Monday, 2017-06-19

*** tovin07_ has joined #openstack-swift00:54
*** tovin07 has joined #openstack-swift00:56
*** cshastri has joined #openstack-swift01:10
*** tovin07 has quit IRC01:27
*** tanner_ has joined #openstack-swift01:28
tanner_I want to insert this code into swift01:28
tanner_<script src="typed.js"></script> <script>   document.addEventListener('DOMContentLoaded', function(){'.element', {         strings: ["First sentence.", "Second sentence."],         typeSpeed: 0       });   }); </script> ...  <div class="element"></div>01:29
kota_good morning01:30
*** tanner_ has quit IRC01:43
*** jerrygb has joined #openstack-swift02:33
*** jerrygb has quit IRC02:35
*** jerrygb has joined #openstack-swift02:46
openstackgerritHieu LE proposed openstack/swift master: OSprofiler in OpenStack Swift
*** gkadam has joined #openstack-swift03:02
*** tovin07_ has quit IRC03:15
*** SkyRocknRoll has joined #openstack-swift03:17
*** tovin07_ has joined #openstack-swift03:23
*** jerrygb has quit IRC03:33
*** SkyRocknRoll has quit IRC03:34
*** klrmn has joined #openstack-swift03:39
*** jerrygb has joined #openstack-swift03:54
*** mingyu has joined #openstack-swift03:56
*** jerrygb has quit IRC03:59
*** gkadam has quit IRC04:01
*** mingyu has quit IRC04:02
*** jerrygb has joined #openstack-swift04:03
*** gkadam has joined #openstack-swift04:03
*** gkadam has quit IRC04:06
*** gkadam has joined #openstack-swift04:06
mahaticgood morning04:08
*** gkadam has quit IRC04:09
*** gkadam has joined #openstack-swift04:10
*** klrmn has quit IRC04:17
*** gkadam_ has joined #openstack-swift04:23
*** gkadam has quit IRC04:26
*** yuan621 has joined #openstack-swift04:27
yuan621I'm trying install swift + keystone v3 in devstack, is there any documents for this?04:29
*** winggundamth has joined #openstack-swift04:31
*** jerrygb has quit IRC04:44
*** yuan621 has quit IRC04:58
*** qwertyco has joined #openstack-swift05:00
*** jerrygb has joined #openstack-swift05:06
*** pcaruana has joined #openstack-swift05:14
mahaticyuan621: might help05:14
*** jerrygb has quit IRC05:21
*** jerrygb has joined #openstack-swift05:32
*** ChubYann has quit IRC05:39
*** tonanhngo has joined #openstack-swift05:53
*** tonanhngo has quit IRC05:54
*** mingyu has joined #openstack-swift06:01
*** chsc has joined #openstack-swift06:18
*** chsc has joined #openstack-swift06:18
*** rcernin has joined #openstack-swift06:19
*** winggundamth has quit IRC06:20
*** chsc has quit IRC06:22
*** jerrygb has quit IRC06:33
*** skudlik has joined #openstack-swift06:42
*** hseipp has joined #openstack-swift06:43
*** gkadam_ is now known as gkadam06:44
*** silor has joined #openstack-swift06:54
*** tesseract has joined #openstack-swift07:05
openstackgerritKazuhiro MIYAHARA proposed openstack/swift master: [Don't Merge]: Automated Tiering Discssion Organization Patch
*** winggundamth has joined #openstack-swift07:21
*** jaosorior has joined #openstack-swift07:26
*** Dinesh_Bhor has joined #openstack-swift07:31
acolesgood morning07:32
*** oshritf has joined #openstack-swift07:32
*** jerrygb has joined #openstack-swift07:33
*** rcernin has quit IRC07:34
*** rcernin has joined #openstack-swift07:37
*** rcernin has quit IRC07:37
*** rcernin has joined #openstack-swift07:37
*** jerrygb has quit IRC07:38
*** openstackgerrit has quit IRC07:48
mattoliverauacoles: morning07:54
*** jerrygb has joined #openstack-swift08:02
*** oshritf has quit IRC08:04
*** jerrygb has quit IRC08:06
*** openstackgerrit has joined #openstack-swift08:09
openstackgerritVan Hung Pham proposed openstack/swift master: Stop using deprecated 'message' attribute in Exception
*** hseipp has quit IRC08:20
mahaticacoles: good morning08:20
mahaticmattoliverau: o/08:21
acolesmahatic: mattoliverau: o/08:21 proposed openstack/swift master: Make a hacking way for OpenStack changes
*** qwertyco has quit IRC08:23
*** qwertyco has joined #openstack-swift08:24
*** cbartz has joined #openstack-swift08:34
cbartztimburke: tdasilva: I've uploaded a new patchset for patch 414232 . Could you pls review it again? Thx.08:36
patchbot - swift - Exclude containers for account quota08:36
*** psachin has joined #openstack-swift08:41
*** mingyu has quit IRC08:43
*** ujjain has quit IRC08:59
*** oshritf has joined #openstack-swift09:00
*** ujjain has joined #openstack-swift09:00
*** ujjain has joined #openstack-swift09:00
*** mingyu has joined #openstack-swift09:03
*** ujjain has quit IRC09:03
*** ujjain has joined #openstack-swift09:11
*** ujjain has joined #openstack-swift09:11
*** jerrygb has joined #openstack-swift09:19
*** mingyu_ has joined #openstack-swift09:19
*** jerrygb has quit IRC09:23
*** mingyu has quit IRC09:23
*** tanee has quit IRC09:24
*** tanee has joined #openstack-swift09:25
*** saltsa has quit IRC09:26
*** saltsa has joined #openstack-swift09:27
*** geaaru has joined #openstack-swift09:39
*** joeljwright has joined #openstack-swift09:42
*** ChanServ sets mode: +v joeljwright09:42
*** mingyu_ has quit IRC09:44
*** mingyu has joined #openstack-swift09:47 proposed openstack/swift master: Make a hacking way for OpenStack changes
*** tovin07_ has quit IRC10:02
*** cschwede has quit IRC10:02
*** cschwede has joined #openstack-swift10:07
*** cshastri has quit IRC10:12
*** kei_yama has quit IRC10:17
*** mingyu has quit IRC10:19
*** skudlik has quit IRC10:28
*** gkadam has quit IRC10:28
*** gkadam has joined #openstack-swift10:30
*** qwertyco has quit IRC10:56
*** qwertyco has joined #openstack-swift10:57
*** winggundamth has quit IRC11:07
*** jerrygb has joined #openstack-swift11:20
*** jerrygb has quit IRC11:25
*** hseipp has joined #openstack-swift11:30
*** links has joined #openstack-swift11:40
*** links has quit IRC11:41
*** skudlik has joined #openstack-swift11:51
*** bob_cheesey has left #openstack-swift11:52
*** ujjain has quit IRC12:00
*** ujjain has joined #openstack-swift12:10
*** ujjain has joined #openstack-swift12:10
*** spiette has quit IRC12:12
*** MVenesio has joined #openstack-swift12:16
*** jerrygb has joined #openstack-swift12:29
*** gkadam has quit IRC12:31
*** jerrygb has quit IRC12:33 proposed openstack/swift master: Make a hacking way for OpenStack changes
*** jerrygb has joined #openstack-swift12:47
*** lucasxu has joined #openstack-swift12:59
*** catintheroof has joined #openstack-swift13:09
*** catintheroof has quit IRC13:14
*** catintheroof has joined #openstack-swift13:14
*** klamath has joined #openstack-swift13:31
*** klamath has quit IRC13:31
*** klamath has joined #openstack-swift13:31
openstackgerritAlistair Coles proposed openstack/swift master: Experimental swift-ring-composer CLI to build composite rings
*** hseipp has quit IRC13:56
*** hseipp has joined #openstack-swift13:57
*** qwertyco has quit IRC13:57
*** hseipp has quit IRC14:00
*** hseipp has joined #openstack-swift14:00
*** aselius has joined #openstack-swift14:01
*** mingyu has joined #openstack-swift14:01
*** klamath has quit IRC14:22
*** klamath has joined #openstack-swift14:23
*** psachin has quit IRC14:36
*** mingyu_ has joined #openstack-swift14:40
*** mingyu has quit IRC14:42
*** skudlik has quit IRC14:48
*** mvk has quit IRC14:53
*** cnf has quit IRC14:53
*** mingyu_ has quit IRC14:54
*** qwertyco has joined #openstack-swift14:56
*** cnf has joined #openstack-swift14:58
*** klrmn has joined #openstack-swift15:00
*** chlong has joined #openstack-swift15:03
*** catintheroof has quit IRC15:07
*** catintheroof has joined #openstack-swift15:08
timburkegood morning15:12
*** oshritf has quit IRC15:12
tdasilvatimburke: o/15:25
acolestimburke: tdasilva o/15:25
notmynamegood morning15:27
mahaticnotmyname: timburke: tdasilva: good morning15:29
tdasilvamahatic, notmyname o/15:29
*** rcernin has quit IRC15:30
*** pcaruana has quit IRC15:33
notmynamejoeljwright: still around today?15:48
notmynamejoeljwright: on friday I was looking at because you asked for API design review15:50
patchbotpatch 365371 - swift - Add Preamble and Postamble to SLO and SegmentedIte...15:50
notmynamejoeljwright: I'm really interested in your plans for "TLO". I can imagine a few ways it might work15:51
notmynamewill it be something the client makes (a special "spelling" of an SLO) that just happens to end up being a tarfile on read?15:51
joeljwrightnotmyname: I'm still here, but on the phone atm15:51
notmynamewill it be something that the client sets on upload that the server side translates into the special spelling of SLO?15:52
notmynamewill it be something the client asks for on read that takes a set of data and happens to internally use the new SLO attributes to send to SegmentedIterable (Therefore a "TLO" is only a logical thing and not an acutall stored thing)15:53
joeljwrightmy plan is to provide a second middleware that translates a TLO request with manifest containing size/tar_name into an SLO request with pre/post ambles15:53
notmynameon read or write?15:53
joeljwrightso it creates the appropriate SLO manifest on TLO manifest upload15:54
notmynamewhich then on read is an SLO that happens to be formatted in the response body as a valid tarfile?15:54
joeljwrightcan you think of a better/more elegant way to handle this?15:55
notmynamewere you considering the existing bulk upload functionality to create this? since it already reads tarfiles on PUT15:55
joeljwrightI hadn't considered bulk upload, but it's certainly an interesting option as an addition15:58
*** cbartz has quit IRC15:58
notmynameI'd imagine that the TLO functionality is interesting iff the individual components are accessed too15:58
notmynamethat is, if the user only wants a tarfile, then just upload the tarfile15:59
notmynamethinking out loud follows.... if the user has discrete objects they want to download in one request as a tarfile, then the trick is telling the server *which* objects they want16:03
notmynamecould be done as a listing response, but that has eventual consistency and ordering issues (ie the requested objects would have to be named some contiguous list)16:03
notmynamecould be a "normal" SLO, but that means a "normal" request to it would not return a valid tarfile and would need some other flag on request like ?as_tar for "return this but insert the tarball bits in the right places"16:04
notmynameon the other hand, the "tarball bits" includes filesystem metadata (eg owner, permissions), so that might be best generated dynamically at request time based on the requestor?16:05
*** gyee has joined #openstack-swift16:06
notmynamebut my assumption here on the use case is that there will be one TLO creator and many readers. that is, the writer knows that there are several discrete sets of tarballs to be downloaded given the total set of available objects16:08
joeljwrightnotmyname: sorry I had to disappear for a few minutes16:08
notmynameotherwise, it seems a *lot* simpler to simply upload the objects individually and also upload a tarball16:09
notmynamejoeljwright: no worries. allowed be to rant a little ;-)16:09
joeljwrightjust catching up16:09
*** pcaruana has joined #openstack-swift16:11
joeljwrightso, tar is not my only potential use case for the pre/post amble16:12
notmynameit seems like an important property is that the vast majority of clients won't need to be changed in order to consume TLOs16:12
notmynamejoeljwright: good! I was getting to that point :-)16:12
joeljwrightand I'm interested in making tar files downloadable using a tempurl for people without swift accounts16:12
joeljwrightwe're also looking into the possibility of building video/audio containers using the raw data + SLO pre/post ambles16:13
joeljwrightalthough I have to admin this is little more than an idea atm16:13
joeljwrightbeing able to download many objects from swift and maintain/create an appropriate file structure is also important to my ideas16:14
joeljwrightthe SLO style creation of a tarball means we can control not only the permissions, but how the data appears to the end user when downloaded16:15
joeljwright(it's not just a reflection of how the data is stored in swift)16:15
notmynameyeah the pre/post amble idea is a very generic and therefore powerful idea. means that any format that has discrete parts with small envelope data around those parts (or separators) can be done16:15
notmynameI don't know enough about streaming video to know if it fits there or not16:16
joeljwrightaside from zip being a PITA I'd agree ;)16:16
notmynameso you're going to have a bunch of discrete objects and you want to create a set of "TLOs" that are all different bundlings of those objects? with different metadata too?16:17
joeljwrightit would be possible to be MORE generic16:17
joeljwrightand just allow segments specified as data in the manifest, but I didn't want to just allow base64 encoded data in the manifest!16:18
*** klrmn has quit IRC16:19
*** dja has quit IRC16:19
notmynamewhy not?16:19
notmyname(said with a smile, but serious question)16:19
joeljwrightyes, multiple TLOs referencing the same data, but providing a different structure is also something I'd like to be able to do16:20
*** dja has joined #openstack-swift16:20
joeljwrightre. just base64 encoded data in manifests, I was trying to at least force the use of objects when dealing with an object store ;)16:20
joeljwrightit was an API for added structure that would negatively impact performance16:21
joeljwrightrather than an alternative way to store data...16:21
joeljwrightone reason this exists is that the download performance of tarballs as SLOs isn't great when your objects are not large enough16:22
notmynameright. because SLO rate limits small files and it's only a single stream16:22
joeljwrightadding a 1k object and a ~256byte object either side of your data in an SLO in a pain16:22
notmynameah, right. because the overhead of fetching additional small objects in an SLO is relatively high16:23
joeljwrightyeah, that and the rate limiting combine to kill performance16:23
notmynamethat makes sense. if the objects will also be accessed individually16:23
joeljwrightthere';s no reason we couldn't allow data only segments as a future update16:24
*** hseipp has quit IRC16:24
notmyname"good sense" may be a reason we don't do that in the future ;-)16:24
notmynameso I end up with 1000 cat pics that users access individually. and I *also* create several SLO-with-ambles so that different subsets can be downloaded as tarballs16:25
joeljwrightyes :)16:26
notmynameis that how I describe this to (non-sohonet) users?16:26
notmynameI like the power that *LOs-as-mutators provides. eg cat'ing objects (as exists today), providing structure (as proposed with TLO), providing bit torrent seeds, providing compression, even potentially transcoding, etc16:28
notmynamethat all sounds *really* complicated16:28
joeljwrightsorry, another phone call16:28
notmynameI can still type :-)16:28
notmynameso I wonder if the pre/post amble is something we want to support long term. is it the simplest way to provide tarfiles? simplest might be to tell users to upload a tarball, too.16:30
notmynameif the goal is just tarballs, is this the right way?16:30
notmynameif the goal is more than tarballs but just structured formats, is this the right way?16:30
joeljwrightthe problem with 'just upload tarballs' is it doesn't really help you if the data is already in swift :)16:30
notmynameif the goal is more than structured formats, but includes actual manipulation of the data, is this the right way?16:30
notmynameso the base question is "what is the goal we're trying to support?"16:32
joeljwrightI was certainly aiming at 'the data is in swift, provide different structured presentations of it'16:32
notmynameI think that's very interesting (as in, I like the sound of it, and I'm really interested in what it would take to see it done)16:34
notmynameand so far, the structured formats you need are (1) tar (2) there is no 216:34
notmynameI'm a little worried about inventing something to solve a bunch of use cases that we make up :-)16:35
joeljwrighttar was the proof of concept and most immediate need16:35
joeljwrightI was sure torgomatic had another use case for pre/postambles,16:38
notmynamehe's got use cases for those being passed in to SegmentedIterable16:38
joeljwrightaha, but not necessarily exposed in the API16:38
joeljwrightmaybe I need to prove that making an MKV from raw video/audio data is possible…16:40
notmynamethat would be awesome (even independently of the TLO discussion)16:41
notmynameI want to be clear, since we're doing this over text. I'm not opposed to your patch. just trying to understand it and what the implications are16:41
joeljwrightIt was just an idea, I need to learn a lot more about video container formats now!16:41
joeljwrightyeah, sure16:41
joeljwrightI understand16:41
joeljwrightmaking this part of the API means we have to support it16:41
*** ChubYann has joined #openstack-swift16:43
notmynamethere's such a huge difference between "this patch is good enough to land" and "this patch should land" :-)16:43
*** saint_ has joined #openstack-swift16:45
*** caiobrentano has joined #openstack-swift16:46
joeljwrightyou'll get no argument from me, but I genuinely believe that it's a useful and performant way to offer SLO as tar without huge negative performance impacts16:46
joeljwrightif nothing else16:46
joeljwrightno need to upload loads of nasty tiny tar header/footer objects16:47
notmynameif this pre/post amble method allows for more data encapsulation than just tar (more things that people will use), then many of my questions go away16:48
notmynamebut if it's just tar, then it's a question of if this is the best way to support just tar16:49
joeljwrightthanks for explaining your thoughts16:50
joeljwrighteven for tar though, I still think the pre/postamble approach is useful16:51
joeljwrighteven if we don't expose it through the API16:51
*** geaaru has quit IRC16:51
joeljwrightand bake tar functionality into SLO16:51
joeljwrightright, I need to run, I will try to order my thoughts some more before Wednesday evening :)16:52
notmynameme too :-)16:52
notmynamehave a good night16:52
joeljwrightyou too16:52
*** chlong has quit IRC16:56
*** tesseract has quit IRC16:59
*** lucasxu has quit IRC17:00
*** geaaru has joined #openstack-swift17:04
*** joeljwright has quit IRC17:10
*** chlong has joined #openstack-swift17:11
*** klrmn has joined #openstack-swift17:17
*** rcernin has joined #openstack-swift17:18
*** lucasxu has joined #openstack-swift17:33
*** zaitcev has joined #openstack-swift17:36
*** ChanServ sets mode: +v zaitcev17:36
*** tonanhngo has joined #openstack-swift17:50
*** cschwede has quit IRC18:01
*** vint_bra has joined #openstack-swift18:02
*** cschwede has joined #openstack-swift18:02
*** geaaru has quit IRC18:03
*** zigo has quit IRC18:05
*** zigo has joined #openstack-swift18:09
*** klrmn1 has joined #openstack-swift18:11
*** cargonza_ has joined #openstack-swift18:12
*** dosaboy_ has joined #openstack-swift18:15
*** kencjohnston_ has joined #openstack-swift18:18
*** torgomatic_ has joined #openstack-swift18:18
*** klrmn has quit IRC18:19
*** dja has quit IRC18:19
*** wasmum has quit IRC18:19
*** kencjohnston has quit IRC18:19
*** cargonza has quit IRC18:19
*** torgomatic has quit IRC18:19
*** jamielennox has quit IRC18:19
*** d0ugal has quit IRC18:19
*** dosaboy has quit IRC18:19
*** alecuyer has quit IRC18:19
*** honga has quit IRC18:19
*** fungi has quit IRC18:19
*** cargonza_ is now known as cargonza18:19
*** alecuyer has joined #openstack-swift18:21
*** jamielennox has joined #openstack-swift18:21
*** hseipp has joined #openstack-swift18:22
*** dja has joined #openstack-swift18:22
*** wasmum has joined #openstack-swift18:22
*** ujjain has quit IRC18:23
*** onovy has quit IRC18:23
*** onovy has joined #openstack-swift18:24
*** honga has joined #openstack-swift18:24
*** ujjain has joined #openstack-swift18:25
*** ujjain has quit IRC18:25
*** ujjain has joined #openstack-swift18:25
*** d0ugal has joined #openstack-swift18:26
*** fungi has joined #openstack-swift18:26
*** skudlik has joined #openstack-swift18:27
*** hseipp has quit IRC18:29
*** pcaruana has quit IRC18:46
*** swift_guy_- has joined #openstack-swift18:59
swift_guy_-hello Swift Experts, anyone with experience with TempURL middleware for Swift? I'm trying to determine what headers are valid to be used in the TempURL configuration (incoming_allow_headers , incoming_remove_headers, outgoing_allow_headers, outgoing_remove_headers) and I haven't found a lot of information about that on the documentation or googling it19:02
swift_guy_-Are only object metadata headers valid? starting with x-object-meta prefix? or also for container and account?19:03
*** saint_ has quit IRC19:27
*** lucasxu has quit IRC19:32
*** lucasxu has joined #openstack-swift19:47
*** silor has quit IRC19:54
*** lucasxu has quit IRC20:15
*** _notme has quit IRC20:37
*** swift_guy_- has quit IRC20:39
*** chlong has quit IRC20:41
*** chlong has joined #openstack-swift20:55
*** MVenesio has quit IRC20:55
jrichlidarn, swift_guy_- left.  was gonna give this
*** mvk has joined #openstack-swift21:03
*** skudlik has quit IRC21:12
*** rledisez has quit IRC21:21
*** mgagne has quit IRC21:21
*** asettle has quit IRC21:21
*** Anticimex has quit IRC21:21
*** stew925 has quit IRC21:21
*** jcook has quit IRC21:21
*** timur has quit IRC21:21
*** Anticimex has joined #openstack-swift21:21
*** rledisez_ has joined #openstack-swift21:21
*** timur has joined #openstack-swift21:21
*** jcook has joined #openstack-swift21:21
*** mgagne has joined #openstack-swift21:22
*** Guest39557 has joined #openstack-swift21:22
*** chlong has quit IRC21:46
*** rcernin has quit IRC22:21
*** jerrygb has quit IRC22:25
*** tonanhngo has quit IRC22:32 proposed openstack/swift master: Make a hacking way for OpenStack changes
*** klamath has quit IRC23:00
*** qwertyco has quit IRC23:06
*** kei_yama has joined #openstack-swift23:30

Generated by 2.15.3 by Marius Gedminas - find it at!