| *** ctina has joined #openstack-glance | 00:03 | |
| *** ducttape_ has joined #openstack-glance | 00:07 | |
| *** harshs has quit IRC | 00:08 | |
| *** Ctina_ has quit IRC | 00:11 | |
| *** TravT_ has joined #openstack-glance | 00:13 | |
| *** TravT has quit IRC | 00:15 | |
| *** TravT has joined #openstack-glance | 00:21 | |
| *** takedakn has quit IRC | 00:23 | |
| *** TravT_ has quit IRC | 00:24 | |
| *** tsekiyama has quit IRC | 00:26 | |
| *** mtanino has quit IRC | 00:32 | |
| *** ducttape_ has quit IRC | 00:34 | |
| *** dims has joined #openstack-glance | 00:34 | |
| *** ducttape_ has joined #openstack-glance | 00:40 | |
| *** ducttape_ has quit IRC | 00:40 | |
| *** ducttape_ has joined #openstack-glance | 00:41 | |
| openstackgerrit | Tomoki Sekiyama proposed openstack/glance: Importing oslo.rootwrap into Glance https://review.openstack.org/186201 | 00:46 |
|---|---|---|
| *** ducttape_ has quit IRC | 00:48 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 00:49 | |
| *** ctina has quit IRC | 01:03 | |
| *** ctina has joined #openstack-glance | 01:07 | |
| *** ducttape_ has joined #openstack-glance | 01:10 | |
| *** mtanino has joined #openstack-glance | 01:14 | |
| *** takedakn has joined #openstack-glance | 01:42 | |
| *** takedakn has quit IRC | 01:42 | |
| *** junhongl has joined #openstack-glance | 01:46 | |
| *** ducttape_ has quit IRC | 02:02 | |
| *** haomaiwang has joined #openstack-glance | 02:06 | |
| *** mtanino has quit IRC | 02:11 | |
| *** dims has quit IRC | 02:30 | |
| *** baojg has joined #openstack-glance | 02:31 | |
| *** gberginc has joined #openstack-glance | 02:32 | |
| *** bkopilov has quit IRC | 02:32 | |
| *** harshs has joined #openstack-glance | 02:42 | |
| *** harshs_ has joined #openstack-glance | 03:00 | |
| *** haomaiwang has quit IRC | 03:01 | |
| *** harshs has quit IRC | 03:01 | |
| *** harshs_ is now known as harshs | 03:01 | |
| *** baojg has quit IRC | 03:02 | |
| *** haomaiwa_ has joined #openstack-glance | 03:02 | |
| *** baojg has joined #openstack-glance | 03:15 | |
| *** changbl has quit IRC | 03:30 | |
| *** sdake has joined #openstack-glance | 03:32 | |
| *** sdake_ has quit IRC | 03:36 | |
| *** sdake_ has joined #openstack-glance | 03:42 | |
| *** sdake has quit IRC | 03:45 | |
| *** baojg has quit IRC | 03:49 | |
| *** bkopilov has joined #openstack-glance | 03:58 | |
| *** haomaiwa_ has quit IRC | 04:01 | |
| *** haomaiwa_ has joined #openstack-glance | 04:02 | |
| *** baojg has joined #openstack-glance | 04:16 | |
| *** sdake_ is now known as sdake | 04:22 | |
| *** ayoung has quit IRC | 04:25 | |
| *** lakshmiS has joined #openstack-glance | 04:36 | |
| *** haomaiwa_ has quit IRC | 05:01 | |
| *** haomaiwa_ has joined #openstack-glance | 05:02 | |
| *** vijendar has quit IRC | 05:14 | |
| *** harshs has quit IRC | 05:15 | |
| *** dims has joined #openstack-glance | 05:20 | |
| *** dims has quit IRC | 05:26 | |
| *** gberginc has quit IRC | 05:26 | |
| *** lakshmiS has quit IRC | 05:52 | |
| *** haomaiwa_ has quit IRC | 06:01 | |
| *** haomaiwang has joined #openstack-glance | 06:02 | |
| *** gberginc has joined #openstack-glance | 06:04 | |
| *** groen692 has joined #openstack-glance | 06:05 | |
| *** tpeoples has quit IRC | 06:11 | |
| *** GB21 has joined #openstack-glance | 06:17 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/glance: Imported Translations from Transifex https://review.openstack.org/214022 | 06:27 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/glance_store: Imported Translations from Transifex https://review.openstack.org/214026 | 06:38 |
| *** openstackgerrit_ has joined #openstack-glance | 06:42 | |
| *** haomaiwang has quit IRC | 07:01 | |
| *** haomaiwang has joined #openstack-glance | 07:02 | |
| *** GB21 has quit IRC | 07:02 | |
| *** peristeri has joined #openstack-glance | 07:11 | |
| *** abhishekk has joined #openstack-glance | 07:13 | |
| *** markus_z has joined #openstack-glance | 07:14 | |
| *** exploreshaifali has joined #openstack-glance | 07:36 | |
| *** dims has joined #openstack-glance | 07:48 | |
| *** sgotliv has joined #openstack-glance | 07:48 | |
| *** dims has quit IRC | 07:53 | |
| *** jistr has joined #openstack-glance | 07:55 | |
| *** haomaiwang has quit IRC | 08:01 | |
| *** haomaiwang has joined #openstack-glance | 08:02 | |
| openstackgerrit | Flavio Percoco proposed openstack/glance: Add the trigger field to the tasks table https://review.openstack.org/197899 | 08:10 |
| openstackgerrit | Flavio Percoco proposed openstack/glance: Add task trigger model https://review.openstack.org/197890 | 08:10 |
| *** exploreshaifali has quit IRC | 08:29 | |
| openstackgerrit | Merged openstack/glance: Imported Translations from Transifex https://review.openstack.org/214022 | 08:48 |
| *** d0ugal has joined #openstack-glance | 08:54 | |
| *** haomaiwang has quit IRC | 09:01 | |
| *** shikel has quit IRC | 09:01 | |
| *** haomaiwang has joined #openstack-glance | 09:02 | |
| *** sh1kel has joined #openstack-glance | 09:06 | |
| *** MattMan has joined #openstack-glance | 09:06 | |
| *** MattMan has quit IRC | 09:07 | |
| *** MattMan has joined #openstack-glance | 09:08 | |
| *** haomaiwang has quit IRC | 10:01 | |
| *** haomaiwang has joined #openstack-glance | 10:02 | |
| *** baojg has quit IRC | 10:06 | |
| *** baojg has joined #openstack-glance | 10:06 | |
| *** baojg has quit IRC | 10:07 | |
| *** baojg has joined #openstack-glance | 10:07 | |
| *** baojg has quit IRC | 10:07 | |
| *** baojg has joined #openstack-glance | 10:07 | |
| *** pbourke has quit IRC | 10:11 | |
| *** exploreshaifali has joined #openstack-glance | 10:12 | |
| *** pdb has joined #openstack-glance | 10:14 | |
| openstackgerrit | Alexander Tivelkov proposed openstack/glance: Artifacts are now properly filtered by dict props https://review.openstack.org/207374 | 10:14 |
| *** sdake has quit IRC | 10:15 | |
| *** baojg has quit IRC | 10:17 | |
| *** baojg has joined #openstack-glance | 10:18 | |
| *** baojg has quit IRC | 10:23 | |
| *** sayali has quit IRC | 10:24 | |
| *** sayali has joined #openstack-glance | 10:25 | |
| *** exploreshaifali has quit IRC | 10:43 | |
| *** sh1kel has quit IRC | 10:48 | |
| *** dims has joined #openstack-glance | 10:51 | |
| *** haomaiwang has quit IRC | 10:54 | |
| *** delattec has quit IRC | 11:10 | |
| *** cdelatte has quit IRC | 11:10 | |
| *** exploreshaifali has joined #openstack-glance | 11:13 | |
| *** ankit_ag has joined #openstack-glance | 11:15 | |
| *** marcusvrn has joined #openstack-glance | 11:27 | |
| *** dims_ has joined #openstack-glance | 11:28 | |
| *** dims has quit IRC | 11:30 | |
| *** cdelatte has joined #openstack-glance | 11:32 | |
| *** smatzek has joined #openstack-glance | 11:36 | |
| *** DuncanT has quit IRC | 11:38 | |
| *** marcusvrn has quit IRC | 11:38 | |
| *** bkopilov has quit IRC | 11:58 | |
| *** dims has joined #openstack-glance | 12:00 | |
| *** dims has quit IRC | 12:00 | |
| *** dims_ has quit IRC | 12:00 | |
| *** dims has joined #openstack-glance | 12:01 | |
| *** dims has quit IRC | 12:05 | |
| openstackgerrit | Alexander Tivelkov proposed openstack/glance: Artifacts are now properly filtered by dict props https://review.openstack.org/207374 | 12:11 |
| *** TravT_ has joined #openstack-glance | 12:15 | |
| *** dims has joined #openstack-glance | 12:15 | |
| openstackgerrit | Alexander Tivelkov proposed openstack/glance: Artifacts are now properly filtered by dict props https://review.openstack.org/207374 | 12:15 |
| *** TravT has quit IRC | 12:16 | |
| *** ducttape_ has joined #openstack-glance | 12:20 | |
| openstackgerrit | Alexander Tivelkov proposed openstack/glance: Artifacts are now properly filtered by dict props https://review.openstack.org/207374 | 12:20 |
| *** annegentle has joined #openstack-glance | 12:24 | |
| *** ducttape_ has quit IRC | 12:37 | |
| *** marcusvrn has joined #openstack-glance | 12:38 | |
| *** sgotliv_ has joined #openstack-glance | 12:38 | |
| *** ducttape_ has joined #openstack-glance | 12:38 | |
| *** ducttape_ has quit IRC | 12:38 | |
| *** sgotliv has quit IRC | 12:40 | |
| *** dims_ has joined #openstack-glance | 12:41 | |
| *** dims has quit IRC | 12:41 | |
| *** ctina has quit IRC | 12:43 | |
| *** ctina has joined #openstack-glance | 12:44 | |
| *** edmondsw has joined #openstack-glance | 12:47 | |
| *** exploreshaifali has quit IRC | 12:48 | |
| *** gberginc has quit IRC | 12:49 | |
| *** marcusvrn has quit IRC | 12:49 | |
| *** chlong has joined #openstack-glance | 12:52 | |
| *** GB21 has joined #openstack-glance | 12:56 | |
| *** GB21_ has joined #openstack-glance | 12:56 | |
| *** GB21_ has quit IRC | 12:56 | |
| *** dims has joined #openstack-glance | 13:01 | |
| *** dims_ has quit IRC | 13:02 | |
| *** haomaiwang has joined #openstack-glance | 13:03 | |
| openstackgerrit | Darja Shakhray proposed openstack/glance: Artifacts are now properly filtered by dict props https://review.openstack.org/207374 | 13:06 |
| *** dims_ has joined #openstack-glance | 13:15 | |
| *** exploreshaifali has joined #openstack-glance | 13:16 | |
| *** dims has quit IRC | 13:17 | |
| *** changbl has joined #openstack-glance | 13:19 | |
| *** julim has joined #openstack-glance | 13:24 | |
| edmondsw | kragniz flaper87 jokke_ this should be ready for another look: https://review.openstack.org/#/c/203242/ | 13:27 |
| *** tpeoples has joined #openstack-glance | 13:27 | |
| openstackgerrit | Alexander Tivelkov proposed openstack/glance: Fixed version unequality artifact filtering https://review.openstack.org/214185 | 13:31 |
| *** dims_ has quit IRC | 13:31 | |
| openstackgerrit | Mike Fedosin proposed openstack/glance-specs: Swift driver with multithreading support https://review.openstack.org/207075 | 13:31 |
| *** dims has joined #openstack-glance | 13:31 | |
| openstackgerrit | Alexander Tivelkov proposed openstack/glance: Fixed version unequality artifact filtering https://review.openstack.org/214185 | 13:31 |
| *** ayoung has joined #openstack-glance | 13:31 | |
| kragniz | edmondsw: thanks, I'll take a look when I'm near my computer | 13:33 |
| edmondsw | tx | 13:33 |
| *** ankit_ag has quit IRC | 13:38 | |
| openstackgerrit | Mike Fedosin proposed openstack/glance-specs: Swift driver with multithreading support https://review.openstack.org/207075 | 13:39 |
| *** GB21 has quit IRC | 13:40 | |
| *** ducttape_ has joined #openstack-glance | 13:40 | |
| *** dims has quit IRC | 13:40 | |
| *** changbl has quit IRC | 13:41 | |
| openstackgerrit | Alexander Tivelkov proposed openstack/glance: Fixed version unequality artifact filtering https://review.openstack.org/214185 | 13:43 |
| *** dims has joined #openstack-glance | 13:43 | |
| *** smatzek has quit IRC | 13:45 | |
| *** pdb is now known as pbourke | 13:50 | |
| *** tpeoples has quit IRC | 13:50 | |
| *** pbourke has quit IRC | 13:51 | |
| *** pbourke has joined #openstack-glance | 13:51 | |
| *** bkopilov has joined #openstack-glance | 13:52 | |
| *** marcusvrn_ has joined #openstack-glance | 13:52 | |
| *** haomaiwang has quit IRC | 14:01 | |
| *** haomaiwang has joined #openstack-glance | 14:02 | |
| *** abhishekk has quit IRC | 14:02 | |
| *** DuncanT has joined #openstack-glance | 14:03 | |
| *** dims has quit IRC | 14:04 | |
| *** dims has joined #openstack-glance | 14:05 | |
| *** smatzek has joined #openstack-glance | 14:06 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 14:11 | |
| openstackgerrit | Alexander Tivelkov proposed openstack/glance: Fixed version unequality artifact filtering https://review.openstack.org/214185 | 14:14 |
| *** gberginc has joined #openstack-glance | 14:17 | |
| *** peristeri has quit IRC | 14:20 | |
| *** annegentle has quit IRC | 14:20 | |
| *** zul has joined #openstack-glance | 14:29 | |
| flaper87 | o/ | 14:30 |
| nikhil_k_ | o/ | 14:30 |
| flaper87 | rosmaita: don't run | 14:30 |
| flaper87 | I know you're there | 14:30 |
| flaper87 | :P | 14:30 |
| rosmaita | i am here | 14:30 |
| flaper87 | nikhil_k_: rosmaita sigmavirus24 https://review.openstack.org/#/c/188388/ | 14:31 |
| flaper87 | rosmaita: the last PS just adds some API examples of how it'd work from a user perspective | 14:31 |
| *** sdake has joined #openstack-glance | 14:31 | |
| nikhil_k_ | you know I feel discussing overarching questions on this seems more important than the details.. | 14:31 |
| rosmaita | flaper87: great ... i need a few min to read through | 14:31 |
| flaper87 | rosmaita: okis | 14:32 |
| nikhil_k_ | so, the way I looked at it is this. not sure if I get it correctly: | 14:32 |
| flaper87 | nikhil_k_: yup, sounds go to me | 14:32 |
| *** tpeoples has joined #openstack-glance | 14:32 | |
| nikhil_k_ | 1. User uploads image-data, expects image to go to active. we either create a new status or state the expected behavior may change with deployment | 14:34 |
| nikhil_k_ | the second one doesn't fit the interop purposes truly but might be good to ask wider audience | 14:35 |
| flaper87 | mmh, you mean the user expects the image to go active after going through all the tasks ? | 14:35 |
| nikhil_k_ | first one would change the expected behavior for image-creates | 14:35 |
| nikhil_k_ | right | 14:35 |
| nikhil_k_ | because if we set it to active | 14:35 |
| *** dims_ has joined #openstack-glance | 14:35 | |
| nikhil_k_ | and then chance the container and disk format | 14:36 |
| nikhil_k_ | change* | 14:36 |
| flaper87 | Who changes the expected behavior? | 14:36 |
| flaper87 | We ? or OPs? | 14:36 |
| nikhil_k_ | it's not truly active | 14:36 |
| nikhil_k_ | I am looking at this from a API stability POV | 14:36 |
| flaper87 | ok, that helps | 14:36 |
| *** spzala has joined #openstack-glance | 14:36 | |
| flaper87 | So, may I describe a scenario? | 14:36 |
| nikhil_k_ | Either the API scrutiny is unnecessarily high or there are some real use cases for this | 14:37 |
| flaper87 | One related to the use-case I've been using as an example | 14:37 |
| nikhil_k_ | (this == keeping expected API behavior) | 14:37 |
| *** dims has quit IRC | 14:37 | |
| * nikhil_k_ listens | 14:37 | |
| *** sdake_ has joined #openstack-glance | 14:37 | |
| flaper87 | coolio | 14:37 |
| flaper87 | So, OPs actions: | 14:37 |
| flaper87 | 1) The OP creates a trigger task that will convert images to raw format when they are uploaded | 14:38 |
| flaper87 | Users: | 14:38 |
| flaper87 | 1) User creates an image | 14:38 |
| flaper87 | 2) User uploads the image-data | 14:38 |
| flaper87 | 3) The image is uploaded, tasks for the "image.upload" action are loaded and executed | 14:39 |
| flaper87 | 4) After tasks are executed the image is set to active | 14:39 |
| flaper87 | 5) Profit? | 14:39 |
| flaper87 | :P | 14:39 |
| flaper87 | From a user perspective, I think the API is still behaving as expected | 14:39 |
| rosmaita | flaper87: so this is 1 image created, or 2? | 14:40 |
| flaper87 | Tasks could fail, yes. But that's true for a world w/o tasks too | 14:40 |
| flaper87 | rosmaita: 1 image | 14:40 |
| nikhil_k_ | a traditional upload operation would put image in active | 14:40 |
| flaper87 | It's not a clone but a conversion | 14:40 |
| *** sdake has quit IRC | 14:40 | |
| *** exploreshaifali has quit IRC | 14:40 | |
| flaper87 | nikhil_k_: mmh, I see what you mean | 14:40 |
| flaper87 | sorry, it took me a bit | 14:41 |
| nikhil_k_ | the workflow for image upload changes | 14:41 |
| nikhil_k_ | cool | 14:41 |
| flaper87 | How would you feel about an intering status? | 14:41 |
| flaper87 | interin | 14:41 |
| flaper87 | not even sure if that's a word | 14:41 |
| flaper87 | :P | 14:41 |
| flaper87 | For example, 'processing' | 14:41 |
| flaper87 | 'baking' | 14:42 |
| flaper87 | :P | 14:42 |
| rosmaita | what about only using this on import tasks? then it's just part of the import workflow | 14:42 |
| flaper87 | 'in-da-oven' | 14:42 |
| *** spzala has quit IRC | 14:42 | |
| nikhil_k_ | iterim might be what you intended but that's what I was thinking if that would become a necessity | 14:42 |
| flaper87 | nikhil_k_: yup, that's what I meant. Thanks | 14:42 |
| nikhil_k_ | rosmaita: :) | 14:42 |
| *** fnordahl has quit IRC | 14:42 | |
| *** vijendar has joined #openstack-glance | 14:42 | |
| flaper87 | rosmaita: that doesn't really solve the problem. What I keep hearing from OPs is that the import tasks and the tasks we have nwo are not useful at all if they have to run them manually | 14:43 |
| nikhil_k_ | no uploads allowed, ever! :P :P | 14:43 |
| rosmaita | flaper87: but they don't have to run them manually, the end-user does it | 14:43 |
| flaper87 | Uploading an image is probably the most common way to create an image in Glance | 14:43 |
| nikhil_k_ | flaper87: I can see your point on usability | 14:43 |
| flaper87 | rosmaita: right but the end user doesn't care if the OP has ceph or not or whether raw would be a better format | 14:43 |
| *** annegentle has joined #openstack-glance | 14:44 | |
| rosmaita | flaper87: right, so that's why the op constructs an import workflow to get the data into the appropriate format | 14:44 |
| rosmaita | i may be thinking public cloud too much here | 14:44 |
| flaper87 | rosmaita: yup :P | 14:44 |
| flaper87 | also | 14:44 |
| flaper87 | the import workflow will likely create a new image in Glance with a different ID | 14:45 |
| flaper87 | which is not idea if you created the image and got the ID from there | 14:45 |
| flaper87 | And we currently just allow HTTP in our import workflow, which means the OP has to find a way to set that up | 14:45 |
| flaper87 | it's just not friendly | 14:45 |
| flaper87 | I know leseb is one of the OPs interested in this | 14:46 |
| flaper87 | hope he's around to provide better feedback | 14:46 |
| flaper87 | and other use-cases | 14:46 |
| flaper87 | Here's another idea: | 14:46 |
| flaper87 | What if instead of messing with the specific actions - image.create, image.saved, image.blah - we do it on status changes ? | 14:47 |
| nikhil_k_ | flaper87: one thing this changes drastically is ability to mutate base properties. and now I understand why you were complaining about domain model :P | 14:47 |
| *** GB21 has joined #openstack-glance | 14:47 | |
| flaper87 | nikhil_k_: :D | 14:47 |
| flaper87 | nikhil_k_: that is true. However, it does it internally in glance and it doesn't expose that to the users | 14:47 |
| sigmavirus24 | lol | 14:48 |
| nikhil_k_ | yeah, it would be a change. Also, it can be a point of exploit for advanced complicated tasks that do mutation regularly. which I dm not sure if is good or bad thing | 14:48 |
| *** harshs has joined #openstack-glance | 14:49 | |
| sigmavirus24 | so I think consistency here will be hard to achieve | 14:49 |
| nikhil_k_ | well expressed | 14:49 |
| flaper87 | agreed | 14:49 |
| sigmavirus24 | also, I wonder what the impact of this would be on glance servers under a heavy load | 14:49 |
| nikhil_k_ | that was my 2. :)) | 14:50 |
| sigmavirus24 | let's say an operator sets up an auto-triggered task to convert new images | 14:50 |
| rosmaita | it would kill them | 14:50 |
| sigmavirus24 | right | 14:50 |
| flaper87 | it'll require workers but that's why we're consuming taskflow to begin with | 14:50 |
| sigmavirus24 | right | 14:50 |
| nikhil_k_ | with no handle to queue, may be? | 14:50 |
| sigmavirus24 | rosmaita: but even if it doesn't topple the cloud (which flaper87 would assure wouldn't happen) | 14:50 |
| nikhil_k_ | we may have issues with scheduling for this on a large scale deployment | 14:50 |
| sigmavirus24 | I'm imagining a large queue of images that users can't use | 14:51 |
| *** julim_ has joined #openstack-glance | 14:51 | |
| nikhil_k_ | yup | 14:51 |
| sigmavirus24 | prior to auto-triggered tasks, users can upload an image and use it almost immediately | 14:51 |
| sigmavirus24 | now, they ahve to wait for the tasks to complete before being able to use it | 14:51 |
| sigmavirus24 | which makes the case for 2 images but 2 images is a storage nightmare possibly | 14:51 |
| rosmaita | it would make upload similar to image import | 14:51 |
| nikhil_k_ | or replace it | 14:52 |
| flaper87 | yup but it really depends on the tasks configured | 14:52 |
| sigmavirus24 | eww | 14:52 |
| flaper87 | and we go back to sigmavirus24's comment w.r.t consistency | 14:52 |
| nikhil_k_ | or create a hybrid | 14:52 |
| sigmavirus24 | replacing it seems like the worst option | 14:52 |
| sigmavirus24 | then the image data is not immutable | 14:52 |
| sigmavirus24 | if someone launches an instance from that image prior to being replaced and then resizes, what would that do nikhil_k_ ? | 14:52 |
| nikhil_k_ | that's true even today with tasks but here it's upload+task | 14:52 |
| flaper87 | I've always thought of the image data to be immutable after it's been stored | 14:53 |
| flaper87 | because that's the only moment when we can say: "We've got this" | 14:53 |
| nikhil_k_ | flaper87: yeah, that's one point people don't seem to agree on and scares me with our mission statement as we cannot gurantee upload delivery consistency | 14:53 |
| nikhil_k_ | sigmavirus24: I think flaper87 is proposing a new state for such cases (analogous to 'interim') | 14:54 |
| sigmavirus24 | so, I keep getting mixed messages, but would glance_store in Nova allow people to upload directly to a backing store? | 14:54 |
| nikhil_k_ | may be not on the spec | 14:54 |
| sigmavirus24 | nikhil_k_: yeah, I grok that | 14:54 |
| nikhil_k_ | sigmavirus24: that's still a open question | 14:54 |
| flaper87 | sigmavirus24: nope | 14:54 |
| *** zul has quit IRC | 14:54 | |
| nikhil_k_ | some people like and some don't | 14:54 |
| sigmavirus24 | I'm just imaging queued being in the task queue, and interim being part of a task | 14:54 |
| sigmavirus24 | flaper87: okay so that rules out that nastiness | 14:55 |
| nikhil_k_ | sigmavirus24: even today if the direct location url is exposed, one may be able to set the image location fter uploading the data from Nova. please correct me if I am wrong | 14:55 |
| *** mtanino has joined #openstack-glance | 14:55 | |
| flaper87 | tbh, I don't think the 'processing' status between queued and active is bad | 14:56 |
| sigmavirus24 | flaper87: I don't either | 14:56 |
| sigmavirus24 | It's very transparent which I like | 14:56 |
| sigmavirus24 | but | 14:56 |
| openstackgerrit | Niall Bunting proposed openstack/glance: Tasks now try the schema before the database https://review.openstack.org/214232 | 14:56 |
| sigmavirus24 | So, rosmaita represents our public cloud needs | 14:56 |
| sigmavirus24 | I'm aware of very large private clouds | 14:56 |
| nikhil_k_ | sigmavirus24: no, I think we need a new image state as the regular upload workflow would change and the users expecting a image to go active after upload completes would need to know some details on what's happening to their image | 14:56 |
| flaper87 | I want to represent our chaos | 14:56 |
| flaper87 | please, please, please | 14:56 |
| sigmavirus24 | lol | 14:57 |
| sigmavirus24 | right | 14:57 |
| sigmavirus24 | I think private cloud operators would like this | 14:57 |
| sigmavirus24 | 100% agree with you flaper87 | 14:57 |
| *** pbourke_ has joined #openstack-glance | 14:57 | |
| sigmavirus24 | I guess we'll need someway to allow auto-triggered task authors to skip certain events | 14:57 |
| sigmavirus24 | Like "see X in proposed task; bail out early" and I can't recall if we give them enough information | 14:58 |
| flaper87 | yeah | 14:58 |
| sigmavirus24 | And I'm not even sure what those conditions might be | 14:58 |
| sigmavirus24 | So | 14:58 |
| flaper87 | nikhil_k_: the way I've seen this done in Nova/Cinder is by having a task status | 14:58 |
| flaper87 | image state is 'processing' and the task status is 'converting' | 14:58 |
| nikhil_k_ | that people have strongly suggested against | 14:58 |
| sigmavirus24 | heh | 14:58 |
| flaper87 | I know, I'm just throwing it out there :P | 14:59 |
| nikhil_k_ | I think the Nova task status is racy and inconsistent is what I was told | 14:59 |
| flaper87 | The status being racy has nothing to do with the use-case but the code :P | 14:59 |
| *** pbourke has quit IRC | 14:59 | |
| *** julim has quit IRC | 14:59 | |
| *** wznoinsk has quit IRC | 14:59 | |
| nikhil_k_ | flaper87: that was my intial idea but it's very risky to introduce that status as it will be confusing | 14:59 |
| flaper87 | That said, if we aim to have parallel tasks, that's probably not a good idea | 14:59 |
| nikhil_k_ | if tasks are independent of the image workflow, it can be okay. | 15:00 |
| flaper87 | I just don't like the idea of having to maintain a gazillion of states for the images because tasks are independant | 15:00 |
| nikhil_k_ | because we won't expose the image until all sub-tasks are donw | 15:00 |
| flaper87 | and OPs can have their own downstream tasks | 15:00 |
| johnthetubaguy | we have quite a few "fake" status codes in Nova, not getting the context from the scroll back very quickly right now | 15:00 |
| *** haomaiwang has quit IRC | 15:01 | |
| nikhil_k_ | johnthetubaguy: we are discussing if image.status + image.task_status would be good idea or not | 15:01 |
| flaper87 | johnthetubaguy: tl;dr: How should we keep the status of tasks that are in execution without messing with the image status | 15:01 |
| flaper87 | ? | 15:01 |
| johnthetubaguy | ah I see | 15:01 |
| nikhil_k_ | and I was suggesting that it will be racy and inconsistent | 15:01 |
| flaper87 | nikhil_k_: damn, you were faster | 15:01 |
| *** dims has joined #openstack-glance | 15:01 | |
| flaper87 | :P | 15:01 |
| nikhil_k_ | :P | 15:01 |
| *** haomaiwang has joined #openstack-glance | 15:02 | |
| johnthetubaguy | so honestly, we don't do this well right now, we have a vague plan for "tasks" to change how we do all that (to support parallel, etc) | 15:02 |
| nikhil_k_ | just borrowing the idea from nova vm.state and some cinder state | 15:02 |
| johnthetubaguy | but ignoring that I can share what we do now | 15:02 |
| rosmaita | i really like the current model, where task status is completely independent of image status | 15:02 |
| johnthetubaguy | basically we have added substates that are hidden from the API users | 15:02 |
| johnthetubaguy | to track specific bits of progress that are important | 15:03 |
| *** dims_ has quit IRC | 15:03 | |
| johnthetubaguy | an example is where we spot the compute node got restarted half way through an operation | 15:03 |
| johnthetubaguy | but we keep the API states looking the same they always did (like a state alias) | 15:03 |
| johnthetubaguy | would that help you out with what you need? | 15:03 |
| flaper87 | I believe it would | 15:04 |
| nikhil_k_ | johnthetubaguy: we are discussing https://review.openstack.org/#/c/188388/6/specs/liberty/task-automatic-triggering.rst | 15:04 |
| *** julim has joined #openstack-glance | 15:04 | |
| nikhil_k_ | so the shadow state would give information about image being processed | 15:04 |
| nikhil_k_ | but the processing would have processing.upload, processing.convert etc | 15:04 |
| nikhil_k_ | I am just saying it could be that and not would be that | 15:05 |
| nikhil_k_ | situation is hypothetical for now | 15:05 |
| johnthetubaguy | now one was is to convert every API call to a task, so its just dealt with by tasks | 15:05 |
| johnthetubaguy | oh wait, I see, custom triggers | 15:06 |
| flaper87 | johnthetubaguy: yup | 15:06 |
| flaper87 | :D | 15:06 |
| nikhil_k_ | I think we have a issue in tasks currently, the image delivery is underpromised by task while the regular upload somewhat overpromises | 15:06 |
| *** julim_ has quit IRC | 15:06 | |
| johnthetubaguy | yeah, I mean I quite like the upload API creating a task to do its work | 15:07 |
| nikhil_k_ | tasks don't expose image completely until finished and regular api sets even the immutable meta during the upload | 15:07 |
| flaper87 | nikhil_k_: but that depends on the task too, really. | 15:08 |
| nikhil_k_ | johnthetubaguy: I can see how this would solve the defcore issue but it may open more doors. some which we discussed earlier :P | 15:08 |
| flaper87 | ok, lets stick to the convertion example | 15:08 |
| nikhil_k_ | oops, I am missing a webinar... | 15:08 |
| flaper87 | to avoid confusions | 15:08 |
| nikhil_k_ | lemme do some multi-tasking | 15:08 |
| * nikhil_k_ slows down typing and looks more at chats now | 15:09 | |
| flaper87 | I'm wondering if we can start narrowing the discussion down to more specific things | 15:09 |
| flaper87 | what I mean is, can we start answering some questions like: | 15:09 |
| johnthetubaguy | flaper87: +1 to that | 15:09 |
| *** zul has joined #openstack-glance | 15:09 | |
| flaper87 | 1) Do we agree this is useful and we want it? | 15:09 |
| flaper87 | 2) Do we agree on separating the models? | 15:10 |
| flaper87 | 3) Do we agree on having the interim status? or do we prefer a different workflow? | 15:10 |
| johnthetubaguy | oh, so I am thinking, its an easer conversation if you just try to fix one of the use cases, like implementing security checks on images or converting images | 15:10 |
| flaper87 | That way we can stop talking hypotetically and start talking with more realistic cases (like the ones proposed) | 15:11 |
| johnthetubaguy | yeah, I would just talk about something concrete, it will be easier | 15:11 |
| flaper87 | johnthetubaguy: yeah, that's where I was getting at :P | 15:11 |
| johnthetubaguy | flaper87: cool | 15:11 |
| *** wznoinsk has joined #openstack-glance | 15:12 | |
| flaper87 | The one proposed in the spec is automatic image conversion and it's the one I'm more interested at because that's the one we've gotten more feedback from OPs | 15:12 |
| flaper87 | I think I mentioned a couple more but well... | 15:12 |
| flaper87 | In Vancouver, over email, over IRC and even over twitter (no kidding) | 15:12 |
| johnthetubaguy | so the problem I have interest in is how do you unify the two different upload methods, or maybe its really, how do you do automatic image conversion using both | 15:12 |
| flaper87 | I'd say yes: The way I've envision this is by triggering tasks based on actions. | 15:13 |
| flaper87 | For example: image.upload would trigger tasks that need to convert the image to some other format | 15:13 |
| flaper87 | Those tasks would run before the image is saved in the store | 15:13 |
| flaper87 | and that's something that happens regardless of the way you import the image | 15:13 |
| rosmaita | you have 2 kinds of uploads, snapshots from nova and end-user uploads | 15:14 |
| rosmaita | prob don't need conversion on snapshots | 15:14 |
| flaper87 | It's still an upload | 15:14 |
| flaper87 | Glance doesn't care about that | 15:14 |
| rosmaita | well, maybe glance should ... that's why we have an import task | 15:14 |
| flaper87 | it gets the image, it checks the format, if it's not the expected format it converts it | 15:14 |
| flaper87 | rosmaita: mmh, what would be the benefit of differentiating these 2 uploads? | 15:15 |
| johnthetubaguy | so the conversion would be based on the type being uploaded, so more than likely nova would just upload a type that doesn't have a conversion trigger registered, its probably just coming out in the wash | 15:15 |
| rosmaita | well, import could have a different workflow | 15:15 |
| nikhil_k_ | yep | 15:16 |
| sigmavirus24 | like this is all up to operators too | 15:16 |
| johnthetubaguy | rosmaita: but why have two API for the same operation? | 15:16 |
| flaper87 | sigmavirus24: exactly | 15:16 |
| *** fnordahl has joined #openstack-glance | 15:16 | |
| sigmavirus24 | so operators absolutely could register a task for import that's different than upload | 15:16 |
| sigmavirus24 | which would only hurt their users | 15:16 |
| rosmaita | johnthetubaguy: because they're not really the same operation | 15:16 |
| flaper87 | I don't understand why they are different | 15:17 |
| flaper87 | what makes them different uploads? | 15:17 |
| rosmaita | trusted source vs. untrusted source | 15:17 |
| flaper87 | TBH, the way I see the import task is as an async copy_from | 15:17 |
| johnthetubaguy | there is a level of trust difference, but thats really an RBAC issue, in my head | 15:18 |
| sigmavirus24 | trusted source (e.g., nova), untrusted source (e.g., users of a public cloud) | 15:18 |
| flaper87 | rosmaita: but you can use RBAC or different users for that | 15:18 |
| rosmaita | not really | 15:18 |
| flaper87 | not saying it's not a real issue | 15:18 |
| johnthetubaguy | you could add a nova service token to the import so you can skip some import checks, for example | 15:18 |
| flaper87 | it's just that it's still an upload operation | 15:18 |
| sigmavirus24 | glance's current policy is a bit ... simplistic as it is | 15:18 |
| sigmavirus24 | the number of rules we have is very tiny and is more based around actions | 15:19 |
| sigmavirus24 | if actions had sub-actions we'd be more capable of having finer-grained policies that would eliminate the need for separate endpoints | 15:19 |
| flaper87 | sigmavirus24: sure | 15:19 |
| flaper87 | ok ok, but lets not digress | 15:19 |
| flaper87 | :D | 15:19 |
| flaper87 | can we go back to tasks triggering ? | 15:19 |
| flaper87 | :) | 15:19 |
| *** tsekiyama has joined #openstack-glance | 15:19 | |
| johnthetubaguy | import vs upload is I think hitting on the image conversion problem | 15:20 |
| johnthetubaguy | I think they should be a single API, that both support conversion, if required | 15:20 |
| rosmaita | well, the upload is synchronous whereas import is not | 15:20 |
| rosmaita | that's kind of a big difference | 15:21 |
| johnthetubaguy | rosmaita: thats just because there are two APIs right? | 15:21 |
| flaper87 | johnthetubaguy: tbh, they all go down the same path when it comes to storing the data | 15:21 |
| flaper87 | johnthetubaguy: yes, one uses the tasks api | 15:21 |
| flaper87 | the other is just the g'old upload | 15:21 |
| johnthetubaguy | flaper87: they don't quite though, some move data, some just do a metadata operation, as I understood it | 15:21 |
| johnthetubaguy | rosmaita: my thinking was the upload operation completes, but the image is still not Active until its completed the conversion | 15:22 |
| johnthetubaguy | rosmaita: now thats a semantic API change, that causes problems, but it does unify the two operations | 15:22 |
| johnthetubaguy | the upload operation could register a task as a way to report its progress, even if its not actually started via an explicit task API | 15:23 |
| johnthetubaguy | in those cases, you have a single way to register the conversion, I think | 15:23 |
| johnthetubaguy | (waves hands about widely) | 15:23 |
| nikhil_k_ | haha | 15:23 |
| nikhil_k_ | yup, so to me this is a battle of use-case priority | 15:24 |
| nikhil_k_ | and I feel this needs to tie into the interop discussion for a good solid API | 15:24 |
| johnthetubaguy | nikhil_k_: which use cases are competing here? | 15:24 |
| nikhil_k_ | tasks make it that you can deploy clouds that make the image useable across such | 15:24 |
| rosmaita | nova and untrusted end user | 15:24 |
| rosmaita | johnthetubaguy: ^^ the 2 use cases | 15:25 |
| nikhil_k_ | and api consistency make it such that scripted users benefit from this | 15:25 |
| * flaper87 back | 15:25 | |
| flaper87 | sorry | 15:25 |
| nikhil_k_ | which one is more introp and interop-friendly | 15:26 |
| nikhil_k_ | that would be a deciding factor to take the auto-task trigger approach yes or no route, I think | 15:26 |
| flaper87 | mmh, mind elaborate on that? | 15:26 |
| flaper87 | I feel like I lost something | 15:27 |
| nikhil_k_ | because, both these approaches seem like interop centric | 15:27 |
| nikhil_k_ | flaper87: which part, I may have thrown out a digest of info in a small words so if you can point out which one to elaborate more it would be useful | 15:27 |
| flaper87 | "nikhil_k_ tasks make it that you can deploy clouds that make the image useable across such" | 15:28 |
| nikhil_k_ | ok | 15:28 |
| nikhil_k_ | so why import and upload needs to be two different operation is governed by that ^ | 15:28 |
| nikhil_k_ | if you have an explicit independent task API, you can use your image across different deployments using the tasks provided | 15:29 |
| *** dims has quit IRC | 15:29 | |
| johnthetubaguy | except you could make a single API that does both, and standardise around that | 15:29 |
| flaper87 | I'm honestly missing the point why the automatic task triggering interfeers with that | 15:29 |
| nikhil_k_ | advanced operations can be performed and fine tuning to optimized it using the operator deployed scripts is possible | 15:29 |
| flaper87 | you could even have a glance-api node that is used internally without any tasks configured | 15:29 |
| flaper87 | to which nova would upload stuff | 15:29 |
| johnthetubaguy | but what does the end user see, they just see something that gets uploaded correctly and works in that cloud? | 15:30 |
| *** dims has joined #openstack-glance | 15:30 | |
| johnthetubaguy | end user, I mean the API user doing the upload | 15:30 |
| flaper87 | yes | 15:30 |
| nikhil_k_ | johnthetubaguy: but end users are also expecting a standardized workflow on the image right? | 15:30 |
| nikhil_k_ | like today, the definite result of an upload is active image that is ready to use | 15:31 |
| johnthetubaguy | nikhil_k_: they don't need to see the workflow, thats an administrator thing, they are just uploading a supported image | 15:31 |
| nikhil_k_ | given all the correct params ar provided and data is good | 15:31 |
| flaper87 | johnthetubaguy: ++ | 15:31 |
| nikhil_k_ | yup, the key is supported image | 15:31 |
| nikhil_k_ | there's no way to know what is a supported image if the upload workflow can be dynamic | 15:32 |
| rosmaita | this is starting to look like a deployment nightmare to me | 15:32 |
| johnthetubaguy | an upload could go to the Error state if image validation fails after the upload API has finished right? | 15:32 |
| * flaper87 will have to step out in 10 mins but this has been very useful conversation | 15:32 | |
| * rosmaita need to go too, can we set up another meeting to discuss again? | 15:32 | |
| johnthetubaguy | rosmaita: why, you just get Nova to add a service token, and skip steps only if you get a valid service token (wsgi middlewear already supports all that) | 15:32 |
| nikhil_k_ | johnthetubaguy: I think we are asking for more than validation. like conversion on image data after upload is successful | 15:32 |
| rosmaita | i'm worried about load on the nodes | 15:33 |
| johnthetubaguy | nikhil_k_: its the same difference though, I think | 15:33 |
| flaper87 | I don't think it's a real deployment mess. TBH, OPs are not even required to have tasks if they don't want it | 15:33 |
| johnthetubaguy | rosmaita: it should not be done on the API nodes, thats fine | 15:33 |
| rosmaita | our current import tasks don't do a lot, and practically kill th3em as it is | 15:33 |
| johnthetubaguy | rosmaita: yes, thats crazy, you need workers for this that balance their load | 15:33 |
| flaper87 | rosmaita: nikhil_k_ johnthetubaguy will you guys be around tomorrow at this same time? | 15:34 |
| flaper87 | would love to keep going with this discussion | 15:34 |
| * johnthetubaguy checks... | 15:34 | |
| flaper87 | johnthetubaguy: fwiw, we're already using taskflow, which should help with the workers aspect of this | 15:34 |
| rosmaita | flaper87: what time do you mean UTC? | 15:34 |
| nikhil_k_ | johnthetubaguy: today users set disk and container format on the image and it's immutable as it's a known deployment scenario. with upload users would be kind of in the blind until after image has been processed by a task identifier that's unknown and converted into a completely different format | 15:34 |
| flaper87 | 14/15 UTC ? | 15:34 |
| nikhil_k_ | works for me flaper87 | 15:35 |
| flaper87 | nikhil_k_: awesome, thanks | 15:35 |
| rosmaita | 14:00 to 15:00 is good | 15:35 |
| flaper87 | awesome | 15:35 |
| nikhil_k_ | flaper87: the reason for delay in input is this, now you know. it's a complex change that involves affect to other parties | 15:35 |
| nikhil_k_ | so, good to have people around for interactive convo | 15:35 |
| johnthetubaguy | yeah, ping me, I should be able to jump in | 15:35 |
| flaper87 | sounds good | 15:36 |
| *** DuncanT has quit IRC | 15:36 | |
| *** marcusvrn_ has quit IRC | 15:36 | |
| nikhil_k_ | personally, I think we could go either way. Just need to know the definite priority use case for this change. | 15:36 |
| nikhil_k_ | defining* | 15:37 |
| nikhil_k_ | something that ties into the wider openstack goal | 15:37 |
| flaper87 | I had asked the folks that pinged me through different mediums to chime in on the spec with their +1's and whatnot | 15:37 |
| flaper87 | to be entirely honest | 15:37 |
| rosmaita | i think the question is, do we complicate tasks, or do we complicate images? | 15:37 |
| rosmaita | flaper87: i think the functionality is definitely important, jsut a matter of where to put it | 15:38 |
| flaper87 | I think our tasks feature without auto-triggering is really limited and close to not having enough use cases to make it worth maintaining | 15:38 |
| flaper87 | rosmaita: totally agreed | 15:38 |
| nikhil_k_ | flaper87: I see, wondering if they would be willing to shed some linght on the usecases too then | 15:38 |
| flaper87 | I'll do my homework and try to think about a different solution | 15:38 |
| nikhil_k_ | rosmaita: I think tasks are bound to be complicated irrespective of images | 15:39 |
| nikhil_k_ | today it's end user tasks | 15:39 |
| flaper87 | nikhil_k_: leseb is one of them, I think I can find the guy that pinged me on twitter and I'll comment on the spec too | 15:39 |
| nikhil_k_ | tomorrow we will have RBAC on tasks too if we need to differential between end user triggers and admin triggered | 15:39 |
| nikhil_k_ | flaper87: thank you!! | 15:39 |
| nikhil_k_ | rosmaita: also, we will have to identify which task is being run on the upload trigger (provide info to end user) | 15:40 |
| flaper87 | On a closing note, I kinda think tasks should be an internal thing and definitely an undercloud feature | 15:40 |
| nikhil_k_ | oops.. | 15:41 |
| nikhil_k_ | so that they can know the error message and some result info | 15:42 |
| nikhil_k_ | flaper87: sure | 15:43 |
| nikhil_k_ | if that's what the use-cases are of priority | 15:43 |
| *** gberginc has quit IRC | 15:43 | |
| nikhil_k_ | I kinda like what jcook says, we are developing towards use-cases / business needs. So, the differentiating factor here is what is that and what does openstack want | 15:44 |
| *** sayali has quit IRC | 15:45 | |
| *** dims has quit IRC | 15:45 | |
| *** zul has quit IRC | 15:46 | |
| *** spzala has joined #openstack-glance | 15:46 | |
| *** dims has joined #openstack-glance | 15:47 | |
| *** groen692 has quit IRC | 15:54 | |
| openstackgerrit | Alexander Tivelkov proposed openstack/glance: Fixed version unequality artifact filtering https://review.openstack.org/214185 | 15:56 |
| openstackgerrit | Alexander Tivelkov proposed openstack/glance: Fixed version unequality artifact filtering https://review.openstack.org/214185 | 15:58 |
| *** marcusvrn_ has joined #openstack-glance | 15:58 | |
| *** haomaiwang has quit IRC | 16:01 | |
| *** haomaiwang has joined #openstack-glance | 16:01 | |
| *** sayali has joined #openstack-glance | 16:03 | |
| *** haomaiwang has quit IRC | 16:04 | |
| openstackgerrit | Kairat Kushaev proposed openstack/glance: Show properties with empty values https://review.openstack.org/214255 | 16:05 |
| *** gberginc has joined #openstack-glance | 16:06 | |
| *** jecarey has joined #openstack-glance | 16:06 | |
| jcook | you said use cases / business needs, so I thought it might be useful to share this: https://drive.google.com/folderview?id=0BxtM4AiszlEyfm9UTW5LMEQ5cUhHbmFsSkd5WFNfdTMwVFIwRUM1TVFXSHhhWHl6VHlpRzg&usp=sharing | 16:07 |
| jcook | Product Working Group is at mid-cycle this week and they are trying to get more involved in helping drive use cases upstream | 16:07 |
| *** jistr has quit IRC | 16:08 | |
| openstackgerrit | Chris Buccella proposed openstack/python-glanceclient: Allow using image name for image-show using v2 https://review.openstack.org/214258 | 16:08 |
| sigmavirus24 | jcook: that'd be awesome for them to help with this kind of stuff | 16:11 |
| *** DuncanT has joined #openstack-glance | 16:12 | |
| jcook | I've gotten involved a bit with the group, but mostly from the standpoint of having performance and reliability tests across various configurations (scale and elasticity) | 16:14 |
| jcook | I know the RPC product managers from Rackspace are pretty involved and there is pretty good representation across companies (Intel, IBM, ...). Not sure if I saw a HP PM or not. I'm new to the community | 16:15 |
| sigmavirus24 | jcook: that's good to know | 16:15 |
| sigmavirus24 | After http://www.rackspace.com/blog/newsarticles/rackspace-collaborates-with-intel-to-accelerate-openstack-enterprise-feature-development-and-adoption/ I'm wondering if Intel has involved itself yet with the Product WG | 16:16 |
| *** exploreshaifali has joined #openstack-glance | 16:18 | |
| *** pbourke_ is now known as pbourke | 16:21 | |
| openstackgerrit | GB21 proposed openstack/glance: Add a soft delete functionality for tasks. https://review.openstack.org/209255 | 16:22 |
| *** exploreshaifali has quit IRC | 16:22 | |
| *** gberginc has quit IRC | 16:28 | |
| *** gberginc has joined #openstack-glance | 16:28 | |
| *** gberginc has quit IRC | 16:28 | |
| *** exploreshaifali has joined #openstack-glance | 16:31 | |
| *** sdake_ is now known as sdake | 16:34 | |
| *** dims has quit IRC | 16:36 | |
| *** dims has joined #openstack-glance | 16:39 | |
| *** annegentle has quit IRC | 16:41 | |
| *** dims has quit IRC | 16:54 | |
| *** dims has joined #openstack-glance | 16:57 | |
| *** sgotliv_ has quit IRC | 16:59 | |
| *** dims_ has joined #openstack-glance | 17:02 | |
| openstackgerrit | Tomoki Sekiyama proposed openstack/glance_store: Implement get, add and delete for cinder store https://review.openstack.org/166414 | 17:03 |
| *** dims has quit IRC | 17:04 | |
| *** markus_z has quit IRC | 17:07 | |
| *** smatzek has quit IRC | 17:08 | |
| *** dims has joined #openstack-glance | 17:08 | |
| *** dims_ has quit IRC | 17:09 | |
| *** dims_ has joined #openstack-glance | 17:11 | |
| *** dims has quit IRC | 17:14 | |
| *** ayoung has quit IRC | 17:17 | |
| *** dims_ has quit IRC | 17:20 | |
| *** boris-42 has quit IRC | 17:20 | |
| *** dims has joined #openstack-glance | 17:20 | |
| *** gberginc has joined #openstack-glance | 17:20 | |
| *** dims has quit IRC | 17:20 | |
| *** dims has joined #openstack-glance | 17:21 | |
| *** MattMan has quit IRC | 17:34 | |
| *** TravT has joined #openstack-glance | 17:41 | |
| hogepodge | nikhil_k_: We're at the operators meetup right now | 17:42 |
| hogepodge | nikhil_k_: there's a lot of talk about glance and feature requests. Think you can be available to join a hangout for feedback? | 17:42 |
| hogepodge | johnthetubaguy: ^^ | 17:42 |
| *** TravT_ has quit IRC | 17:43 | |
| johnthetubaguy | hogepodge: I could answer nova question | 17:43 |
| johnthetubaguy | s | 17:43 |
| *** changbl has joined #openstack-glance | 17:46 | |
| hogepodge | johnthetubaguy: sure, that would be cool. Lots of PTLs in the room, and virtual attendance would make the state of awesome even more awesome | 17:47 |
| nikhil_k_ | hogepodge: I can join | 17:49 |
| *** achanda has joined #openstack-glance | 17:51 | |
| *** harshs_ has joined #openstack-glance | 17:55 | |
| achanda | Hey guys, can I use cinder as a backend for glance in Kilo? | 17:56 |
| *** harshs has quit IRC | 17:56 | |
| *** harshs_ is now known as harshs | 17:56 | |
| *** ayoung has joined #openstack-glance | 18:03 | |
| *** chlong has quit IRC | 18:06 | |
| *** chlong has joined #openstack-glance | 18:09 | |
| *** openstackgerrit_ has quit IRC | 18:11 | |
| cpallares | achanda: no, but hopefully sometime soon with this patch https://review.openstack.org/#/c/166414/ | 18:18 |
| sigmavirus24 | achanda: it's a think you can set in glance.conf, but it's not something anyone would suggest you set ;) | 18:19 |
| sigmavirus24 | (or what cpallares said before I realized she said it better than I did) | 18:19 |
| sigmavirus24 | that patch isn't the only one iirc | 18:20 |
| sigmavirus24 | we also need to add oslo.rootwrap to glance | 18:20 |
| sigmavirus24 | which is ... far from ideal | 18:20 |
| achanda | thanks guys :) | 18:20 |
| sigmavirus24 | There's apparently a better alternative around the corner that bknudson mentioned in #openstack-security last week but I've already forgotten the name | 18:21 |
| achanda | sigmavirus24: that sounds interesting! | 18:26 |
| sigmavirus24 | oslo.rootwrap is ... not secure by default | 18:26 |
| achanda | let me try to look | 18:26 |
| sigmavirus24 | which is why I left the security docs about it on the review that imports oslo.rootwrap into glance | 18:27 |
| achanda | not secure and often slow | 18:27 |
| achanda | we had problems in neutron | 18:27 |
| achanda | ended up using sudo | 18:27 |
| sigmavirus24 | O_O | 18:28 |
| sigmavirus24 | that sounds even worse | 18:28 |
| achanda | I guess the dameon mode will help with slowness though | 18:29 |
| *** ducttape_ has quit IRC | 18:29 | |
| sigmavirus24 | achanda: https://review.openstack.org/#/c/204073/ | 18:29 |
| sigmavirus24 | (from bknudson in #openstack-keystone) | 18:30 |
| achanda | thanks sigmavirus24 | 18:31 |
| sigmavirus24 | You're welcome | 18:31 |
| achanda | about the perf issues: http://lists.openstack.org/pipermail/openstack-dev/2014-March/029017.html | 18:32 |
| sigmavirus24 | achanda: that's over a year old, surely that's still not a problem =P | 18:34 |
| sigmavirus24 | /s | 18:34 |
| achanda | true | 18:34 |
| achanda | honestly, I haven't looked at the daemon mode | 18:35 |
| achanda | need to try that out | 18:35 |
| *** ducttape_ has joined #openstack-glance | 18:47 | |
| nikhil_k_ | thanks jcook | 18:52 |
| jcook | sigmavirus24: I think the moderator / chair of the meeting I was in was someone from intel | 18:54 |
| *** ducttape_ has quit IRC | 18:54 | |
| jcook | sigmavirus24: I think there were at least 2 intel people on list, let me pull up etherpad | 18:54 |
| *** ducttape_ has joined #openstack-glance | 18:55 | |
| jcook | sigmavirus24: yeah, Carol: https://etherpad.openstack.org/p/User_Story_Discussion_1 https://etherpad.openstack.org/p/User_Story_Discussion_2 | 18:56 |
| *** smatzek has joined #openstack-glance | 18:56 | |
| jcook | sigmavirus24: she seems to be leading many of the discussions and organization | 18:56 |
| *** sdake_ has joined #openstack-glance | 18:56 | |
| *** sdake has quit IRC | 18:59 | |
| *** annegentle has joined #openstack-glance | 19:06 | |
| *** e0ne has joined #openstack-glance | 19:07 | |
| *** ayoung has quit IRC | 19:08 | |
| *** smatzek has quit IRC | 19:15 | |
| *** sdake_ is now known as sdake | 19:16 | |
| *** GB21 has quit IRC | 19:21 | |
| *** annegentle has quit IRC | 19:21 | |
| *** achanda has quit IRC | 19:22 | |
| sigmavirus24 | jcook: that's awesome | 19:22 |
| *** annegentle has joined #openstack-glance | 19:26 | |
| *** harshs has quit IRC | 19:34 | |
| *** belmoreira has joined #openstack-glance | 19:36 | |
| *** e0ne has quit IRC | 19:44 | |
| *** belmoreira has quit IRC | 19:51 | |
| *** ayoung has joined #openstack-glance | 19:51 | |
| *** annegentle has quit IRC | 19:57 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 19:57 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 19:58 | |
| *** annegentle has joined #openstack-glance | 20:00 | |
| *** gberginc has quit IRC | 20:04 | |
| *** boris-42 has joined #openstack-glance | 20:19 | |
| flaper87 | nikhil_k_: https://review.openstack.org/#/c/197594/ don't forget | 20:20 |
| *** achanda has joined #openstack-glance | 20:22 | |
| *** achanda has quit IRC | 20:22 | |
| *** achanda has joined #openstack-glance | 20:23 | |
| *** zul has joined #openstack-glance | 20:28 | |
| *** achanda has quit IRC | 20:28 | |
| *** achanda has joined #openstack-glance | 20:28 | |
| *** changbl has quit IRC | 20:40 | |
| *** ctina has quit IRC | 20:49 | |
| openstackgerrit | Abhishek Chanda proposed openstack/glance: Pass CONF to logging setup https://review.openstack.org/214342 | 20:50 |
| *** cdelatte has quit IRC | 20:57 | |
| *** julim has quit IRC | 21:00 | |
| *** zul has quit IRC | 21:00 | |
| *** vijendar has quit IRC | 21:01 | |
| *** vijendar has joined #openstack-glance | 21:01 | |
| *** jecarey has quit IRC | 21:17 | |
| *** zul has joined #openstack-glance | 21:22 | |
| *** ducttape_ has quit IRC | 21:26 | |
| openstackgerrit | Abhishek Chanda proposed openstack/glance: Pass CONF to logging setup https://review.openstack.org/214342 | 21:27 |
| *** ducttape_ has joined #openstack-glance | 21:27 | |
| *** sileht has quit IRC | 21:43 | |
| *** edmondsw has quit IRC | 21:44 | |
| *** zul has quit IRC | 21:49 | |
| cpallares | nikhil_k_: Can I get your input on this? https://review.openstack.org/#/c/197594/ | 21:50 |
| nikhil_k_ | cpallares: surely, today Iam planning to approve it. Just need to note down points for me to amend the spec | 21:51 |
| nikhil_k_ | cpallares: I am doing that for all proposed and merged this cycle | 21:52 |
| nikhil_k_ | will update and merge them in bunch :P | 21:52 |
| cpallares | nikhil_k_: Thanks :) | 21:52 |
| achanda | how well supported or used is glance-replicator in production? | 21:55 |
| achanda | there is not much info on it around... | 21:56 |
| *** wokuma has joined #openstack-glance | 22:08 | |
| *** vijendar has quit IRC | 22:11 | |
| sigmavirus24 | achanda: I have it on my list to fix a ton of bugs | 22:13 |
| sigmavirus24 | But I have not had time to work on glance or anything else usptream lately | 22:13 |
| sigmavirus24 | so, sorry, and you're welcome =P | 22:13 |
| achanda | :) | 22:13 |
| *** harshs has joined #openstack-glance | 22:13 | |
| wokuma | sigmavirus24: Hi Ian, I would like to discuss i18n work I need to do for metadef JSON files...who would be the best person to ask about this in Glance? | 22:14 |
| wokuma | or can i ask you? :) | 22:14 |
| sigmavirus24 | I'm certainly not the most qualified person to ask :D | 22:15 |
| sigmavirus24 | Like, actually I'm probably the least qualified person | 22:15 |
| wokuma | ok...:) | 22:15 |
| sigmavirus24 | nikhil_k_: can easily direct your questions to the right people with better certainty | 22:15 |
| wokuma | ok...thanks. | 22:15 |
| wokuma | nikhil_k:: any tips | 22:16 |
| sigmavirus24 | nikhil_k_: ^ | 22:17 |
| *** annegentle has quit IRC | 22:17 | |
| wokuma | sigmavirus: guess he's not around right now... | 22:18 |
| *** sigmavirus24 is now known as sigmavirus24_awa | 22:18 | |
| *** tpeoples has quit IRC | 22:32 | |
| *** tpeoples has joined #openstack-glance | 22:39 | |
| *** ducttape_ has quit IRC | 22:50 | |
| *** exploreshaifali has quit IRC | 23:07 | |
| *** annegentle has joined #openstack-glance | 23:08 | |
| *** annegentle has quit IRC | 23:09 | |
| *** annegentle has joined #openstack-glance | 23:10 | |
| *** harshs has quit IRC | 23:11 | |
| *** dims_ has joined #openstack-glance | 23:20 | |
| *** achanda has quit IRC | 23:22 | |
| *** dims has quit IRC | 23:22 | |
| *** chlong has quit IRC | 23:29 | |
| *** achanda has joined #openstack-glance | 23:30 | |
| *** achanda has quit IRC | 23:30 | |
| *** achanda has joined #openstack-glance | 23:31 | |
| *** changbl has joined #openstack-glance | 23:31 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!