Thursday, 2015-12-03

*** dims has quit IRC00:04
*** shakamunyi has joined #openstack-glance00:10
*** chlong has quit IRC00:12
*** ducttape_ has quit IRC00:15
*** ninag has joined #openstack-glance00:26
*** RichardRaseley has joined #openstack-glance00:31
*** dims has joined #openstack-glance00:33
RichardRaseleyDoes this ( seem like the same issue reported here ( ?00:36
openstackLaunchpad bug 1476770 in python-glanceclient "_translate_from_glance fails with "AttributeError: id" in grenade" [High,In progress] - Assigned to Flavio Percoco (flaper87)00:36
RichardRaseleyI get that first error in the nova-api logs whenever I try to launch an instance with an image from glance or try to access the image details.00:37
*** rcernin has quit IRC00:37
RichardRaseleyThough, I can successfully do an `openstack image list` and `openstack image show <image_id>` from the CLI and the later returns the ID property of the image without issue.00:38
*** e0ne has joined #openstack-glance00:41
*** kebray has quit IRC00:42
*** ninag has quit IRC00:44
*** ninag has joined #openstack-glance00:44
*** e0ne has quit IRC00:45
*** ninag has quit IRC00:49
*** mingdang1 has joined #openstack-glance00:50
*** mingdang1 has quit IRC00:50
*** mingdang1 has joined #openstack-glance00:50
*** RichardRaseley has quit IRC01:01
*** shakamunyi has quit IRC01:02
*** shakamunyi has joined #openstack-glance01:03
*** flaper87 has left #openstack-glance01:15
*** flaper87 has joined #openstack-glance01:15
*** sgotliv__ has quit IRC01:22
openstackgerritMerged openstack/glance: Move store config opt to glance_store section
*** dims has quit IRC01:30
*** mancdaz has quit IRC01:39
*** mingdang1 has quit IRC01:39
*** mancdaz has joined #openstack-glance01:40
*** mtanino has quit IRC01:47
*** chlong has joined #openstack-glance01:48
*** chlong has quit IRC01:51
*** dims has joined #openstack-glance02:00
*** chlong has joined #openstack-glance02:00
*** chlong has quit IRC02:06
*** ninag has joined #openstack-glance02:16
*** chlong has joined #openstack-glance02:20
*** ninag has quit IRC02:27
*** ninag has joined #openstack-glance02:28
*** ninag has quit IRC02:28
*** ninag has joined #openstack-glance02:29
*** ninag has quit IRC02:33
*** chlong has quit IRC02:35
*** mingdang1 has joined #openstack-glance02:40
*** mingdang1 has quit IRC02:45
*** chlong has joined #openstack-glance02:49
*** houming has joined #openstack-glance02:59
*** ayoung has joined #openstack-glance03:01
*** ayoung has quit IRC03:01
*** ayoung has joined #openstack-glance03:03
*** e0ne has joined #openstack-glance03:06
*** jerrygb has quit IRC03:19
*** chlong has quit IRC03:21
*** ducttape_ has joined #openstack-glance03:29
*** azbiswas has quit IRC03:35
*** e0ne has quit IRC03:35
*** azbiswas has joined #openstack-glance03:42
*** ducttape_ has quit IRC04:12
*** jerrygb has joined #openstack-glance04:20
*** spzala has quit IRC04:23
*** jerrygb has quit IRC04:26
*** mdnadeem has left #openstack-glance04:38
*** dims has quit IRC04:52
*** GB21 has joined #openstack-glance04:52
*** links has joined #openstack-glance05:06
*** e0ne has joined #openstack-glance05:08
openstackgerritMerged openstack/glance: Add the Docker container format
*** changbl has quit IRC05:13
*** e0ne has quit IRC05:14
*** ianbrown has joined #openstack-glance05:14
openstackgerritMerged openstack/glance: Remove version from setup.cfg
*** e0ne has joined #openstack-glance05:16
*** haomaiwang has joined #openstack-glance05:16
*** ianbrown has quit IRC05:21
*** GB21 has quit IRC05:29
*** changbl has joined #openstack-glance05:30
*** azbiswas has quit IRC05:40
*** TravT has joined #openstack-glance05:42
*** e0ne has quit IRC06:11
*** melwitt has quit IRC06:19
*** melwitt has joined #openstack-glance06:22
*** jerrygb has joined #openstack-glance06:22
*** e0ne has joined #openstack-glance06:24
*** e0ne has quit IRC06:24
*** jerrygb has quit IRC06:26
*** e0ne has joined #openstack-glance06:29
*** e0ne has quit IRC06:35
*** mingdang1 has joined #openstack-glance06:36
*** e0ne has joined #openstack-glance06:43
*** e0ne has quit IRC06:45
*** GB21 has joined #openstack-glance06:53
*** azbiswas has joined #openstack-glance06:57
*** e0ne has joined #openstack-glance07:02
*** e0ne has quit IRC07:06
*** teju has joined #openstack-glance07:08
*** bpoulos has quit IRC07:09
*** sgotliv__ has joined #openstack-glance07:16
*** dshakhray has joined #openstack-glance07:17
*** rcernin has joined #openstack-glance07:19
*** rcernin_ has joined #openstack-glance07:24
*** e0ne has joined #openstack-glance07:35
*** e0ne has quit IRC07:37
openstackgerritBrian Rosmaita proposed openstack/glance-specs: Image Import Refactor
*** flaper87 has quit IRC07:44
*** flaper87 has joined #openstack-glance07:49
*** chlong has joined #openstack-glance07:55
openstackgerritAtsushi SAKAI proposed openstack/python-glanceclient: Add docker to image_schema on glance v2 cli
*** wxy_ has joined #openstack-glance07:59
*** wxy__ has quit IRC08:02
*** e0ne has joined #openstack-glance08:03
*** belmoreira has joined #openstack-glance08:04
*** markus_z has joined #openstack-glance08:16
*** e0ne has quit IRC08:24
openstackgerritMerged openstack/python-glanceclient: Fix help for image-create
*** MattMan has quit IRC08:24
*** MattMan has joined #openstack-glance08:25
*** Guest66073 has joined #openstack-glance08:26
*** chlong has quit IRC08:29
*** teju has quit IRC08:34
*** azbiswas has quit IRC08:34
*** GB21 has quit IRC08:43
*** rcernin has quit IRC08:50
*** rcernin_ has quit IRC08:51
*** GB21 has joined #openstack-glance09:02
*** belmoreira has quit IRC09:04
*** aix has joined #openstack-glance09:29
openstackgerritKairat Kushaev proposed openstack/python-glanceclient: Support docker images in client
*** houming has quit IRC09:49
*** rcernin has joined #openstack-glance09:59
openstackgerritKairat Kushaev proposed openstack/python-glanceclient: Run py34 env first when launching tests
*** osho_ has joined #openstack-glance10:02
openstackgerritKairat Kushaev proposed openstack/glance: Run py34 env first when launching tests
*** marekd has quit IRC10:05
*** jistr has joined #openstack-glance10:09
*** marekd has joined #openstack-glance10:13
*** marekd has quit IRC10:15
*** marekd has joined #openstack-glance10:16
openstackgerritzhangsong proposed openstack/glance_store: Sheepdog: Change storelocation format
*** tsufiev has quit IRC10:22
*** osho_ has quit IRC10:23
*** tsufiev has joined #openstack-glance10:29
*** ativelkov has quit IRC10:29
openstackgerritDarja Shakhray proposed openstack/glance: Fix Glance doesn't catches UnicodeDecodeError exception
*** ativelkov has joined #openstack-glance10:32
*** ninag has joined #openstack-glance10:33
*** e0ne has joined #openstack-glance10:33
*** e0ne has quit IRC10:36
*** ninag has quit IRC10:37
*** belmoreira has joined #openstack-glance10:38
*** haomaiwang has quit IRC10:38
*** Guest66073 has quit IRC10:39
*** venkatesh_ has joined #openstack-glance11:01
*** esheffield_ has joined #openstack-glance11:01
*** chlong has joined #openstack-glance11:02
*** wxy_ has quit IRC11:02
*** rosmaita_ has joined #openstack-glance11:03
*** wxy has joined #openstack-glance11:03
*** dims_ has joined #openstack-glance11:03
*** venkatesh has quit IRC11:03
*** esheffield has quit IRC11:03
*** rosmaita has quit IRC11:03
*** aix has quit IRC11:18
*** mclaren has quit IRC11:34
kairatflaper87, ping11:40
kairatflaper87, I have a question. If I assigned bug to myself but someone applied the patch for this bug before me (without referencing to the bug)11:41
kairatflaper87, that's more proper, abandon my patch I assign the bug to the person who applied the patch11:42
kairat*what is11:42
buntingThats what i would probably do11:43
*** houming has joined #openstack-glance11:43
kairatI think the same also11:43
kairatJust would like to know because I may be in the same situation if I apply the patch before other person=)11:45
buntingYeah probably comment mentioning the bug number11:46
*** houming has quit IRC11:48
*** houming has joined #openstack-glance11:51
buntingkairat: ping11:57
kairatbunting, pong11:58
buntingI'm just trying to get my head around your 4th comment on
buntingjust a bit confused what you mean11:59
kairatI am wondering why you need to parse keystone auth address12:00
kairatIf you have auth_version in configs12:00
kairatI am not sure that it is right decision12:00
openstackLaunchpad bug 1507610 in glance_store "Keystone v3 incompatable with keystone v2" [Undecided,In progress] - Assigned to Niall Bunting (niall-bunting)12:00
buntingIf they have the old type and change it to keystone v312:00
buntingit won't work12:01
buntingby old type i mean the url stored in the database12:01
kairathave you updated swift_store_auth_version?12:02
kairatto 3?12:02
kairatalthough it won't work12:02
kairatBecause it is hard-coded:)12:02
buntingin the configs?12:02
kairatBut why do we need to parse keystone auth address, why can't we initialize the version with swift_store_auth_version?12:03
kairatin glance-api.conf or conig reference file in case of swift+config12:03
buntingBecause if the database contains a url that is something like keystone:5000/v2 if you try to use auth version 3 it wont work12:04
buntingdue to keystone url changing from /v2.0/tokens to /v3/auth/tokens12:05
*** houming has quit IRC12:08
*** houming has joined #openstack-glance12:08
*** houming has quit IRC12:09
*** houming has joined #openstack-glance12:09
kairatJust a second12:09
kairatbunting, I have a meeting, will be available in 1 hour12:10
kairatI have several questions about your case then:)12:10
kairatFirst question, should we update swift_url address if config and version in keystone url is different12:11
kairatinstead of parsing version from keystone url12:12
kairatbunting, ^12:12
buntingHmmn but what happens if this uses a diffrent keystone? from the one in the configs?12:16
*** flaper87 has quit IRC12:17
*** flaper87 has joined #openstack-glance12:17
kairatSo if you updated config then glance_store need to use the new version12:18
flaper87kairat: yeah, that's what I would do12:19
kairatflaper87, thanks12:19
buntingSo we should just ignore the url in the database?12:20
kairatThat's the question12:20
kairatI don't mean url but keystone version in url12:20
buntingBecause we currently use it, that leads to this problem12:20
kairatit sounds like a question for meeting12:21
kairator letter to openstack-dev=)12:21
kairatI am in favor of this12:21
kairatBut there could be other opinions12:21
buntingYeah! I'll put it on the agenda for now12:22
*** jaypipes has joined #openstack-glance12:25
*** GB21 has quit IRC12:29
*** openstackgerrit has quit IRC12:32
*** openstackgerrit has joined #openstack-glance12:32
*** aix has joined #openstack-glance12:39
*** haomaiwa_ has joined #openstack-glance12:43
*** ducttape_ has joined #openstack-glance12:47
*** ninag has joined #openstack-glance12:59
*** ninag has quit IRC12:59
*** ninag has joined #openstack-glance13:00
*** ninag has quit IRC13:04
*** ducttape_ has quit IRC13:14
*** openstackgerrit has quit IRC13:17
*** openstackgerrit has joined #openstack-glance13:17
*** smatzek has joined #openstack-glance13:18
*** jerrygb has joined #openstack-glance13:21
*** ninag has joined #openstack-glance13:21
*** ninag_ has joined #openstack-glance13:22
*** spzala has joined #openstack-glance13:23
*** bdossant has joined #openstack-glance13:25
*** ninag has quit IRC13:26
*** ctina_ has joined #openstack-glance13:32
openstackgerritMerged openstack/python-glanceclient: Fix tests for image-create
openstackgerritNiall Bunting proposed openstack/glance_store: Make glance_store use the correct keystone version
*** haomaiwa_ has quit IRC13:40
*** bpoulos has joined #openstack-glance13:45
*** links has quit IRC13:51
*** e0ne has joined #openstack-glance14:04
*** e0ne has quit IRC14:05
*** julim has joined #openstack-glance14:06
*** e0ne has joined #openstack-glance14:06
*** rosmaita_ is now known as rosmaita14:07
*** RichardRaseley has joined #openstack-glance14:09
*** edmondsw has joined #openstack-glance14:11
*** takedakn has joined #openstack-glance14:15
*** takedakn has quit IRC14:20
*** mingdang1 has quit IRC14:20
kairatkrotscheck, can I use wildcards or reg-exp for list of headers exposed/allowed?14:21
RichardRaseleyHello Glance team. Quick question to help me in orienting my troubleshooting. Does this ( ) error I am experiencing seem like a manifestation of this (  ) already reported bug? Running OpenStack Kilo from RDO packaging. Looking at PIP I have requests (2.8.1) and urllib3 (1.12) installed.14:23
openstackLaunchpad bug 1476770 in python-glanceclient "_translate_from_glance fails with "AttributeError: id" in grenade" [High,In progress] - Assigned to Flavio Percoco (flaper87)14:23
*** bdossant_ has joined #openstack-glance14:30
*** bdossant has quit IRC14:31
nikhilRichardRaseley: apparently so (to me)14:33
*** e0ne_ has joined #openstack-glance14:34
*** houming has quit IRC14:34
RichardRaseleynikhil : Thank you.14:34
*** smatzek has quit IRC14:35
*** e0ne has quit IRC14:36
*** wxy__ has joined #openstack-glance14:45
*** wxy has quit IRC14:48
*** bdossant_ has quit IRC14:53
*** bdossant has joined #openstack-glance14:54
*** sigmavirus24_awa is now known as sigmavirus2414:57
*** ducttape_ has joined #openstack-glance15:00
*** mclaren has joined #openstack-glance15:00
*** smatzek has joined #openstack-glance15:02
*** julim has quit IRC15:03
kairatbunting, I think you can write a letter to openstack-dev15:03
kairatIMO, it would be more effective than waiting a week for next meeting15:03
*** houming has joined #openstack-glance15:04
buntingkairat: Yeah, i agree with you15:05
*** julim has joined #openstack-glance15:06
flaper87bunting: kairat ++ for using the mailing list15:07
*** bdossant has quit IRC15:16
*** bdossant has joined #openstack-glance15:18
*** houming has quit IRC15:20
openstackgerritKairat Kushaev proposed openstack/glance_store: Change approach to request storage url for multi-tenant store
kairatbunting, yes15:25
*** houming has joined #openstack-glance15:25
buntingJust as i write this email do you mean that we would take the version off the end of the url?15:25
buntingand use the one in the config?15:25
kairatI mean we should use the value from config15:25
kairatAFAIK, it should be defined by deployer15:25
kairatBy default the value is 215:26
kairatThe problem is that the version in config and url could contradict15:26
kairatSo what should we do in this case15:26
kairatStick to config or15:26
kairatuse url15:26
kairatI think we need to stick to config15:27
buntingwell if we stick to config we can't use the full url15:27
kairatso we need to update it15:27
buntingif it is something like  swift:redacted@
kairatthat's one of option15:27
buntingthen we would need to change that to v315:27
kairatthe second option is yours15:27
buntingor else it would not work if there is a 3 in the config15:27
kairatbut why do we need config at all then15:27
kairatI think we need to choose the single approach15:28
buntingbecause the 'new' style is swift:ref/glance15:28
buntingit replaces the ref15:28
*** bpoulos has quit IRC15:28
kairatwhen ref is parsed the attribute value would be keystone auth address15:28
kairatso there should not be a problem15:29
buntingYeah in that case it uses all the values in the config15:29
*** dshakhray has quit IRC15:30
buntingkairat: I just don't have a clear picture in my head what values you propose to use15:31
*** mancdaz has quit IRC15:31
*** julim has quit IRC15:31
*** mancdaz has joined #openstack-glance15:31
kairatok, we have swift_store_auth_version or auth_version in reference config file15:32
*** julim has joined #openstack-glance15:32
kairatThat option defines what version need to be used as auth_version in swift15:33
buntingwell with that example url above15:33
buntingthe auth_url would be
buntingso we could not use that url if both the configs are set to 315:34
kairatSo you can change this also15:34
kairatAFAIK these options don't need to specified15:34
kairatswift_store_auth_version is deprecated at least for single-tenant store15:34
kairatbecause it is expose creds to users15:35
buntingYeah it is deprecated, but some users may still have full locations in the database15:35
kairatyep, that's the problem15:35
kairatMy question is, should we support authenticating with old keystone versions if config changes15:36
kairator we can use the same version for all images15:37
buntingSo you think if we change the config, we could potentially mean users can't use there images any more?15:37
kairatif old keystone api will be dropped15:37
kairatthen yes15:37
kairatand there would be no way except db update to fix that15:38
kairatI am not sure if it is appropriate15:38
kairatI would like to have the single source of auth_version information15:39
kairatBut of course there could be some problems15:39
buntingokay i'll try to get your stance across in the email as the other option15:40
kairatthanks bunting15:41
*** kebray has joined #openstack-glance15:41
openstackgerritKairat Kushaev proposed openstack/glance: Import i18 functions directly
* flaper87 dives into the import process spec15:43
kairatbunting, if you stick with your implementation15:44
kairatthen you don't need to consider config at all15:44
kairatIt becomes useless15:44
mclarenflaper87: I'm around for a bit if you want to ping me on anything15:44
kairatso anyway the patch might need to be updated15:44
flaper87mclaren: I do, I'll need you to elaborate a bit more on your comment15:45
flaper87mclaren: what are you proposing there?15:45
mclarenWell I'm just throwing it out there to see what folks think...15:45
mclarenwe could separate the bikeshed from the image15:46
flaper87mclaren: yup, is that a separate endpoint from images ?15:46
mclarenyeah, so basically the swift case works pretty nicely15:46
mclarenso I was thinking what could we do to make the glance case more like that15:46
mclarentotally separate endpoint for bikesheds15:47
mclarena fairly dumb thing to support PUT, list, DELETE bikesheds independently of images15:48
mclarenIt feels like we may need much of that anyway (although I'm prepared to accept I'm wrong on that)15:49
mclarenand I wonder about spaghettification if we try to put it into the existing code paths15:49
flaper87mclaren: mmh, I think you're wrong on that. I don't see the difference between the `/imaves/$id/bikeshed` and `/bokesheds` other than I don't see the need of list/deletes in the former15:50
mclarenone example I can think of is v1 delete: it would need to become bikeshed aware if a bikeshed is part of an image15:50
flaper87mclaren: that's a good point, though15:50
mclarenimaging two users Alice and Bob15:51
mclarenthere's 100 GB of storage, each gets a quota of 50GB.15:51
mclarenBob shouldn't be able to use Alice's quota15:51
mclarenBob should be able to see what's using his quota15:51
*** haomaiwa_ has joined #openstack-glance15:52
mclarenAre those fair assumptions?15:52
flaper87sure, but that can be done with `/images/$id/bikeshed` without adding lists/deletes15:53
flaper87I think the same rules for `/file` apply to `/bikeshed`15:53
flaper87with the difference that `/bikeshed` will have to be processed15:53
flaper87same rules with regard to quota, I mean.15:53
mclarenThe bikeshed needs a checksum/size and to co-exist with /file -- agreed?15:54
flaper87mmh, I think so... yes.15:55
mclarenand you need to be able to list them?15:55
flaper87nope, that's what the image is for15:55
flaper87we should have that info in the image15:56
mclarenSo that't the kind of thing that worries me slightly15:56
krotscheckkairat: No, they have to be explicitly defined.15:57
flaper87mclaren: anything that changes the state of things worries you <half-joke> :D15:57
mclarena bikeshed isn't necessarily a natural thing that an image has15:57
kairatkrotscheck, ok, thanks, we have some metadata in headers for v115:58
mclarenif v1 show should show me my image, should it show me the bikeshed?15:58
krotscheckkairat: Gotcha. Annotate my patch and I'll update it after my meeting this morning :)15:58
kairatkrotscheck, these headers can be user-defined15:59
*** RichardRaseley has quit IRC15:59
krotscheckkairat: Huhn.15:59
krotscheckkairat: Looked at the meeting logs - I'm guessing v2 support only is acceptable.15:59
kairatkrotscheck, so it would not be good idea to use ity with v1)15:59
mclarenwe could say that image show is just different in v1 and v2, but separating /bikeshed avoids that15:59
flaper87fwiw, image-show is already different in v1 and v216:00
*** haomaiwa_ has quit IRC16:00
*** julim has quit IRC16:01
openstackgerritKairat Kushaev proposed openstack/glance: Validate empty location value for v1 api
flaper87TBH, I'm not that worried about this16:01
*** 18VAAEQWZ has joined #openstack-glance16:01
mclarenDo you see and advantage in keeping the /bikeshed as part of an image?16:01
*** aix has quit IRC16:01
flaper87Separating bikeshed will likely endup being like the tasks endpoint we have16:01
*** 18VAAEQWZ has quit IRC16:02
mclarenin what sense?16:02
flaper87yes, it sends a clear message that it's related to the image and that, whenever the image is changed/deleted/whatever that'll likely affect the bikeshed as we;;16:02
*** haomaiwa_ has joined #openstack-glance16:02
*** haomaiwa_ has quit IRC16:03
*** haomaiwa_ has joined #openstack-glance16:03
*** julim has joined #openstack-glance16:03
flaper87The close relation between the place the image is upload to (when using direct-upload) and the image resource is important16:03
*** haomaiwa_ has quit IRC16:04
mclarenon the other hand it's more consitent with the import from swift approach16:04
flaper87I see bikeshed the same way I see /file16:04
*** haomaiwang has joined #openstack-glance16:04
*** haomaiwang has quit IRC16:05
flaper87but swift is an external resource, it's like the old copy-from16:05
*** haomaiwa_ has joined #openstack-glance16:05
*** haomaiwa_ has quit IRC16:06
*** haomaiwang has joined #openstack-glance16:06
mclarenI see the bikeshed as a staging area for unvalidated bytes16:06
mclarenlike the swift case16:06
flaper87and like the /file endpoint16:06
*** haomaiwang has quit IRC16:07
flaper87the bytes there are unvalidated16:07
mclarendo you not see any potential for the implementation to be cleaner?16:07
flaper87I guess the difference is that I see the /bikeshed endpoint as something internal and strictly related to the image process16:07
*** haomaiwang has joined #openstack-glance16:07
*** haomaiwang has quit IRC16:08
flaper87I think the latest PS made it smaller by removing the swift part and a set of resources that aren't needed16:08
flaper87the implementation can still re-use existing code and be cleaner16:08
*** haomaiwa_ has joined #openstack-glance16:08
*** haomaiwa_ has quit IRC16:09
mclarenwhich PS?16:09
flaper87what worries me more is the process itself as I don't think bikshed being under /images is a problem16:09
flaper87it made it smaller compared to PS516:09
*** haomaiwa_ has joined #openstack-glance16:09
*** haomaiwa_ has quit IRC16:10
*** 5EXAADMHT has joined #openstack-glance16:10
mclarenI think bikeshed under /images has the potential to make the code more complex/entangled than it needs to be. You're going to have to handle more dependencies.16:10
*** 5EXAADMHT has quit IRC16:11
flaper87Having it separted is moving the dependencies to the user, basically16:11
*** haomaiwang has joined #openstack-glance16:11
*** haomaiwang has quit IRC16:12
mclarenbecause the number of potential internal states an image can be in will increase. And failure for each operation on each of those states will need to be handled. For v1 and v2.16:12
*** 5EXAADMKT has joined #openstack-glance16:12
flaper87I think you still need them. The separate /bikeshed resource is just moving them somewhere else16:12
mclarenI diagree.16:12
*** rcernin has quit IRC16:12
flaper87we have to agree to disagree :D16:12
*** 5EXAADMKT has quit IRC16:13
flaper87Until now, the user just cared about uploading stuff to the glance upload path16:13
flaper87now the user needs to worry about /bikeshed16:13
*** 5EXAADML5 has joined #openstack-glance16:13
flaper87list it, figure out if it's using space and whatnot16:13
mclarenYou have two things, each of which have a small set of states. Rather than one thing which has a large set. The number of permutations is much larger in the latter case.16:13
*** 5EXAADML5 has quit IRC16:13
flaper87Right, but, as I said, you still need them. You just moved those states and responsibility somewhere else16:14
flaper87I don't think this is something we should offload to the user, fwiw.16:14
mclarenit's like the difference between two decks of cards with 10 cards each or one deck of cards with 20 cards. The number of states is much bigger with one larger deck.16:14
mclarenmathematically and shit :-)16:15
flaper87I understand that and I'm not really worried about that, tbh.16:15
flaper87but that's me16:15
flaper87I'm more worried about what this means for the user16:15
flaper87and how this impacts the process16:15
flaper87which is exactly what we're trying to solve here16:15
mclarenFrom a user's point of view the process becomes the same for swift or glance: upload the bits to some store (provided by glance or swift) and then import the bits16:17
mclarenyou can reuse your object if you want in both cases16:17
mclarenyou can choose to delete it or not in both cases16:17
*** bdossant has quit IRC16:18
mclarenyou have a command to list your objects in both cases16:18
flaper87no no, it's not the same. Until now, the user didn't have to worry about that and I don't think the user should worry about that. Not when using the direct-upload to glance16:18
flaper87The bikeshed endpoint as you described it will be just an http server16:19
mclarenHmm, so I'm not sure that the need for any of that goes away if the bikeshed goes into the image16:19
tsufievnikhil, speaking of multipart/form-data...16:19
mclarenuser's will still need to 'manage' their bikesheds16:19
nikhiltsufiev: hi16:20
flaper87it does. The user just uploads the image and the used doesn't have to worrye about managing the bikeshed16:20
tsufievnikhil, hello :)16:20
nikhiltsufiev: let's chat somewhere else16:20
flaper87no, because the bikeshed will be strictly related to the image16:20
mclarenusers will need to know where their bikesheds are right? They'll need a call (whatever it looks like) "show me my bikesheds"16:21
flaper87if re-using image bits is a requirement, we can very well implement image-copy in the future16:21
flaper87no, why do users need that?16:21
flaper87Do they need a "show me my files" call now ?16:21
mclarenif you have loads of images and a few bikesheds how you do find them?16:22
mclaren(that's probably a fairly common scenario)16:22
flaper87uh crap, meeting in 8mins.16:22
flaper87the info is in the image metadata16:22
flaper87I'll comment on the spec with my thoughts16:23
mclarensure, but you've got to search a large number of images for a small number of bikesheds16:23
flaper87just filter images with size != 0 ?16:23
flaper87or something like that16:23
flaper87whatever filters there are that can be used16:24
mclarenyou mean an 'images with bikesheds' filter?16:24
flaper87I guess16:24
* flaper87 shrugs16:24
mclarenbecause that's kind of listing bikesheds :-)16:24
flaper87I knew you were going to say that but I still believe it's not the same16:24
flaper87as the image is the resource16:25
flaper87the bikeshed is just some data associated with that image16:25
flaper87Once the image is processed, the bikeshed is going away as it'll become the processed, imag data16:26
flaper87ok, gtg now.16:26
flaper87I'll comment on the spec16:26
mclarenok, thanks16:26
flaper87np, thank you for throwing more ideas out. It's better to know we're exploring as many ideas as we can16:27
*** kebray has quit IRC16:37
*** davemc_afk has quit IRC16:37
*** dshakhray has joined #openstack-glance16:39
*** belmoreira has quit IRC16:42
*** rcernin has joined #openstack-glance16:54
*** jaypipes has quit IRC16:59
*** ninag_ has quit IRC17:00
*** ninag_ has joined #openstack-glance17:04
*** ninag_ has quit IRC17:09
*** houming has quit IRC17:09
openstackgerritMerged openstack/glance_store: Put py34 first in the env order of tox
*** RichardRaseley has joined #openstack-glance17:19
*** jistr has quit IRC17:21
*** bpoulos has joined #openstack-glance17:25
*** mtanino has joined #openstack-glance17:28
bpouloshemanthm: you up for discussing the sha-2 stuff further?17:32
bpoulosthe signature_utils only supports SHA-2 hashes, but there isn't a default in place -- the user has to specify which hash method to use (this is the hash for the signature verification, not for the glance checksum)17:34
bpoulosand this uses the cryptography library, which appears to work on the latest version of openssl 0.9.8:
bpoulosand no hashes are done unless all of the signature properties are provided during upload17:35
*** ninag has joined #openstack-glance17:36
*** ninag has quit IRC17:36
hemanthmbpoulos: o/17:36
bpouloshemanthm: hello17:37
hemanthmlet me get the error I see with openssl 0.9.817:37
bpoulosok, great17:37
hemanthmbpoulos ^17:39
bpouloshemanthm: thanks17:39
hemanthmessentially, openssl 0.9.8 doesn't support SHA-2 out of the box17:39
bpoulosI hesitate to allow support for SHA-1 (adding it to the list of hash algorithm options supported) since it's considered out-of-date and insecure.  It sounds like the signature_utils code should catch the unsupported algorithm exception, and log that signature verification isn't supported on platforms that don't support sha-217:45
*** aix has joined #openstack-glance17:49
bpouloshemathm: thank you for bringing this up -- I think it deserves a separate bug report, where how to handle it can be discussed (i.e., support SHA-1 or add log messages/don't support signature verification on older platforms)17:51
bpouloshemanthm: I'll put together the bug report, unless you'd like to17:51
hemanthmbpoulos: I'm not necessarily saying we should support SHA-117:52
hemanthmjust that glance shouldn't break it can't use SHA-217:52
hemanthm*when it can't use SHA-217:52
hemanthmbpoulos: feel free to do the bug report, you can probably add more detail to it than I can17:54
bpouloshemanthm: ok, will do, thanks again!17:54
hemanthmthanks for looking into it, bpoulos!17:54
*** ninag has joined #openstack-glance17:56
*** aix has quit IRC18:09
bpouloshemanthm: I added the bug report ( -- feel free to add any further details if you see something missing18:10
openstackLaunchpad bug 1522525 in Glance "test_signature_utils tests fail when openssl is older than 1.0.1" [Undecided,New] - Assigned to Brianna Poulos (brianna-poulos)18:10
*** bpoulos has quit IRC18:11
*** aix has joined #openstack-glance18:13
hemanthmthank you, bpoulos18:14
*** markus_z has quit IRC18:16
openstackgerritMike Fedosin proposed openstack/glance: Implement trust support for api v2
*** azbiswas has joined #openstack-glance18:22
*** davideagnello has quit IRC18:26
*** davideagnello has joined #openstack-glance18:26
openstackgerritHemanth Makkapati proposed openstack/glance: Rename glance-store to glance_store
hemanthmnikhil ^18:36
*** ctina_ has quit IRC18:53
*** RichardRaseley has quit IRC19:03
*** edmondsw has quit IRC19:06
*** ninag has quit IRC19:06
*** ninag has joined #openstack-glance19:10
*** ninag_ has joined #openstack-glance19:11
*** ninag has quit IRC19:14
*** ninag_ has quit IRC19:16
*** ninag has joined #openstack-glance19:19
*** ninag has quit IRC19:25
*** ninag has joined #openstack-glance19:26
*** ninag has quit IRC19:27
*** ninag has joined #openstack-glance19:28
*** ninag has quit IRC19:30
*** ninag has joined #openstack-glance19:31
*** azbiswas has quit IRC19:49
*** davideagnello has quit IRC19:50
*** ninag has quit IRC19:50
*** davideagnello has joined #openstack-glance19:50
*** ninag has joined #openstack-glance19:52
*** azbiswas_ has joined #openstack-glance19:53
*** chlong has quit IRC19:57
*** edmondsw has joined #openstack-glance19:58
*** chlong has joined #openstack-glance20:16
*** RichardRaseley has joined #openstack-glance20:25
*** changbl has quit IRC20:26
*** ninag has quit IRC20:31
*** avarner has quit IRC20:32
openstackgerritlifeless proposed openstack/glance: Stop using tearDown in v1/
flaper87hemanthm: pls20:40
openstackgerritlifeless proposed openstack/glance: Stop using tearDown in v1/
*** changbl has joined #openstack-glance20:44
*** openstackgerrit has quit IRC20:47
*** openstackgerrit has joined #openstack-glance20:47
*** zul has quit IRC20:48
*** azbiswas_ has quit IRC20:50
*** azbiswas has joined #openstack-glance20:54
*** azbiswas has joined #openstack-glance20:54
*** mancdaz has quit IRC20:58
*** mancdaz has joined #openstack-glance21:01
*** rcernin has quit IRC21:05
flaper87sabari: pls21:06
*** delatte has quit IRC21:07
sabariflaper87 in a meeting :( but will do asap21:08
sabariflaper87 that was an easy one - sorry I even gave an excuse for it :D21:15
sabari*nods in the meeting room though has no clue what's going on*21:17
*** smatzek has quit IRC21:23
*** julim has quit IRC21:24
flaper87sabari: lol, that's the story of my life21:25
flaper87sabari: actually, I lied. #lifeofaremotee21:25
*** e0ne_ has quit IRC21:26
*** spzala has quit IRC21:26
*** changbl has quit IRC21:34
*** cdelatte has joined #openstack-glance21:38
*** diegows has joined #openstack-glance21:46
*** changbl has joined #openstack-glance21:47
openstackgerritDrew Varner proposed openstack/glance_store: Add reno for release notes management in glance_store
*** edmondsw has quit IRC22:05
*** agireud has quit IRC22:47
*** burgerk has quit IRC22:48
*** maliniB has joined #openstack-glance22:53
*** chlong has quit IRC23:08
*** spzala has joined #openstack-glance23:09
*** jerrygb has quit IRC23:12
*** diegows has quit IRC23:22
*** azbiswas has quit IRC23:28
*** azbiswas has joined #openstack-glance23:28
*** azbiswas has quit IRC23:32
*** spzala has quit IRC23:43
*** agireud has joined #openstack-glance23:52
*** ducttape_ has quit IRC23:53
*** sigmavirus24 is now known as sigmavirus24_awa23:57

Generated by 2.14.0 by Marius Gedminas - find it at!