Wednesday, 2015-08-19

« Tuesday, 2015-08-18 Index Thursday, 2015-08-20 »
*** achanda_ has joined #openstack-glance00:02
*** achanda has quit IRC00:05
*** achanda_ has quit IRC00:07
*** harshs has joined #openstack-glance00:13
*** spzala has quit IRC00:24
*** tsekiyam_ has joined #openstack-glance00:29
*** mtanino has quit IRC00:31
*** tsekiyama has quit IRC00:32
*** tsekiyam_ has quit IRC00:34
*** dims has joined #openstack-glance00:46
*** dims_ has quit IRC00:47
*** ducttape_ has joined #openstack-glance00:49
*** dims has quit IRC00:56
*** dims has joined #openstack-glance00:56
*** wxy__ has joined #openstack-glance01:07
*** julim has joined #openstack-glance01:24
*** ducttape_ has quit IRC01:55
*** haomaiwang has joined #openstack-glance02:05
*** baojg has joined #openstack-glance02:15
*** junhongl has joined #openstack-glance02:25
*** annegentle has quit IRC02:25
*** wokuma has left #openstack-glance02:26
*** annegentle has joined #openstack-glance02:26
*** annegentle has quit IRC02:32
*** achanda has joined #openstack-glance02:32
*** baojg has quit IRC02:35
*** mingdang1 has joined #openstack-glance02:36
*** ducttape_ has joined #openstack-glance02:41
*** ducttape_ has quit IRC02:47
*** baojg has joined #openstack-glance02:48
*** ducttape_ has joined #openstack-glance02:55
*** dims has quit IRC02:55
*** ducttape_ has quit IRC02:57
*** haomaiwang has quit IRC03:01
*** gberginc has joined #openstack-glance03:01
*** haomaiwang has joined #openstack-glance03:02
*** rodrigods has quit IRC03:16
*** rodrigods has joined #openstack-glance03:19
*** raginbajin has quit IRC03:27
*** raginbajin has joined #openstack-glance03:29
*** harshs has quit IRC03:32
openstackgerritwangxiyuan proposed openstack/python-glanceclient: Add more information show in v2  https://review.openstack.org/21443703:41
*** julim has quit IRC03:59
openstackgerritwangxiyuan proposed openstack/python-glanceclient: Add name support in image-show with v2  https://review.openstack.org/21444004:00
*** haomaiwang has quit IRC04:01
*** haomaiwang has joined #openstack-glance04:02
*** ayoung has quit IRC04:05
*** raginbajin has quit IRC04:10
*** lakshmiS has joined #openstack-glance04:21
*** raginbajin has joined #openstack-glance04:28
*** sdake has quit IRC04:29
openstackgerritwangxiyuan proposed openstack/python-glanceclient: Add name support in image-show with v2  https://review.openstack.org/21444004:54
*** baojg has quit IRC04:59
*** haomaiwang has quit IRC05:01
*** haomaiwa_ has joined #openstack-glance05:02
*** gberginc has quit IRC05:13
*** boris-42 has quit IRC05:20
*** peristeri has joined #openstack-glance05:22
*** harshs has joined #openstack-glance05:22
*** harshs has quit IRC05:26
*** sgotliv_ has joined #openstack-glance05:29
*** sgotliv_ has quit IRC05:52
*** gberginc has joined #openstack-glance05:54
*** baojg has joined #openstack-glance05:58
*** haomaiwa_ has quit IRC06:01
*** haomaiwa_ has joined #openstack-glance06:02
*** peristeri has quit IRC06:03
*** peristeri has joined #openstack-glance06:08
*** groen692 has joined #openstack-glance06:14
openstackgerritOpenStack Proposal Bot proposed openstack/glance_store: Imported Translations from Transifex  https://review.openstack.org/21402606:20
*** ankit_ag has joined #openstack-glance06:28
*** markus_z has joined #openstack-glance06:30
*** achanda has quit IRC06:34
*** lakshmiS has quit IRC06:41
*** haomaiwa_ has quit IRC07:01
*** haomaiwang has joined #openstack-glance07:01
*** mingdang1 has quit IRC07:08
*** mingdang1 has joined #openstack-glance07:09
*** ajayaa has joined #openstack-glance07:10
*** sileht has joined #openstack-glance07:15
*** peristeri has quit IRC07:18
*** belmoreira has joined #openstack-glance07:28
*** boris-42 has joined #openstack-glance07:58
*** haomaiwang has quit IRC08:01
*** haomaiwang has joined #openstack-glance08:02
*** MattMan has joined #openstack-glance08:14
*** GB21 has joined #openstack-glance08:21
*** exploreshaifali has joined #openstack-glance08:24
*** GB21 has quit IRC08:33
*** GB21 has joined #openstack-glance08:33
*** jistr has joined #openstack-glance08:34
*** tpeoples has quit IRC08:46
*** achanda has joined #openstack-glance08:48
*** aix has joined #openstack-glance08:54
*** sgotliv has joined #openstack-glance08:57
openstackgerritMerged openstack/glance_store: Imported Translations from Transifex  https://review.openstack.org/21402608:59
*** haomaiwang has quit IRC09:01
*** haomaiwa_ has joined #openstack-glance09:02
*** achanda has quit IRC09:17
*** gberginc has quit IRC09:18
*** abhishekk has joined #openstack-glance09:18
*** ishant has joined #openstack-glance09:19
*** gberginc has joined #openstack-glance09:38
*** GB21 has quit IRC09:46
*** jaypipes has quit IRC09:46
*** exploreshaifali has quit IRC09:57
*** haomaiwa_ has quit IRC10:01
*** haomaiwang has joined #openstack-glance10:02
*** mingdang1 has quit IRC10:02
*** exploreshaifali has joined #openstack-glance10:02
*** sigmavirus24_awa has quit IRC10:06
*** eglute has quit IRC10:06
*** d34dh0r53 has quit IRC10:07
*** eglute has joined #openstack-glance10:08
*** belmoreira has quit IRC10:08
*** d34dh0r53 has joined #openstack-glance10:08
*** belmoreira has joined #openstack-glance10:09
*** sigmavirus24_awa has joined #openstack-glance10:09
*** belmoreira has quit IRC10:14
*** belmoreira has joined #openstack-glance10:15
*** baojg has quit IRC10:26
*** baojg has joined #openstack-glance10:27
*** baojg has quit IRC10:31
openstackgerritGorka Eguileor proposed openstack/glance_store: Allow random read in RBD driver (get method)  https://review.openstack.org/15732510:41
*** belmoreira has quit IRC10:48
*** dims has joined #openstack-glance10:56
*** peristeri has joined #openstack-glance10:58
*** dims_ has joined #openstack-glance11:03
*** dims has quit IRC11:07
*** ishant has quit IRC11:10
*** haomaiwang has quit IRC11:12
openstackgerritMerged openstack/glance-specs: Add HTTP proxy support to S3 store  https://review.openstack.org/19759411:12
*** boris-42 has quit IRC11:20
*** dims has joined #openstack-glance11:29
*** cdelatte has joined #openstack-glance11:31
*** delattec has joined #openstack-glance11:31
*** dims_ has quit IRC11:31
*** dims_ has joined #openstack-glance11:31
*** wxy__ has quit IRC11:32
*** dims has quit IRC11:34
*** kairat has quit IRC11:42
*** kairat has joined #openstack-glance11:44
*** jaypipes has joined #openstack-glance11:53
openstackgerritNiall Bunting proposed openstack/glance: Tasks now try the schema before the database  https://review.openstack.org/21423212:05
*** julim has joined #openstack-glance12:10
*** ducttape_ has joined #openstack-glance12:12
*** belmoreira has joined #openstack-glance12:19
*** belmoreira has quit IRC12:19
*** chlong has joined #openstack-glance12:27
*** edmondsw has joined #openstack-glance12:27
openstackgerritMerged openstack/glance_store: Add proxy support to S3 Store  https://review.openstack.org/21029412:33
*** haomaiwang has joined #openstack-glance12:39
openstackgerritDarja Shakhray proposed openstack/glance: Fixed version unequality artifact filtering  https://review.openstack.org/21418512:41
*** julim has quit IRC12:41
*** julim has joined #openstack-glance12:42
openstackgerritOlena Logvinova proposed openstack/glance: Domain model section  https://review.openstack.org/20751012:50
*** exploreshaifali has quit IRC12:54
*** ducttape_ has quit IRC13:00
*** smatzek has joined #openstack-glance13:00
openstackgerritAlexander Tivelkov proposed openstack/glance: Fixed the output of list artifacts API calls  https://review.openstack.org/21460013:02
*** junhongl has quit IRC13:06
openstackgerritMerged openstack/glance: Fix Request ID has a double 'req-' at the start  https://review.openstack.org/20652113:12
*** annegentle has joined #openstack-glance13:13
*** GB21 has joined #openstack-glance13:15
*** spzala has joined #openstack-glance13:21
*** dims_ has quit IRC13:29
*** dims has joined #openstack-glance13:29
*** annegentle has quit IRC13:37
*** GB21 has quit IRC13:39
*** sgotliv has quit IRC13:44
*** sgotliv has joined #openstack-glance13:45
*** jecarey has joined #openstack-glance13:47
*** abhishekk has quit IRC13:56
*** sdake has joined #openstack-glance13:56
*** sdake_ has joined #openstack-glance13:58
*** sdake has quit IRC14:01
rosmaitaflaper87: hello14:02
openstackgerritKairat Kushaev proposed openstack/glance: Fixed the output of list artifacts API calls  https://review.openstack.org/21460014:03
*** ankit_ag has quit IRC14:03
flaper87rosmaita: :D14:04
rosmaitaguess we should wait for a quorum14:04
rosmaitanikhil_k_: good morning14:04
rosmaitasigmavirus24_awa: wakey wakey14:04
nikhil_k_morning rosmaita14:04
nikhil_k_flaper87:14:05
flaper87helloooooooooo :D14:05
flaper87I thought we had said 15 UTC, I'm probably wrong14:05
rosmaitamaybe, i thought it was 14-15 utc14:06
rosmaitameaning starting at 14 ending at 1514:06
rosmaitai can come back at 15, but have to leave by 15:3014:06
*** ajayaa has quit IRC14:08
flaper87ah no, you're right14:10
flaper87ok, I'm here14:10
rosmaitaok, i just want to throw this out there ...14:11
* flaper87 sits down14:11
flaper87:P14:11
rosmaitathe reason i want to keep two separate workflows, one for upload and one for import14:11
rosmaitais that since nova uses the upload workflow, we need it to be rock solid14:11
*** mingdang1 has joined #openstack-glance14:11
rosmaitai think the custom task configuration will be complicated and could be subject to error until a deployer gets it right14:12
rosmaitaand we don't want to break nova image-creation while that's happening14:12
rosmaitathe tasks are configurable using taskflow14:12
rosmaitaso fit really well for this14:12
rosmaitai'm just afraid if we mix the "trusted" upload path with the "untrusted" path, it will be a mess14:13
flaper87I totally agree with the upload process needing to be rock-solid. My argument against that is that you don't need to have those tasks configured everywhere. It's possible to disable tasks in certain nodes (the ones nova uses) and have them active in other nodes (public)14:13
*** sdake_ is now known as sdake14:13
flaper87The problem I see with the import workflow is that it doesn't solves the other use-case we keep getting feedback/requests on14:14
rosmaitagood, as long as we keep it in mind14:14
johnthetubaguyflaper87: I would actually rather we found a way to add a service token to requests from Nova, to skip checks that are not needed when thats present, so we don't force folks into more complicated setups, but you are right, that could totally work14:14
flaper87cool14:14
flaper87johnthetubaguy: sounds good to me14:14
flaper87I did my homework and started thinking about a different way to do this14:15
flaper87or for it to have a lower impact in its first version14:15
flaper87Here it goes:14:15
nikhil_k_yay!14:15
*** boris-42 has joined #openstack-glance14:16
flaper87The solution I'm thinking about is similar to the one proposed with the only difference that it allows to run tasks automatically just after the image has been saved and marked active.14:16
flaper87It doesn't sound very useful like that but here's the missing part:14:16
flaper87An OP could have a task that would convert the image to the desired format and store it in the desired store14:17
flaper87The question I haven't answered is:14:17
flaper871) Should that task create a new image? or should it just add a new location that will be then used by nova?14:17
flaper87Extra points to this current proposal are:14:18
flaper871) It allows for other tasks like image replication to be written14:18
flaper872) It allows for backport tasks to be written14:18
*** mtanino has joined #openstack-glance14:18
flaper873) It allows for verification tasks to be written14:18
flaper87etc14:18
flaper87It's just a first step14:18
flaper87And it does not change the upload workflow14:19
rosmaitaas far as your question goes, i think create new image, i don't like locations that point to different payloads14:19
flaper87thoughts? all crap?14:19
flaper87rosmaita: agreed, the problem with that is that the user will keep booting the original image14:19
nikhil_k_if it's a different image format and data/metadata then it's a different image14:19
rosmaitaunless you delete it!14:19
flaper87the user won't boot the other one14:19
johnthetubaguywhat about doing it before the image goes Active?14:19
flaper87but then the user will be: Wait what? where's my image?14:20
nikhil_k_it's checksum is different and so will be other Glance-image-meta as per it's formats14:20
*** ducttape_ has joined #openstack-glance14:20
flaper87johnthetubaguy: that's the original proposal that rosmaita doesn't like14:20
flaper87I mean, it's not rosmaita doesn't like it14:20
flaper87that came out wrong14:20
flaper87sorry14:20
rosmaita:)14:20
flaper87I meant to say, it changes the upload workflow and there are concerns around it14:21
johnthetubaguyflaper87: it does break the API semantics, unless we get the user to opt in to the new behavior14:21
flaper87rosmaita: sorry :)14:21
*** jecarey has quit IRC14:21
flaper87johnthetubaguy: that's my point in favor for it14:21
flaper87I'd be all for doing it before the image is active and have a 'processing' step in between14:21
*** edmondsw has quit IRC14:22
*** sigmavirus24_awa is now known as sigmavirus2414:22
rosmaitawhich tasks already have14:22
sigmavirus24sorry rosmaita14:22
nikhil_k_I think this will make the tasks synchronous or will create the requirement for them to be so14:23
rosmaitathe idea of tasks was that we didn't want to pollute the image statuses with irrelevant stuff14:23
flaper87We have image introspection and image conversion14:23
flaper87nikhil_k_: mmh, why?14:23
nikhil_k_flaper87: it's very likely that the next feature request from OP will be to process the in-stream data14:23
flaper87rosmaita: I seem to remember something different but I might be wrong. I remember in Atlanta talkig about tasks as a way to implement image conversion too14:23
nikhil_k_and that needs to be passed to tasks (basically, pass the data stream handle to the task)14:23
nikhil_k_and it becomes synchronous14:24
flaper87nikhil_k_: which doesn't mean it has to happen14:24
*** ducttape_ has quit IRC14:24
flaper87That's were we as a community need to be opinionated bsaed on what makes sense14:24
flaper87the reason I'm pushing this is because I believe the use case makes sense and I see how it can be useful14:24
rosmaitaflaper87: yes, but the idea is you can't do in situ image conversion because it violates immutability, so the task would grab and existing image and convert it into a new image14:24
*** ducttape_ has joined #openstack-glance14:25
*** ducttape_ has quit IRC14:25
flaper87but that's were we differ on opinions. For me, the image becomes immutable once it's saved because anything could happen in between14:25
johnthetubaguyrosmaita: so in situ image conversion, would be an import I guess, just from an existing image rather than data you pass in14:25
rosmaitajohnthetubaguy: exactly14:25
johnthetubaguyrosmaita: although thats starting to mix the semantics up a bit, but the idea seems sound ish14:26
* sigmavirus24 reads scrollback14:26
johnthetubaguyflaper87: +1 on being more opinionated, it makes people lives simpler14:26
nikhil_k_the api behavior is changed in both cases, not necessarily the semantics (or I haven't researched the anticipated api changes)14:26
johnthetubaguy(when we get the correct opinion, lol)14:27
rosmaitai think there's a role for both end-user tasks and admin tasks, question is just where to put them14:27
flaper87johnthetubaguy: :P14:27
rosmaitai mean end-user initiated tasks14:27
johnthetubaguyrosmaita: I think just concentrating on image conversion should shake this all out, getting the use cases written down,  some formats are only accepted after conversion, some formats have optional conversions (that create new images, probably)14:29
nikhil_k_I think the proposal is more than just conversion14:30
sigmavirus24yeah14:30
flaper87The proposal mentions other use cases, yes.14:30
nikhil_k_1. Automatic image conversion to the format that works best with the hypervisor14:30
sigmavirus24So on the whole I'm +0.5 on the proposal14:30
nikhil_k_.4414:30
nikhil_k_4514:31
nikhil_k_2. Automatic image replication to other stores.4614:31
nikhil_k_4714:31
nikhil_k_the question is14:31
flaper87FWIW, I was asked to write more use-cases down :P14:31
nikhil_k_things like : Automatic, in-house, image backup outside glance.14:31
nikhil_k_Glance is and should not be responsible for this14:31
flaper87nikhil_k_: agreed, the proposal is not to add a task for that14:32
nikhil_k_we are upload and discoverability service14:32
flaper87it just enables for that use-case14:32
flaper87I can clarify that on the spec14:32
nikhil_k_flaper87: yeah, but we don't want people to use glance for this14:32
nikhil_k_and set wrong expectations14:32
flaper87to use glance for what?14:32
nikhil_k_Automatic, in-house, image backup outside glance.14:33
nikhil_k_I think it's worth having a task for "Automatic image replication to other stores"14:33
nikhil_k_but it can be a end-user level admin operated task14:33
flaper87That's what's being proposed there14:33
flaper87I think it's just a misunderstanding14:33
flaper87but anyway, nothing prevents folks to write their own task that backsup images outside14:34
rosmaitaso i guess we dont want to do away with the tasks framework?14:34
rosmaitai mean the one connected to the /v2/tasks API14:35
nikhil_k_flaper87: sorry if I conveyed this wrong. But if these use cases are coming from OPs, I wanted to know what is the motivation and likely design of the proposal. If it's just one use case, I think we can narrow down the scope and discuss that way but if it's changing the image-workflow and tasks-workflow then we need to think hard on the proposal. that was the motivation for asking the related use cases.14:35
*** vijendar has joined #openstack-glance14:35
nikhil_k_rosmaita: we cannot, it's a public api that needs to be supported14:36
flaper87(brb sorry)14:36
johnthetubaguynikhil_k_: I am really just thinking focus in one one case (image conversion), enumerate out the cases, see how it looks, then add in the other use cases, might help make things clearer, then look at how to stage the implementation14:36
*** harshs has joined #openstack-glance14:36
rosmaitajohnthetubaguy: +114:36
sigmavirus24johnthetubaguy: right, but I feel like that use case stands out as vastly different from the others flaper87 is talking about14:36
*** edmondsw has joined #openstack-glance14:37
nikhil_k_johnthetubaguy: yeah, that would be a good plan if it works for all those who are asking for it. So, yeah, thanks for asking that14:37
sigmavirus24the others (replication to other regions/stores) don't involve modification of the image flow14:37
rosmaitaflaper87: i agree with nikhil about being clear on whether we should modify image-workflow or tasks-workflow14:37
rosmaitathe tasks-workflow is designed to be deployer-mucked-with14:37
sigmavirus24image conversion seems to need a modification of image-workflow14:37
sigmavirus24rosmaita: right14:37
rosmaitai don't know that we want to mess with the image-workflow14:37
nikhil_k_I think not14:37
rosmaitasigmavirus24: unless you get image first, then use task to convert14:38
rosmaitamy feeling is that it's easier to keep tasks separate14:38
sigmavirus24== rosmaita, nikhil_k_ (re modifying image-workflow)14:38
nikhil_k_also, it would be excellent to get some feedback from the v3 supportability POV. ativelkov ^ (pretty sure this will result into that effort being out of sync and wasteful)14:38
sigmavirus24rosmaita: so make an image, convert and create new image and....?14:38
nikhil_k_sigmavirus24: I think that makes sense14:39
johnthetubaguyso from the defcore perspective, we need a way to have an API (lowest common denominator one, maybe) that works everywhere, thats a slightly separate case here, but I think thats what makes me push into something like that14:39
sigmavirus24so from a v3 POV, the converted image would be a new version of an artifact14:39
sigmavirus24It's not clear to me how that translates to images14:39
nikhil_k_sigmavirus24: right, my concern was his effort to mock v2 usign v314:39
sigmavirus24right14:39
rosmaitasigmavirus24: or it could be part of an artifact, where artifact is "my db image" and the components are "my db VHD" , "my DB qcow", etc14:40
* sigmavirus24 isn't certain that's how artifacts works14:40
nikhil_k_like johnthetubaguy mentioned, we are proposing/going in that direction where v2 would be the standard defcore api. in that case the v3 would need to work alongside v2. it means this feature would need to work in that code14:40
*** sdake_ has joined #openstack-glance14:41
* rosmaita isn't sure either14:41
johnthetubaguynikhil_k_: thats cool, I was meaning specifically how do we get the image import acceptable to all users14:42
rosmaitajohnthetubaguy: good point14:42
rosmaitathe tasks api happened before defcore was a "thing"14:42
rosmaitaat least the design of it did14:42
nikhil_k_johnthetubaguy: right, and in thie new workflow. the defcore ask would be to not enable the image-conversion as it would need to be consistent across clouds14:42
flaper87(so sorry folks, almost back on the convo)14:43
nikhil_k_but I need to check14:43
nikhil_k_also, this would likely affect image-federation if the private and public clouds have different working image-formats for their clouds14:43
nikhil_k_and if you ask for image-sharing across federated clouds14:44
*** ducttape_ has joined #openstack-glance14:44
*** sdake has quit IRC14:44
johnthetubaguynikhil_k_: its the validation that worries me in those cases14:45
johnthetubaguynikhil_k_: but yes, that too14:45
* nikhil_k_ brb. sorry14:46
*** tpeoples has joined #openstack-glance14:46
*** annegentle has joined #openstack-glance14:46
*** annegentle has quit IRC14:47
*** annegentle has joined #openstack-glance14:47
*** GB21 has joined #openstack-glance14:49
*** mingdang1 has quit IRC14:51
*** zxen has joined #openstack-glance14:54
*** pbourke has quit IRC14:57
*** pbourke has joined #openstack-glance14:57
*** MarcF has quit IRC15:00
*** dims has quit IRC15:01
*** dims has joined #openstack-glance15:01
*** peristeri has quit IRC15:01
*** dims_ has joined #openstack-glance15:05
*** dims has quit IRC15:06
*** zul has joined #openstack-glance15:07
*** bdossant has joined #openstack-glance15:10
*** bpoulos has joined #openstack-glance15:16
*** zxen has quit IRC15:18
*** dims_ has quit IRC15:19
*** dims has joined #openstack-glance15:19
jokke_Sorry to chip in late rosmaita, nikhil_k_, sigmavirus24 & flaper87: but personally I really dislike the idea of modifying image that breaks the only reason images are provided by glance, immutability. If you cannot trust that image and the hash you uploaded for your image id X is that for the image id X because deployer decided to set task to fiddle around images, damn you all15:22
*** zul has quit IRC15:23
jokke_Image conversion from existing image, perfectly fine, you get notification of new image created, you transform that to what ever you like and you create new image out of that15:23
jokke_but you don't claim to the user that it is the same image what the user uploaded15:24
rosmaitajokke_: you are preaching to the choir as far as i'm concerned!15:24
jokke_thanks ... that's all then :D15:25
*** burgerk has joined #openstack-glance15:25
flaper87jokke_: the point w.r.t security, signing, etc is a good one15:25
flaper87That's also why I proposed earlier today to just do it after the image has been saved15:25
flaper87and keep that as an internal detail of the deployment15:25
jokke_so do we have problem then to have autotrigger tasks if we agree that the autotrigger tasks are still not allowed to change existing images? :P15:26
flaper87I'm not sure how the booting bits of that would work, honesly15:26
flaper87jokke_: I'd say they are not allowed to "mutate images"15:26
flaper87nikhil_k_: rosmaita jokke_ sigmavirus24 I'm really sorry for having disapeared in the middle of our chat15:27
flaper87johnthetubaguy: ^15:27
jokke_np gave me good time to read through and catch up :D15:27
johnthetubaguyflaper87: is this the "on demand" conversion of images, that you have a question about?15:28
jokke_flaper87: I don't think we should encourage any behavior nova booting another image than what it was requested15:28
flaper87johnthetubaguy: no, I was just saying I'm sorry for dissapearing :)15:28
johnthetubaguyflaper87: oh, no worries15:28
johnthetubaguyjokke_: it sounds fairly evil15:29
jokke_If user A uploads image with hash X to image ID Y and requests nova to boot image Y nova should boot image which hash is X. If OP wants to promote certain image formats, they should do that auto-conversation, create new image Z out of it and notify the user that it was done and they would prefer user booting from Z rather than Y15:31
jokke_conversion even15:31
burgerkIs there a reason that the glance DB does not accept 4 byte unicode characters for property values ( image name for example ) ?   https://bugs.launchpad.net/glance/+bug/1370954  moved the error from a 500 to a 400 error, but nova DB for example allows 4 byte for VM name, etc.   Is it just a matter of changing encoding from utf8 to utf8mb4 or something more behind that ?15:32
openstackLaunchpad bug 1370954 in Glance "glance 500's when passed image name with a 4-byte utf-8 character" [Medium,Fix released] - Assigned to Inessa Vasilevskaya (ivasilevskaya)15:32
johnthetubaguyjokke_: flaper87 : rosmaita: honestly, I kinda want to hear the use case around the auto image conversion, I am not quite seeing that properly right now15:33
sigmavirus24burgerk: we maybe using utf8 instead of utf8mb4 because on mysql those are two very different things15:34
openstackgerritMerged openstack/glance: Fix the db_sync problem in 039 for db2  https://review.openstack.org/16710115:34
burgerksigmavirus24: so it is a conscious decision to limit to 3 byte utf-8 in the glance DB vs. something unintentional ?15:36
sigmavirus24burgerk: I didn't say that15:36
sigmavirus24I'm saying the root cause may be a mistaken in the encoding for tables and I think flaper87 fixed this recently15:37
sigmavirus24at least for master15:37
burgerkI understand ... that was just my next question :)15:37
sigmavirus24or flaper87 + someone else15:37
sigmavirus24looking at the bug it was apparently fixed in kilo15:38
sigmavirus24I feel like we just merged something in the past few weeks though to fix this for other tables15:38
*** tsekiyama has joined #openstack-glance15:38
*** bdossant has quit IRC15:38
sigmavirus24most people who use mysql don't quite grok the difference between utf8 and utf8mb415:38
jokke_johnthetubaguy: lets say you run pure kvm house and would prefer all your images being qcow2 for operations simplicity. You could tell your endusers that qcow2 is only supported format in your cloud and auto-conversion is provided from other formats.15:39
*** haomaiwang has quit IRC15:40
*** jecarey has joined #openstack-glance15:40
burgerksigmavirus24: are you referencing the bug I pointed to that was fixed in kilo, or another ?15:41
flaper87rosmaita: nikhil_k_ jokke_ sigmavirus24 to be honest, I don't feel comfortable pushing this in Liberty anymore. The fact that there are many concerns worries me and I'd prefer us to take more time to think this through. I'll work on a summit session where we can discuss this a bit better15:42
nikhil_k_thanks flaper87 !15:42
jokke_johnthetubaguy: totally hypothetical from my end, but I can see bunch of such reasons why you would want to run task X, Y, Z when there is new image uploaded15:42
flaper87We'd have had enough time if we'd had agreed on the proposal but that's not the case so, lets push it back to Mitaka15:42
nikhil_k_session/feedback sounds good15:42
flaper87nikhil_k_: rosmaita sigmavirus24 jokke_ johnthetubaguy thank you all for the feedback and time15:43
*** groen692 has quit IRC15:43
flaper87I'll put all those back into the proposal15:43
*** harshs has quit IRC15:43
rosmaitaflaper87: that sounds very reasonable to me, i think the extra time will be good15:44
johnthetubaguyjokke_: agreed with those ones you say about, I was getting the impression there were more obscure ones15:44
jokke_tbh one of the biggest benefits I'd see for autotasks is automated antivirus runs ... upload image, triggers antivirus run, if flagged, disabled until admin interaction15:44
rosmaitaflaper87: i think everyone agrees that the general idea is a good one, it's just a matter of working out the details in a good way15:44
johnthetubaguyflaper87: happy to help, this stuff is important to get right15:44
jokke_johnthetubaguy: I don't think we should block funtionality just based on that there might be obscure use cases that might benefit from it ;)15:45
*** zul has joined #openstack-glance15:45
sigmavirus24jokke_: so I think the detail we want to work out is "Should tasks be able to disturb the image workflow"15:46
nikhil_k_well put15:47
nikhil_k_determining precedent on the image centric way of glance15:48
jokke_sigmavirus24: my take on that is hopefully pretty clear above ... I'm happy to provide -2 services to any proposal trying to do otherwise :P15:48
nikhil_k_or are the new use cases strong enough to establish otherwsise15:48
sigmavirus24jokke_: I generally agree but I'm open to be persuaded otherwise provided a good argument with reasonable details15:48
sigmavirus24My bias is against modifying the workflow though15:48
*** dims has quit IRC15:49
*** sdake_ is now known as sdake15:49
*** dims has joined #openstack-glance15:49
johnthetubaguyjokke_: totally, thats basically what I was trying to say15:51
johnthetubaguyjokke_: i just said it badly15:51
*** dims_ has joined #openstack-glance15:53
jokke_:)15:53
*** dims has quit IRC15:54
*** zul has quit IRC15:55
*** takedakn has joined #openstack-glance15:57
burgerklooking at master it still seems @no_4byte_params decorating is still being used, blocking the passing of 4 byte values16:00
*** exploreshaifali has joined #openstack-glance16:06
*** spzala_ has joined #openstack-glance16:07
*** spzala has quit IRC16:11
*** spzala_ is now known as spzala16:11
*** ayoung has joined #openstack-glance16:13
*** takedakn has quit IRC16:20
*** spzala has quit IRC16:22
nikhil_k_yeah, we block the 4byte val in db as things can be messy depending on the DB used16:27
nikhil_k_we test things against standard mysql for most cases, so that's what the default & in this case expected behavior is set to16:27
nikhil_k_burgerk: ^16:27
openstackgerritBrianna Poulos proposed openstack/glance: Add image signing verification  https://review.openstack.org/18313716:29
*** jistr has quit IRC16:30
burgerknikhal_k: other DBs such as nova do not have a similar block so just seemed inconsistent.  I can have a 4 byte VM name but not a 4 byte Image name.  That is why I was questioning the need for the limitation in glance16:31
burgerkI understand 4 byte can get complicated/messy16:33
nikhil_k_burgerk: I am not sure why operators for nova DB haven't complained as they did got glance. it's a validation added after some of that feedback16:36
*** dims_ has quit IRC16:39
*** dims has joined #openstack-glance16:39
*** zul has joined #openstack-glance16:40
*** zul has quit IRC16:47
*** zul has joined #openstack-glance16:49
*** sgotliv has quit IRC17:09
*** markus_z has quit IRC17:14
*** sdake_ has joined #openstack-glance17:17
*** sdake has quit IRC17:20
*** spzala has joined #openstack-glance17:21
*** zul has quit IRC17:23
openstackgerritdane-fichter proposed openstack/glance: Add unit tests for signature_utils class  https://review.openstack.org/21472617:31
*** sdake has joined #openstack-glance17:37
*** exploreshaifali has quit IRC17:38
*** sdake_ has quit IRC17:40
*** achanda has joined #openstack-glance17:42
*** sdake has quit IRC17:47
*** sdake has joined #openstack-glance17:51
*** achanda_ has joined #openstack-glance17:54
*** aix has quit IRC17:55
*** ericpeterson has joined #openstack-glance17:55
*** achanda has quit IRC17:55
*** harshs has joined #openstack-glance17:56
*** ducttape_ has quit IRC17:57
*** tpeoples has quit IRC17:57
*** annegentle has quit IRC18:04
*** ericpeterson has quit IRC18:09
*** tpeoples has joined #openstack-glance18:11
*** ayoung has quit IRC18:15
openstackgerritVictor Howard proposed openstack/glance: Removing sample conf showing options that are depreciated  https://review.openstack.org/17542318:17
*** danieru has quit IRC18:26
*** bapalm has quit IRC18:31
*** bapalm has joined #openstack-glance18:32
*** jwang has quit IRC18:34
*** GB21 has quit IRC18:38
*** jwang has joined #openstack-glance18:40
*** exploreshaifali has joined #openstack-glance18:42
*** kebray has joined #openstack-glance18:46
*** MattMan has left #openstack-glance18:46
*** zul has joined #openstack-glance18:50
*** ducttape_ has joined #openstack-glance19:02
*** danieru has joined #openstack-glance19:12
*** annegentle has joined #openstack-glance19:13
*** nikhil_k_ has quit IRC19:13
*** nikhil has joined #openstack-glance19:14
*** nikhil is now known as nikhil_k19:16
*** achanda_ has quit IRC19:37
*** e0ne has joined #openstack-glance19:45
*** e0ne has quit IRC19:47
*** belmoreira has joined #openstack-glance19:50
*** achanda has joined #openstack-glance19:51
*** alejandrito has joined #openstack-glance19:59
*** annegentle has quit IRC19:59
*** alejandrito_ has joined #openstack-glance20:00
*** alejandrito has quit IRC20:00
*** e0ne has joined #openstack-glance20:02
*** e0ne has quit IRC20:18
*** alejandrito_ has quit IRC20:18
*** r-daneel has joined #openstack-glance20:24
*** julim has quit IRC20:25
*** e0ne has joined #openstack-glance20:29
*** e0ne has quit IRC20:35
*** e0ne has joined #openstack-glance20:36
*** exploreshaifali has quit IRC20:41
*** belmoreira has quit IRC20:44
*** annegentle has joined #openstack-glance20:45
*** achanda has quit IRC20:47
*** achanda has joined #openstack-glance20:52
*** annegentle has quit IRC20:53
*** annegentle has joined #openstack-glance20:54
*** zul has quit IRC21:00
*** achanda has quit IRC21:06
*** e0ne has quit IRC21:15
*** e0ne has joined #openstack-glance21:19
*** smatzek has quit IRC21:28
*** e0ne has quit IRC21:28
openstackgerritKent Wang proposed openstack/glance: Allows OVA/OVF package upload to Glance  https://review.openstack.org/21481021:44
*** sgotliv has joined #openstack-glance21:51
*** burgerk has quit IRC22:02
*** bpoulos has quit IRC22:03
*** sgotliv has quit IRC22:13
*** vijendar has quit IRC22:14
*** annegentle has quit IRC22:17
*** tpeoples has quit IRC22:44
*** r-daneel has quit IRC22:49
*** jecarey has quit IRC22:59
*** zul has joined #openstack-glance23:07
*** ducttape_ has quit IRC23:11
*** zul has quit IRC23:12
*** chlong has quit IRC23:28
*** flwang1 has quit IRC23:32
*** achanda has joined #openstack-glance23:33
*** dims_ has joined #openstack-glance23:33
*** dims has quit IRC23:36
*** dims_ has quit IRC23:39
*** dims has joined #openstack-glance23:39
*** zul has joined #openstack-glance23:41
*** dims has quit IRC23:44
*** kebray has quit IRC23:47
*** ayoung has joined #openstack-glance23:51
*** edmondsw has quit IRC23:56
*** flwang1 has joined #openstack-glance23:57
« Tuesday, 2015-08-18 Index Thursday, 2015-08-20 »

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!