has quit IRC | 20:47 | |
| *** rmcall has quit IRC | 20:55 | |
| *** HenryG has quit IRC | 20:55 | |
| *** annegent_ has quit IRC | 20:56 | |
| *** tsg has quit IRC | 21:09 | |
| openstackgerrit | Samuel Merritt proposed a change to openstack/swift: Fix ring-builder crash. https://review.openstack.org/125793 | 21:26 |
| openstackgerrit | Richard Hawkins proposed a change to openstack/swift: Fix metadata overall limits bug https://review.openstack.org/125360 | 21:30 |
| NM | notmyname: Actually I could name you. Your name is pretty common here :) | 21:31 |
| notmyname | heh. also, uh oh. | 21:31 |
| *** mikehn has quit IRC | 21:39 | |
| *** corvus has quit IRC | 21:40 | |
| *** mikehn has joined #openstack-swift | 21:40 | |
| *** corvus has joined #openstack-swift | 21:41 | |
| *** tsg has joined #openstack-swift | 21:52 | |
| *** tsg has quit IRC | 21:57 | |
| *** mlauter has quit IRC | 21:58 | |
| openstackgerrit | John Dickinson proposed a change to openstack/swift: Ensure sys.exit called in fork_child after exception https://review.openstack.org/125197 | 21:59 |
| notmyname | I only fixed the spelling error. and approved it | 22:00 |
| openstackgerrit | Richard Hawkins proposed a change to openstack/swift: Fix metadata overall limits bug https://review.openstack.org/125360 | 22:03 |
| NM | Well, I think I can leave now. Everything seens fine…. | 22:04 |
| *** tsg has joined #openstack-swift | 22:05 | |
| notmyname | NM: nice | 22:12 |
| NM | Tomorrow I'll fix the other storage nodes. I think it's a good idea to commit our ring files at git. | 22:13 |
| notmyname | NM: yeah, I know another swift deployer that uses VCS for storing configs (and rings are really just config files. sortof.) | 22:14 |
| mattoliverau | Morning all | 22:14 |
| notmyname | hi mattoliverau | 22:14 |
| NM | Thank you guys! See you tomorrow! And congrats for the hackathon! | 22:15 |
| mattoliverau | Yay some chatter in channel, its been very quiet with a hackathon going on! | 22:16 |
| *** CaioBrentano has quit IRC | 22:16 | |
| mattoliverau | How's the hackathon going? Solved all the things yet? | 22:17 |
| *** NM has quit IRC | 22:18 | |
| *** lcurtis has quit IRC | 22:19 | |
| notmyname | mattoliverau: just about. looking at the last patch now | 22:29 |
| notmyname | it's been a very production week | 22:29 |
| notmyname | *productive | 22:30 |
| *** sandywalsh has quit IRC | 22:33 | |
| mattoliverau | Awesome, wish I could've been there, hopefully next time. | 22:33 |
| *** sandywalsh has joined #openstack-swift | 22:36 | |
| notmyname | hurricanerix: https://bugs.launchpad.net/swift/+bug/1365350 | 22:36 |
| *** annegent_ has joined #openstack-swift | 22:37 | |
| openstackgerrit | Richard Hawkins proposed a change to openstack/swift: Fix metadata overall limits bug https://review.openstack.org/125360 | 22:38 |
| *** annegent_ has quit IRC | 22:41 | |
| *** zul has joined #openstack-swift | 22:45 | |
| *** ZBhatti has joined #openstack-swift | 22:48 | |
| *** sandywalsh has quit IRC | 22:53 | |
| *** sandywalsh has joined #openstack-swift | 22:53 | |
| *** tdasilva has quit IRC | 22:53 | |
| *** ppai has quit IRC | 22:55 | |
| *** dmsimard is now known as dmsimard_away | 22:56 | |
| *** ZBhatti has quit IRC | 23:05 | |
| *** sandywalsh has quit IRC | 23:05 | |
| *** nosnos has joined #openstack-swift | 23:06 | |
| *** mlauter has joined #openstack-swift | 23:10 | |
| *** HenryG has joined #openstack-swift | 23:10 | |
| *** mlauter has quit IRC | 23:17 | |
| *** annegent_ has joined #openstack-swift | 23:19 | |
| *** annegent_ has quit IRC | 23:25 | |
| *** annegent_ has joined #openstack-swift | 23:28 | |
| *** annegent_ has quit IRC | 23:31 | |
| *** echevemaster has joined #openstack-swift | 23:34 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!