Friday, 2020-07-17

*** vishakha has quit IRC00:02
*** hamalq_ has quit IRC00:05
*** deiter has quit IRC00:19
*** rcernin has quit IRC00:34
*** rcernin has joined #openstack-cinder00:39
*** zhanglong has joined #openstack-cinder01:01
*** Lucas_Gray has quit IRC01:04
*** Liang__ has joined #openstack-cinder01:10
*** yaawang has quit IRC01:33
*** yaawang has joined #openstack-cinder01:33
*** yaawang has quit IRC01:59
*** yaawang has joined #openstack-cinder02:02
*** markvoelker has joined #openstack-cinder02:13
*** markvoelker has quit IRC02:18
*** zhanglong has quit IRC02:21
*** zhanglong has joined #openstack-cinder02:22
*** sapd__x has joined #openstack-cinder02:23
*** zhanglong has quit IRC02:39
*** zhanglong has joined #openstack-cinder02:41
*** psachin has joined #openstack-cinder03:37
*** zhanglong has quit IRC03:47
*** markvoelker has joined #openstack-cinder03:48
*** markvoelker has quit IRC03:53
*** vishalmanchanda has joined #openstack-cinder04:01
openstackgerritXinliang Liu proposed openstack/cinder master: Add ceph-iscsci CI job  https://review.opendev.org/66710804:18
*** rcernin has quit IRC04:47
*** Liang__ has quit IRC04:54
*** KeithMnemonic has quit IRC04:55
*** Liang__ has joined #openstack-cinder04:55
*** rcernin has joined #openstack-cinder04:57
*** rcernin has quit IRC04:57
*** udesale has joined #openstack-cinder04:57
*** rcernin has joined #openstack-cinder04:57
*** Liang__ has quit IRC05:01
*** Liang__ has joined #openstack-cinder05:02
*** yaawang has quit IRC05:04
*** yaawang has joined #openstack-cinder05:05
*** rcernin has quit IRC05:12
*** suryasingh has joined #openstack-cinder05:18
*** rcernin has joined #openstack-cinder05:19
*** zhanglong has joined #openstack-cinder05:44
*** yaawang has quit IRC05:45
*** yaawang has joined #openstack-cinder05:46
*** Liang__ has quit IRC05:47
*** Liang__ has joined #openstack-cinder05:48
*** markvoelker has joined #openstack-cinder05:49
*** markvoelker has quit IRC05:54
openstackgerritfutaotao proposed openstack/cinder master: Due to Fusionstorage rename to OceanStor 100D, there is a necessary to support  OceanStor 100D with Fusionstorage Cinder Driver.(dsware)  https://review.opendev.org/74157906:16
*** baojg has quit IRC06:18
*** rcernin has quit IRC06:19
*** baojg has joined #openstack-cinder06:19
*** rcernin has joined #openstack-cinder06:32
*** yaawang has quit IRC06:32
*** yaawang has joined #openstack-cinder06:33
openstackgerritfutaotao proposed openstack/cinder master: Due to Fusionstorage rename to OceanStor 100D, there is a necessary to support OceanStor 100D with Fusionstorage Cinder Driver.(dsware)  https://review.opendev.org/74157906:36
openstackgerritfutaotao proposed openstack/cinder master: Fusionstorage Cinder Driver Support OceanStor 100D Storage.(dsware)  https://review.opendev.org/74157906:39
*** rcernin has quit IRC06:46
*** rcernin has joined #openstack-cinder06:52
*** zhanglong has quit IRC07:04
*** yaawang has quit IRC07:04
*** yaawang has joined #openstack-cinder07:05
*** zhanglong has joined #openstack-cinder07:05
*** rcernin has quit IRC07:06
*** udesale has quit IRC07:17
*** deiter has joined #openstack-cinder07:21
*** Liang__ has quit IRC07:25
*** brinzhang0 has joined #openstack-cinder07:26
*** brinzhang_ has quit IRC07:29
*** tosky has joined #openstack-cinder07:33
*** rcernin has joined #openstack-cinder07:34
*** rcernin has quit IRC07:42
*** rcernin has joined #openstack-cinder07:46
openstackgerritLuigi Toscano proposed openstack/cinder stable/ussuri: zuul: collect cinderlib logs from tempest node(s) only  https://review.opendev.org/74159107:46
*** baojg has quit IRC07:47
*** zhanglong has quit IRC07:50
*** rcernin has quit IRC07:50
*** zhanglong has joined #openstack-cinder07:51
*** udesale has joined #openstack-cinder07:57
*** brinzhang_ has joined #openstack-cinder07:59
*** brinzhang0 has quit IRC08:02
*** Lucas_Gray has joined #openstack-cinder08:03
*** ociuhandu has joined #openstack-cinder08:08
*** baojg has joined #openstack-cinder08:10
openstackgerritNaoki Saito proposed openstack/cinder master: NEC driver: fix a snapshot detach error  https://review.opendev.org/74159708:11
*** brinzhang0 has joined #openstack-cinder08:47
*** spatel has joined #openstack-cinder08:49
*** brinzhang_ has quit IRC08:49
openstackgerritNaoki Saito proposed openstack/cinder master: NEC driver: fix live-migration failure with FC  https://review.opendev.org/74160208:51
*** spatel has quit IRC08:53
*** zhanglong has quit IRC09:04
*** zhanglong has joined #openstack-cinder09:05
*** brinzhang_ has joined #openstack-cinder09:09
*** brinzhang0 has quit IRC09:12
*** sapd__x has quit IRC09:12
*** martinkennelly has joined #openstack-cinder09:14
*** zhanglong has quit IRC09:14
*** zhanglong has joined #openstack-cinder09:17
*** sapd1 has joined #openstack-cinder09:18
*** martinkennelly has quit IRC09:19
*** rcernin has joined #openstack-cinder09:29
*** zhanglong has quit IRC09:30
*** zhanglong has joined #openstack-cinder09:31
*** Liang__ has joined #openstack-cinder09:36
*** Liang__ has quit IRC09:41
*** raghavendrat has joined #openstack-cinder09:42
raghavendrathi geguileo: are you online ?09:45
geguileoraghavendrat: hi09:45
raghavendratthis is regarding backport of "Kaminario: Fix unique_fqdn_network option"09:46
raghavendrathttps://review.opendev.org/#/c/741428/09:47
raghavendratwhenever you have time, it would be great if you can have a look. Thanks.09:47
geguileook09:47
*** rcernin has quit IRC09:48
*** baojg has quit IRC09:49
*** tkajinam has quit IRC09:52
*** e0ne has joined #openstack-cinder09:53
*** udesale_ has joined #openstack-cinder10:01
*** yaawang has quit IRC10:02
*** zhanglong has quit IRC10:02
*** zhanglong has joined #openstack-cinder10:03
*** udesale has quit IRC10:04
*** yaawang has joined #openstack-cinder10:04
*** yaawang has quit IRC10:14
*** zhanglong has quit IRC10:15
*** yaawang has joined #openstack-cinder10:21
*** brinzhang0 has joined #openstack-cinder10:22
*** brinzhang_ has quit IRC10:25
*** brinzhang0 has quit IRC10:30
*** baojg has joined #openstack-cinder10:56
*** martinkennelly has joined #openstack-cinder10:57
*** sapd1 has quit IRC10:57
*** martinkennelly has quit IRC11:01
*** raghavendrat has quit IRC11:04
*** pcaruana has quit IRC11:33
*** pcaruana has joined #openstack-cinder11:44
*** raghavendrat has joined #openstack-cinder12:05
*** vishakha has joined #openstack-cinder12:11
*** baojg has quit IRC12:35
*** raghavendrat has quit IRC12:46
rosmaitawhoami-rajat: left a response for you on https://review.opendev.org/#/c/733555/12:48
*** spatel has joined #openstack-cinder12:50
*** spatel has quit IRC12:55
openstackgerritIvan Pchelintsev proposed openstack/cinder master: Add Cinder driver for Dell EMC PowerStore  https://review.opendev.org/74115813:14
*** Lucas_Gray has quit IRC13:25
*** Lucas_Gray has joined #openstack-cinder13:29
whoami-rajat__rosmaita: yep saw it, so for volume create, the response cannot be changed now right?13:30
rosmaitano, i don't think so13:31
rosmaitai guess we could microversion it, but i don't see the point13:31
rosmaitaso 2 ways to look at this: return 404 to be consistent but wrong, or return 409 and be inconsistent, but with the idea that (a) this has its own microversion, and (b) it's got a different URL structure /v3/resource vs /v3/{project_id}/resource, so you could arguably expect it to behave differently13:31
whoami-rajat__yeah, it's been like that since a long time13:31
rosmaitathat's why i'm being a PITA about this ... we will be stuck with whatever we come up with for eternity13:32
whoami-rajat__rosmaita: i think for project id not accessible, we've HTTPUnauthorized or HTTPForbidden13:35
rosmaitayes, but this new api is a bit different because of where the project_id occurs in the URL13:36
rosmaitai mean, if you make the policies admin-only, then a non admin will get 401 if keystone doesn't validate the request token, or 403 if the token is OK but the policy rejects the request13:38
rosmaitaso you can still get Unauthorized or Forbidden13:38
rosmaitain this case, an admin passes the token and policy checks, and then the only issue left is whether the project_id in  /v3/default-types/{project_id} matches an existing project13:39
whoami-rajat__rosmaita:  hmm, but in both cases a project id is being validated, first it's the context and second it's the passed project_id13:41
rosmaitawell, the way to look at this is what is the response telling me to do?  401 - i need to get a new token, 403 - i need to contact someone about getting my permissions elevated, 404 - resource described in the URL is not found13:43
whoami-rajat__rosmaita: so if the user doesn't has access to the requested project, keystone sends a unauthorized back, what should we do in this case? and if the context project isn't a parent or root of the given project_id then the response is forbidden, what to do there13:43
whoami-rajat__404 is correct13:44
whoami-rajat__but for this case, i'm not sure about 401 and 40313:44
whoami-rajat__because we've 2 validations here13:44
rosmaitathose are handled by the middleware and policies13:44
whoami-rajat__rosmaita:  i'm talking about the code part13:45
whoami-rajat__the target project validation13:46
rosmaitaso you mean the project_id in /v3/default-types/{project_id} ?13:46
whoami-rajat__yep13:46
whoami-rajat__https://review.opendev.org/#/c/737707/5/cinder/api/v3/default_types.py13:46
openstackgerritSimon O'Donovan proposed openstack/cinder master: PowerMax Driver - sync_check_snapshot_delete_fix  https://review.opendev.org/74163913:47
whoami-rajat__rosmaita: one thing i want to mention, geguileo has enough details mentioned on the spec that i usually don't see in specs, is there a possibility to merge that and finalize the discussion on the implementation?13:49
whoami-rajat__i mean we can't expect the specs to have 100% detail about the implementation, it already has a lot13:51
rosmaitageguileo wrote a great spec, but as far as API changes go, we need to know exactly how they are defined13:51
rosmaitai'm not arguing about how it gets implemented, i'm arguing about what the API should be like13:51
whoami-rajat__yep i agree13:53
rosmaitaif we don't get this settled up front, then you wind up with the weirdnesses that we currently have13:53
whoami-rajat__that's also true, it's an important part since it's a unique one13:54
rosmaitabut to answer your earlier question, once the policy check is done, i think you can assume that the caller is authorized to set a default-type13:55
rosmaitaso then it's just a matter of does that project_id exist in keystone or not13:55
*** tobias-urdin is now known as tobias-urdin|pto13:55
whoami-rajat__rosmaita: there are a lot of cases in that13:56
whoami-rajat__rosmaita: the policy can be changed and can be set to non-admins as well, so the keystone project get command can fail13:56
whoami-rajat__rosmaita: if the user is admin, does it has access to the target project13:57
whoami-rajat__rosmaita: does the target project come under the hierarchy of the context project13:57
whoami-rajat__i think the second one might be not a case13:58
whoami-rajat__but the other 2 are valid cases13:58
rosmaitawhoami-rajat__: ok, so what are the open questions at this point?14:02
whoami-rajat__rosmaita: the question is, there are total 2 validations for 401 and 403, two are done by the middleware and two are done inside the API code14:07
whoami-rajat__rosmaita: so we should respond with the same error codes, we can have different messages for both cases14:07
whoami-rajat__rosmaita: or should we handle the second case and return a different response14:08
rosmaitai think you don't need to worry about 401 internally, because at that point you have an authenticated user14:08
rosmaitaonly question is are they authorized14:09
rosmaitabut i also think you need to check the extent to which that juno spec was implemented in keystone14:09
whoami-rajat__rosmaita: i think i can note down all the possible cases in an etherpad, then we can discuss on that?14:11
rosmaitathat would be helpful14:11
whoami-rajat__ok14:11
whoami-rajat__rosmaita: basic cases i've added here, please add any case which i missed https://etherpad.opendev.org/p/default_type_overrides14:21
rosmaitawill do, thanks14:21
*** baojg has joined #openstack-cinder14:27
*** enriquetaso has joined #openstack-cinder14:47
*** pcaruana has quit IRC14:58
*** Lucas_Gray has quit IRC15:03
*** pcaruana has joined #openstack-cinder15:11
*** Lucas_Gray has joined #openstack-cinder15:13
*** ociuhandu_ has joined #openstack-cinder15:20
*** ociuhandu has quit IRC15:24
*** ociuhandu_ has quit IRC15:24
*** udesale_ has quit IRC15:37
*** hamalq has joined #openstack-cinder15:45
*** hamalq_ has joined #openstack-cinder15:47
*** hamalq has quit IRC15:50
*** laurent\ has quit IRC15:51
hemnarosmaita, smcginnis so since the ceph-iscsi driver seems be working and we are trying to get it to land in cinder, should we also make the rest client I wrote, another cinder project in gerrit?15:51
hemnahttps://github.com/hemna/rbd-iscsi-client15:51
hemnathat one15:51
smcginnisI wonder long term if that makes sense for this team to own, or if it would be better to keep that separate.15:51
hemnathe ceph-iscsi driver uses it to talk to the rbd-target-api15:51
rosmaitahemna: in a meeting, will look later15:52
hemnathe downside to keeping it where it is, is doing fixes/updates to it.15:52
hemnaI guess we can always pull that in under the cinder umbrella whenever.15:52
smcginnisWould the ceph community want to adopt that? Seems like something they would want to help promote.15:53
hemnaI would hope so, I even point it out in several issues I filed against ceph-iscsi15:54
hemnaI wouldn't be opposed to it, but we would still need to be able to make fixes, etc15:54
hemnaanyway, food for thought.15:55
smcginnisYeah, probably good to get a good home for it.15:56
*** psachin has quit IRC16:00
*** e0ne has quit IRC16:02
*** martinkennelly has joined #openstack-cinder16:06
*** kklimonda has quit IRC16:20
*** kklimonda has joined #openstack-cinder16:20
*** laurent\ has joined #openstack-cinder16:29
*** xyang has quit IRC16:50
*** xyang has joined #openstack-cinder16:51
*** Lucas_Gray has quit IRC16:56
*** markvoelker has joined #openstack-cinder16:59
*** vishalmanchanda has quit IRC17:20
*** andrebeltrami has joined #openstack-cinder17:27
*** martinkennelly has quit IRC17:52
*** NobodyCam has quit IRC18:12
*** NobodyCam has joined #openstack-cinder18:12
*** TheJulia has quit IRC18:22
*** TheJulia has joined #openstack-cinder18:23
*** johnsom has quit IRC18:24
*** johnsom has joined #openstack-cinder18:24
*** sonivi has joined #openstack-cinder18:48
*** baojg has quit IRC18:56
*** baojg has joined #openstack-cinder18:57
*** masayukig has quit IRC18:58
*** masayukig has joined #openstack-cinder18:59
openstackgerritMerged openstack/cinder stable/train: Disable siblings for the cinderlib functional tests  https://review.opendev.org/74026319:30
*** sonivi has quit IRC19:40
openstackgerritRajini Karthik proposed openstack/cinder master: Rebranding of VxFlex OS driver to PowerFlex  https://review.opendev.org/74052219:51
*** geguileo has quit IRC19:54
*** geguileo has joined #openstack-cinder19:55
*** hemna has quit IRC19:59
*** hemna has joined #openstack-cinder19:59
*** Lucas_Gray has joined #openstack-cinder20:11
*** jv_ has quit IRC20:22
*** spatel has joined #openstack-cinder20:34
*** spatel has quit IRC20:34
*** markvoelker has quit IRC20:42
*** vishakha has quit IRC20:48
openstackgerritMerged openstack/cinder stable/ussuri: Kaminario: Fix unique_fqdn_network option  https://review.opendev.org/74142820:59
*** baojg has quit IRC21:12
*** baojg has joined #openstack-cinder21:21
*** hamalq has joined #openstack-cinder21:28
*** hamalq_ has quit IRC21:31
*** baojg has quit IRC21:34
*** baojg has joined #openstack-cinder21:35
*** melwitt is now known as jgwentworth21:49
*** deiter has quit IRC22:14
*** baojg has quit IRC22:15
*** baojg has joined #openstack-cinder22:17
*** markvoelker has joined #openstack-cinder22:22
*** markvoelker has quit IRC22:27
*** ociuhandu has joined #openstack-cinder22:39
*** ociuhandu has quit IRC22:44
*** deiter has joined #openstack-cinder22:46
*** markvoelker has joined #openstack-cinder22:47
*** markvoelker has quit IRC22:52
*** tosky_ has joined #openstack-cinder22:52
*** tosky has quit IRC22:53
*** tosky_ is now known as tosky23:01
*** deiter has quit IRC23:15
*** andrebeltrami has quit IRC23:37
*** tosky has quit IRC23:38
*** hamalq has quit IRC23:38

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!