Thursday, 2015-09-24

*** earlephilhower has quit IRC00:02
*** diogogmt has joined #openstack-cinder00:03
*** zhangjn has quit IRC00:04
*** _fortis has quit IRC00:33
*** tsekiyam_ has joined #openstack-cinder00:41
*** kfox1111 is now known as kfox1111_away00:42
*** _fortis has joined #openstack-cinder00:43
*** tsekiyama has quit IRC00:44
*** tsekiyam_ has quit IRC00:45
*** mtanino has quit IRC00:46
*** edmondsw has quit IRC00:46
*** rhagarty_ has quit IRC00:47
*** rhagarty__ has quit IRC00:47
*** boris-42 has quit IRC00:49
*** zhangjn has joined #openstack-cinder00:52
*** annegentle has joined #openstack-cinder00:54
*** p0rtal has quit IRC00:56
*** leeantho has quit IRC00:57
*** p0rtal has joined #openstack-cinder00:57
*** haomaiwang has joined #openstack-cinder00:57
*** annegentle has quit IRC00:59
*** haomaiwang has quit IRC01:00
*** haomaiwa_ has joined #openstack-cinder01:00
*** haomaiwa_ has quit IRC01:01
*** haomaiwang has joined #openstack-cinder01:01
*** p0rtal has quit IRC01:01
*** akerr is now known as akerr_away01:03
*** haomaiwang has quit IRC01:12
*** Apoorva has quit IRC01:24
*** bill_az has quit IRC01:25
*** Lee1092 has joined #openstack-cinder01:28
*** davechen has joined #openstack-cinder01:30
*** esker has joined #openstack-cinder01:36
*** geguileo has quit IRC01:36
*** davechen1 has joined #openstack-cinder01:37
*** mudassirlatif has quit IRC01:39
*** davechen has quit IRC01:40
*** zhenguo has joined #openstack-cinder01:43
*** davechen has joined #openstack-cinder01:43
*** geguileo has joined #openstack-cinder01:45
*** davechen1 has quit IRC01:46
*** julim has joined #openstack-cinder01:51
*** davechen1 has joined #openstack-cinder01:55
*** davechen has quit IRC01:57
*** haomaiwang has joined #openstack-cinder01:57
*** chenying has joined #openstack-cinder01:58
*** rushil has joined #openstack-cinder01:58
*** haomaiwang has quit IRC02:01
*** markstur has left #openstack-cinder02:01
*** rushil_ has joined #openstack-cinder02:01
*** haomaiwa_ has joined #openstack-cinder02:01
*** chenying11 has joined #openstack-cinder02:02
*** DericHorn-HP has joined #openstack-cinder02:03
*** rushil has quit IRC02:04
openstackgerritLisaLi proposed openstack/cinder: Data transfer between encrypted volumes and images
openstackgerritLisaLi proposed openstack/cinder: Copy encryptors from Nova to Cinder
*** DericHorn-HP has quit IRC02:16
*** DericHorn-HP has joined #openstack-cinder02:18
*** rushil_ has quit IRC02:23
*** rushil has joined #openstack-cinder02:24
*** esker has quit IRC02:27
*** esker has joined #openstack-cinder02:28
*** akerr_away is now known as akerr02:33
*** DericHorn-HP has quit IRC02:33
*** r-daneel has quit IRC02:33
*** rushil has quit IRC02:33
*** DericHorn-HP has joined #openstack-cinder02:38
openstackgerritLisaLi proposed openstack/cinder: Copy encryptors from Nova to Cinder
*** annegentle has joined #openstack-cinder02:43
*** ericksonsantos has quit IRC02:47
*** annegentle has quit IRC02:48
openstackgerritLisaLi proposed openstack/cinder: Data transfer between encrypted volumes and images
*** DericHorn-HP has quit IRC02:52
*** ericksonsantos has joined #openstack-cinder02:53
openstackgerritLisaLi proposed openstack/cinder: Create encrypted volumes from images
*** salv-orl_ has joined #openstack-cinder02:59
*** DericHorn-HP has joined #openstack-cinder03:00
*** haomaiwa_ has quit IRC03:01
*** haomaiwang has joined #openstack-cinder03:01
*** salv-orlando has quit IRC03:02
*** diogogmt has quit IRC03:04
*** akerr_ has joined #openstack-cinder03:09
*** akerr_ has quit IRC03:09
openstackgerritLisaLi proposed openstack/cinder: Prevent uploading encrypted volume to images
*** DericHorn-HP has quit IRC03:10
*** akerr has quit IRC03:12
*** davechen1 is now known as davechen03:20
*** DericHorn-HP has joined #openstack-cinder03:26
*** edtubill has joined #openstack-cinder03:38
*** julim has quit IRC03:41
*** chenying11 has quit IRC03:48
*** harlowja_at_home has joined #openstack-cinder03:48
*** martyturner has joined #openstack-cinder03:48
*** chenying11 has joined #openstack-cinder03:48
*** martyturner has quit IRC03:52
openstackgerritTina Tang proposed openstack/cinder: Clone cg support in VNX driver
*** haomaiwang has quit IRC04:01
*** haomaiwang has joined #openstack-cinder04:01
*** geguileo has quit IRC04:14
*** geguileo has joined #openstack-cinder04:15
*** jwcroppe has joined #openstack-cinder04:26
*** chenying has quit IRC04:30
*** chenying11 has quit IRC04:30
*** annegentle has joined #openstack-cinder04:32
*** annegentle has quit IRC04:36
*** garthb has joined #openstack-cinder04:39
*** fifieldt has joined #openstack-cinder04:52
*** edtubill has quit IRC04:56
*** haomaiwang has quit IRC05:01
*** haomaiwang has joined #openstack-cinder05:01
*** ircuser-1 has quit IRC05:09
*** harlowja_at_home has quit IRC05:11
*** garthb has quit IRC05:13
*** dsariel_ has joined #openstack-cinder05:15
*** yrabl has quit IRC05:24
openstackgerritYuriy Nesenenko proposed openstack/cinder: Small optimization in Block Device driver
openstackgerritYuriy Nesenenko proposed openstack/cinder: Small optimization in Block Device driver
*** ankit_ag has joined #openstack-cinder05:31
*** vgridnev has joined #openstack-cinder05:36
*** mudassirlatif has joined #openstack-cinder05:37
*** vgridnev has quit IRC05:52
*** annasort has quit IRC05:58
*** annasort_ has joined #openstack-cinder05:58
*** annasort_ is now known as annasort05:58
*** jwcroppe has quit IRC05:59
*** jwcroppe has joined #openstack-cinder06:00
*** nkrinner has joined #openstack-cinder06:00
*** haomaiwang has quit IRC06:01
*** haomaiwang has joined #openstack-cinder06:01
*** DericHorn-HP has quit IRC06:02
*** esker has quit IRC06:04
*** esker has joined #openstack-cinder06:04
*** jwcroppe has quit IRC06:04
openstackgerritWilson Liu proposed openstack/cinder: Huawei driver add check before use a QoS
*** p0rtal has joined #openstack-cinder06:07
*** p0rtal has quit IRC06:07
*** p0rtal has joined #openstack-cinder06:07
*** vgridnev has joined #openstack-cinder06:10
*** cfriesen__ has quit IRC06:11
*** vlaza has joined #openstack-cinder06:11
openstackgerritWilson Liu proposed openstack/cinder: Huawei driver add check before use a QoS
openstackgerritVincent Hou proposed openstack/cinder: Implement update_migrated_volume for NFS driver
*** vgridnev has quit IRC06:15
*** chenying has joined #openstack-cinder06:15
*** chenying11 has joined #openstack-cinder06:15
*** ircuser-1 has joined #openstack-cinder06:16
*** bkopilov has quit IRC06:17
*** BharatK has joined #openstack-cinder06:18
*** yrabl has joined #openstack-cinder06:18
*** bkopilov has joined #openstack-cinder06:20
*** annegentle has joined #openstack-cinder06:20
*** vgridnev has joined #openstack-cinder06:25
*** annegentle has quit IRC06:25
*** markvoelker has quit IRC06:27
*** p0rtal_ has joined #openstack-cinder06:30
*** p0rtal has quit IRC06:33
*** boris-42 has joined #openstack-cinder06:34
*** p0rtal_ has quit IRC06:34
*** p0rtal has joined #openstack-cinder06:35
openstackgerritCedric Zhuang proposed openstack/cinder: Add retype logic in manage_existing for VNX
*** bkopilov has quit IRC06:39
*** p0rtal has quit IRC06:40
*** pradipta has joined #openstack-cinder06:43
*** deepakcs has joined #openstack-cinder06:46
openstackgerritWilson Liu proposed openstack/cinder: Encode username and password in Huawei driver
openstackgerritWilson Liu proposed openstack/cinder: Encode username and password in Huawei driver
*** bkopilov has joined #openstack-cinder06:54
*** salv-orl_ has quit IRC06:55
*** salv-orlando has joined #openstack-cinder06:56
*** anshul has joined #openstack-cinder06:57
*** daneyon has quit IRC07:00
*** haomaiwang has quit IRC07:01
*** daneyon has joined #openstack-cinder07:01
*** haomaiwa_ has joined #openstack-cinder07:01
openstackgerritWilson Liu proposed openstack/cinder: Use function capsulation in Huawei driver
openstackgerritVincent Hou proposed openstack/cinder: Implement update_migrated_volume for NFS driver
openstackgerritWilson Liu proposed openstack/cinder: Fix capacity report error in Huawei driver
openstackgerritVincent Hou proposed openstack/cinder: Implement update_migrated_volume for NFS driver
*** vgridnev has quit IRC07:18
*** stevemar has quit IRC07:20
*** stevemar has joined #openstack-cinder07:21
*** stevemar has quit IRC07:24
*** markvoelker has joined #openstack-cinder07:28
*** jwcroppe has joined #openstack-cinder07:28
*** markvoelker has quit IRC07:33
*** xiaolidi has joined #openstack-cinder07:46
*** ronis has joined #openstack-cinder07:47
*** mudassirlatif has quit IRC07:47
*** vgridnev has joined #openstack-cinder07:50
*** jamielennox is now known as jamielennox|away07:53
*** jordanP has joined #openstack-cinder07:56
*** haomaiwa_ has quit IRC08:01
*** haomaiwang has joined #openstack-cinder08:01
*** markus_z has joined #openstack-cinder08:04
*** annegentle has joined #openstack-cinder08:09
*** sam_wan has joined #openstack-cinder08:10
openstackgerritShay Halsband proposed openstack/cinder: XtremIO fix formatting bug
*** annegentle has quit IRC08:14
*** bluex-pl has joined #openstack-cinder08:18
bauzasDuncanT: heya good morning08:22
bauzasDuncanT: are you BST ?08:22
DuncanTMorning. I'm two hours ahead of BST08:23
bauzasDuncanT: so 10.24am now ?08:24
DuncanT11:24 now08:25
bauzasDuncanT: just asking you that because I'm sorry you felt left off when we discussed about AZs last night08:25
bauzasDuncanT: it was not the intent and jgriffith was mentioning your work about AZs08:25
bauzasDuncanT: so, I just replied to you, but something is still unclear in my mind with Cinder AZs and I had no precise answer last night08:26
bauzasDuncanT: so I'd love to get more insights08:26
DuncanTbauzas: Ah, yesterday was a bit religious festival here, so I wasn't working for the day. Not everybody can keep up to the lists every day. I totally agree some things are unclear, I just thing we should take a breath, let operators talk about a) how they have deployed it and b) how they want to deploy it08:26
bauzasDuncanT: given what I understand and what jgriffith said, Cinder AZs are an historical piece of code that was inherited from Nova when it was formerly nova-volumes08:27
DuncanTI'll take a look at your reply now08:27
DuncanTYes, they date from pre-cinder, but they are very much in use for some operators for some very sensible reasons08:27
DuncanTCan't see any emails yet08:27
bauzasDuncanT: like I explained in my email, Nova AZs are logical abstractions that don't necessarly match with network isolation or physical segregation unless you do a very specific setup08:28
DuncanTFor other operators, they just get in the way08:28
bauzasie. cells with distinct nova networks08:28
bauzasDuncanT: the 99% other deployments are sharing the same bus and database08:29
DuncanTbauzas: You can do nova AZs with compute-host level fault isolation without cells for sure. You don't get control plane resiliancy, but you can do network etc08:29
bauzasDuncanT: agreed, the failure domain is the host08:29
DuncanTYou can make your control plane and bus resilient to one of three failure easily enough08:29
DuncanTSo allow any one cell to go down without affecting the others08:29
bauzasDuncanT: but that's because you made a very specific setup08:30
bauzasDuncanT: so that you're considering that your logical abstraction maps with the physical setup you made08:30
DuncanTPeople are absolutely already doing so - breaking them is not a good thing. We should make resiliency easier, not harder08:30
bauzasDuncanT: like I said, I agree with you08:30
DuncanTbauzas: Most admin level decissions in openstack need to be made with knowledge of physical and logical constraints, that isn't new08:31
*** pradipta has quit IRC08:31
bauzasDuncanT: okay, read my email then and reach me later on08:31
DuncanTWill do. The email hasn't come throught to me yet, but I'll keep an eye out. We definitely have work to do in this area08:32
bauzasmy take on that is that I just want to make sure that everybody is on the same page on how to do volume affinity08:33
bauzasDuncanT: but I'm still puzzled on how you can deploy Cinder on a way close to cells08:33
bauzasDuncanT: so I guess the deployment would be to see the same controlplane, but cinder-volumes behind separate logical networks ?08:34
openstackgerritVincent Hou proposed openstack/cinder: Implement update_migrated_volume for NFS driver
bauzasDuncanT: so that's more a discussion about having network isolation rather than resiliency and BCP ?08:35
DuncanTSeparate physical network if you want... Some people run a separate 10gig network for each AZ08:35
DuncanTPeople are doing different things with it08:35
*** zhangjn has quit IRC08:35
DuncanTThey all follow one of the two models I posted though08:36
*** sgotliv has quit IRC08:36
bauzasDuncanT: so, I just think they used a very loosely defined abstraction model for their own purposes without having us properly designing the model08:37
bauzasDuncanT: since it was a piece of information shared between Nova and Cinder, they used in, but not in the sense we planned it08:38
*** zhangjn has joined #openstack-cinder08:38
bauzaswell, we haven't actally planned anything...08:38
DuncanTSince the abstraction works fine for them, I don't see a good reason to rip anything out without carefully planning what we aim to do. We've had the same behaviour for ages, and only now are we getting feedback - this isn't a thing that needs fixing tomorrow08:40
*** sgotliv has joined #openstack-cinder08:41
bauzasDuncanT: agreed again08:41
DuncanTJust got your email. I didn't realise the deprecation is on M not Liberty - that makes me much happier08:42
DuncanTWe can always un-deprecate it after Tokyo if that is the best plan08:42
DuncanTPersonally I think we should just make cinder able to ignore AZs and make nova not care - it always passes in the AZ to cinder, and cinder knows best. Cinder can refuse attaches if it wants, and nova doesn't need to know or care08:43
DuncanTWe want to remove more of the volume decision code from nova this cycle as-is, it being there rather than in cinder is really hampering a bunch of fixes and improvements we want to make.08:43
bauzasDuncanT: okay, I see your point08:44
bauzasDuncanT: that's certainly understandable08:44
bauzasDuncanT: FWIW, we're actively working on refinining our RPC interfaces by passing o.voo objects to the scheduler as resources08:46
bauzas(speaking of the nova-scheduler)08:46
bauzasDuncanT: because lots of operators are asking us to do volume and network affinity for our VMs08:46
*** lpetrut has joined #openstack-cinder08:46
DuncanTIt looks like we haven't done anything that is not easily unwound after Tokyo, which is great. I plan on putting a writeup of operator feedback on the list at least a week before the summit, so that people can digest the ideas before the session.08:46
DuncanTVolume affinity is something that keeps coming up, again and again. The trouble is, different people mean different things by it08:47
bauzasDuncanT: so, I was also planning to call for volunteers at Tokyo to gather feedback from Cinder and Neutron teams about their requirements for passing resources to Nova in order to allow that kind of SLA request08:47
bauzasDuncanT: that, I agree08:48
bauzasDuncanT: hence a discussion about what could be defined as affinity08:48
DuncanTWe really need detailed, worked usecases for each idea, then to work through them and see where they are compatable and incompatable08:48
DuncanTMaybe some up with some new names for the the different ideas08:49
*** lifeless has quit IRC08:49
bauzasDuncanT: fair to say that08:49
bauzasDuncanT: I guess that's the duty of the next PTL to do the planning where the sessions will happen ?08:49
bauzasDuncanT: I guess also Cinder haven't yet identified which sessions to discuss ?08:50
DuncanTI'm beginning to think we should ban the words 'affinity' and 'replication', make people come up with new terms for what they actually want, and see if we make more progress that way08:50
DuncanTWe have an etherpad of proposed sessions, but nothing nailed down yet08:50
bauzasDuncanT: I suppose mostly fishbowls + one day meetup, like Nova ?08:50
DuncanTWe tend to do the actual session planning quite late08:50
DuncanTOne moment, I'll find the etherpad08:50
*** chenying11 has quit IRC08:50
bauzasDuncanT: heh, I won't speak for Nova :p08:50
*** chenying1 has joined #openstack-cinder08:52
openstackgerritWilson Liu proposed openstack/cinder: Report *real* free capacity in Huawei driver
*** chenying1 has quit IRC08:54
*** chenying has quit IRC08:54
*** chenying has joined #openstack-cinder08:55
bauzasDuncanT: coolness, will try to follow08:59
*** haomaiwang has quit IRC09:01
*** haomaiwang has joined #openstack-cinder09:01
*** lifeless has joined #openstack-cinder09:03
*** BharatK has quit IRC09:06
*** daneyon has quit IRC09:06
*** esker has quit IRC09:06
*** daneyon has joined #openstack-cinder09:07
*** esker has joined #openstack-cinder09:07
openstackgerritVincent Hou proposed openstack/cinder: Remove the destination volume check in delete_volume
*** BharatK has joined #openstack-cinder09:20
*** rakesh_hs has joined #openstack-cinder09:23
*** rakesh_hs has left #openstack-cinder09:23
*** gouthamr has joined #openstack-cinder09:25
*** gouthamr_ has joined #openstack-cinder09:26
*** ndipanov has joined #openstack-cinder09:27
*** markvoelker has joined #openstack-cinder09:29
*** aix has quit IRC09:29
*** gouthamr has quit IRC09:30
*** markvoelker has quit IRC09:34
*** lprice has joined #openstack-cinder09:44
*** esker has quit IRC09:44
*** lprice1 has quit IRC09:44
*** BharatK has quit IRC09:49
*** davechen has left #openstack-cinder09:54
*** IanGovett has joined #openstack-cinder09:55
*** annegentle has joined #openstack-cinder09:58
*** haomaiwang has quit IRC10:01
*** haomaiwang has joined #openstack-cinder10:01
*** aix has joined #openstack-cinder10:02
*** annegentle has quit IRC10:03
openstackgerritWilson Liu proposed openstack/cinder: Fix update Huawei driver issue
*** BharatK has joined #openstack-cinder10:06
*** takedakn has joined #openstack-cinder10:20
*** zhangjn has quit IRC10:21
*** takedakn has quit IRC10:22
*** e0ne has joined #openstack-cinder10:23
*** e0ne has quit IRC10:24
*** gouthamr has joined #openstack-cinder10:24
*** anshul has quit IRC10:26
*** gouthamr_ has quit IRC10:26
*** EinstCrazy has joined #openstack-cinder10:34
*** ndipanov has quit IRC10:35
*** salv-orlando has quit IRC10:40
*** salv-orlando has joined #openstack-cinder10:40
*** anshul has joined #openstack-cinder10:42
*** zhenguo has quit IRC10:42
*** gouthamr has quit IRC10:44
*** haomaiwang has quit IRC10:47
*** haomaiwang has joined #openstack-cinder10:48
*** haomaiwa_ has joined #openstack-cinder10:48
*** haomaiwang has quit IRC10:48
*** stevemar has joined #openstack-cinder10:52
*** stevemar has quit IRC10:55
*** anshul has quit IRC10:56
*** julim has joined #openstack-cinder10:58
*** chenying has quit IRC10:59
*** haomaiwa_ has quit IRC11:01
*** haomaiwa_ has joined #openstack-cinder11:01
*** gouthamr has joined #openstack-cinder11:03
*** anshul has joined #openstack-cinder11:12
*** timcl has joined #openstack-cinder11:23
*** zhangjn has joined #openstack-cinder11:27
*** cdelatte has joined #openstack-cinder11:28
*** markvoelker has joined #openstack-cinder11:30
*** ociuhandu has quit IRC11:30
*** markvoelker has quit IRC11:34
*** alexpilotti has joined #openstack-cinder11:41
*** bswartz has quit IRC11:43
*** sam_wan has quit IRC11:44
*** takedakn has joined #openstack-cinder11:44
openstackgerritWilson Liu proposed openstack/cinder: Encode username and password in Huawei driver
*** annegentle has joined #openstack-cinder11:47
*** porrua has joined #openstack-cinder11:48
*** chenying has joined #openstack-cinder11:49
*** annegentle has quit IRC11:52
*** takedakn has quit IRC11:52
*** chenying has quit IRC11:53
*** takedakn has joined #openstack-cinder11:56
*** haomaiwa_ has quit IRC12:01
*** haomaiwang has joined #openstack-cinder12:01
*** bkopilov has quit IRC12:03
*** ankit_ag has quit IRC12:06
*** haomaiwang has quit IRC12:09
openstackgerritWilson Liu proposed openstack/cinder: Report *real* free capacity in Huawei driver
*** bswartz has joined #openstack-cinder12:13
*** sayali has quit IRC12:17
*** dims__ has quit IRC12:19
*** sayali has joined #openstack-cinder12:20
*** dims_ has joined #openstack-cinder12:20
*** annegentle has joined #openstack-cinder12:22
*** raildo-afk is now known as raildo12:29
*** markvoelker has joined #openstack-cinder12:31
*** cdelatte has quit IRC12:31
*** edmondsw has joined #openstack-cinder12:32
*** ociuhandu has joined #openstack-cinder12:33
*** annegentle has quit IRC12:34
*** jamielennox|away is now known as jamielennox12:37
*** esker has joined #openstack-cinder12:38
*** esker has quit IRC12:38
*** esker has joined #openstack-cinder12:39
*** zhangjn has quit IRC12:46
*** EinstCrazy has quit IRC12:46
*** martyturner has joined #openstack-cinder12:50
*** deepakcs has quit IRC12:51
*** akerr has joined #openstack-cinder12:53
*** stevemar has joined #openstack-cinder12:53
*** toabctl has joined #openstack-cinder12:54
*** BharatK has quit IRC12:54
*** tpsilva has joined #openstack-cinder12:55
*** takedakn has quit IRC12:56
*** stevemar has quit IRC12:57
*** takedakn has joined #openstack-cinder12:59
*** cdelatte has joined #openstack-cinder13:03
*** markstur has joined #openstack-cinder13:04
*** markstur has left #openstack-cinder13:05
*** larsks has quit IRC13:07
*** sayali has quit IRC13:10
*** bill_az has joined #openstack-cinder13:10
*** EinstCrazy has joined #openstack-cinder13:12
*** sayali has joined #openstack-cinder13:13
*** alejandrito has joined #openstack-cinder13:14
*** mriedem_away is now known as mriedem13:18
*** jungleboyj_ has joined #openstack-cinder13:20
*** takedakn has quit IRC13:23
*** Yogi1 has joined #openstack-cinder13:24
*** merooney has joined #openstack-cinder13:26
*** jgregor has joined #openstack-cinder13:27
*** bluex-pl has quit IRC13:29
*** xyang1 has joined #openstack-cinder13:29
*** mriedem is now known as mriedem_away13:30
*** mriedem_away is now known as mriedem_meeting13:30
*** takedakn has joined #openstack-cinder13:33
*** dustins has joined #openstack-cinder13:36
*** bluex-pl has joined #openstack-cinder13:37
erlondavid-lyle: ping13:37
*** tomh_ has joined #openstack-cinder13:39
*** esker has quit IRC13:41
*** haomaiwang has joined #openstack-cinder13:43
*** gouthamr has quit IRC13:43
*** zhangjn has joined #openstack-cinder13:45
*** zhangjn has quit IRC13:46
*** zhangjn has joined #openstack-cinder13:47
*** lprice1 has joined #openstack-cinder13:47
*** zhangjn has quit IRC13:47
*** zhangjn has joined #openstack-cinder13:48
*** zhangjn has quit IRC13:49
*** lprice has quit IRC13:50
*** cdelatte has quit IRC13:54
*** timcl has quit IRC13:54
*** eharney has joined #openstack-cinder13:56
*** EinstCrazy has quit IRC13:57
*** julim has quit IRC13:58
*** zhangjn has joined #openstack-cinder13:59
*** haomaiwang has quit IRC14:01
*** haomaiwang has joined #openstack-cinder14:01
jgriffithbauzas: You're still digging on that?14:02
bauzasjgriffith: I'd rather say that AZs are haunting me :)14:02
*** lprice1 has quit IRC14:03
jgriffithbauzas: I stil stand by what I said yesterday.  On the cinder side they pretty much do nothing14:03
bauzasjgriffith: yeah I see14:03
jgriffithbauzas: we have lots of great unit tests for it though :)14:04
smcginnisIf we want to change the concept (and probably the name) their could be some value there. But as it is now... meh.14:04
bauzasjgriffith: since DuncanT is proposing to revive those in Cinder, I just want to understand if all of that goes in the righr direction14:04
jgriffithI'm still trying to parse DuncanT statements and figure out what I'm missing14:04
bauzasjgriffith: my only problem is that I definitely miss some Cinder mojo :)14:04
*** martyturner has quit IRC14:05
jgriffithI don't know of anywhere that it's used in a volume-service, all api checks AFAIK and that's about it.  arbitrary14:05
jgriffithbauzas: well load up devstack and get it back :)14:05
*** rhagarty has joined #openstack-cinder14:05
*** rhagarty_ has joined #openstack-cinder14:05
jgriffithstart tracing that code :)14:05
*** Yogi1 has quit IRC14:05
bauzasjgriffith: heh14:06
*** martyturner has joined #openstack-cinder14:06
jgriffithbauzas: :)14:06
*** mriedem_meeting is now known as mriedem14:07
* DuncanT starts on the spec earlier than planned to hopefully get across what people are trying to do. TLDR: We definitely need a way to tell cinder to get AZs completely and utterly out of the way. This is a very common desire and seems like the way the majority of ops want to deploy. Great. But a buncho f other people want AZs to mean something, for a good reason,14:07
* DuncanT and we should support that too. Preferably as transparently as possible.14:07
*** mudassirlatif has joined #openstack-cinder14:07
DuncanTDraft spec up soon14:07
jgriffithI think14:08
*** diogogmt has joined #openstack-cinder14:09
bauzasDuncanT: feel free to CC me when it's up, I'd be glad to sneak peek on it14:09
DuncanTbauzas: Will do14:09
*** breitz has quit IRC14:10
*** breitz has joined #openstack-cinder14:11
*** dims__ has joined #openstack-cinder14:11
*** j_king has quit IRC14:12
*** BharatK has joined #openstack-cinder14:13
*** j_king has joined #openstack-cinder14:13
*** dims_ has quit IRC14:14
*** takedakn has quit IRC14:15
*** jaypipes has joined #openstack-cinder14:19
*** amoturi has joined #openstack-cinder14:19
*** mc_nair has joined #openstack-cinder14:21
guitarzanhey, did devstack used to use sqlite and now use mysql?14:23
*** diogogmt has quit IRC14:23
smcginnisguitarzan: If I remember right, unit tests use sqlite but devstack uses mysql.14:24
guitarzansmcginnis: it does now, I'm just wondering if it used to use sqlite14:24
guitarzanI thought I remembered /etc/cinder/state/cinder.sqlite14:25
smcginnisguitarzan: Not since I've been using it. But maybe.14:25
eharneyif it did it must have been quite a while ago, because i don't recall ever seeing that14:25
guitarzanyeah, it's no problem14:25
guitarzanI had a dev environment that ran sqlite and it's running into lots of sqlite lock problems14:26
*** jamielennox is now known as jamielennox|away14:26
*** ndipanov has joined #openstack-cinder14:26
*** dsariel_ has quit IRC14:27
guitarzanbut it's probably the case where no one runs sqlite so no one else has this problem :)14:27
*** martyturner has quit IRC14:27
*** jwcroppe has quit IRC14:28
*** cdelatte has joined #openstack-cinder14:29
*** mc_nair has quit IRC14:29
*** gouthamr has joined #openstack-cinder14:29
*** thingee has joined #openstack-cinder14:30
*** edtubill has joined #openstack-cinder14:31
openstackgerritShay Halsband proposed openstack/cinder: XtremIO fix remapping bug
jgriffithguitarzan: no14:31
guitarzanyeah, I might be the only one who has run sqlite then :)14:32
*** lprice1 has joined #openstack-cinder14:32
*** timcl has joined #openstack-cinder14:33
*** mc_nair has joined #openstack-cinder14:36
*** julim has joined #openstack-cinder14:38
*** martyturner has joined #openstack-cinder14:39
*** zhangjn has quit IRC14:39
*** EinstCrazy has joined #openstack-cinder14:40
*** zhangjn has joined #openstack-cinder14:40
*** zhangjn has quit IRC14:40
*** garthb has joined #openstack-cinder14:40
*** salv-orlando has quit IRC14:41
*** vlaza has left #openstack-cinder14:41
*** zhangjn has joined #openstack-cinder14:41
*** zhangjn has quit IRC14:41
*** salv-orlando has joined #openstack-cinder14:41
*** zhangjn has joined #openstack-cinder14:42
*** zhangjn has quit IRC14:42
*** zhangjn has joined #openstack-cinder14:43
*** zhangjn has quit IRC14:43
*** zhangjn has joined #openstack-cinder14:44
*** ronis has quit IRC14:44
*** cfriesen__ has joined #openstack-cinder14:45
*** mtanino has joined #openstack-cinder14:45
bswartzsqlite isn't a real database14:47
openstackgerritVincent Hou proposed openstack/cinder: Implement update_migrated_volume for NFS driver
*** yrabl has quit IRC14:49
*** cdelatte has quit IRC14:51
*** Yogi1 has joined #openstack-cinder14:51
*** stevemar has joined #openstack-cinder14:54
*** cdelatte has joined #openstack-cinder14:56
*** tsekiyama has joined #openstack-cinder14:56
guitarzanbswartz: I don't think anyone is arguing that :)14:56
bswartzokay good14:57
*** timcl has quit IRC14:59
*** haomaiwang has quit IRC15:01
*** haomaiwang has joined #openstack-cinder15:01
*** annegentle has joined #openstack-cinder15:01
*** haomaiwang has quit IRC15:03
*** crose has joined #openstack-cinder15:03
*** fifieldt has quit IRC15:05
*** erhudy has joined #openstack-cinder15:06
openstackgerritwanghao proposed openstack/cinder: CG creation should be scheduled on backend level
*** annegentle has quit IRC15:09
*** zhangjn_ has joined #openstack-cinder15:09
*** annegentle has joined #openstack-cinder15:09
*** cdelatte has quit IRC15:11
*** timcl has joined #openstack-cinder15:12
*** zhangjn has quit IRC15:13
*** jordan_ has joined #openstack-cinder15:16
*** jordanP has quit IRC15:16
*** annasort has quit IRC15:17
*** annasort has joined #openstack-cinder15:17
*** diogogmt has joined #openstack-cinder15:21
*** annasort has quit IRC15:22
*** alexpilotti has quit IRC15:23
*** kfox1111_away is now known as kfox111115:26
*** zhangjn_ has quit IRC15:26
*** zhangjn has joined #openstack-cinder15:27
*** zhangjn has quit IRC15:27
*** zhangjn has joined #openstack-cinder15:28
*** BharatK has quit IRC15:28
*** zhangjn has quit IRC15:28
*** salv-orlando has quit IRC15:28
*** zhangjn has joined #openstack-cinder15:28
*** kevincarr1991 has joined #openstack-cinder15:33
kevincarr1991I am trying to install cinder on each compute node and point them to a nfs. is this possible?15:33
*** esker has joined #openstack-cinder15:34
*** zhangjn_ has joined #openstack-cinder15:35
*** earlephilhower has joined #openstack-cinder15:36
*** zhangjn has quit IRC15:37
*** hemnafk is now known as hemna15:37
*** esker has quit IRC15:40
tbarronkevincarr1991: yes, it is possible15:40
kevincarr1991tbarron: do you know of any good documentation for this?15:41
*** krtaylor has quit IRC15:41
tbarronkevincarr1991: are you using the generic nfs driver?15:41
*** esker has joined #openstack-cinder15:41
kevincarr1991tbarron: yes15:41
tbarronkevincarr1991: in that case, the regular openstack doc should cover it.  I haven't worked with the generic driver much, but what issues are you having?15:42
kevincarr1991tbarron: well the regular install has the setup for the controller and the cinder server. I dont have a cinder server to work with. I only have 3 compute nodes, network node, and a controller node. I have a storage array that I want to mount using cinder. When I try the install guide I am unsure of what do for Cinder Server portion.15:44
tbarronkevincarr1991: one approach would be to get cinder running w/o nfs (with lvm) on one of your compute nodes, then modify it to do nfs.15:46
tbarronkevincarr1991: divide and conquer15:46
*** zhangjn has joined #openstack-cinder15:46
kevincarr1991tbarron: would the basic install guide be sufficient for that?15:46
tbarronkevincarr1991: I think so, but honestly I'm a cinder developer who works in simplified devstack environment and am not the best15:47
tbarronperson to advise on how to setup cinder in your multinode production environment.15:48
kevincarr1991tbarron: well I appreciate the help and will give it a shot15:48
tbarronkevincarr1991: my point is that that issue is really orthogonal to getting cinder to work with nfs15:48
tbarronfor that you just need whatever node that runs cinder to have net connectivity and nfs access to the share or shares15:48
tbarronused to back the cinder volumes.15:49
*** zhangjn_ has quit IRC15:49
tbarronkevincarr1991: the basic, first thing to do is to manually mount the nfs shares running as the openstack user and ensure that you can read/write to them.15:49
tbarronif the manual mount and access won't work, then openstack won't be able to do it either.15:50
kevincarr1991tbarron: I am able to mount the nfs15:50
tbarronkevincarr1991: so then your main problem is getting cinder running.  After that, I think the docs will show you how to modify cinder.conf and restart services.15:51
*** _cjones_ has quit IRC15:51
*** annasort has joined #openstack-cinder15:52
kevincarr1991tbarron: understood. Am I able to install the entire cinder service on the controller?15:53
*** esker has quit IRC15:53
*** krtaylor has joined #openstack-cinder15:54
mc_nairhey all - question about string formatting...  Is there a preference between using "{} string {}".format("my", "replace") vs. "%(my)s string %(replace)s" % ("my": "my", "replace": "replace")?15:54
smcginnismc_nair: My preference is "%(stuff)s" % {'stuff': val} because it's used in over 90% of the code.15:55
*** diogogmt has quit IRC15:55
smcginnismc_nair: Just my preference though.15:55
*** diogogmt has joined #openstack-cinder15:56
tbarronkevincarr1991: as I say, I'm not an expert on how to deploy multinode openstack in production, but that is probably what I would do.15:57
guitarzanmc_nair: for translation reasons, the dict method is absolutely preferred15:58
tbarronkevincarr1991: your 'controller' is I assume the node running your centralized services and multiple compute nodes consume these.15:58
kevincarr1991tbarron: that is correct15:59
mc_nairsmcginnis, guitarzan - cool, thanks.  I noticed that was the trend but didn't know that was the reason.  So then guessing it's kosher to -1 a patch for using ".format" a bunch instead of the dict method?15:59
tbarronkevincarr1991: yes, that's what I would do in your shoes then.15:59
tbarronkevincarr1991: I have to run off to a meeting for an hour or so15:59
tbarronkevincarr1991: but will check back in later15:59
kevincarr1991tbarron: OK, thank you for your help15:59
*** ronis has joined #openstack-cinder16:02
*** cdelatte has joined #openstack-cinder16:08
*** ndipanov has quit IRC16:11
*** _cjones_ has joined #openstack-cinder16:12
*** timcl has quit IRC16:15
jgriffithkevincarr1991: yes, many people commonly do install c-vol services on the controller node.  Especially when using an external backend device16:16
*** cdelatte has quit IRC16:17
*** esker has joined #openstack-cinder16:17
kevincarr1991jgriffith: Do you know of any good documentation? Or would i just follow the basic guide doing all of the commands on the controller node.16:18
jgriffithjungleboyj_: when you get a moment, probably ready to lift our -2's16:18
*** p0rtal has joined #openstack-cinder16:18
jgriffithjungleboyj_:  and
jgriffithmaybe others... not sure16:19
jgriffithkevincarr1991: depending on what you're goal is I'd go with the openstack-docs16:19
jgriffithkevincarr1991: there is no shortage of opinions on the web about how you should deploy :)16:19
jgriffithkevincarr1991: but they all boil down to "what you might need"16:20
jungleboyj_Ok, I have that on the todo list for today.  I will start pulling those off.16:20
*** markvoelker_ has joined #openstack-cinder16:21
kevincarr1991jgriffith: well i mainly just need to be able to have my compute nodes store instances on the nfs and i am not finding many helpful docs, but i might be looking in the wrong places16:22
*** nkrinner has quit IRC16:23
*** cdelatte has joined #openstack-cinder16:23
*** ociuhandu has quit IRC16:24
*** afazekas has quit IRC16:24
*** markvoelker has quit IRC16:24
*** eikke has quit IRC16:24
*** wolsen has quit IRC16:24
*** pckizer_ has joined #openstack-cinder16:24
*** pckizer_ has quit IRC16:24
*** pckizer_ has joined #openstack-cinder16:24
*** ronis has quit IRC16:24
*** kbyrne has quit IRC16:24
*** pckizer has quit IRC16:24
*** leeantho has joined #openstack-cinder16:24
hemnamriedem, jgriffith re: boot from volume source=image16:24
*** ronis_ has joined #openstack-cinder16:24
hemnaI was under the impression that the momentum for users/operators is to start using the openstackclient ?16:24
*** leeantho has quit IRC16:24
*** SergeyLukjanov has quit IRC16:24
*** afazekas has joined #openstack-cinder16:24
*** SergeyLukjanov has joined #openstack-cinder16:24
*** wolsen has joined #openstack-cinder16:24
*** kbyrne has joined #openstack-cinder16:24
*** eikke has joined #openstack-cinder16:25
jgriffithhemna: I think some of that whole thread is a bit intertwined16:25
hemnayah, there are a few threads of discussion going on16:25
jgriffithhemna: also openstack client works by calling nova-client, cinder-client etc :)16:25
hemnathe AZ stuff and the Nova orchestration stuffs, and the timeouts, etc.16:26
jgriffithhemna: it just aggreagtes them for you, and gives you a better shell etc, long term that may be different16:26
hemnabut I was curious, if we could simply deprecate the nova orchestration, and then migrate that into the openstack client.16:26
jgriffithhemna: yeah... possibly16:26
*** dustins is now known as dustins|afk16:26
hemnaask the openstack client to create the volume from image, and deal w/ the waiting16:26
*** leeantho_ has joined #openstack-cinder16:26
thingeejgriffith, hemna lots of great stuff for Cinder came out of the 1.6.0 release of os-client
hemnathen have it simply pass in the volume id to the nova client to boot.16:27
hemnaand get nova out of the job of orchestration for this piece.16:27
jgriffithhemna: turns out there are more folks that use that than I thought (and have made their own fixes etc)16:27
jgriffiththingee: indeed!16:27
*** porrua has quit IRC16:27
openstackgerritEinst Crazy proposed openstack/cinder: Gluster: extand volume to the right path
jgriffithIt's a cool project16:27
hemnayah I saw the several responses of "I need this! we use it!" comments16:27
hemnahence deprecation16:27
hemnaand not outright removal16:28
hemnaand see if we can get those operators to migrate to the openstack client eventually ?16:28
hemnamaybe I'm drunk16:28
jgriffiththingee: I'm still trying to decide if I like cliff or not :)  Some things about it are great, some things newish enough that I'm having a hard time figuring it out16:28
jgriffithhemna: sure16:29
jgriffithhemna: the other thing is make some minor mods if/where needed to avoid the timeout issue altogether :)16:29
hemnayah if that's possible16:29
thingeehemna: keystone is doing it. deprecation warnings when you use the keystone client cli directly16:29
openstackgerritScott DAngelo proposed openstack/cinder-specs: cinder-api-microversions
hemnaso if we get the novaclient to wait for cinder, don't we run into possible nova api -> compute manager timeouts as well ?16:30
thingeehemna: dunno when the cut is though for them16:30
hemnait could be a cascade of timeout issues16:30
hemnaI'd just hate to see us put in a lot of hacks to make it work16:31
*** ociuhandu has joined #openstack-cinder16:31
jgriffithhemna: TBF it does "work"16:31
hemnawhen the long term is to just use an external (non novaclient) orchestration.16:31
jgriffithhemna: there are just some edge cases/timing issues.  BUT, it also appears based on what mriedem posted yesterday there have been some fixes to that problem already16:31
hemnaas long as the volume/image is small enough for it to work right ?16:32
jgriffithhemna: smaller image, faster storage etc16:32
jgriffithhemna: but they have a configurable timeout in their now16:33
jgriffithso that solves a lot of the problem16:33
jgriffithI think16:33
jgriffithhemna: certainly a good thing to look at... I just never liked that wrapper :)  But, regardless... that doesn't really matter16:34
*** EinstCrazy has quit IRC16:34
jgriffithhemna: also keep in mind, your idea of OpenStack client is fine... but you also have to remember that the functionality is in Horizon.16:34
scottdaleeantho_: hemna Was there ever a POC patch to remove the Volume manager locks? or just a spec?16:35
hemnaI just wonder if the cinder timeout was long enough, that the novaclient itself might timeout to the nova api, etc.  dunno.16:35
jgriffithhemna: I know david-lyle was fixing some things up on that side already16:35
hemnascottda, Anthony did have a poc patch16:35
leeantho_scottda, there was an initial poc patch from a while ago.  I can find it if you want it16:35
scottdaOK, couldn't find it in Launchpad...16:35
hemnascottda, that patch showed all the tempest failures caused by returning the VolumeIsBusy16:35
hemnajgriffith, yah I get that.   The good thing about Horizon is that it doesn't have to use novaclient to do it all, it could call cinder directly to create the volume from image then pass in the volume id to nova boot.16:36
scottdaleeantho_: That'd be great if you could find it. I was going to try to make it work with cinder-api-microversions16:37
*** jordan_ has quit IRC16:37
leeantho_scottda, here it is.  not update for a long time16:38
*** Zhongjun has joined #openstack-cinder16:38
scottdathanks. I'll see if I can merge it up locally.16:38
*** BharatK has joined #openstack-cinder16:40
jgriffithbswartz: tbarron
openstackLaunchpad bug 1499334 in Manila "Regression: Proprietary code needed to use NetApp with Cinder and Manila" [Undecided,In progress] - Assigned to Thomas Bechtold (toabctl)16:41
jgriffith^^ ?16:41
jgriffithbswartz: tbarron I'm wondering if there's a mistake/misunderstanding there?16:42
jgriffithbswartz: tbarron also noticed that I can't seem to find the source for the library either?16:42
toabctljgriffith: I'm the guy who filled the bug. And I talked to bswartz already. and also to eharney16:43
jgriffithbswartz: tbarron the license is a bit of a problem I think16:43
jgriffithtoabctl: ahh... ok.  So what's the story?16:43
ameadethe source is in the pypi package, it's python16:43
jgriffithameade: that's not the point16:44
toabctljgriffith: well - what I wrote in the bug. some apache-2.0 licensed code was replaced with netapp_lib which is proprietary16:44
toabctljgriffith: which is a problem for distributions (and I guess deployers) because we can not ship the needed code16:44
*** raildo is now known as raildo-afk16:45
jgriffithtoabctl: indeed16:45
toabctljgriffith: I think there are other vendors doing similar things but in this case it's a regression because there is already working code released under the apache license16:45
toabctland imho that's not acceptable in a opensource project.16:45
jgriffithtoabctl: sadly there are quite a few vendors using that model, BUT all the ones that I've looked at are using OSS licenses16:46
jgriffithtoabctl: and I completely agree, which is why I pinged bswartz and tbarron :)16:46
*** BharatK has quit IRC16:47
*** anshul has quit IRC16:47
*** raildo-afk is now known as raildo16:48
jgriffithtoabctl: I agree with the proposal for a revert, but need to get input from them first16:48
*** esker has quit IRC16:50
bswartzjgriffith: just got back from lunch, what's up?16:52
openstackLaunchpad bug 1499334 in Manila "Regression: Proprietary code needed to use NetApp with Cinder and Manila" [Undecided,In progress] - Assigned to Thomas Bechtold (toabctl)16:53
bswartzoh yeah toabctl talked to me about this this morning16:53
jgriffithbswartz: yeah, he filled me in :)16:53
bswartzI've recused myself on that bug due to a conflict of interest16:53
jgriffithbswartz: I always knew you were a smart person :)16:53
openstackgerritJohn Griffith proposed openstack/cinder: Revert "Update NetApp Drivers to use netapp_lib"
openstackgerritJohn Griffith proposed openstack/cinder: Revert "Update NetApp Drivers to use netapp_lib"
*** jordanP has joined #openstack-cinder16:56
*** porrua has joined #openstack-cinder16:57
mriedemjungleboyj_: fyi, thorst expressed interest in a nova call this morning on helping with multiattach stuff in mitaka16:57
mriedemsince he's done some of that already16:57
tbarronjgriffith: in my role as netapp employee I think i can definitively say "no comment" :-)16:59
*** p0rtal_ has joined #openstack-cinder16:59
jgriffithtbarron: fair enough16:59
*** zhangjn_ has joined #openstack-cinder16:59
*** zhangjn has quit IRC16:59
* tbarron runs to a meeting :-)17:00
*** ronis_ has quit IRC17:01
*** p0rtal has quit IRC17:01
jungleboyj_mriedem: Cool.  Thanks for letting me know.17:02
*** p0rtal_ has quit IRC17:03
*** markus_z has quit IRC17:03
*** mriedem has quit IRC17:03
*** ronis_ has joined #openstack-cinder17:04
scottdahemna: I'm going to push out taking Cinder-Nova API changes to today's Nova meeting if we don't have a Nova spec. Is that OK with you?17:05
hemnaI had started on the spec last night17:06
jgriffithhemna: jungleboyj_ something to consider on reviews in the future :)17:06
hemnascottda, I just hadn't finished it yet17:06
hemnascottda, it's always something that pulls me away.  *sigh*17:06
jgriffithso the problem here is we can't do a simple revert of just that patch due to changes after the merge17:06
scottdaOK, I'm not  wanting to nag you, just looking at the clock and all.17:06
hemnascottda, ok, it might be worth pushing it to next week17:06
jgriffithI hate to propose complete removal of the driver, but I also don't think I should have to go in and try and fix merge conflicts on the netapp drivers17:07
hemnathat'll give us more time to churn on it and iron some issues out17:07
hemnascottda, sorry man :(17:07
* jgriffith is talking to himself :)17:07
scottdathingee: We had said we'd bring cinder-nova API changes to the Nova meeting this week, but hemna and I think it'd be better to wait....17:07
scottdaWe need a Nova spec.17:07
scottdaAnd We're likely to have some POC code next week (I'm working on it now).17:07
hemnajgriffith, crap, what did I miss ?17:08
scottdahemna: It's cool. I want to keep The Boss happy. Even if he is a lame duck...17:08
hemnajgriffith, oh the license17:08
openstackLaunchpad bug 1499334 in Cinder "Regression: Proprietary code needed to use NetApp with Cinder and Manila" [Critical,In progress] - Assigned to John Griffith (john-griffith)17:08
*** zhangjn has joined #openstack-cinder17:08
*** zhangjn_ has quit IRC17:09
hemnahrmm yah17:09
hemnathat does sux0rs17:09
hemnawhen we created our lib we intentionally used the same license as openstack to avoid such issues.17:09
*** mriedem has joined #openstack-cinder17:13
hemnascottda, do you have a cinderclient poc yet ?17:13
scottdaNo, I'm working on the lock removal stuff first.17:14
scottdacinderclient is next...maybe later today but I've a few other things queued up.17:14
xyang1jgriffith: are we open to merge new features now since Mitaka is open or should we wait until 10/15?17:15
hemnaxyang1, Mitaka is open afaik.17:15
jgriffithxyang1: master is opened back up17:16
xyang1jgriffith: hemna so those -2 should be removed now?17:16
jgriffithameade: you want to fix the merge conflicts on that revert proposal or should I just remove the netapp drivers altogether?17:16
jgriffithxyang1: yes17:16
xyang1jungleboyj_: ^17:16
xyang1jungleboyj_: can you remove -2's?17:17
hemnaxyang1, yes I think so17:17
jungleboyj_Sorry, in a meeting.  Will tackle as soon as I am off.17:17
xyang1jungleboyj_: no problem17:17
jungleboyj_xyang1: THanks for patience.17:18
*** zhangjn has quit IRC17:23
*** raildo is now known as raildo-afk17:24
*** cdelatte has quit IRC17:25
*** salv-orlando has joined #openstack-cinder17:29
*** ronis_ has quit IRC17:32
*** raildo-afk is now known as raildo17:34
*** rajinir has quit IRC17:38
*** bluex-pl has quit IRC17:38
*** eharney has quit IRC17:39
*** cdelatte has joined #openstack-cinder17:40
*** delattec has joined #openstack-cinder17:40
*** yrabl has joined #openstack-cinder17:42
*** ociuhandu has quit IRC17:43
*** eharney has joined #openstack-cinder17:43
*** Yogi1 has quit IRC17:50
*** Yogi1 has joined #openstack-cinder17:52
*** aix has quit IRC17:52
*** leseb has quit IRC17:54
*** dustins|afk is now known as dustins17:56
*** annegentle has quit IRC17:57
patrickeasthey, so i have a linux+python question... anyone know how to check if a disk is writable for the root user from python code running as not the root user?17:58
akerrcouldn't you ls to check the permissions?  Or is it not even viewable as non-root?17:58
patrickeastback to chasing ghosts with and am looking for a way to catch this type of error early in os-brick
openstackLaunchpad bug 1495701 in OpenStack Compute (nova) "Sometimes Cinder volumes fail to attach with error "The device is not writable: Permission denied"" [Undecided,New]17:58
patrickeastakerr: ah yea thats a good idea17:59
patrickeastoh derp, os.stat() is what i was looking for i think18:03
*** angela-s has joined #openstack-cinder18:03
jgriffithpatrickeast: :)18:05
*** ociuhandu has joined #openstack-cinder18:06
*** p0rtal has joined #openstack-cinder18:06
eharneypatrickeast: keep in mind that i think libvirt will (usually) temporarily change permissions when attaching devices to instances18:07
eharneythough maybe that's only for files18:08
patrickeasteharney: oh, i did not know that... do you know what it does? ie restricts or adds more?18:08
openstackgerritKendall Nelson proposed openstack/cinder: Missing configuration opts from cinder.sample.conf
eharneypatrickeast: changes user/group ownership from root/root to things like qemu/qemu -- you can see by checking when the device is attached18:08
patrickeastahh gotcha18:09
eharneythey call this "dynamic ownership" i think18:09
diablo_rojogeguileo: I updated the commit message to remove the comment so it's ready for your +2 :)18:09
patrickeasthmm well a bug with that would explain the libvirt issues for that bug, but not the c-vol and c-bak ones18:10
geguileodiablo_rojo: Ok, I'll do it now :-)18:10
patrickeastafaik we don't change ownership18:10
diablo_rojogeguileo: THank you!18:10
*** timcl has joined #openstack-cinder18:10
geguileodiablo_rojo: Thank you for fixing that!!18:10
diablo_rojogeguileo: No problem :) After I get the hacking check all squared away I will get another patch out with the suggested refactor from hemna and clean up a few more things :)18:11
geguileodiablo_rojo: Awesome!!18:11
*** cdelatte has quit IRC18:11
*** delattec has quit IRC18:11
eharneypatrickeast: i'm really confused looking at one of the logs in that bug.  screen-n-cpu.txt seems to be a Nova log?18:13
jgriffithHoly netapp commits during feature freeze BATMAN!!!18:14
eharneypatrickeast: er, never mind.  need more coffee18:14
*** delattec has joined #openstack-cinder18:14
*** cdelatte has joined #openstack-cinder18:14
patrickeasteharney: yea the issue happens sometimes in n-cpu when attaching the volume to an instance, sometimes when writing backing data in c-bak and sometimes when doing image -> volume operations in c-vol18:14
patrickeastkind of anywhere we need to attach and write to a disk18:14
eharneyi see these on lio and pure jobs, i wonder if they've hit the LVM tgtd jobs18:15
patrickeastits very sparse in the gate, there are like 2 instances of the n-cpu one in the last 7 days18:15
patrickeastbut its wrecking my FC ci jobs18:15
patrickeastlike 2/3 fail from it18:15
*** salv-orlando has quit IRC18:15
patrickeasteharney: gate-tempest-dsvm-neutron-dvr uses LVM18:17
patrickeastits the first one in the list for the last 7d now
eharneyahh, ok18:18
*** Apoorva has joined #openstack-cinder18:20
tsekiyamahemna: are you around?18:24
hemnatsekiyama, no18:24
tsekiyamahemna: lol18:24
tsekiyamahemna: want to clarify somthing about
hemnaok shoot18:24
*** eharney_ has joined #openstack-cinder18:24
tsekiyamahemna: was FC+multipath+encryption working before?18:24
hemnaheh good question.  I don't believe I've tested encrypted volumes much.18:25
tsekiyamahemna: I'm wondering it was not working if cinder was returning a path starting with /dev/mapper18:25
tsekiyamahemna: because it conflicts with crypt device name nova specified.18:25
hemnabut I believe if we relied on paths existing in /dev/mapper, then there are problems with that for sure18:25
patrickeastit doesn't look like it would have been working18:25
hemnapatrickeast, +118:25
hemnafriendly names should die a fiery death18:26
*** eharney has quit IRC18:26
patrickeastshould just add code into brick to disable friendly names18:26
hemnaredhat 7 changed to default friendly names on.18:27
hemnathat caused no end of problems18:27
tsekiyamahemna: hmm18:27
tsekiyamahemna: so .. If it is not working, we don't have to care about upgrading from that version, right?18:27
hemnaso the only thing that's safe really is to know the WWN of the volume18:27
hemnatsekiyama, well I dunno, maybe a fallback mechanism as a best effort?18:28
hemnawhich is what I try to do for FC multipath devices in os-brick18:28
hemnatries 3 different mechanisms to find the mpath device before it punts18:29
hemnaand each location is different :(18:29
tsekiyamahemna: what I should do in that patch is a way to estimate how crypt device name is determined in previous version. not multipath device..18:29
hemnaand ignore the fact that multipath daemon sometimes fails to ever create the mpath device at all, on busy systems.18:29
tsekiyamahemna: it was determined from multipath device name, but was broken if the multipath device was placed on /dev/mapper/. We shouldn't have to care about that case18:30
hemnajust depends on how robust you want to be.18:31
hemnamy faith in the mpath daemon isn't overly good at the moment.18:31
stevemarthingee: oh i like this conversation about using osc :)18:32
stevemari was going to approach you all at the summit, but irc works just fine :)18:32
tsekiyamahemna: yah. maybe additional checking for /dev/mapper/<wwn> would be make things a bit more robust. For friendly name ..18:33
tsekiyamahemna: I don't know how I can estimate the old name18:33
stevemarthingee: hemna so, aside from a few of the newer v2 cinder commands, openstackclient should have pretty decent parity with the current command set18:34
stevemarfor both v1 and v218:34
stevemaras as thingee said, the keystone team started to print out deprecation warnings for their CLI, and refused any patches to it that weren't security fixes / critical18:35
hemnatsekiyama, friendly names will show up as /dev/mapper/mpathX   :(18:35
tsekiyamahemna: and X is impossible to guess :(18:36
morganit made for a pretty easy transition to the common client (though we still get asked about keystoneclient supporting keystone v3 CLI fairly often)18:36
hemnatsekiyama, correct18:36
hemnatsekiyama, only via calling multipath -ll <device> can you get it.18:36
stevemarmorgan: that peaked around kilo18:36
hemnaand running multipath -ll sux18:36
morganstevemar: I still tell people "no" but it's getting less frequent18:37
stevemarmorgan: ever since the official docs went to OSC there have been fewer and fewer questions about it18:37
tsekiyamahemna: right18:37
*** r-daneel has joined #openstack-cinder18:37
hemnastevemar, I really like the idea of having the openstack client do the orchestration for the nova boot from volume source=image18:37
hemnastevemar, where osc will call cinder to create the volume from image and deal with the waiting, then pass the volid to nova18:37
DuncanThemna: The disadvantage there is that there are many clients, not just the official python one, and plenty of room for bugs18:38
tsekiyamahemna: but I believe attaching encryption deviece with friendly multipath fails now (without my patch)18:38
hemnaDuncanT, well bugs have to get fixed either way, but I think it's less than ideal that nova coordinates creating cinder volumes.18:38
DuncanThemna: I see the best solution as being to spilt the orcestration bits (compute, vol and net) out of nova into its own service18:39
stevemarhemna: its got a lot of upside18:39
hemnastevemar, yah18:39
stevemarhemna: it'll also handle all the authentication bits in one common spot, instead of the clients struggling with that18:39
*** timcl has quit IRC18:39
DuncanThemna: That can be agile, like cinder was to start with, and talk to clean, simple apis for the individual services18:39
hemnaDuncanT, yah that's I think the best long term solution18:40
*** eharney_ is now known as eharney18:40
hemnadeprecate for now, push operators to use the osc, and at some time in the future deal with removing it from novaclient.18:40
hemnaeasier said than done :)18:40
stevemarhemna: fwiw, you can see the entire command set here:
*** jordanP has quit IRC18:41
hemnastevemar, nice18:41
DuncanTYou can keep a pass-through proxy to the new service in nova for as long as they feel the need - I don't know how much of the volume one is still there, but it was certainly there for many releases18:41
hemnafor a time I had cinder client puking in my vagrant VMs but osc worked18:41
stevemarhemna: for keystone it worked out nicely, since we were transitioning to v3 and creating a completely new set of commands18:42
*** Apoorva has quit IRC18:42
*** Apoorva_ has joined #openstack-cinder18:42
stevemarit provided a nice clean slate to work from18:42
hemnayah, very nice.  I'm a fan of osc18:42
morganhemna: yeah especially if "new hotness" stuff only appears in OSC it is an easy move18:43
stevemarhemna: i'm not sure what the plans for cinder are in the near future? new API version or just iterating on v2 for a while longer?18:43
morganhemna: fix the docs/devstack and deprecate cli in clinderclient -2 any "new features/commands" and it forces a move18:43
morganstevemar: i don't think it matters.18:43
hemnastevemar, so we are going to stick with v2, and then implement microversions.18:43
morganstevemar: if new support only goes into OSC it's good.18:43
hemnaat least that's the current in flight plan18:44
morganand if old client only does base microversion -- even easier ;)18:44
morganeffectively the same as the ksc move18:44
*** Adriano_ has quit IRC18:45
*** adrianofr has quit IRC18:45
tsekiyamahemna: If encryption was broken before, there is no possibility that /dev/mapper/mpathX is an encryption device. So _is_crypt_device_available('mpathX') will always return False, we don't have to check this.18:45
morganhemna: so I'd use the microversions as the push for OSC CLI vs cinderclient CLI - but that is just from experience in keystone working18:45
hemnatsekiyama, ok18:46
*** Apoorva_ has quit IRC18:47
hemnamorgan, well, for now we are going to have to do microversions in cinderclient, as it's the only way, that I know of, that nova talks to cinder18:47
*** Apoorva has joined #openstack-cinder18:47
hemnacinderclient as the python class vs. cmdln18:47
morganhemna: the library yes. just from the CLI don't support anything but the base level18:47
morganhemna: keystoneclient as a library fully supports keystone v318:48
morganwe just never did a CLI for v318:48
*** delattec has quit IRC18:48
*** cdelatte has quit IRC18:48
morganhave OSC support the "new features" that are part of the microversions for the CLI perspective18:48
hemnamorgan, so the one caveat to this is that we are working on removing volume manager locks in cinder to allow HA18:49
hemnathat requires a new microversion bump18:49
hemnaand as a result the cinder api will return a 409 error code when the volume is busy doing something already18:50
morganhemna: so out of curiosity, what happens if an older client makes a request against the lower version and that state is happening?18:50
hemnaIf/when we get this in it would be really nice if the clients used that.18:50
hemnamorgan, we are ironing that out18:50
hemnabut we are required to support the existing functionality18:50
hemnawhich is going to leave the locks in place for the older version.18:51
hemnaand that means folks won't get active/active c-vol18:51
hemnaso be it18:51
*** esker has joined #openstack-cinder18:51
morganhemna: I think that is a fair reason to push to OSC if you can do it in the same timeframe18:51
morganhemna: but this is just a personal view18:51
morgani'm not a cinder dev/deployer at the moment18:51
hemnaall of it hinges on this stuff getting in.18:52
hemnawhich is non-trivial18:52
*** esker has quit IRC18:52
hemnascottda, is doing most of the heavy lifting so far.18:52
morganactive/active c-vol would be a nice push to "stop using the old CLI" :)18:52
morganwell hope you can land it all :)18:52
hemname too, but it has to work :)18:52
*** esker has joined #openstack-cinder18:52
* morgan looks for a rock to lurk under again :P18:53
*** hemna is now known as hemnalunch18:53
*** Zhongjun has quit IRC18:57
*** lcurtis has joined #openstack-cinder18:58
*** aix has joined #openstack-cinder19:01
openstackgerritDuncan Thomas proposed openstack/cinder-specs: Cinder AZ rework spec
DuncanTjgriffith: hemna: for the start of an AZs spec. I'll try to work on the proposed solution section tomorrow19:04
*** thangp has joined #openstack-cinder19:04
*** krtaylor has quit IRC19:15
*** timcl has joined #openstack-cinder19:18
*** thangp has quit IRC19:25
*** krtaylor has joined #openstack-cinder19:28
*** ronis_ has joined #openstack-cinder19:28
*** martyturner has quit IRC19:29
*** lpetrut has quit IRC19:29
*** martyturner has joined #openstack-cinder19:30
*** hodos has joined #openstack-cinder19:31
*** bnemec has quit IRC19:34
*** rhagarty_ has quit IRC19:37
*** rhagarty has quit IRC19:37
jgriffithmgagne: I find the whole thing interesting and maybe kinda scary19:48
jgriffithmgagne: even though I don't like the feature, I don't know how one would propose deprecatoin/removal of something that's known to be in use without having an equivalent replacement19:49
jgriffithmgagne: and to your point, nova does networking and other 'things' as well19:49
jgriffithmgagne: I owe an apology for voicing my opinions about that whole thing19:50
jgriffithto you19:50
mgagnejgriffith: I fortunately happen to have (somewhat small) time to voice my opinion19:50
mgagnejgriffith: what I find unfortunate is the trend to deprecate stuff and not proposing a well tested and feature complete alternative. We stop developing feature and propose something else that is not even half finished or better and still WIP for months/years.19:52
mgagnejgriffith: that's not fair to the end users and us operators supporting those people19:52
*** alejandrito has quit IRC19:53
*** timcl has quit IRC19:54
kevincarr1991Can anyone point me to some good documentation on how to add cinder to the controller and use nfs mount19:58
openstackgerritJohn Griffith proposed openstack/cinder: Remove NetApp drivers from Cinder
jgriffithmgagne: I completely agree... ala keystone client and openstackclient etc20:01
jgriffithmgagne: there's a disconnect between developers and users that's actually the source of quite a bit of trouble20:02
*** martyturner has quit IRC20:02
mgagnejgriffith: that's what I implicitly referred to. see my reply on #openstack-nova20:02
jgriffithahh... yes, I see20:03
smcginnisjgriffith: ping20:03
jgriffithsmcginnis: hi20:03
smcginnisAbout the netapp thing...20:04
jgriffithsmcginnis: why are you the only one every around and watching what's going on ?20:04
*** martyturner has joined #openstack-cinder20:04
smcginnisThe shadow knows...20:04
smcginnisThat's just a dependency on a proprietary library, right?20:04
smcginnisNot wrong license for Cinder code.20:04
jgriffithsmcginnis: correct20:05
jgriffithsmcginnis: but it seems there's some issues for folks that do distros20:05
*** jwcroppe has joined #openstack-cinder20:05
*** timcl has joined #openstack-cinder20:05
smcginnisI was under the assumption that we had other drivers like that.20:05
jgriffithsmcginnis: where having that import in the code may or may not represent some sort of issue for them20:05
smcginnisjgriffith: That the driver could be distributed, but if someone wanted to actually use it they would need to download the proprietary bits to make it run.20:06
jgriffithmore importantly though IMHO is the principle of the matter in terms of what OpenStack has traditionally held up as values20:06
jgriffithsmcginnis: Yes, but lawyers and licenses are funny20:06
smcginnisjgriffith: Don't get me wrong, I don't really like it - but I could see why someone may need to do it that way. :/20:06
jgriffithsmcginnis: things like "you referenced it in your code.. now I own your ass" etc20:06
*** hemnalunch is now known as hemna20:06
eharneyit actually is a big problem for us, since we have customers using that driver, and can't distribute the new lib... but you can ignore all of that and figure it's just wrong from the upstream community POV anyway20:06
jgriffitheharney: +120:07
smcginnisjgriffith, eharney: OK, fair enough.20:07
xyang1jgriffith: smcginnis i think one problem is the code was in driver in kilo with apache2 license20:07
xyang1So now it was in a lib with proprietory license20:07
smcginnisxyang1: My impression was there was code in kilo, but not _that_ code.20:07
xyang1It is that code20:07
jgriffithxyang1: yes!20:08
xyang1The code was moved to a lib20:08
*** tpsilva has quit IRC20:08
smcginnisxyang1: Doesn't one of the EMC drivers have a similar dependency on an outside lib or command tool that needs to be installed separately?20:08
jgriffithxyang1: so they took the code contributed up part way into L and just copy/pasted it into a lib and slapped a proprietary license on it20:08
jgriffithpoor form20:08
xyang1So seems moving from open source to non open source20:08
jgriffithas it includes contributions from all of us20:08
cfriesen__downside of apache license?20:09
smcginnisjgriffith, xyang1: If that is the case, then I'm all for the proposed direction.20:09
smcginnisBut do we know that for sure?20:09
xyang1smcginnis: emc does not move open source code to emc license20:09
jgriffithand as I said before, regardless of legal issues, distribution problems etc, IMO it's just poor form20:09
eharneysmcginnis: see the notes in the LP bug20:09
xyang1It is the otherway around20:09
xyang1So that seems odd to me20:09
smcginniseharney: I read that. I thought it said it was just switching to a different library. I didn't think it was the same code repackaged.20:09
jgriffithsmcginnis: so the issue is not moving code to libs....20:09
* smcginnis rereads...20:10
jgriffithsmcginnis: the issue is moving it to a lib with a proprietary license20:10
jgriffithsmcginnis: so folks like EMC, Pure and HP that do that and use the Apache license are completely different deal20:10
*** breitz has quit IRC20:10
smcginnistbarron, ameade: Any input? Or you need to wait for the lawyers? :)20:11
*** breitz has joined #openstack-cinder20:11
xyang1smcginnis: we still have lib that is not open source, but we did not change the same code from apache2 to EMC license20:11
smcginnisThe point I would like clarified is - is the code in the proprietary lib the code that was formerly Apache? Or is it a new set of code.20:11
jgriffithsmcginnis: it appears it's the same code that was in place before20:12
smcginnisxyang1: Hence my point of needing clarification.20:12
openstackLaunchpad bug 1499334 in Cinder "Regression: Proprietary code needed to use NetApp with Cinder and Manila" [Critical,In progress] - Assigned to John Griffith (john-griffith)20:12
xyang1It is the same code to my understanding20:12
tbarronsmcginnis: jgriffith: all: we know there is a problem here.  It's not a technical issue, but something we are pushing up the management/lawyer chain, etc.20:12
xyang1Netapp folks can clarify of course20:12
jgriffithtbarron: push faster! :)20:12
tbarronsmcginnis: jgriffith: yes, it's the same code20:12
* jgriffith is working on the backport of that patch 20:13
smcginnistbarron: Thanks for the clarification.20:13
jgriffithto rc20:13
*** Lee1092 has quit IRC20:13
jgriffitherr... stable/liberty as it were20:13
hemnagah, it's the same code?20:15
*** leseb has joined #openstack-cinder20:15
jgriffithhemna: well.. you approved it :)20:16
cfriesen__   <---  netap lib20:16
cfriesen__someone more familiar with that code should be able to see if it's the same20:16
xyang1So to jgriffith's point, the code could contain contribution from non-NetApp people also20:16
*** Yogi1 has quit IRC20:16
hemnajgriffith, *sigh* yah I know, I guess I should have looked into installing the lib and looking at the source.....20:16
jgriffithxyang1: and does, as pointed out by the link eharney provided20:16
xyang1Because it is been out for a while20:16
jgriffithhemna: TBH I most likely would've missed it too20:16
jgriffithhemna: this is good though, because in the future it's something we'll think about probably20:17
eharneyeven if it wasn't the same code, you can't just slip in a change to require non open-source licensed code on people who are already using this driver20:17
hemnajgriffith, yah, learn from the mistake.20:17
jgriffitheharney: yeah... there are a number of levels that make it wrong20:17
xyang1eharney: if you have a need to support new feature, maybe20:18
jgriffithanyway... hopefully we can get it fixed up20:18
xyang1eharney: to support replication, install this lib20:18
jgriffithxyang1: then it needs to be a new driver20:18
*** rhagarty_ has joined #openstack-cinder20:18
*** rhagarty has joined #openstack-cinder20:18
*** crose has quit IRC20:18
jgriffithxyang1: it's a case of upgrading existing deployments20:18
xyang1jgriffith: right20:18
cfriesen__interesting...the license on that netapp-lib package is non-assignable and non-transferable....might have implications.20:19
hemnaand the NetAppLibException looks quite similar to CinderException.....20:19
*** bswartz has quit IRC20:19
hemnabasically a copy/paste20:19
*** salv-orlando has joined #openstack-cinder20:19
*** Yogi1 has joined #openstack-cinder20:20
*** merooney has quit IRC20:21
*** salv-orlando has quit IRC20:22
openstackgerritAngela Smith proposed openstack/cinder-specs: Spec for FCZM Zone Driver Friendly Zone names
*** bnemec has joined #openstack-cinder20:24
openstackgerritAngela Smith proposed openstack/cinder-specs: Spec for FCZM Zone Driver Friendly Zone names
hemnaangela-s, :)20:26
*** wyatt22 has joined #openstack-cinder20:27
openstackgerritAngela Smith proposed openstack/cinder-specs: Spec for Brocade Zone Driver Virtual Fabrics support
*** eduardo has quit IRC20:34
*** nate1 has joined #openstack-cinder20:34
*** salv-orlando has joined #openstack-cinder20:35
*** salv-orlando has quit IRC20:35
*** salv-orlando has joined #openstack-cinder20:36
openstackgerritLisaLi proposed openstack/cinder: Create encrypted volumes from images
openstackgerritAngela Smith proposed openstack/cinder-specs: Spec for Brocade Zone Driver Virtual Fabrics support
*** martyturner has quit IRC20:37
angela-shemna: :D20:38
*** mudassirlatif has quit IRC20:40
*** timcl has quit IRC20:40
*** wyatt22 has quit IRC20:43
*** nate1 has quit IRC20:43
*** jgriffith1 has joined #openstack-cinder20:51
*** jgriffith has quit IRC20:51
*** jgriffith1 has quit IRC20:52
*** akerr has quit IRC20:52
*** jgriffith has joined #openstack-cinder20:55
*** raildo is now known as raildo-afk20:56
*** martyturner has joined #openstack-cinder20:58
*** ronis_ has quit IRC20:59
*** jgriffith has quit IRC21:00
*** mtanino has quit IRC21:01
*** hemna is now known as hemnafk21:01
openstackgerritxing-yang proposed openstack/cinder: Fix VMAX live migration problem
*** edmondsw has quit IRC21:02
jungleboyj_xyang1: Which summit was it that was in China?  Was that for Icehouse?21:03
xyang1jungleboyj_: yes21:03
jungleboyj_xyang1: Thank you.21:03
xyang1jungleboyj_: hong kong21:03
jungleboyj_Ah, thanks!21:03
jungleboyj_I was thinking Beijing.21:03
jungleboyj_And yes, I will remove my -2's still.  :-)21:03
xyang1jungleboyj_: that was olympics:)21:03
*** angela-s has quit IRC21:04
*** daneyon has quit IRC21:04
*** hemna has joined #openstack-cinder21:05
*** daneyon has joined #openstack-cinder21:05
*** timcl has joined #openstack-cinder21:08
*** mudassirlatif has joined #openstack-cinder21:08
*** mtanino has joined #openstack-cinder21:11
*** angela-s has joined #openstack-cinder21:12
*** mdenny has quit IRC21:15
angela-sthingee: for Brocade CI, we have addressed all the issues you mentioned.  Thank you for bringing those items to our attention.21:16
angela-sthingee: since CI was the blocker, could you please remove the -2s from these reviews so that I can work on them for M?21:17
*** martyturner has quit IRC21:19
*** jgriffith has joined #openstack-cinder21:21
*** lprice1 has quit IRC21:22
*** eharney has quit IRC21:26
*** jungleboyj_ has quit IRC21:27
thingeeesker: RE from #openstack-manila, I'm happy to hear the attribution issue will be resolved immediately for netapp_lib.21:29
*** dims__ has quit IRC21:29
thingeejgriffith, smcginnis ^21:29
*** cbader has quit IRC21:30
*** dustins has quit IRC21:31
thingeescottda, hemnafk definitely agree. wait for the spec to be done.21:31
scottdacool, wanted to keep you in the loop.21:31
smcginnisthingee, esker: Good to hear.21:32
thingeehemnafk, stevemar I'm definitely for us moving towards the openstack-client... I think it's good to help along with keystone influencing others to do the same.21:32
*** Guest99 has joined #openstack-cinder21:33
thingeeI know some of us here have some reservations with how things work in it, but maybe we could bring those thoughts to the project itself.21:33
jgriffiththingee: I don't know that anybody opposed that at all actually21:33
jgriffiththingee: do you know what reservations people have?21:34
stevemarjgriffith: i thought you had some of the reservations :)21:34
thingeejgriffith: well you mentioned cliff. I know that's a dependency deep, but..21:34
stevemarwhat with it being based on cliff and all21:34
thingeestevemar: :)21:34
jgriffithI have reservations about removing existing functionality that deployers are using  yes21:34
jgriffithstevemar: please don't adlib what I stated earlier21:35
stevemarjgriffith: no one is saying to remove it, just a long ass deprecation cycle21:35
jgriffithstevemar: that statement was specifically in resonse to someobdy stating they hadn't used it much, and my response was that I have been poking at it21:35
jgriffithstevemar: and that I think it's cool and the right direction21:35
jgriffithstevemar: but that I was having trouble with figuring out some of the inricancies of cliff21:35
jgriffithstevemar: the only other statement that I made, that I will make again very CLEARLY21:36
stevemarjgriffith: ah, didn't mean to adlib - but proceed, its all good21:36
jgriffithstevemar: is that I do NOT believe you can/should deprecate an existing API or interface UNTIL a replacement is avaialable and solid and NOT in beta etc21:36
jgriffithstevemar: I apologize if my words gave a different impression21:37
stevemarjgriffith: its all good, happens over text, plus i wasn't paying the most attention in the morning21:37
*** changbl has quit IRC21:37
* morgan never pays attention21:38
stevemarjgriffith: re: your concern, so it's fair and sensible. someone (maybe the cinder team?) needs to evaluate osc and see if it's a viable replacement? FWIW osc is *not* in beta21:38
stevemarthat sounds like a lot of work though21:38
*** changbl has joined #openstack-cinder21:39
stevemarbut i agree, marking something as deprecated when the solution is half baked is silly pant21:39
stevemarwhen the alternative solution*...21:39
jgriffithstevemar: yes, sorry beta being a poor word there21:39
jgriffithstevemar: there is some fundamental misunderstanding here though I think21:39
*** diablo_rojo has quit IRC21:40
jgriffithstevemar: the fact is that osc doesn't replace any of what was being discussed when the topic came up21:40
*** eikke has quit IRC21:40
jgriffithstevemar: the nova-boot-create-volume is completely different issue21:40
jgriffithstevemar: and I don't think people seem to understand that "for now" at least osc leverages all of the individual clients21:40
*** eikke has joined #openstack-cinder21:41
jgriffithstevemar: the focus on "new" right now is an awesome shell21:41
jgriffithI don't know too much more than that, so I won't say anything that will be used against me in a court of law later :)21:41
thingeeyeah it won't be easy. I think we agree on that. But we're evolving from that choice back when it was just a few projects and it was ok for people to have their own clients.21:41
*** alejandrito has joined #openstack-cinder21:41
jgriffiththingee: it has nothign to do with easy21:41
jgriffiththingee: it's reality.. right now it's a shell no?21:42
openstackgerritAngela Smith proposed openstack/cinder: Adds friendly zone name support
thingeejgriffith: sure21:42
stevemarjgriffith: right, the long term goal is to use openstacksdk, but right now we call the python bindings of the python-*client libraries21:42
*** dtroyer has joined #openstack-cinder21:42
jgriffiththingee: I'm not sure why it's a topic here actually.  So I'm a HUGE fan of OSC and what's going on there21:42
thingeejgriffith: I disagree that it being difficult does play into it though in someway.21:42
jgriffiththingee: I'm not sure why there's this impression that I'm not21:42
jgriffithdtroyer: :)21:43
stevemaras the sdk matures, we will drop some of the more troublesome client libs (swift and glance for starters) :)21:43
dtroyerstevemar: rang the bell21:43
jgriffithdtroyer: just trying to clear up a misinterpretation of something I typed earlier21:43
jgriffithdtroyer: I LOVE osc!!!21:43
stevemarjgriffith: lol21:43
jgriffiththink it's a great project and the right direction21:43
jgriffithfor some reason somebody is spreading rumors that I said something bad abou tit21:44
thingeejgriffith: I was talking about your specific mention of cliff, which is a dependency of osc. I know your opinions on osc are great.21:44
*** vgridnev has quit IRC21:44
jgriffithI think cliff is fantastic as well... I simply said that it was new and I was stuck trying to figure out a couple of details around it21:44
jgriffiththingee: I've been trying to write my own code that uses it21:44
thingeeok, no problem!21:45
jgriffiththingee: and for the most part it's been super awesome and fast learning curve21:45
dtroyerI saw that in the logs…cliff is interesting for sure…and was written for OSC so isn't always intuitive to use elsewhere21:45
jgriffithbut I got hung up on some things with reading in subcommands that I haven't figure out, not because there's anything wrong with it, but because it's different than anything I've used, and things like osc and nuetron use it differently than I am trying to21:46
jgriffithdtroyer: thanks21:46
dtroyerDON'T look at neutronclient as an example21:46
dtroyerfor anything21:46
jgriffithI noticed that it is infact VERY different21:46
dtroyerway over-normalized21:46
*** kevincar_ has joined #openstack-cinder21:46
dtroyerOSC may be the other extreme, we copy-n-paste a good bit, but in the name of not fouling up older versions and other commands21:47
dtroyerso cliff gets most of our factored-out bits at the command-parsing level21:47
*** jgregor has quit IRC21:47
jgriffithdtroyer: yeah, I was messing with trying to have base command files if that makes sense21:48
jgriffithdtroyer: so like: command/, etc etc21:48
jgriffithdtroyer: the idea being "myshell volumes -h"21:48
dtroyerso not version-specific?21:48
jgriffithdtroyer: hadn't gotten to that part yet :)21:49
mriedemcinder peeps - would appreciate review of this for anyone that was involved in that ML thread
dtroyeroh, no, I see21:49
jgriffithdtroyer: but that was going to be next21:49
jgriffithdtroyer: so I don't want a class per method, I want a class per resource21:49
jgriffithand methods/sub-commands under that21:49
jgriffithit might be stupid, don't know21:49
dtroyerI think we have cliff doing partial command helps where it lists the matching remainig bits now21:49
*** daneyon has quit IRC21:49
jgriffithdtroyer: yes, I played with osc and it's almost exactly what I want actually21:49
*** daneyon has joined #openstack-cinder21:50
jgriffithdtroyer: couple of things that I just haven't figured out yet but honestly haven't looked at it in a week or two21:50
jgriffithdtroyer: it was more in some of the "magic" stuff that cliff reads in and does21:50
*** kevincarr1991 has quit IRC21:50
dtroyeryeah, you need to know the flow through to get it well21:51
jgriffithdtroyer: yeah.. I do think it's super cool21:51
dtroyerfor the setup, the commands themselves are pretty simple21:51
jgriffithdtroyer: coming from using things like argparse, click and codegangsta-cli I really like it21:51
*** kevincar_ has quit IRC21:51
jgriffithJust need more time and probably examples to figure it out :)21:52
jgriffithcompletely that is21:52
dtroyerFWIW, I've been toying with the idea of making stand-alone subcommands (if that makes sense) to help with the object name overload we're getting in to.  that might be helpful for what you are thinking about too21:52
*** dims_ has joined #openstack-cinder21:52
jgriffithdtroyer: YES!21:53
dtroyerit's mostly for plugins, but I don't want to lose the all-in-one idea either21:53
*** yrabl has quit IRC21:53
jgriffithdtroyer: if we're on the same page (and I think we are) that makes sense21:53
*** mriedem has quit IRC21:53
dtroyerit would also let you build any random cli with just the clientmanager (wihtout keystone)/cliff/shell bits easily21:54
dtroyerI used github as the original plugin sample21:54
*** timcl has quit IRC21:55
*** e0ne has joined #openstack-cinder21:56
*** changbl has quit IRC22:00
thingeeangela-s: ack, will get around to it today to take a closer look22:03
*** edtubill has quit IRC22:03
angela-sthingee: thanks!22:03
*** annasort has quit IRC22:04
*** changbl has joined #openstack-cinder22:04
*** jwcroppe has quit IRC22:07
*** jwcroppe has joined #openstack-cinder22:07
*** zhangjn has joined #openstack-cinder22:09
*** hemna has quit IRC22:10
*** zhangjn has quit IRC22:11
*** jwcroppe has quit IRC22:12
*** zhangjn has joined #openstack-cinder22:12
*** zhangjn has quit IRC22:12
*** zhangjn has joined #openstack-cinder22:13
*** Guest99 has quit IRC22:14
*** garthb has quit IRC22:15
*** e0ne has quit IRC22:17
*** zhangjn has quit IRC22:17
*** Guest88 has joined #openstack-cinder22:19
*** hemnafk is now known as hemna22:22
*** amoturi has left #openstack-cinder22:25
*** Guest88 has left #openstack-cinder22:25
*** Guest88 has joined #openstack-cinder22:26
*** stevemar has quit IRC22:30
scottdaIf anyone is interested in helping with a Tokyo Summit talk on contributing to OpenStack, I've 3 questions for you....22:36
*** zhangjn has joined #openstack-cinder22:36
*** zhangjn has quit IRC22:37
*** zhangjn has joined #openstack-cinder22:38
*** mc_nair has quit IRC22:40
*** sghanekar has joined #openstack-cinder22:40
*** lcurtis has quit IRC22:44
*** jgriffith has quit IRC22:55
*** jgriffith has joined #openstack-cinder22:55
*** bswartz has joined #openstack-cinder22:55
*** Guest88 has quit IRC22:57
*** jgriffith has quit IRC22:57
*** zhangjn_ has joined #openstack-cinder22:57
*** jgriffith has joined #openstack-cinder23:00
*** IanGovett has quit IRC23:00
*** zhangjn has quit IRC23:01
*** alejandrito has quit IRC23:03
*** markvoelker_ has quit IRC23:04
*** dims_ has quit IRC23:05
*** gouthamr has quit IRC23:11
*** Yogi1 has quit IRC23:13
*** diogogmt has quit IRC23:16
*** stevemar has joined #openstack-cinder23:31
*** stevemar has quit IRC23:33
*** erhudy has quit IRC23:39
openstackgerritJohn Griffith proposed openstack/cinder: Remove NetApp drivers from Cinder
*** jamielennox|away is now known as jamielennox23:56
*** rex_lee_ has quit IRC23:56
*** jaypipes has quit IRC23:56
*** harlowja has quit IRC23:56
*** sayali has quit IRC23:56
*** thrawn01 has quit IRC23:56
*** r-daneel has quit IRC23:57
*** sage has quit IRC23:57
*** harlowja has joined #openstack-cinder23:57
*** xyang1 has quit IRC23:57
*** thrawn01 has joined #openstack-cinder23:58
*** sage has joined #openstack-cinder23:58
*** thingee has quit IRC23:59
*** r-daneel has joined #openstack-cinder23:59

Generated by 2.14.0 by Marius Gedminas - find it at!