| openstackgerrit | paul luse proposed openstack/swift: Erasure Code Reconstructor https://review.openstack.org/131872 | 00:05 |
|---|---|---|
| *** joeljwright has joined #openstack-swift | 00:10 | |
| *** yuanzz has quit IRC | 00:12 | |
| *** joeljwright has quit IRC | 00:15 | |
| *** thumpba_ has quit IRC | 00:15 | |
| *** Gues_____ has joined #openstack-swift | 00:19 | |
| *** Guest___ has quit IRC | 00:19 | |
| *** Gues_____ has quit IRC | 00:21 | |
| *** dmorita has quit IRC | 00:29 | |
| *** shri1 has left #openstack-swift | 00:30 | |
| *** kota_ has joined #openstack-swift | 00:34 | |
| *** dmorita has joined #openstack-swift | 00:53 | |
| *** jkugel has joined #openstack-swift | 01:04 | |
| *** jkugel1 has joined #openstack-swift | 01:06 | |
| *** jkugel has quit IRC | 01:08 | |
| *** jkugel1 has left #openstack-swift | 01:10 | |
| *** tsg has joined #openstack-swift | 01:42 | |
| *** nellysmitt has joined #openstack-swift | 01:43 | |
| *** nellysmitt has quit IRC | 01:47 | |
| zaitcev | I'm wondering if we have many people trying to submit (with git-review) from long-living branches. https://review.openstack.org/158877 | 01:57 |
| zaitcev | I'm not into branches in general, but I know many people are | 01:57 |
| zaitcev | MKJ is trying to implement a whole new workflow basically. | 01:58 |
| *** haomaiwang has joined #openstack-swift | 02:32 | |
| openstackgerrit | paul luse proposed openstack/swift: wip: ec probe test https://review.openstack.org/164291 | 03:03 |
| *** gyee has quit IRC | 03:15 | |
| *** nellysmitt has joined #openstack-swift | 03:44 | |
| *** nellysmitt has quit IRC | 03:48 | |
| *** dmorita has quit IRC | 03:57 | |
| *** kota_ has quit IRC | 04:05 | |
| *** ppai has joined #openstack-swift | 04:22 | |
| *** tsg has quit IRC | 04:30 | |
| *** kota_ has joined #openstack-swift | 05:00 | |
| *** nshaikh has joined #openstack-swift | 05:16 | |
| *** ho has joined #openstack-swift | 05:20 | |
| *** ho has quit IRC | 05:20 | |
| *** panbalag has quit IRC | 05:27 | |
| openstackgerrit | Merged openstack/swift: Add the missing "file" keyword for print https://review.openstack.org/164740 | 05:38 |
| *** nellysmitt has joined #openstack-swift | 05:44 | |
| *** nellysmitt has quit IRC | 05:49 | |
| *** SkyRocknRoll has joined #openstack-swift | 05:55 | |
| *** bkopilov has joined #openstack-swift | 05:55 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Imported Translations from Transifex https://review.openstack.org/164964 | 06:11 |
| *** nellysmitt has joined #openstack-swift | 06:32 | |
| *** silor has joined #openstack-swift | 06:59 | |
| openstackgerrit | Prashanth Pai proposed openstack/swift: Refactor server side copy as middleware https://review.openstack.org/156923 | 07:03 |
| *** Trozz has joined #openstack-swift | 07:05 | |
| *** aswadr has joined #openstack-swift | 07:14 | |
| *** aix has quit IRC | 07:20 | |
| *** nellysmi_ has joined #openstack-swift | 07:21 | |
| *** nellysmitt has quit IRC | 07:22 | |
| *** nellysmi_ has quit IRC | 07:26 | |
| *** nellysmitt has joined #openstack-swift | 07:28 | |
| *** joeljwright has joined #openstack-swift | 07:55 | |
| *** Bsony has joined #openstack-swift | 07:58 | |
| *** jordan_ has joined #openstack-swift | 08:18 | |
| *** jistr has joined #openstack-swift | 08:22 | |
| *** geaaru has joined #openstack-swift | 08:23 | |
| *** Trozz is now known as zzTrozz | 08:28 | |
| *** zzTrozz is now known as Trozz | 08:28 | |
| *** silor has quit IRC | 08:30 | |
| *** uschreiber_ has joined #openstack-swift | 08:31 | |
| *** uschreiber_ has quit IRC | 08:33 | |
| *** uschreiber_ has joined #openstack-swift | 08:38 | |
| *** uschreiber_ has quit IRC | 08:39 | |
| *** uschreiber_ has joined #openstack-swift | 08:41 | |
| *** uschreiber_ has quit IRC | 08:42 | |
| *** uschreiber_ has joined #openstack-swift | 08:42 | |
| *** uschreiber_ has quit IRC | 08:42 | |
| *** ppai has quit IRC | 08:54 | |
| *** kota_ has quit IRC | 08:56 | |
| *** ppai has joined #openstack-swift | 09:08 | |
| *** ho has joined #openstack-swift | 09:30 | |
| *** ho has quit IRC | 09:31 | |
| *** ppai has quit IRC | 10:04 | |
| *** nellysmitt has quit IRC | 10:13 | |
| openstackgerrit | Christian Schwede proposed openstack/swift: Diskfile decides if durable is written based on policy https://review.openstack.org/165208 | 10:16 |
| *** ppai has joined #openstack-swift | 10:17 | |
| *** acoles_away is now known as acoles | 10:24 | |
| *** haomaiwang has quit IRC | 10:26 | |
| *** haomaiwang has joined #openstack-swift | 10:27 | |
| acoles | cschwede: hi! how's things? | 10:27 |
| acoles | cschwede: thanks for fixing my commit message - i was pretty tired when i wrote that !) | 10:27 |
| cschwede | acoles: fine, thanks! and for you? you’re working a lot on the EC stuff right now? | 10:27 |
| cschwede | acoles: you’re welcome - i didn’t want to nitpick, and because there was no other +2 yet i thought i just do it | 10:28 |
| acoles | cschwede: yes, trying to help out there, and some crypto stuff going on in the background too | 10:28 |
| acoles | cschwede: absolutely, v helpful to just have it fixed! | 10:28 |
| cschwede | yeah, i’m playing with the EC stuff as well, going to add some comments today - some of the patches require quite a bit of time if one needs read > 1000 LOC ;) | 10:29 |
| acoles | cschwede: yes :( | 10:29 |
| *** haomaiwang has quit IRC | 10:31 | |
| acoles | cschwede: be aware that there are two 'alternative' patch chains for some of the diskfile stuff - one below patch 165208 (the one you reviewed) | 10:32 |
| patchbot | acoles: https://review.openstack.org/#/c/165208/ | 10:32 |
| acoles | cschwede: and one below patch 162717 | 10:32 |
| patchbot | acoles: https://review.openstack.org/#/c/162717/ | 10:32 |
| acoles | cschwede: so avoid duplicating your review effort! | 10:33 |
| *** nellysmitt has joined #openstack-swift | 10:35 | |
| acoles | jrichli: hi, good morning, i had an idea re not being able to send 'footers' on crypto branch yet - ping me when you are around | 10:35 |
| cschwede | acoles: oh, ok thx - so what is the plan on this? the 165208 is newer, thus i assume the idea for that patch is kind of „better“, or more sophisticated? | 10:35 |
| acoles | cschwede: the reason for the two is that clayg (and i) are trying to break the diskfile changes into per-policy-type diskfile implementations (sub-classing legacy diskfile) | 10:36 |
| *** nellysmitt has quit IRC | 10:37 | |
| acoles | cschwede: so patch 165125 for example gets us EC specific changes in an ECDiskFile impl, with a relatively unchanged 'legacy' DiskFile class | 10:38 |
| patchbot | acoles: https://review.openstack.org/#/c/165125/ | 10:38 |
| cschwede | yes, that’s the thing i like about this patch | 10:39 |
| acoles | cschwede: i think there is a meeting in swiftstack later today when peluse at al will discuss/decide | 10:39 |
| cschwede | ok, good to know | 10:40 |
| acoles | cschwede: 165208 and 162717 are basically the same but based on different underlying approaches | 10:40 |
| acoles | cschwede: maybe we'll get an update on th status of those approaches in today's meeting | 10:42 |
| cschwede | ok, so i’m going back to 159637 and see if i can contribute a helpful review on that large thing | 10:44 |
| acoles | cschwede: *some* of 159637 is duplicated in 165188 (based on the per-policy subclasses) - clayg was going to move the rest of 159637 but i don't see a patch for that yet | 10:48 |
| * cschwede needs to draw a map for all the ec patches | 10:49 | |
| acoles | cschwede: (sorry, i realise it's got a bit confusing) | 10:49 |
| cschwede | no worries, need to get my head around it | 10:49 |
| acoles | cschwede: its like erasure coded gerrit reviews - split up the changes, introduce some redundancy, and push multiple patches :) as long as 10/14 patches land we'll be good - no matter which 10 :) | 10:53 |
| *** Bsony_ has joined #openstack-swift | 10:54 | |
| *** nellysmitt has joined #openstack-swift | 10:55 | |
| *** Bsony has quit IRC | 10:57 | |
| *** ppai has quit IRC | 11:00 | |
| tdasilva | cschwede: Hey, if you are looking for patches to review, check out patch 156825 :P | 11:03 |
| patchbot | tdasilva: https://review.openstack.org/#/c/156825/ | 11:03 |
| cschwede | tdasilva: will do :) | 11:04 |
| *** nellysmitt has quit IRC | 11:08 | |
| *** ppai has joined #openstack-swift | 11:13 | |
| *** nellysmitt has joined #openstack-swift | 11:33 | |
| openstackgerrit | Prashanth Pai proposed openstack/swift: Refactor server side copy as middleware https://review.openstack.org/156923 | 11:39 |
| *** nellysmitt has quit IRC | 11:43 | |
| *** erlon has joined #openstack-swift | 11:53 | |
| *** david-lyle_afk is now known as david-lyle | 12:00 | |
| *** ppai has quit IRC | 12:04 | |
| openstackgerrit | Merged openstack/swift: Allow swift-object-info to inspect .meta and .ts files https://review.openstack.org/162306 | 12:15 |
| *** silor has joined #openstack-swift | 12:17 | |
| *** jistr is now known as jistr|biab | 12:39 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: Test swift-object-info opens meta and ts files https://review.openstack.org/165395 | 12:49 |
| *** silor has quit IRC | 13:03 | |
| *** rdaly2 has joined #openstack-swift | 13:13 | |
| openstackgerrit | Merged openstack/swift: Imported Translations from Transifex https://review.openstack.org/164964 | 13:24 |
| *** bkopilov has quit IRC | 13:27 | |
| *** nshaikh has quit IRC | 13:34 | |
| *** jistr|biab is now known as jistr | 13:39 | |
| *** mahatic has joined #openstack-swift | 13:49 | |
| openstackgerrit | Thiago da Silva proposed openstack/swift: Refactoring the PUT method https://review.openstack.org/156825 | 13:50 |
| *** fifieldt has quit IRC | 13:51 | |
| tdasilva | clayg: ^ removed the versioning dependency | 13:52 |
| *** tsufiev_ has joined #openstack-swift | 13:58 | |
| tsufiev_ | hello, there! | 13:58 |
| tsufiev_ | could anybody tell me whether it's possible to create hidden objects in Swift? | 13:58 |
| *** Gues_____ has joined #openstack-swift | 14:06 | |
| *** nellysmitt has joined #openstack-swift | 14:15 | |
| *** jrichli has joined #openstack-swift | 14:17 | |
| *** elmo has joined #openstack-swift | 14:19 | |
| *** SkyRocknRoll has quit IRC | 14:19 | |
| *** tsg_ has joined #openstack-swift | 14:30 | |
| *** tsg_ has quit IRC | 14:32 | |
| tdasilva | tsufiev_: hi, i'm not really sure what you mean by hidden object, can you describe your use case more? | 14:35 |
| *** Fin1te has joined #openstack-swift | 14:36 | |
| tsufiev_ | tdasilva, I'm trying to fix https://bugs.launchpad.net/horizon/+bug/1403129 So far the most logical solution seems to immediately stream the file from the Horizon UploadFileHandler class (Django's entity) to the swift, and once the whole image is uploaded to swift, pass it to glance | 14:38 |
| openstack | Launchpad bug 1403129 in OpenStack Dashboard (Horizon) "Glance with swift backend - upload of an image from horizon is not streaming" [High,In progress] - Assigned to Timur Sufiev (tsufiev-x) | 14:38 |
| tsufiev_ | tdasilva, currently I'm thinking of making this temporary file in swift invisible to other actors | 14:38 |
| tdasilva | tsufiev_: I'm probably missing something here (not very familiar with glance) but if you uploading an object to swift, I don't think there's a need to upload with a tmp name and then rename. In fact swift does not have a rename funcionality | 14:44 |
| tdasilva | tsufiev_: The object is not really available for download until you have completed uploading it to swift | 14:44 |
| tdasilva | tsufiev_: just FYI...I saw in the bug report that these images can be pretty large (~40 GB), so you would need to break them up and use slo or dlo | 14:46 |
| tdasilva | tsufiev_: the manifest file would be the last thing you upload, so that the whole image is not really available for download until the manifest file is there | 14:47 |
| tsufiev_ | tdasilva, it's not the period from starting the file upload to swift till its finish that bothers me, but the period from the object successfully uploaded to swift till the moment it's uploaded to glance | 14:48 |
| tsufiev_ | tdasilva, what if someone deletes it before it's fully uploaded to glance? | 14:48 |
| portante | tsufiev_: why is the object uploaded to swift and glance/ | 14:50 |
| portante | ? | 14:50 |
| tsufiev_ | portante, you mean why I need to upload it Swift during upload to Glance? | 14:51 |
| portante | yes, I suppose, I just don't know the workflow well, hence the question | 14:52 |
| *** osapryc has joined #openstack-swift | 14:53 | |
| tsufiev_ | portante, I need some storage place for the large image before it's passed to Glance. It could be either stored at web-server, which poses a problem because Controller node disk space could be limited (and that's why the bug was filed), or some other storage service. I'm trying the approach with Swift as this storage service | 14:54 |
| portante | I thought Glance can use swift as a backing store, so was wondering if by using that setup you'd avoid the double upload | 14:54 |
| portante | tsufiev_: ^^^ | 14:55 |
| tsufiev_ | portante, yes, the double upload will take place, it's unavoidable IMO | 14:55 |
| *** Bsony has joined #openstack-swift | 14:55 | |
| tsufiev_ | portante, the thing I'm trying to avoid is storing the image on web-server | 14:57 |
| *** Bsony_ has quit IRC | 14:57 | |
| portante | tsufiev_: where does your image originate from? | 14:57 |
| acoles | tsufiev_: i don;t want to confuse matters, but FYI swift recently added a feature to support services uploading objects that 'normal' users cannot delete unilaterally | 14:58 |
| acoles | tsufiev_: http://docs.openstack.org/developer/swift/overview_auth.html#openstack-service-using-composite-tokens | 14:58 |
| portante | created on the fly dynamically, or is it a file? | 14:58 |
| portante | tsufiev_: ^ | 14:58 |
| acoles | portante: hi! | 14:58 |
| portante | acoles: hello | 14:58 |
| portante | hope all is well with you and yours | 14:58 |
| acoles | portante: good thanks, and you? keeping busy? | 14:59 |
| *** reed has joined #openstack-swift | 14:59 | |
| portante | good, yes, very busy | 14:59 |
| portante | lots of D3/nvd3 work these days | 14:59 |
| tsufiev_ | tdasilva, portante, acoles: gentlemen, thanks for your help, I've just realized that I was making the task more difficult for myself :). Instead of dancing around Django guts I'll change the Horizon form for uploading images at client side, to first create an image, and pass the file contents with second request | 15:03 |
| tsufiev_ | this will allow to not upload the same file to Swift 2 times | 15:04 |
| *** Trozz has quit IRC | 15:10 | |
| *** Trozz has joined #openstack-swift | 15:11 | |
| portante | tsufiev_: great hear, hope that works out | 15:11 |
| *** Fin1te has quit IRC | 15:13 | |
| *** welldannit has joined #openstack-swift | 15:14 | |
| *** pcaruana has joined #openstack-swift | 15:19 | |
| *** Gues_____ has quit IRC | 15:30 | |
| *** nellysmitt has quit IRC | 15:32 | |
| *** Gues_____ has joined #openstack-swift | 15:35 | |
| clayg | morning! | 16:07 |
| *** mmcardle has joined #openstack-swift | 16:07 | |
| clayg | tdasilva: acoles: I'm sure you guys notied I didn't manage to get anything done yesterday :'( | 16:07 |
| clayg | acoles: I'm still working on the fi-hash-suffix stuff - I was agonizing over the base class Manager implemenation and the life cycle of hash_cleanup_listdir | 16:08 |
| acoles | clayg: morning! np i have been catching up on other stuff, like feature/crypto | 16:09 |
| clayg | tdasilva: did you see my note about maybe flipping the order of the PUT method extraction and versioned writes middleware patches on master? is that a bad idea? | 16:09 |
| clayg | acoles: ok - i'm really torn on the unregistered manager - the ReplicatedDiskFileManager or BaseDiskFileManager - whatever you want to call it | 16:09 |
| clayg | like i think it's a good idea long term - but i'm not sure how much it adds unless we're ready to refactor the base DiskFile to use it? | 16:10 |
| acoles | clayg: understand. it was less work for me to leave it there than squash all the 'new' into ECDiskFileManager. it would have been going backwards. | 16:11 |
| clayg | but since you wrote it - i guess you're thinking it's better to get ahead of it - or even that come next friday we'll be merge to master with the refactor | 16:11 |
| clayg | *ah* - you pulled some of that from the other extraction you were working on - got it | 16:11 |
| acoles | clayg: yup. and i hoped it would illustrated where this could be headed. but i understand that right now its extra baggage | 16:12 |
| acoles | clayg: thats the problem with forward thinking :) | 16:12 |
| *** Gues_____ has quit IRC | 16:12 | |
| clayg | acoles: ok - i'll continue to agonize over if we want it on the merge to master or after - sounds like you won't feel slighted if sam/notmyname/peluse/me sorta think one way or the other | 16:13 |
| clayg | acoles: you wanna vote? | 16:13 |
| peluse | clayg, where are you? | 16:13 |
| *** Bsony has quit IRC | 16:13 | |
| *** Gues_____ has joined #openstack-swift | 16:13 | |
| clayg | i have a 10am appointment - I can't even leave the city until after that - so I'll see you guys at lunch ;) | 16:13 |
| clayg | which is *technically* not much later than I get into the sf office on a normal day - sorry my for my schedule | 16:13 |
| peluse | ahhh, OK | 16:14 |
| clayg | it's the kids late start school (I lie to myself) | 16:14 |
| peluse | yeah, its spring break for us right now | 16:14 |
| clayg | ok, i gotta go | 16:14 |
| clayg | peluse: i'll see you later | 16:14 |
| acoles | clayg: i don't mind what goes to master, it *may* be helpful for anyone who wants to understand wtf its all about in mean time. idk | 16:15 |
| clayg | acoles: think about the bagage - sometimes having that extra stuff makes you feel like a boyscott if it turns out it's exactly what you need - sometimes it just weights you down | 16:15 |
| peluse | ya... so I didn't keep up on conversations yesterday but I do see a nice chain from acoles. Which way is the wind currently blowing wrt getting the functionlaity of MFI in there (acoles patch set, something else?) | 16:15 |
| clayg | yeah... the vision... I've been thinking about the end game too... | 16:15 |
| * peluse wonders if he has time to hit The Pink Poodle before the meeting.... | 16:16 | |
| clayg | peluse: it's there except for the fi-hash-suffix stuff - which *I* was supposed to port over yesterday - but am still going to do asap - then i'm going to see if i can help port the reconstructor branch over to the new interface | 16:16 |
| acoles | clayg: so do you need to me to take a crack at moving the hash_suffix pieces on top of per-policy or do you have it in hand? | 16:17 |
| clayg | peluse: if that looks like it's going to work we're gunna do it | 16:17 |
| acoles | peluse: ^^ thats the answer to your wuestion :) | 16:17 |
| acoles | s/w/q/ | 16:17 |
| clayg | acoles: i'm halfway done and not pushed - the hcl life cycle is really tricky | 16:17 |
| peluse | heh, OK. So acoles patch chain + pending refactor of hash_suffix? | 16:17 |
| clayg | acoles: honestly i like most of the code you write better than the code i write - so if you wanna take a stab at it knock yourself out - my experience yesterday was that it's going to be tricky | 16:18 |
| peluse | BTW, I do need the FI list per suffiex easily accessible but I do not need the entire 2nd dict that includes the hashes | 16:18 |
| clayg | i really gotta go - but peluse and i will talk about it highbandwidth later | 16:18 |
| peluse | so that 2nd dict can simply be a list instead | 16:18 |
| peluse | OK, later | 16:18 |
| clayg | acoles: my gut says don't sweat the fi-suffix-hash until peluse and I can talk about it | 16:19 |
| acoles | clayg: okay. i may get on the phone later too. not sure | 16:19 |
| acoles | clayg: my gut agrees | 16:19 |
| peluse | acoles, I can get on the horn in about 20 min if you want to do some pre-talk and and least get you and I on the same page before clayg shows up? | 16:19 |
| acoles | peluse: sure, using the bridge or direct? | 16:20 |
| peluse | and BTW anyone else out there in swift land is welcome to call in too... 671642307 (passcode) 1(916)356-2663 (or your local bridge access #) Choose bridge 5. | 16:20 |
| acoles | peluse: so where are you? | 16:20 |
| peluse | acoles, I'm at Intel in Santa Clara waiting on a conference room | 16:20 |
| peluse | maybe 15 more min | 16:21 |
| peluse | notmyname, is on his way, probably be here in 30-40 min. torgomatic is probably just waking up :) | 16:21 |
| *** wer has quit IRC | 16:21 | |
| acoles | peluse: ping here when you're ready to conf | 16:21 |
| peluse | will do, going to walk up to room now and kick some ass if someone is squating! | 16:22 |
| acoles | cschwede: ^^^ may be interested in getting some context | 16:22 |
| cschwede | acoles: yeah, thanks, definetely! | 16:23 |
| cschwede | peluse: is it possible to join by some webinterface? | 16:23 |
| cschwede | nevermind, found a link (trying) | 16:24 |
| peluse | OK, I'll start it up now | 16:27 |
| peluse | OK, I'm on the call now. Here's a link that might work: https://meet.intel.com/paul.e.luse/SDBQH0G1 | 16:30 |
| *** Fin1te has joined #openstack-swift | 16:31 | |
| peluse | brb | 16:31 |
| acoles | calling | 16:33 |
| cschwede | i joined the meeting, but need some kind of ack? | 16:34 |
| peluse | cschwede, hmmm, let me see if I need to also join on my laptop | 16:35 |
| acoles | i somehow joined a lync meeting and can see peluse and cschwede as participants but no audio | 16:37 |
| acoles | i've no idea how i did that! | 16:37 |
| cschwede | acoles: lower left corner, i had to click one of the buttons | 16:37 |
| peluse | yeah, just start clicking buttons :) | 16:38 |
| peluse | acoles, I see you on the call and it doesnt show you muted or anything | 16:38 |
| *** Bsony has joined #openstack-swift | 16:42 | |
| *** gyee has joined #openstack-swift | 16:43 | |
| *** Gues_____ has quit IRC | 16:45 | |
| *** wer has joined #openstack-swift | 16:49 | |
| *** openstackgerrit has quit IRC | 16:59 | |
| *** openstackgerrit has joined #openstack-swift | 16:59 | |
| cschwede | peluse: we lost you | 17:00 |
| cschwede | you’re still shown on the call, but we can’t hear you anymore | 17:01 |
| peluse | hmm | 17:02 |
| peluse | can you guys hear each other? | 17:02 |
| cschwede | yes | 17:02 |
| peluse | nice :) | 17:02 |
| peluse | calling in again | 17:02 |
| *** jistr has quit IRC | 17:08 | |
| *** dmorita has joined #openstack-swift | 17:08 | |
| *** Fin1te has quit IRC | 17:11 | |
| tdasilva | clayg: hey! so for now I removed the dependency in the refactor and just rebased it against master. I thought it was a good idea. | 17:14 |
| *** jrichli_ has joined #openstack-swift | 17:14 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: FakeFooters middleware https://review.openstack.org/165517 | 17:16 |
| *** jrichli has quit IRC | 17:16 | |
| acoles | jrichli: ^^ | 17:16 |
| *** dmorita_ has joined #openstack-swift | 17:25 | |
| *** dmorita has quit IRC | 17:25 | |
| *** rdaly2 has quit IRC | 17:29 | |
| *** bkopilov has joined #openstack-swift | 17:30 | |
| *** vinsh has quit IRC | 17:36 | |
| *** setmason has joined #openstack-swift | 17:38 | |
| setmason | Can I present to tenants two different backends (radosGW) and Openstack Swift? | 17:39 |
| openstackgerrit | Richard Hawkins proposed openstack/swift: Add additional func tests for TempURLs https://review.openstack.org/155985 | 17:41 |
| *** Gue______ has joined #openstack-swift | 17:41 | |
| *** nellysmitt has joined #openstack-swift | 17:44 | |
| *** jordan_ has quit IRC | 17:48 | |
| *** vinsh has joined #openstack-swift | 17:55 | |
| *** vinsh has joined #openstack-swift | 17:56 | |
| *** dencaval has quit IRC | 18:04 | |
| openstackgerrit | Lorcan Browne proposed openstack/swift: Add swift-recon feature to track swift-drive-audit error count https://review.openstack.org/163889 | 18:05 |
| notmyname | jrichli_: are you still looking at the functional tests being able to run with an arbitrary policy? | 18:08 |
| jrichli_ | just got back from lunch | 18:17 |
| acoles | jrichli_: the right answer is 'i'll trade for reviews on feature/crypto' ;) | 18:17 |
| jrichli_ | notmyname: I haven't gotten very far. but - I think I can do some more work on that if somebody could approve Alistair's reviews on crypto branch :-) | 18:18 |
| acoles | lol | 18:18 |
| jrichli_ | notmyname: I would still like to work on it, but will need some help. I know that env vars were a suggestion | 18:19 |
| jrichli_ | I can still experiment on my own, but it will take longer that way. any info that somebody could pass my way would help | 18:21 |
| notmyname | jrichli_: one question is how many functional tests use the swift client vs how many don't | 18:22 |
| jrichli_ | right, which I know it was stated that not all of them do | 18:22 |
| *** CaioBrentano has joined #openstack-swift | 18:23 | |
| jrichli_ | you want to know the number? perhaps to see if those that dont could be changed? | 18:23 |
| acoles | jrichli_: options i have heard are : monkey patch swiftclient and test_client to force a header in every container PUT | 18:23 |
| acoles | and/or add a feature to swiftclient to set a default header from env var (doesn't cover all tests) and/or change all tests to use swiftclient | 18:24 |
| jrichli_ | right. that would be a way for the tests using swiftclient. but then we would still need another approach for the others anyway, right? | 18:24 |
| jrichli_ | ah | 18:25 |
| jrichli_ | ok, I will take a look with those options in mind. thx! | 18:29 |
| *** geaaru has quit IRC | 18:29 | |
| *** rdaly2 has joined #openstack-swift | 18:32 | |
| notmyname | acoles: https://review.openstack.org/#/c/162717/ | 18:33 |
| *** thumpba has joined #openstack-swift | 18:33 | |
| *** CaioBrentano has quit IRC | 18:34 | |
| acoles | jrichli_: i'll try to take a look at func tests tomorrow and see if i can give you any ideas | 18:41 |
| jrichli_ | acoles: thanks. I will be getting familiar with things today and learning about the use of swiftclient and the container class | 18:43 |
| *** CaioBrentano has joined #openstack-swift | 18:44 | |
| *** CaioBrentano has quit IRC | 18:44 | |
| *** shri has joined #openstack-swift | 18:47 | |
| *** mmcardle has quit IRC | 18:48 | |
| shri | Hi.. question about rings. I created rings using part_power of 8. Expected partitions is (2 ^ 8) = 256. However, when I run swift-ring-builder /etc/swift/object.builder, it shows that partitions as 64. What gives? | 18:48 |
| shri | oh.. partitions per device is 64.. I see! | 18:49 |
| shri | sorry for the confusion | 18:49 |
| shri | I have 8 disks | 18:49 |
| notmyname | swift team meeting in #openstack-meeting in 10 minutes | 18:50 |
| *** mahatic has quit IRC | 18:50 | |
| glange | shri: I think you want a bigger part_power than that | 18:50 |
| MooingLemur | yeah, I was gonna say that. If you expect this thing to grow beyond a tiny test instance, you'll want enough partitions to reasonably balance your data across dozens or hundreds of disks. | 18:51 |
| *** mahatic has joined #openstack-swift | 18:52 | |
| glange | https://answers.launchpad.net/swift/+question/211929 | 18:52 |
| glange | that/sam suggests at least 100 partitions per disk | 18:52 |
| glange | but you need to take into account possible growth too | 18:52 |
| MooingLemur | 100 partitions per disk at the point where you will end up growing :) | 18:52 |
| glange | yeah | 18:52 |
| shri | glange, MooingLemur… yup! Just trying out some config right now. | 18:52 |
| *** cutforth has joined #openstack-swift | 18:52 | |
| glange | oh, ok | 18:52 |
| *** kota_ has joined #openstack-swift | 18:53 | |
| mattoliverau | morning | 18:59 |
| kota_ | mattoliverau: morning | 19:00 |
| mattoliverau | kota_: morning to you too :) | 19:00 |
| notmyname | meeting time | 19:00 |
| *** mahatic has quit IRC | 19:01 | |
| notmyname | in #openstack-swift | 19:01 |
| *** mahatic has joined #openstack-swift | 19:03 | |
| *** nellysmitt has quit IRC | 19:12 | |
| *** jrichli__ has joined #openstack-swift | 19:13 | |
| *** silor has joined #openstack-swift | 19:15 | |
| *** jrichli_ has quit IRC | 19:16 | |
| glange | will encryption work with EC when both are done? just curious | 19:23 |
| torgomatic | if not, then we screwed up bad somewhere | 19:24 |
| peluse | it will encrypt, just not decrypt :) | 19:24 |
| glange | hah | 19:25 |
| acoles | glange: wot torgomatic said | 19:25 |
| acoles | yup, the beta will be PUT path only :) | 19:25 |
| notmyname | lol | 19:25 |
| glange | ok, because encryption will sort of just be a middleware filter on the stream of data going in and out? | 19:25 |
| notmyname | patch 156825 | 19:25 |
| patchbot | notmyname: https://review.openstack.org/#/c/156825/ | 19:25 |
| torgomatic | that's the idea, yeah | 19:25 |
| acoles | glange: mostly middleware. | 19:26 |
| glange | maybe not filter, but data transform or something like that | 19:26 |
| acoles | glange: gory detail is here https://review.openstack.org/#/c/154318/ | 19:28 |
| glange | acoles: yeah, I read through that earlier, I was just sort of confirming what I thought by asking that question | 19:29 |
| glange | it'll be interesting to see how much extra cpu an encrypted request will require on the proxy | 19:30 |
| notmyname | glange: and now you know why peluse likes swift so much ;-) | 19:33 |
| *** cutforth has quit IRC | 19:33 | |
| *** kota_ has quit IRC | 19:33 | |
| torgomatic | glange: yes, definitely it will... the hope is that newer Intel CPUs have magic instructions that make it not take a whole lot of work | 19:33 |
| torgomatic | or rather, it does take the work, but there's a pile of transistors specialized to do the work, so it's fast | 19:34 |
| notmyname | now we need someone from AMD to start contributing to swift ;-) | 19:34 |
| torgomatic | yep, gotta support their specialty instructions too | 19:34 |
| torgomatic | AES-NIH :) | 19:34 |
| *** aswadr has quit IRC | 19:34 | |
| mattoliverau | lol | 19:35 |
| tdasilva | acoles: now it might be peluse who is kicking notmyname :P | 19:36 |
| acoles | tdasilva: hehe | 19:37 |
| *** jrichli__ has quit IRC | 19:38 | |
| *** jrichli has joined #openstack-swift | 19:39 | |
| *** classicsnail has quit IRC | 19:39 | |
| *** classics1 has joined #openstack-swift | 19:39 | |
| *** erlon has quit IRC | 19:39 | |
| *** zacksh has quit IRC | 19:39 | |
| *** zacksh has joined #openstack-swift | 19:40 | |
| mattoliverau | Well I might go find some breakfast. | 19:40 |
| acoles | mattoliverau: you want cookies? | 19:41 |
| *** erlon has joined #openstack-swift | 19:41 | |
| mattoliverau | acoles: look at me... of course I do :P | 19:41 |
| jrichli | mattoliverau: stay away from the cookies with the strange "american" sugar on it | 19:41 |
| *** gyee has quit IRC | 19:41 | |
| mattoliverau | jrichli: lol, yeah, crazy americans and putting salt on everything :P | 19:41 |
| *** lpabon has joined #openstack-swift | 19:42 | |
| *** acoles is now known as acoles_away | 19:44 | |
| glange | torgomatic: we are putting magic in swift now, nice! :) | 19:46 |
| mattoliverau | Swift: Storing your objects with a little bit of magic! | 19:48 |
| *** silor has quit IRC | 19:51 | |
| *** tab____ has joined #openstack-swift | 19:54 | |
| *** CaioBrentano has joined #openstack-swift | 19:55 | |
| *** devlaps has joined #openstack-swift | 19:55 | |
| *** zigo__ has joined #openstack-swift | 19:56 | |
| *** zigo has quit IRC | 19:56 | |
| *** Gue______ has quit IRC | 19:59 | |
| *** mahatic has quit IRC | 20:06 | |
| MooingLemur | what's the config option and default for expiry of .ts files? | 20:06 |
| *** CaioBrentano has quit IRC | 20:08 | |
| MooingLemur | or does replication take care of it? | 20:12 |
| notmyname | MooingLemur: ya, replication cleans that up. | 20:13 |
| MooingLemur | sweet, thanks | 20:14 |
| notmyname | relaim_time IIRC. or something similar | 20:14 |
| notmyname | reclaim_time* | 20:14 |
| *** openstackgerrit has quit IRC | 20:14 | |
| *** openstackgerrit has joined #openstack-swift | 20:15 | |
| MooingLemur | reclaim_age looks like 1w default.. works for me :) | 20:18 |
| notmyname | MooingLemur: ya, that sounds right | 20:19 |
| *** zhill has joined #openstack-swift | 20:20 | |
| *** Fin1te has joined #openstack-swift | 20:29 | |
| *** Fin1te has quit IRC | 20:30 | |
| *** shri1 has joined #openstack-swift | 20:32 | |
| *** shri1 has quit IRC | 20:32 | |
| *** rdaly2 has quit IRC | 20:32 | |
| *** shri has quit IRC | 20:32 | |
| *** Bsony has quit IRC | 20:34 | |
| peluse | anyone know how to override a .conf setting in a probe test when firing up a server with the Manager? | 20:37 |
| *** Bsony has joined #openstack-swift | 20:38 | |
| peluse | OK, think I just found it | 20:38 |
| *** mmcardle has joined #openstack-swift | 20:45 | |
| *** Bsony has quit IRC | 21:06 | |
| *** Guest__ has joined #openstack-swift | 21:25 | |
| zaitcev | peluse: years from now someone is going to see this IRC log and this will happen: http://xkcd.com/979/ | 21:27 |
| *** Guest__ has quit IRC | 21:29 | |
| *** jamielennox is now known as jamielennox|away | 21:31 | |
| *** devlaps has quit IRC | 21:43 | |
| *** jamielennox|away is now known as jamielennox | 22:12 | |
| *** Bsony has joined #openstack-swift | 22:13 | |
| *** mmcardle has quit IRC | 22:16 | |
| *** Bsony has quit IRC | 22:18 | |
| *** tab____ has quit IRC | 22:21 | |
| *** welldannit has quit IRC | 22:22 | |
| *** osapryc has quit IRC | 22:33 | |
| *** jrichli has quit IRC | 22:47 | |
| *** gyee has joined #openstack-swift | 22:48 | |
| *** echevemaster has joined #openstack-swift | 22:54 | |
| *** setmason has quit IRC | 23:03 | |
| *** joeljwright has quit IRC | 23:28 | |
| notmyname | for people not looking at feature/ec yet, please look at getting this patch reviewed and landed on master will (1) make today's code better and (2) means that it's one less thing to consider when doing the feature/ec merge | 23:36 |
| notmyname | getting this patch reviewed and landed on master will (1) make today's code better and (2) means that it's one less thing to consider when doing the feature/ec merge | 23:36 |
| notmyname | bah | 23:36 |
| notmyname | for people not looking at feature/ec yet, please look at https://review.openstack.org/#/c/156825/ | 23:37 |
| notmyname | there | 23:37 |
| notmyname | the refactoring there is (or will be) in the feature/ec branch. so landing it on master in the next 7 days means that it (1) makes the feature/ec merge easier and (2) will more clearly show where the feature/ec code does and does not touch existing replication code | 23:38 |
| notmyname | dfg: glange: redbo: cschwede: mattoliverau: zaitcev: ^^ | 23:39 |
| mattoliverau | notmyname: having pasting issues :P | 23:40 |
| notmyname | mattoliverau: yup | 23:40 |
| notmyname | point is, look at that patch. it will make life easier for people working on feature/ec, and, more importantly, it will make your life easier :-) | 23:41 |
| *** thumpba has quit IRC | 23:42 | |
| zaitcev | notmyname: okay | 23:43 |
| notmyname | thanks :-) | 23:43 |
| mattoliverau | notmyname: already reviewed it yesterday, and plan to re-review today now that a new patchset is up :) Ahh good it's not dependant on obj_versioning anymore, tho that will be another clean up of the PUT path. | 23:44 |
| notmyname | great | 23:44 |
| notmyname | right, that's the key | 23:44 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!