| mattoliverau | Morning swift land... I've gone and got me the flu (winter/flu season here in Oz).. so am around but definitely not a full strength.. just FYI | 00:11 |
|---|---|---|
| *** ho has joined #openstack-swift | 00:14 | |
| ho | good morning! | 00:15 |
| *** jrichli has quit IRC | 01:04 | |
| *** jrichli has joined #openstack-swift | 01:04 | |
| mattoliverau | ho: morning | 01:29 |
| jrichli | mattoliverau: I hope you get to feeling better soon! | 01:29 |
| mattoliverau | jrichli: ta :).. And hello! | 01:30 |
| jrichli | hello! I am making new reviews out of code I have already done :-) | 01:31 |
| jrichli | splitting up the encryption stuff into smaller reviews - I hope it helps to get some eyes on it | 01:31 |
| wbhuber | jrichli: splitting into smaller reviews sounds like a good plan. hopefully that brings in a more acute focus. =] | 01:40 |
| jrichli | wbhuber: thx. Hope you are having a good weekend | 01:41 |
| wbhuber | jrichi: it's been awrite. my son is battling some kind of stomach bug just like mattoliverau. =] we hope it's not as serious as it looks. | 01:42 |
| jrichli | wbhuber: I hope he feels better soon as well. | 01:43 |
| wbhuber | jrichli: hope yours has been bug-free! | 01:43 |
| jrichli | no sickness, anyway! :-) | 01:43 |
| wbhuber | then you're more swift than mattoliverau and i are (puns intended). | 01:44 |
| mattoliverau | Lol | 01:46 |
| *** coreyfarrell has joined #openstack-swift | 01:49 | |
| *** coreyfarrell has left #openstack-swift | 01:54 | |
| *** klrmn has left #openstack-swift | 02:00 | |
| openstackgerrit | Kuo-tung Kao proposed openstack/swift: Use testtools for base test class https://review.openstack.org/192960 | 02:03 |
| *** linkedinyou has quit IRC | 02:06 | |
| ho | mattoliverau: hello! | 02:23 |
| *** jrichli has quit IRC | 02:50 | |
| *** blmartin has joined #openstack-swift | 02:55 | |
| blmartin | Hello all | 02:56 |
| *** silor has joined #openstack-swift | 03:05 | |
| openstackgerrit | Kuo-tung Kao proposed openstack/swift: Use testtools for base test class https://review.openstack.org/192960 | 03:07 |
| *** asettle has joined #openstack-swift | 03:44 | |
| openstackgerrit | Merged openstack/swift: Allow 1+ object-servers-per-disk deployment https://review.openstack.org/184189 | 03:45 |
| *** wbhuber has quit IRC | 03:52 | |
| *** blmartin has quit IRC | 04:17 | |
| asettle | notmyname, you around? :) | 04:25 |
| *** ppai has joined #openstack-swift | 04:41 | |
| notmyname | asettle: what's up | 04:55 |
| asettle | notmyname, just wondering where in Austin the meet up will be held? | 04:55 |
| notmyname | asettle: I don't have any details publicly yet. I expect to have that this week | 04:56 |
| asettle | notmyname, that's cool :) thank you. If you're sending out an email, could you please CC me on the list? :) | 04:59 |
| notmyname | sure. definitely bug me if you haven't seen anything in a day or two | 05:00 |
| asettle | Thanks notmyname :) I'm just coming up with a really basic itinerary list for Lana | 05:07 |
| asettle | Also I had a question notmyname about the Swift docs | 05:09 |
| notmyname | ok | 05:09 |
| asettle | I got SAIO up and working (thanks to mattoliverau ) however noticing the docs only build within a Swift environment is a bit of an issue. Is there a workaround for this or is this what is at present? | 05:09 |
| notmyname | well...that's what it is, but I'm not sure how much of a bad thing it is | 05:10 |
| notmyname | that is, the docs are generated in part from the source files, so it kinda makes sense to need to evaluate them. and that involves, at times, looking at dependencies | 05:11 |
| notmyname | that being said, if you know a way to general the .rst files and the docstrings from the code into static docs like we have now, that would be great | 05:12 |
| * asettle sits in silence | 05:13 | |
| asettle | Lemme get back to you | 05:13 |
| asettle | I have a handy docs PTL lying around and I happen to know where she lives, too :p | 05:13 |
| notmyname | asettle: I do realize I haven't given you that setup doc I mentioned (yes I know mattoliverau helped you out) | 05:18 |
| notmyname | I did finish it for it's original audience, so let me see if I can make it more generic with some judicious copy/paste | 05:19 |
| asettle | That sounds great :) thanks notmyname . Definitely be helpful to have | 05:19 |
| *** silor1 has joined #openstack-swift | 05:20 | |
| *** silor has quit IRC | 05:20 | |
| notmyname | asettle: https://gist.githubusercontent.com/notmyname/88b29888e21dfc8372cb/raw/ee6b43951cf7016831c61292cb46981f8e47eef8/gistfile1.txt | 05:25 |
| notmyname | rough, but the idea is there | 05:25 |
| asettle | notmyname, looks good so far :) I use command line every day, but I think keeping it at a basic level is a good idea :) | 05:27 |
| notmyname | yeah, the person I was originally writing that for isn't familiar with the cmd line | 05:27 |
| notmyname | I'll probably turn that into a blog post (hopefully soon) as a tool for new contributors | 05:28 |
| asettle | Great idea notmyname | 05:29 |
| mattoliverau | Love the typey typey reference | 05:30 |
| asettle | mattoliverau, lives! | 05:33 |
| notmyname | :-) | 05:33 |
| mattoliverau | In some definition of lives :p | 05:34 |
| asettle | So, notmyname with regards to your 'IF you know a way to general the .rst files and docstrings" - I'm going to send an email to Lana (Docs PTL), and get her in on this one. I think her insight into the process would be very helpful and invaluable. I can do the typey typey, and whatever you want me to do, but tool chains is not necessarily in my jurisdiction | 05:35 |
| notmyname | yeah, that makes sense | 05:36 |
| notmyname | I'm not looking to add new tools for these docs | 05:36 |
| notmyname | they are maintained by the dev contributors, and I think that's where the primary responsibility lives and should remain | 05:36 |
| notmyname | ie for those people most responsible, having a swift dev environment isn't a barrier | 05:37 |
| notmyname | however, if there are ways to make it simpler for others to come in "just" to make docs fixes, I'd love to have them | 05:37 |
| asettle | Yes, of course. Which does make sense. I don't necessarily think we should move it, but I think increasing the accessibility of the documentation would be invaluable to the future of Swift documentation. | 05:38 |
| notmyname | I'm not trivializing docs patches (note the scare quotes around one of my least favorite words) | 05:38 |
| notmyname | yes! I completely agree! | 05:38 |
| asettle | Wonderful, we're all on the same page \o/c | 05:38 |
| asettle | Hands in the air like you just communicated effectively! | 05:38 |
| asettle | notmyname, did you want anyone in particular CC'd on this email | 05:39 |
| notmyname | I'd appreciate being on it | 05:39 |
| asettle | notmyname, haha besides you :P you're definitely going to be on it | 05:40 |
| notmyname | guess I can't get out of it ;-) | 05:44 |
| asettle | Oh no, you're sucked in now | 05:46 |
| *** proteusguy has quit IRC | 05:55 | |
| *** proteusguy has joined #openstack-swift | 05:56 | |
| *** SkyRocknRoll has joined #openstack-swift | 05:56 | |
| asettle | email = sent notmyname - feel free to add anything if you think i've missed something or have something incorrect | 05:59 |
| notmyname | asettle: looks fine. thanks. I won't comment further until Lana has her say :-) | 06:00 |
| asettle | Thanks :) | 06:02 |
| *** asettle has quit IRC | 06:10 | |
| notmyname | I'm off to bed. talk to you tomorrow | 06:14 |
| notmyname | mattoliverau: hope you get to feeling better | 06:14 |
| *** SkyRocknRoll_ has joined #openstack-swift | 06:16 | |
| *** ianbrown has quit IRC | 06:35 | |
| mattoliverau | notmyname: thanks, night | 07:01 |
| *** rledisez has joined #openstack-swift | 07:14 | |
| *** Guest87092 has quit IRC | 07:23 | |
| *** d0ugal has joined #openstack-swift | 07:24 | |
| *** d0ugal is now known as Guest14875 | 07:24 | |
| *** foexle has joined #openstack-swift | 07:24 | |
| *** Guest14875 is now known as d0ugal | 07:25 | |
| *** d0ugal has quit IRC | 07:25 | |
| *** d0ugal has joined #openstack-swift | 07:25 | |
| *** chlong has quit IRC | 07:39 | |
| *** mmcardle has joined #openstack-swift | 07:52 | |
| *** haomaiwa_ has quit IRC | 07:56 | |
| *** km has quit IRC | 08:09 | |
| *** km has joined #openstack-swift | 08:09 | |
| *** strycore has quit IRC | 08:12 | |
| *** strycore has joined #openstack-swift | 08:16 | |
| *** SkyRocknRoll has quit IRC | 08:19 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 08:28 |
| *** SkyRocknRoll has joined #openstack-swift | 08:33 | |
| *** aix has joined #openstack-swift | 08:36 | |
| *** SkyRocknRoll has quit IRC | 08:36 | |
| *** SkyRocknRoll has joined #openstack-swift | 08:37 | |
| *** silor1 has quit IRC | 08:52 | |
| *** marzif has joined #openstack-swift | 08:56 | |
| *** joeljwright has joined #openstack-swift | 09:00 | |
| *** ChanServ sets mode: +v joeljwright | 09:00 | |
| *** linkedinyou has joined #openstack-swift | 09:02 | |
| *** okdas_ is now known as okdas | 09:19 | |
| *** okdas has joined #openstack-swift | 09:19 | |
| *** haypo has joined #openstack-swift | 09:38 | |
| haypo | cschwede: hi. i had to rebase my xrange patch one more time :-p https://review.openstack.org/#/c/185460/ | 09:39 |
| *** esker has quit IRC | 09:42 | |
| openstackgerrit | Victor Stinner proposed openstack/swift: Get StringIO and cStringIO from six.moves https://review.openstack.org/185457 | 09:46 |
| openstackgerrit | Victor Stinner proposed openstack/swift: Replace StringIO with BytesIO in ssync https://review.openstack.org/186073 | 09:46 |
| openstackgerrit | Victor Stinner proposed openstack/swift: Replace StringIO with BytesIO for file https://review.openstack.org/186072 | 09:46 |
| openstackgerrit | Victor Stinner proposed openstack/swift: Replace StringIO with BytesIO for WSGI input https://review.openstack.org/186071 | 09:46 |
| cschwede | haypo: hello! Thanks for the rebase, will have a look later | 09:47 |
| haypo | cschwede: i'm rebasing my other patche | 09:47 |
| haypo | s | 09:47 |
| *** SkyRocknRoll has quit IRC | 09:47 | |
| joeljwright | cschwede: if you have time I could do with some help working out why https://review.openstack.org/#/c/189815/ is failing gate-tempest-dsvm-neutron-src-python-swiftclient | 09:48 |
| joeljwright | it seems to fail consistently on the same tests | 09:49 |
| joeljwright | but I can't for the life of me work out why it would cause that to fail! | 09:49 |
| openstackgerrit | Victor Stinner proposed openstack/swift: Use six to fix imports on Python 3 https://review.openstack.org/185453 | 09:50 |
| *** SkyRocknRoll has joined #openstack-swift | 09:52 | |
| openstackgerrit | Hisashi Osanai proposed openstack/swift: Enable Object Replicator's failure count in recon https://review.openstack.org/138342 | 09:53 |
| cschwede | joeljwright: i had a look last week, was wondering too why this fails all the time. i didn’t found a reason so far :/ i was even checking other openstack clients if they maybe use a named arg (for example get_object(„container“, obj=„objname“). That is the only reason i can think why it fails all the time | 09:56 |
| joeljwright | cschwede: I'm glad I'm not the only one who's confused | 09:59 |
| joeljwright | cschwede: I rechecked a few times to make sure the fails were consistent | 10:01 |
| joeljwright | cschwede: there's certainly a 'real' issue with those tests and this change, but it still seems wrong | 10:02 |
| cschwede | joeljwright: fully agree on the issue thing. i was worried in the beginning that it might break some 3rd party utils that use swiftclient; i didn’t expect this on the gate though.... | 10:03 |
| ho | cschwede: I found your report in http://techs.enovance.com/7501/openstack-swift-vancouver-summit-report the url in Changing policies goes to Container sharding :-) | 10:21 |
| *** jamielennox is now known as jamielennox|away | 10:25 | |
| *** rsFF has quit IRC | 10:36 | |
| *** ho has quit IRC | 10:36 | |
| *** km has quit IRC | 10:41 | |
| *** rsFF has joined #openstack-swift | 10:42 | |
| *** d0ugal has quit IRC | 10:43 | |
| *** proteusguy has quit IRC | 10:43 | |
| *** jlk has quit IRC | 10:43 | |
| *** kcaj has quit IRC | 10:43 | |
| *** MooingLemur has quit IRC | 10:43 | |
| *** xnox has quit IRC | 10:43 | |
| *** wasmum has quit IRC | 10:43 | |
| *** zacksh has quit IRC | 10:43 | |
| *** Guest11697 has quit IRC | 10:43 | |
| *** fbo_ has quit IRC | 10:43 | |
| *** zigo has quit IRC | 10:43 | |
| *** Trozz_ has quit IRC | 10:43 | |
| *** jroll has quit IRC | 10:43 | |
| *** Trozz_ has joined #openstack-swift | 10:43 | |
| *** xnox has joined #openstack-swift | 10:43 | |
| *** jlk has joined #openstack-swift | 10:43 | |
| *** jlk has quit IRC | 10:43 | |
| *** jlk has joined #openstack-swift | 10:43 | |
| *** d0ugal has joined #openstack-swift | 10:43 | |
| *** zacksh has joined #openstack-swift | 10:43 | |
| *** proteusguy has joined #openstack-swift | 10:43 | |
| *** jroll has joined #openstack-swift | 10:43 | |
| *** d0ugal is now known as Guest29011 | 10:43 | |
| *** mgagne has joined #openstack-swift | 10:43 | |
| *** mgagne is now known as Guest81202 | 10:43 | |
| *** MooingLe1ur has joined #openstack-swift | 10:43 | |
| *** MooingLe1ur has joined #openstack-swift | 10:43 | |
| *** fbo_ has joined #openstack-swift | 10:43 | |
| *** kcaj has joined #openstack-swift | 10:44 | |
| *** zigo has joined #openstack-swift | 10:44 | |
| *** Guest29011 is now known as d0ugal | 10:44 | |
| *** d0ugal has quit IRC | 10:44 | |
| *** d0ugal has joined #openstack-swift | 10:44 | |
| *** wasmum has joined #openstack-swift | 10:46 | |
| *** MooingLe1ur has quit IRC | 10:50 | |
| *** MooingLemur has joined #openstack-swift | 10:51 | |
| *** SkyRocknRoll_ has joined #openstack-swift | 10:53 | |
| *** silor has joined #openstack-swift | 11:00 | |
| *** linkedinyou has quit IRC | 11:04 | |
| *** kei_yama has quit IRC | 11:14 | |
| *** daemontool_ has joined #openstack-swift | 11:16 | |
| *** marzif has quit IRC | 11:18 | |
| *** cdelatte has joined #openstack-swift | 11:23 | |
| *** SkyRocknRoll_ has quit IRC | 11:23 | |
| *** daemontool_ has quit IRC | 11:38 | |
| *** marzif has joined #openstack-swift | 11:39 | |
| *** foexle has quit IRC | 11:41 | |
| *** marzif_ has joined #openstack-swift | 11:42 | |
| *** proteusguy has quit IRC | 12:05 | |
| *** proteusguy has joined #openstack-swift | 12:22 | |
| *** chlong has joined #openstack-swift | 12:25 | |
| *** chlong has quit IRC | 12:28 | |
| *** SkyRocknRoll has quit IRC | 12:28 | |
| openstackgerrit | Christian Schwede proposed openstack/swift: Add note about updatedb to the docs https://review.openstack.org/194126 | 12:50 |
| cschwede | ho: thx for the note, I’ll update it! | 12:59 |
| *** ppai has quit IRC | 13:00 | |
| *** CaioBrentano has joined #openstack-swift | 13:27 | |
| *** mwheckmann has joined #openstack-swift | 13:36 | |
| *** jroll has quit IRC | 13:36 | |
| *** jroll has joined #openstack-swift | 13:36 | |
| *** SkyRocknRoll has joined #openstack-swift | 13:40 | |
| *** zigo has quit IRC | 13:47 | |
| *** SkyRocknRoll has quit IRC | 13:53 | |
| *** breitz has joined #openstack-swift | 13:56 | |
| *** ruind has joined #openstack-swift | 14:01 | |
| *** blmartin has joined #openstack-swift | 14:03 | |
| *** zigo has joined #openstack-swift | 14:05 | |
| ruind | I googled around but couldnt find the answer to this.... What is a swift object? | 14:07 |
| CaioBrentano | Do you mean what is a object in "object storage concept"? | 14:21 |
| CaioBrentano | ruind: Do you mean what is a object in "object storage concept"? | 14:21 |
| *** tellesnobrega_ is now known as tellesnobrega | 14:22 | |
| *** tong_ has joined #openstack-swift | 14:23 | |
| ruind | yes | 14:23 |
| ruind | i mean what is an object in the storage sense | 14:23 |
| *** acampbell has joined #openstack-swift | 14:24 | |
| ruind | CaioBrentano, | 14:24 |
| *** wbhuber has joined #openstack-swift | 14:24 | |
| *** tongli has joined #openstack-swift | 14:28 | |
| *** tong_ has quit IRC | 14:28 | |
| CaioBrentano | In my understanding is any unstructured data... Pictures, Movies, static web files, etc | 14:30 |
| *** acampbell has quit IRC | 14:30 | |
| CaioBrentano | ruind, | 14:30 |
| *** Guest78977 has joined #openstack-swift | 14:33 | |
| ruind | CaioBrentano, ok, but how is an object different from a file then | 14:33 |
| haypo | cschwede: we have to synchronize ourself to merge the patch quickly when you agree on it, my xrange() patch is in conflict again! | 14:34 |
| haypo | cschwede: i fixed conflicts this morning :-p | 14:34 |
| *** blmartin_ has joined #openstack-swift | 14:34 | |
| CaioBrentano | ruind: the difference is the way you are storing and accessing it | 14:34 |
| ruind | which is? | 14:35 |
| ctennis | via HTTP | 14:35 |
| ruind | ctennis, I can access files via http | 14:35 |
| ruind | I can download, copy, upload, etc | 14:36 |
| ruind | modify | 14:36 |
| ctennis | right but this is in comparison to a normal filesystem | 14:36 |
| *** blmartin has quit IRC | 14:36 | |
| ctennis | you address and store the object within http, and it worries about where inside of a very large system of harddrives the data is stored and replicated | 14:36 |
| ruind | ctennis, what is abnormal about a filesystem that stores objects? (or should i say objectsystem) | 14:36 |
| ctennis | there's nothing abnormal about it | 14:37 |
| ctennis | this is a different way of doing it | 14:37 |
| ruind | how is it different? | 14:37 |
| hurricanerix | ruind: maybe you should start out here: http://docs.openstack.org/developer/swift/api/object_api_v1_overview.html | 14:37 |
| ctennis | probably need to read and watch some videos | 14:38 |
| ctennis | https://www.youtube.com/watch?v=lYT7PGbribk | 14:38 |
| ctennis | it's just a different design for a way to store things | 14:38 |
| ruind | hurricanerix, ok, this is interesting | 14:39 |
| *** jrichli has joined #openstack-swift | 14:39 | |
| ruind | seems like it is a grouping of files | 14:39 |
| *** marzif_ has quit IRC | 14:39 | |
| *** marzif_ has joined #openstack-swift | 14:39 | |
| ruind | ctennis, I'll check that video. I watched another one by Dickinson and it was too high level for what I was looking for | 14:40 |
| ctennis | you can think of it that way, btu I would say: "what is a file?" | 14:40 |
| ruind | a set of blocks on the filesystem with a header containing metadata about that grouping of blocks like the name, creation time, permissions, etc. | 14:41 |
| ruind | seems like an object is a higher level abstraction | 14:42 |
| ctennis | right, so an object is just a different way of abstracting that. you don't have to worry about blocks, or a filesystem. you have a named object with bytes associated with it, and you upload it to the system | 14:42 |
| ruind | block -> file -> object | 14:42 |
| ctennis | right | 14:42 |
| ruind | ctennis, yes, I dont _have_ to worry about it, but I am interested in it. | 14:43 |
| ctennis | Certainly, I'm just trying to differentiate file to object | 14:43 |
| *** marcusvrn_ has joined #openstack-swift | 14:43 | |
| ruind | well thanks. I think I have a basic understanding, and I'll look at these materials to dive deeper | 14:45 |
| notmyname | ruind: https://dl.dropboxusercontent.com/u/21194/distributed-object-store-principles-of-operation.pdf is a good overview of "what is an object storage system" and "what are the hard parts of building one" | 14:46 |
| CaioBrentano | notmyname: cool, thanks for sharing! | 14:47 |
| *** Guest78977 has quit IRC | 14:53 | |
| *** minwoob has joined #openstack-swift | 14:54 | |
| *** marzif has quit IRC | 14:55 | |
| *** wshao has joined #openstack-swift | 14:55 | |
| *** wshao has left #openstack-swift | 14:55 | |
| *** doxavore has joined #openstack-swift | 15:01 | |
| *** janonymous_ has joined #openstack-swift | 15:10 | |
| *** aix has quit IRC | 15:16 | |
| *** janonymous_ has quit IRC | 15:19 | |
| *** agentle has joined #openstack-swift | 15:19 | |
| openstackgerrit | Janie Richling proposed openstack/swift: Encrypter and Decrypter middleware - object body only https://review.openstack.org/194191 | 15:21 |
| *** jlhinson has joined #openstack-swift | 15:28 | |
| *** haypo has left #openstack-swift | 15:28 | |
| *** acampbell has joined #openstack-swift | 15:30 | |
| *** acampbell has quit IRC | 15:35 | |
| *** Guest81202 is now known as mgagne | 15:36 | |
| *** mgagne has joined #openstack-swift | 15:36 | |
| *** agentle has quit IRC | 15:39 | |
| *** agentle has joined #openstack-swift | 15:40 | |
| *** marzif_ has quit IRC | 15:41 | |
| *** idevops has joined #openstack-swift | 15:41 | |
| *** marzif_ has joined #openstack-swift | 15:41 | |
| *** idevops has quit IRC | 15:42 | |
| *** janonymous_ has joined #openstack-swift | 15:42 | |
| *** janonymous_ has quit IRC | 15:52 | |
| *** petertr7 is now known as petertr7_away | 15:55 | |
| *** acampbell has joined #openstack-swift | 16:00 | |
| *** eranrom has joined #openstack-swift | 16:02 | |
| eranrom | Greetings everyone. I have a general question regarding container sync Vs. the reconciler: | 16:05 |
| eranrom | Would it be correct to state that they both traverse a db sequentially, potentially reading an object from one side and writing it to another side plus they both run locally in the container server hence can produce pretty much the same throughgput? | 16:05 |
| *** gyee has joined #openstack-swift | 16:06 | |
| *** agentle has quit IRC | 16:08 | |
| *** peluse has quit IRC | 16:08 | |
| *** acampbell has quit IRC | 16:09 | |
| *** zhill has joined #openstack-swift | 16:17 | |
| eranrom | Will try my luck later :-) Thanks! | 16:17 |
| *** eranrom has left #openstack-swift | 16:18 | |
| *** david-ly_ is now known as david-lyle | 16:18 | |
| *** nadeem has joined #openstack-swift | 16:22 | |
| *** agentle has joined #openstack-swift | 16:23 | |
| *** bpranswift has joined #openstack-swift | 16:27 | |
| *** proteusguy has quit IRC | 16:31 | |
| *** rledisez has quit IRC | 16:36 | |
| *** bpranswift has quit IRC | 16:43 | |
| *** proteusguy has joined #openstack-swift | 16:44 | |
| *** tongli has quit IRC | 16:46 | |
| *** tongli has joined #openstack-swift | 16:47 | |
| *** petertr7_away is now known as petertr7 | 16:47 | |
| *** peluse has joined #openstack-swift | 16:47 | |
| *** peluse_ has joined #openstack-swift | 16:48 | |
| *** peluse has quit IRC | 16:48 | |
| *** bpranswift has joined #openstack-swift | 16:49 | |
| *** agentle has quit IRC | 16:49 | |
| *** bpranswift has left #openstack-swift | 16:49 | |
| *** agentle has joined #openstack-swift | 16:50 | |
| *** tongli has quit IRC | 16:51 | |
| *** jlhinson has quit IRC | 16:52 | |
| *** doxavore has quit IRC | 16:56 | |
| *** tongli has joined #openstack-swift | 16:59 | |
| *** esker has joined #openstack-swift | 17:03 | |
| *** klrmn has joined #openstack-swift | 17:05 | |
| *** doxavore has joined #openstack-swift | 17:09 | |
| *** zhill has quit IRC | 17:16 | |
| *** peluse has joined #openstack-swift | 17:20 | |
| *** peluse_ has quit IRC | 17:20 | |
| notmyname | peluse: this is why I've been bugging you so much about what we can do in swift to take advantage of flash :-) http://www.networkcomputing.com/storage/ssd-prices-in-a-freefall/a/d-id/1320958 | 17:23 |
| peluse | sweet | 17:24 |
| *** zhill has joined #openstack-swift | 17:24 | |
| *** marcusvrn_ has quit IRC | 17:24 | |
| notmyname | the article is especially poignant when I think of a conversation I had late last week about SMR drives. I'm hoping to start testing them this week or next and I should expect them to be very bad (perf-wise) until we get some sort of native scsi command support in swift (ha!) and manage it all ourselves | 17:27 |
| *** peluse has quit IRC | 17:31 | |
| MooingLemur | we've been using dm-cache, with some benefit | 17:33 |
| MooingLemur | but not as much as with bcache | 17:33 |
| *** joeljwright has quit IRC | 17:33 | |
| MooingLemur | problem with bcache is that it crashes/is unstable for what appears to be > 2TiB block devices. | 17:33 |
| *** MVenesio has joined #openstack-swift | 17:35 | |
| MooingLemur | but the fact that bcache takes writeback data and writes it sequentially to the flash device makes it awesome for write latency and performance | 17:35 |
| MooingLemur | dm-cache for us is simply better than nothing | 17:36 |
| MooingLemur | but if Swift can manage the cache somehow, it seems like we could do things even better. | 17:37 |
| MooingLemur | the most expensive operations seem to be stat() during object replication. If dm-cache can eventually cache the related blocks, object replication speeds up considerably for us, and the load drops | 17:38 |
| *** marzif_ has quit IRC | 17:39 | |
| *** zaitcev has joined #openstack-swift | 17:40 | |
| *** ChanServ sets mode: +v zaitcev | 17:40 | |
| notmyname | MooingLemur: yeah, I think the important word there is "somehow" ;-) | 17:40 |
| *** agentle has quit IRC | 17:41 | |
| *** agentle has joined #openstack-swift | 17:41 | |
| CaioBrentano | I'm having a bad time with the URL format in my company! My PM is demanding a custom URL (without /v1/AUTH_<uuid>) for a specific swift account. Does anyone have done something about this? I was thinking about a nginx proxy, is this a bad aproach? | 17:50 |
| *** agentle has quit IRC | 17:50 | |
| notmyname | CaioBrentano: yes, that's possible. but first let me say that it makes me laugh. (at PMs who want things to always be shiny) ;-) | 17:51 |
| notmyname | CaioBrentano: so there's 2 pieces of middleware included in swift that you can use | 17:51 |
| notmyname | the first is called domain_remap | 17:51 |
| notmyname | what it does is translate subdomains to account and/or container path segments | 17:52 |
| notmyname | eg translates: bar.foo.swift.example.com/obj --> swift.example.com/v1/foo/bar/obj | 17:52 |
| notmyname | so that's something | 17:52 |
| notmyname | but if you also use the cname_lookup middleware, you can whitelabel that whole subdomain | 17:53 |
| notmyname | eg myawesome.site.com/obj -> (via CNAME) bar.foo.swift.example.com/obj --> swift.example.com/v1/foo/bar/obj | 17:54 |
| notmyname | CaioBrentano: make sense? | 17:54 |
| notmyname | note that if you have a heavily trafficked site and you use the cname_lookup middleware, you'll probably want to beef up your DNS infrastructure to handle the extra load | 17:55 |
| CaioBrentano | notmyname: at first look makes sense! I'll check both! thanks! | 17:55 |
| CaioBrentano | notmyname: yeah, it's heavily trafficked! :( | 17:56 |
| CaioBrentano | notmyname: you can bet these decision is more political than tech! ;) | 17:57 |
| notmyname | CaioBrentano: it's just something be aware of. some local caching dns proxies locally should probably be sufficient | 17:57 |
| CaioBrentano | notmyname: thanks again! | 17:59 |
| *** jlhinson has joined #openstack-swift | 18:04 | |
| *** jlhinson_ has joined #openstack-swift | 18:06 | |
| *** jlhinson has quit IRC | 18:09 | |
| *** jlhinson__ has joined #openstack-swift | 18:10 | |
| *** jlhinson_ has quit IRC | 18:12 | |
| *** wbhuber has quit IRC | 18:18 | |
| *** jasondot_ has joined #openstack-swift | 18:25 | |
| *** tongli has quit IRC | 18:27 | |
| *** esker has quit IRC | 18:38 | |
| notmyname | is there any reason we cast some timeout values to int() and some to float()? | 18:42 |
| notmyname | seem the "short" ones are float and the "long" ones are int | 18:42 |
| notmyname | eg look at the object server and the conn_timeout vs node_timeout or client_timeout | 18:42 |
| *** jlhinson has joined #openstack-swift | 18:42 | |
| notmyname | since I'm in that part of the code right now, I'm wondering if there's any reason not to change them all to floats | 18:43 |
| *** jlhinson__ has quit IRC | 18:43 | |
| *** silor has quit IRC | 18:47 | |
| clayg | notmyname: i don't know if it helps - but we can definately *response* with _'s in header names | 18:47 |
| notmyname | clayg: ie if that's what's stored on disk? | 18:47 |
| clayg | notmyname: yay floats! | 18:48 |
| notmyname | if so, that does help, i think | 18:48 |
| clayg | wat? | 18:48 |
| *** amoturi has joined #openstack-swift | 18:48 | |
| notmyname | how'd you get swift to respond with a _ in a header name? | 18:48 |
| clayg | notmyname: no - by the time we get any header that was sent with _'s it looks like the client sent it with -'s | 18:48 |
| notmyname | right | 18:48 |
| clayg | but if we stick a _ in a header response wsgi will leave it that way | 18:48 |
| notmyname | ah ok. that's good | 18:48 |
| clayg | notmyname: i just stuck it in a response in the proxy | 18:48 |
| notmyname | gotcha | 18:48 |
| notmyname | well, that's some good news at least | 18:49 |
| notmyname | for those playing along, we're talking about https://bugs.launchpad.net/swift/+bug/1467155 | 18:49 |
| openstack | Launchpad bug 1467155 in Swift3 "User Metadata keys with Underscore get translated to dash" [Undecided,New] | 18:49 |
| clayg | notmyname: now we just need a "translate-for-crazy-clients" middleware that'll let you configure a list of headers that get duplicated into under's format | 18:49 |
| notmyname | all of them? | 18:50 |
| clayg | well - that *would* be kinda crazy :\ | 18:51 |
| *** jasondot_ has quit IRC | 18:51 | |
| clayg | notmyname: maybe we could do it based on the user-agent? | 18:51 |
| clayg | idk, I feel like this client that's sending the _'s in header key's is sorta asking for trouble - as the bug points out - lots of web services do this - on purpose. | 18:52 |
| notmyname | I don't see how that helps. a user gives us one set of bytes and we give back another. why would that be "fixed" (or broken, depending on the perspective) for one client? | 18:53 |
| notmyname | yeah, but that's what users do, right? ask for trouble? ;-) | 18:53 |
| clayg | liberal in what you accept - conservative in what you send? | 18:53 |
| notmyname | interesting. not forbidden in the rfc, but you have to explicitly enable support in the most popular servers http://stackoverflow.com/questions/22856136/why-underscores-are-forbidden-in-http-header-names | 18:55 |
| clayg | the other option might be to 400 the request down in eventlet land? | 18:55 |
| *** gsilvis_ is now known as gsilvis | 18:55 | |
| clayg | notmyname: there has to be a better use-case than "I *really* just want to store user-metadata with underscores - I just *really* want to!" | 18:57 |
| clayg | otherwise it seems a lot like "docter it hurts when I punch myself in the face" | 18:57 |
| notmyname | requests valid for s3 are broken when sent to swift/swift3 but would otherwise work were it not for this issue. so I think the use case is "better fidelity with s3" | 18:58 |
| *** agentle has joined #openstack-swift | 18:58 | |
| clayg | notmyname: see I'm down with that - except - what header is in the swift3 signature that has the _'s? | 18:59 |
| notmyname | that I'm not sure of | 18:59 |
| clayg | ... because if it's just user metadata ... | 18:59 |
| notmyname | does s3 use the user metadata as part of the hmac auth signature? | 18:59 |
| clayg | notmyname: everything you *send* is part of the signature - yeah | 19:00 |
| notmyname | and if so maybe the x-amz-* headers are translated to x-object-meta-*? and if so we break? | 19:00 |
| clayg | maybe? are there some x-amz-* headers that expect _'s - or we're still in user-meta land? | 19:00 |
| notmyname | AFAIK, still in user-meta land | 19:01 |
| clayg | notmyname: well wtf are we going to do if you send some _'s and some -'s for user meta | 19:01 |
| clayg | sigh | 19:01 |
| *** agentle_ has joined #openstack-swift | 19:02 | |
| notmyname | lol @ "because rfc822 http headers disallow _ in names, in $meta_name two hyphens in a row (--) will be translated to an underscore(_)." | 19:02 |
| notmyname | from http://archive.org/help/abouts3.txt | 19:02 |
| notmyname | http://docs.aws.amazon.com/AmazonS3/latest/dev/UsingMetadata.html#object-metadata is silent on it | 19:04 |
| *** agentle has quit IRC | 19:05 | |
| *** esker has joined #openstack-swift | 19:11 | |
| *** agentle_ has quit IRC | 19:17 | |
| *** agentle has joined #openstack-swift | 19:18 | |
| *** wbhuber has joined #openstack-swift | 19:18 | |
| *** nadeem has quit IRC | 19:23 | |
| *** jasondot_ has joined #openstack-swift | 19:24 | |
| *** agentle has quit IRC | 19:27 | |
| MooingLemur | wuh. | 19:34 |
| MooingLemur | well, most of the perl modules that involve http headers will translate underscores into hyphens for header names | 19:36 |
| MooingLemur | so you can say things like content_type => 'text/html' | 19:36 |
| MooingLemur | rather than having to quote the field name | 19:36 |
| MooingLemur | it sounds like a case of "people started doing things one way, but as a result there's an ambiguity that wasn't forethought by the standard" | 19:38 |
| *** marzif_ has joined #openstack-swift | 19:39 | |
| clayg | notmyname: it looks like we can get ahold of self.headers before they get mangled in get_environ | 19:41 |
| *** peluse has joined #openstack-swift | 19:42 | |
| MooingLemur | perl HTTP::Headers will canonicalize _ to - unless you precede the name with a colon, which tells the module to leave it alone. | 19:42 |
| MooingLemur | so I guess there are use cases for underscores in header names, but you really really have to want it | 19:42 |
| clayg | notmyname: unrelated - i had another patch that was fixing how run_server injected the wsgi.HttpProtocol that already super's get_environ - so it'd be easy enough to add it there and capture the un-managled headers in an environ | 19:42 |
| clayg | after that it's just plumbing :'( | 19:43 |
| clayg | seem annoying tho | 19:43 |
| *** mgarza has joined #openstack-swift | 19:47 | |
| *** peluse has quit IRC | 19:47 | |
| *** agentle has joined #openstack-swift | 19:52 | |
| notmyname | clayg: that's cool | 19:55 |
| *** peluse has joined #openstack-swift | 20:02 | |
| *** agentle has quit IRC | 20:02 | |
| *** marzif_ has quit IRC | 20:07 | |
| *** elmo_ is now known as elmo | 20:08 | |
| *** thurloat is now known as thurloat_isgone | 20:11 | |
| *** petertr7 is now known as petertr7_away | 20:13 | |
| *** petertr7_away is now known as petertr7 | 20:13 | |
| *** blmartin_ has quit IRC | 20:22 | |
| *** amoturi has quit IRC | 20:27 | |
| *** Anticime1 is now known as Anticimex | 20:29 | |
| *** mwheckmann has quit IRC | 20:30 | |
| *** wbhuber_ has joined #openstack-swift | 20:31 | |
| *** jasondot_ has quit IRC | 20:33 | |
| *** wbhuber has quit IRC | 20:34 | |
| *** tongli has joined #openstack-swift | 20:37 | |
| *** peluse has quit IRC | 20:39 | |
| *** tongli has quit IRC | 20:40 | |
| notmyname | OpenStack Mita release | 20:53 |
| zaitcev | I always lose bad at those votes. | 20:56 |
| zaitcev | I swear this is the last time | 20:56 |
| notmyname | I haven't seen anything in official channels yet. just one tweet about it | 20:58 |
| *** jasondotstar has joined #openstack-swift | 20:59 | |
| *** marcusvrn_ has joined #openstack-swift | 21:00 | |
| zaitcev | Oh I'm sure it's going to be like that | 21:01 |
| *** aix has joined #openstack-swift | 21:05 | |
| *** petertr7 is now known as petertr7_away | 21:06 | |
| *** MVenesio has quit IRC | 21:10 | |
| *** wbhuber__ has joined #openstack-swift | 21:10 | |
| *** wbhuber_ has quit IRC | 21:11 | |
| *** wbhuber_ has joined #openstack-swift | 21:11 | |
| *** wbhuber__ has quit IRC | 21:14 | |
| clayg | is it m-eat-a or m-eye-ta | 21:16 |
| notmyname | I think we need kota_/dmorita/ho to tell us | 21:17 |
| si1v3r | me ah ta | 21:18 |
| MooingLemur | ma-ia hii, mi-ia huu, ma-ia hoo, ma-ia haha | 21:18 |
| torgomatic | We've got Rack, WSGI, and PSGI conflating dashes and underscores in HTTP headers. Frameworks: 3, Me: 0 | 21:20 |
| *** marzif_ has joined #openstack-swift | 21:20 | |
| clayg | i love it that PSGI isn't a thing torgomatic had to make up | 21:23 |
| *** jasondotstar has quit IRC | 21:27 | |
| *** peluse has joined #openstack-swift | 21:28 | |
| zaitcev | it's just mee-ta, 2 syllables (三田) | 21:30 |
| *** jasondot_ has joined #openstack-swift | 21:37 | |
| MooingLemur | Lovely Mita, Reter Raid | 21:41 |
| zaitcev | Mikasa would be awesome, but... | 21:41 |
| zaitcev | Not in our age, when we had to rename Quantum. | 21:42 |
| zaitcev | Heck, we might need to rename Swift. | 21:42 |
| notmyname | unlikely at this point :-) | 21:42 |
| *** zigo has quit IRC | 21:43 | |
| *** jlhinson has quit IRC | 21:47 | |
| *** CaioBrentano has quit IRC | 21:48 | |
| *** thurloat_isgone is now known as thurloat | 21:48 | |
| *** zigo has joined #openstack-swift | 21:49 | |
| *** doxavore has quit IRC | 21:50 | |
| *** peluse has quit IRC | 21:51 | |
| *** swat30 has quit IRC | 21:52 | |
| *** proteusguy has quit IRC | 21:53 | |
| *** proteusguy has joined #openstack-swift | 21:54 | |
| *** swat30 has joined #openstack-swift | 21:55 | |
| *** ianbrown has joined #openstack-swift | 21:56 | |
| *** blmartin has joined #openstack-swift | 21:57 | |
| *** mgarza has quit IRC | 22:04 | |
| *** agentle has joined #openstack-swift | 22:06 | |
| *** jrichli has quit IRC | 22:10 | |
| *** jasondot_ has quit IRC | 22:12 | |
| *** blmartin has quit IRC | 22:13 | |
| *** blmartin has joined #openstack-swift | 22:13 | |
| *** zigo has quit IRC | 22:15 | |
| *** peluse has joined #openstack-swift | 22:16 | |
| *** marzif__ has joined #openstack-swift | 22:16 | |
| *** zigo has joined #openstack-swift | 22:17 | |
| *** marzif_ has quit IRC | 22:19 | |
| *** marzif__ has quit IRC | 22:21 | |
| *** peluse has quit IRC | 22:23 | |
| *** thurloat is now known as thurloat_isgone | 22:24 | |
| *** thurloat_isgone is now known as thurloat | 22:33 | |
| *** thurloat is now known as thurloat_isgone | 22:34 | |
| *** aix has quit IRC | 22:41 | |
| *** aix has joined #openstack-swift | 22:41 | |
| *** jasondotstar has joined #openstack-swift | 22:43 | |
| *** agentle has quit IRC | 22:50 | |
| *** agentle has joined #openstack-swift | 22:51 | |
| *** DericHorn-HP_ has joined #openstack-swift | 22:56 | |
| *** wer has quit IRC | 22:58 | |
| *** wer has joined #openstack-swift | 23:03 | |
| *** DericHorn-HP_ is now known as DericHorn-HP | 23:04 | |
| *** km has joined #openstack-swift | 23:10 | |
| *** egon_ is now known as egon | 23:10 | |
| *** jlhinson has joined #openstack-swift | 23:11 | |
| *** pgbridge has quit IRC | 23:13 | |
| *** kei_yama has joined #openstack-swift | 23:14 | |
| *** marcusvrn_ has quit IRC | 23:14 | |
| *** DericHorn-HP has quit IRC | 23:21 | |
| *** DericHorn-HP has joined #openstack-swift | 23:24 | |
| *** jlhinson_ has joined #openstack-swift | 23:25 | |
| *** jlhinson has quit IRC | 23:26 | |
| *** jasondotstar has quit IRC | 23:29 | |
| *** DericHorn-HP has quit IRC | 23:30 | |
| *** ho has joined #openstack-swift | 23:39 | |
| *** agentle has quit IRC | 23:45 | |
| *** blmartin has quit IRC | 23:47 | |
| *** blmartin has joined #openstack-swift | 23:47 | |
| *** gyee has quit IRC | 23:48 | |
| *** zhill has quit IRC | 23:49 | |
| *** jasondotstar has joined #openstack-swift | 23:54 | |
| *** jlhinson_ has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!