Friday, 2015-06-05

*** akerr has quit IRC00:01
*** akerr_ is now known as akerr_away00:05
*** akerr_away is now known as akerr_00:09
*** annegentle has joined #openstack-cinder00:09
openstackgerritSean Chen proposed openstack/cinder: Tintri driver to manage existing backend storage objects
*** alexpilotti has quit IRC00:10
*** dims__ has joined #openstack-cinder00:11
*** dims_ has quit IRC00:12
*** vilobhmm has joined #openstack-cinder00:12
*** akerr has joined #openstack-cinder00:12
*** markvoelker has joined #openstack-cinder00:13
*** annegentle has quit IRC00:14
*** akerr_ has quit IRC00:15
*** dims__ has quit IRC00:17
*** dims_ has joined #openstack-cinder00:17
*** markvoelker has quit IRC00:18
*** merooney has joined #openstack-cinder00:22
*** annegentle has joined #openstack-cinder00:22
*** akerr_ has joined #openstack-cinder00:24
*** merooney has quit IRC00:24
*** merooney has joined #openstack-cinder00:24
*** akerr has quit IRC00:27
*** vilobhmm has quit IRC00:27
*** Sasi has joined #openstack-cinder00:39
*** tsekiyam_ has joined #openstack-cinder00:42
*** tsekiyam_ has quit IRC00:42
*** mtanino has quit IRC00:43
*** tsekiyama has quit IRC00:45
*** raunak has quit IRC00:46
*** _cjones_ has quit IRC00:49
*** merooney has quit IRC00:52
*** rmesta has quit IRC00:52
*** merooney has joined #openstack-cinder00:52
*** chlong has quit IRC00:52
*** salv-orlando has joined #openstack-cinder00:54
*** chlong has joined #openstack-cinder00:54
*** merooney has quit IRC00:56
*** lpabon has joined #openstack-cinder01:03
*** annegentle has quit IRC01:06
*** merooney has joined #openstack-cinder01:06
*** salv-orlando has quit IRC01:10
*** merooney has quit IRC01:10
*** Sasi has quit IRC01:13
*** merooney has joined #openstack-cinder01:18
*** merooney has quit IRC01:19
*** tobe has joined #openstack-cinder01:38
*** Lee1092 has joined #openstack-cinder01:44
*** yamada-h has joined #openstack-cinder01:44
*** vilobhmm has joined #openstack-cinder01:46
*** yamada-h has quit IRC01:49
*** patrickeast has quit IRC01:50
*** lixiaoy11 has joined #openstack-cinder01:58
*** vilobhmm has quit IRC02:01
*** markvoelker has joined #openstack-cinder02:02
tbarronhemnafk: w.r.t., do you think a bug or blueprint is most appropriate?  I was torn between the two.02:02
*** IlyaG has joined #openstack-cinder02:05
*** IlyaG has quit IRC02:06
openstackgerritAlex Meade proposed openstack/cinder: NetApp E-Series driver: Remove caching logic
openstackgerritAlex Meade proposed openstack/cinder: NetApp E-Series: Refactor class structure for FC
openstackgerritAlex Meade proposed openstack/cinder: NetApp E-Series: Add Fibre Channel Support
*** lixiaoy1 has joined #openstack-cinder02:06
*** markvoelker has quit IRC02:07
*** lixiaoy11 has quit IRC02:08
*** salv-orlando has joined #openstack-cinder02:09
*** Tross has joined #openstack-cinder02:14
*** lixiaoy11 has joined #openstack-cinder02:15
*** Tross1 has quit IRC02:15
*** Tross1 has joined #openstack-cinder02:16
*** lixiaoy1 has quit IRC02:17
*** salv-orlando has quit IRC02:17
*** Tross has quit IRC02:19
*** akerr_ has quit IRC02:22
*** akerr has joined #openstack-cinder02:25
*** daneyon_ has quit IRC02:29
smcginnisthingee: ping02:43
smcginnisthingee: Review for this looks good but bp isn't approved yet:
smcginnisthingee: Don't want to push it through until it's green lighted.02:44
smcginnisthingee: Review -
*** yamada-h has joined #openstack-cinder02:45
*** yamada-h has quit IRC02:49
*** akerr has quit IRC02:52
*** Apoorva has quit IRC02:56
*** boris-42 has quit IRC02:58
openstackgerritThang Pham proposed openstack/cinder: WIP: Conversion to volume object
*** dims_ has quit IRC03:07
*** zul has quit IRC03:08
*** xyang1 has quit IRC03:15
*** lpetrut has joined #openstack-cinder03:18
*** zul has joined #openstack-cinder03:21
*** gouthamr has quit IRC03:26
*** gouthamr has joined #openstack-cinder03:27
*** gouthamr has quit IRC03:31
openstackgerritchenzongliang proposed openstack/cinder: Add Chap support for Huawei Storage
*** markvoelker has joined #openstack-cinder03:50
*** Lee1092 has quit IRC03:51
*** salv-orlando has joined #openstack-cinder03:51
*** coolsvap|afk is now known as coolsvap03:53
*** markvoelker has quit IRC03:55
*** marcusvrn has joined #openstack-cinder03:57
openstackgerritRick Chen proposed openstack/cinder: Prophetstor driver needs to return snapshot objects
openstackgerritRick Chen proposed openstack/cinder: Complete switch to snapshot objects
openstackgerritVincent Hou proposed openstack/cinder-specs: Volume migration improvement for L
*** salv-orlando has quit IRC04:01
*** yamada-h has joined #openstack-cinder04:05
*** dims_ has joined #openstack-cinder04:05
*** dims_ has quit IRC04:10
*** lpetrut has quit IRC04:21
*** pradipta has joined #openstack-cinder04:22
*** rushiagr_away is now known as rushiagr04:22
*** marcusvrn has quit IRC04:24
*** yamada-h has quit IRC04:24
*** shyama has joined #openstack-cinder04:26
*** rushiagr is now known as rushiagr_away04:27
*** Longgeek has joined #openstack-cinder04:30
*** Longgeek has quit IRC04:30
*** swamireddy1 has joined #openstack-cinder04:31
*** jamielennox is now known as jamielennox|away04:32
*** yamada-h has joined #openstack-cinder04:32
*** Longgeek has joined #openstack-cinder04:32
*** yamada-h has quit IRC04:36
*** rushiagr_away is now known as rushiagr04:40
*** kaisers1 has quit IRC04:44
*** kaisers1 has joined #openstack-cinder04:45
*** heyun has joined #openstack-cinder04:46
*** Longgeek_ has joined #openstack-cinder04:50
openstackgerritVincent Hou proposed openstack/cinder-specs: Volume migration improvement for L
*** Longgeek has quit IRC04:54
*** yamada-h has joined #openstack-cinder04:55
*** yamada-h has quit IRC05:00
*** BharatK has joined #openstack-cinder05:05
*** spradeepv has quit IRC05:06
*** lpetrut has joined #openstack-cinder05:08
openstackgerritVincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers
*** markvoelker has joined #openstack-cinder05:39
*** mitz has quit IRC05:41
*** tobe has quit IRC05:41
*** tobe has joined #openstack-cinder05:42
*** tobe has quit IRC05:44
*** markvoelker has quit IRC05:44
*** mdbooth has quit IRC05:48
*** nihilifer has joined #openstack-cinder05:49
*** nihilifer has quit IRC05:51
*** nihilifer has joined #openstack-cinder05:51
*** mdbooth has joined #openstack-cinder05:53
*** belmoreira has joined #openstack-cinder05:55
*** yamada-h has joined #openstack-cinder05:59
*** yamada-h has quit IRC06:04
*** vilobhmm has joined #openstack-cinder06:12
*** BharatK has quit IRC06:16
*** lpetrut has quit IRC06:19
*** tobe has joined #openstack-cinder06:19
*** agarciam has joined #openstack-cinder06:19
*** agarciam has quit IRC06:22
*** agarciam has joined #openstack-cinder06:23
*** agarciam has quit IRC06:24
*** agarciam has joined #openstack-cinder06:26
*** agarciam has quit IRC06:27
*** agarciam has joined #openstack-cinder06:28
*** agarciam has quit IRC06:28
*** BharatK has joined #openstack-cinder06:29
*** agarciam has joined #openstack-cinder06:30
*** yamada-h has joined #openstack-cinder06:33
*** yamada-h has quit IRC06:33
*** agarciam has quit IRC06:35
*** agarciam has joined #openstack-cinder06:36
*** vilobhmm has quit IRC06:46
*** Lee1092 has joined #openstack-cinder06:49
*** ociuhandu has joined #openstack-cinder06:50
*** yamada-h has joined #openstack-cinder06:50
*** yamada-h has quit IRC06:50
*** dims_ has joined #openstack-cinder06:54
*** e0ne has joined #openstack-cinder06:54
*** takedakn has joined #openstack-cinder06:56
*** yamada-h has joined #openstack-cinder06:56
*** yamada-h has quit IRC06:57
*** nikeshm has quit IRC06:57
*** jistr has joined #openstack-cinder06:57
*** dims_ has quit IRC06:59
*** takedakn has quit IRC07:00
*** takedakn has joined #openstack-cinder07:01
*** e0ne has quit IRC07:04
*** yamada-h has joined #openstack-cinder07:04
*** yamada-h has quit IRC07:04
*** takedakn has quit IRC07:04
*** chhavi has joined #openstack-cinder07:08
*** markus_z has joined #openstack-cinder07:15
chhaviwinston-d: Need to discuss on are you there07:15
*** takedakn has joined #openstack-cinder07:16
*** takedakn has quit IRC07:17
*** takedakn has joined #openstack-cinder07:18
*** takedakn has quit IRC07:18
*** takedakn has joined #openstack-cinder07:19
*** geguileo has joined #openstack-cinder07:20
*** anshul has joined #openstack-cinder07:21
*** takedakn has quit IRC07:22
*** ronis has joined #openstack-cinder07:24
*** Miouge_ has joined #openstack-cinder07:26
*** ociuhandu has quit IRC07:28
*** markvoelker has joined #openstack-cinder07:28
*** Miouge has quit IRC07:29
*** Miouge_ is now known as Miouge07:29
*** lpetrut has joined #openstack-cinder07:29
*** Miouge has quit IRC07:31
*** markvoelker has quit IRC07:33
*** Miouge has joined #openstack-cinder07:35
openstackgerritMarc Koderer proposed openstack/cinder: Port remote_fs driver to use new driver model
*** jistr is now known as jistr|biab07:47
*** rwsu has joined #openstack-cinder07:48
*** coolsvap is now known as coolsvap|afk07:52
*** yamada-h has joined #openstack-cinder07:56
*** yamada-h has quit IRC08:01
*** Miouge has quit IRC08:05
*** Miouge has joined #openstack-cinder08:08
*** Miouge has quit IRC08:11
*** Miouge has joined #openstack-cinder08:22
*** ociuhandu has joined #openstack-cinder08:32
*** chlong has quit IRC08:34
*** rwsu has quit IRC08:36
*** ndipanov has joined #openstack-cinder08:37
*** shyama has quit IRC08:43
*** shyama has joined #openstack-cinder08:44
*** ankit_ag has joined #openstack-cinder08:45
*** ociuhandu has quit IRC08:47
*** jistr|biab is now known as jistr08:50
*** lennyb has quit IRC08:52
*** coolsvap|afk is now known as coolsvap08:54
*** jistr has quit IRC08:59
*** e0ne has joined #openstack-cinder09:09
openstackgerritYuriy Nesenenko proposed openstack/cinder: Handle incorrect '--config-dir' param
*** markvoelker has joined #openstack-cinder09:17
*** jistr has joined #openstack-cinder09:19
*** markvoelker has quit IRC09:22
*** afazekas_mtg has joined #openstack-cinder09:33
*** aix has joined #openstack-cinder09:34
*** chhavi has quit IRC09:36
*** dims_ has joined #openstack-cinder09:43
*** alexpilotti has joined #openstack-cinder09:45
*** dims__ has joined #openstack-cinder09:47
*** dims_ has quit IRC09:48
*** lixiaoy11 has quit IRC09:53
*** yamada-h has joined #openstack-cinder09:58
openstackgerritwanghao proposed openstack/cinder: Support for force-delete backups
*** yamada-h has quit IRC10:03
*** e0ne is now known as e0ne_10:07
liuxghi all ,Please take some time to review our driver10:11
*** e0ne_ is now known as e0ne10:12
*** mdenny has quit IRC10:17
*** markvoelker has joined #openstack-cinder10:17
*** boris-42 has joined #openstack-cinder10:20
*** markvoelker has quit IRC10:23
openstackgerritPeter Penchev proposed openstack/cinder: Re-add the StorPool distributed storage driver
openstackgerritPeter Penchev proposed openstack/cinder: StorPool: move the testcase to cinder/tests/unit/
openstackgerritPeter Penchev proposed openstack/cinder: StorPool: pass the H402 docstring check
openstackgerritPeter Penchev proposed openstack/cinder: StorPool: fall back to the configured template
openstackgerritPeter Penchev proposed openstack/cinder: StorPool: return volume object in _attach_volume()
openstackgerritPeter Penchev proposed openstack/cinder: StorPool: implement volume retyping and migrations
openstackgerritPeter Penchev proposed openstack/cinder: StorPool: use the ABC metaclasses for volume drivers
openstackgerritchenzongliang proposed openstack/cinder: Support SmartX for Huawei volume driver
liuxg<openstackgerrit> chenzongliang proposed openstack/cinder: Support SmartX for Huawei volume driver
liuxg<openstackgerrit> chenzongliang proposed openstack/cinder: Support SmartX for Huawei volume driver
*** rongze has joined #openstack-cinder10:27
*** BharatK has quit IRC10:27
liuxgPlease to review our driver10:27
liuxg<openstackgerrit> chenzongliang proposed openstack/cinder: Support SmartX for Huawei volume driver
*** BharatK has joined #openstack-cinder10:38
*** SergeyLukjanov has quit IRC10:39
*** pradipta has quit IRC10:43
*** rongze has quit IRC10:46
*** SergeyLukjanov has joined #openstack-cinder10:51
*** coolsvap is now known as coolsvap|afk10:59
*** Lee1092 has quit IRC11:01
*** yamada-h has joined #openstack-cinder11:01
*** yamada-h has quit IRC11:01
*** julim has joined #openstack-cinder11:01
*** rongze has joined #openstack-cinder11:02
*** e0ne is now known as e0ne_11:05
*** e0ne_ is now known as e0ne11:07
openstackgerritMarc Koderer proposed openstack/cinder: Port remote_fs driver to use new driver model
flip214would someone please tell me whether I'm on the right track with
*** markvoelker has joined #openstack-cinder11:18
*** markvoelker has quit IRC11:22
*** rongze has quit IRC11:27
openstackgerritPavel Boldin proposed openstack/cinder: Add iscsi_target_flags configuration option
*** yamada-h has joined #openstack-cinder11:32
*** marcusvrn has joined #openstack-cinder11:38
*** yamada-h has quit IRC11:38
*** tobe has quit IRC11:42
*** rwsu has joined #openstack-cinder11:43
*** heyun has quit IRC11:56
*** marcusvrn has quit IRC11:56
*** sgotliv has quit IRC12:03
*** zul has quit IRC12:16
*** zul has joined #openstack-cinder12:16
*** e0ne is now known as e0ne_12:21
*** e0ne_ is now known as e0ne12:22
*** andymccr has joined #openstack-cinder12:23
*** sc has quit IRC12:23
*** chhavi has joined #openstack-cinder12:25
*** sc has joined #openstack-cinder12:26
*** sasi has joined #openstack-cinder12:31
*** sasi has left #openstack-cinder12:31
*** annegentle has joined #openstack-cinder12:32
*** nihilifer has quit IRC12:33
*** markvoelker has joined #openstack-cinder12:34
*** ndipanov is now known as ndipanoff12:36
*** zhenguo has quit IRC12:37
*** markvoelker has quit IRC12:38
*** ociuhandu has joined #openstack-cinder12:39
*** akerr has joined #openstack-cinder12:43
*** marcusvrn has joined #openstack-cinder12:45
*** bswartz has quit IRC12:49
*** marcusvrn has quit IRC12:49
*** bswartz has joined #openstack-cinder12:50
*** Yogi1 has joined #openstack-cinder12:51
*** sasi has joined #openstack-cinder12:51
*** merooney has joined #openstack-cinder13:06
*** dims__ has quit IRC13:09
*** dims_ has joined #openstack-cinder13:10
*** xyang has joined #openstack-cinder13:16
*** dustins has joined #openstack-cinder13:17
*** boris-42 has quit IRC13:18
*** merooney has quit IRC13:20
*** bswartz has quit IRC13:21
*** cknight has joined #openstack-cinder13:24
*** primechuck has joined #openstack-cinder13:25
*** mriedem has joined #openstack-cinder13:27
*** rushiagr is now known as rushiagr_away13:28
*** avishay has joined #openstack-cinder13:29
*** merooney has joined #openstack-cinder13:30
*** BharatK has quit IRC13:30
*** gouthamr has joined #openstack-cinder13:31
*** avishay has quit IRC13:34
*** thangp has joined #openstack-cinder13:35
*** avishay has joined #openstack-cinder13:36
*** gouthamr has left #openstack-cinder13:37
*** gouthamr has joined #openstack-cinder13:37
*** Lee1092 has joined #openstack-cinder13:37
*** rushil has joined #openstack-cinder13:37
*** e0ne is now known as e0ne_13:39
*** merooney has quit IRC13:39
*** e0ne_ is now known as e0ne13:39
*** afazekas_mtg has quit IRC13:40
*** boris-42 has joined #openstack-cinder13:41
*** jaypipes is now known as leakypipes13:43
*** merooney has joined #openstack-cinder13:43
*** merooney has quit IRC13:45
*** asselin_ has joined #openstack-cinder13:45
*** chhavi has quit IRC13:46
*** asselin__ has joined #openstack-cinder13:47
*** rushil has quit IRC13:47
andymccrRan into the endpoint-type issue on cinderclient which was fixed a few weeks ago, is there any way to get a new release of cinderclient (it got fixed 5 days after 1.2.1) or is there a more "proper" way to go about this?13:49
*** asselin_ has quit IRC13:50
*** eharney has joined #openstack-cinder13:55
openstackgerritTom Barron proposed openstack/cinder: Handle ineffective backup compression
*** afazekas has joined #openstack-cinder13:58
jgriffithtbarron: ping14:01
*** swamireddy1 has quit IRC14:02
*** markvoelker has joined #openstack-cinder14:05
scottdaandymccr: I believe you should ask thingee about a new release of cinderclient. It is independent of cinder/OpenStack release cycle and can be done anytime (by the PTL)14:07
andymccrscottda: thanks! I've seen it done that way with some of the other clients but didn't want to make assumptions :)14:07
*** markvoelker has quit IRC14:09
tbarronjgriffith: pong (just got out of a meeting) wazzup?14:09
*** esker has joined #openstack-cinder14:09
jgriffithtbarron: hey, was just looking at that bug on backup compression14:10
jgriffithtbarron: curious if you have any ideas on what to do with it?14:10
tbarronjgriffith: yeah, I have a review up,
jgriffithOH... excellent14:11
openstackgerritVipin Balachandran proposed openstack/cinder: Fix free virtual space in capacity filter
*** rmesta has joined #openstack-cinder14:11
jgriffithtbarron: so I was  kinda of wondering about that14:11
jgriffithtbarron: so we compress it, compare, then take the lesser?14:11
tbarronjgriffith: as soon as it gets out of our CI, I have another one on the way.14:11
jgriffithtbarron: doesn't that make restore kinda "weird"14:12
tbarronjgriffith: yes, nothing smart in this iteration to detect incompressible data or something in the first place14:12
avishayanyone manage to get cinder working with noauth? i did it once i think but don't remember how i did it...14:12
tbarronjgriffith: the compression algo (including 'none') is already stored with each chunk14:12
jgriffithavishay: I wish!14:12
tbarronand restore looks at that already14:12
jgriffithavishay: please share if/when you get it to work14:12
jgriffithavishay: I *think* guitarzan may have got it to work at one point, but I could be completely wrong14:13
tbarronjgriffith: does that address your concern?14:13
avishayjgriffith: cinder is up but cinderclient isn't happy14:13
jgriffithtbarron: sure14:13
jgriffithtbarron: thanks14:13
avishayguitarzan: here?14:13
tbarronjgriffith: thanks for asking!14:13
jgriffithavishay: oh, interesting14:13
jgriffithavishay: hey... wanna have some fun?14:13
jgriffithavishay: trying downloading latest cinderclient with a Kilo distro14:13
avishayjgriffith: depends on how you define fun :)14:13
jgriffithavishay: :)14:14
avishayjgriffith: why?14:14
guitarzanavishay: oh noauth? hmm14:14
jgriffithosl.conf vs oslo_config deps14:14
guitarzanI'm not sure how well that works anymore14:14
avishayjgriffith: mmm14:14
*** esker has quit IRC14:14
jgriffithavishay: looking this morning, customer seeing it14:15
avishayjgriffith: nice14:15
*** esker has joined #openstack-cinder14:15
avishayguitarzan: i guess i can try curl and take it from there14:15
jgriffithavishay: so you got it all working with Cinder though... that's cool!14:15
jgriffithavishay: I suspect the discovery stuff that got added will break it really good on the client14:15
guitarzanavishay: I think cinder itself may still work with noauth14:16
guitarzanavishay: eventually I gave up and have a little fake auth service14:16
guitarzanglance for example, simply doesn't noauth :)14:16
*** e0ne is now known as e0ne_14:16
guitarzanthe keystoneclient stuff kind of blows up too14:16
jgriffithguitarzan: Oh, that's right, I remember you saying that14:16
jgriffithguitarzan: I wonder if it would be worth having that in Cinder tree as an option?14:17
avishayjgriffith: guitarzan: oh man14:17
guitarzanjgriffith: I'm a big fan of noauth for development14:17
avishayit would really suck if i had to use keystone for tests14:17
jgriffithavishay: +114:17
guitarzanbut as the web gets more webby, it becomes lots harder to do anything useful14:17
jgriffithguitarzan: yeah, and actually I get requests about it A LOT14:17
jgriffithwebby webs... nothing I hate more :)14:18
*** rushiagr_away is now known as rushiagr14:18
jgriffithavishay: we could probably log a bug for cinderclient to get noauth on track14:19
jgriffithavishay: don't know if it'd help, but maybe somebody else with cycles will be interested enough to make it work14:19
jgriffithavishay: personally I'd love to see it14:20
avishayjgriffith: if i can't get it to work i'll probably have to do it :)14:20
jgriffithavishay: I'd love to see Cinder be able to be a stand-alone service some day14:20
jgriffithavishay: :)14:20
kragnizandymccr: you're not the only one who would like a cinderclient release14:22
tbarronjgriffith: v3?14:22
jgriffithtbarron: ?14:23
andymccrkragniz: i imagine not! for the same reason? (endpoint-type var issues)14:24
*** merooney has joined #openstack-cinder14:24
*** rushil has joined #openstack-cinder14:24
kragniznope, the pbr version in the released requirements.txt is breaking the glance_store gate14:24
jgriffithkragniz: is that fix in cinderclient now?14:25
kragnizjgriffith: yup14:25
jgriffithkragniz: k14:26
kragnizmaster requirements is good14:26
*** dhellmann has joined #openstack-cinder14:26
*** e0ne_ has quit IRC14:26
tbarronjgriffith: replication v3 :)14:27
tbarronjgriffith: feel free just to ignore the annoying gadfly14:28
*** ronis has quit IRC14:28
*** mtanino has joined #openstack-cinder14:29
avishayguitarzan: do you have an example of using curl to communicate with cinder on hand?  haven't done it in a while14:30
openstackgerritJohn Griffith proposed openstack/python-cinderclient: Prep for 1.2.2
jgriffithtbarron: LOL14:31
*** salv-orlando has joined #openstack-cinder14:31
jgriffithtbarron: nah, I'm considering just leaving it broken at this point14:31
jgriffithtbarron: Fact that it's been nit picked for almost 6 months indicates nobody will every be happy14:31
tbarronjgriffith: it's a pretty thankless job14:31
jgriffithtbarron: particularly the circular crap: "Person A" says "Must have this"... then 3 months later, same person says "You can't do that"14:32
jgriffithgood times!14:32
tbarronjgriffith: I would like to see the decoupling of the capability from the create path in any case14:32
*** dansmith is now known as superdan14:32
guitarzanavishay: I'm not, I haven't done it in a long time either14:32
guitarzanavishay: er, I don't14:32
avishayguitarzan: ok thanks anyway14:32
jgriffithkragniz: andymccr
tbarronthat was a noble goald14:33
jgriffithtbarron: I'm still going to put my patch up14:33
kragnizjgriffith: awesome!14:33
jgriffithkragniz: once that lands I'll push, assuming I still have permissions14:34
tbarronjgriffith: eventually it may be along the line of what you and thingee are doing w.r.t capabilities14:34
andymccrthanks jgriffith!14:34
tbarron{ replication: True14:34
*** ankit_ag has quit IRC14:34
tbarronvendor_keys: ....}14:34
*** mdenny has joined #openstack-cinder14:34
tbarronand everyone can do it their own way14:35
jgriffithtbarron: EXACTLY!14:35
jgriffithwhich is what I wanted in the first place!14:35
tbarronjgriffith: yeah, I got that.14:35
tbarroncourse I want vendor keys to allow pool_level replication too, and admin-only triggered failover.14:36
*** Tross1 has quit IRC14:37
jgriffithtbarron: of course14:37
*** Tross has joined #openstack-cinder14:37
jgriffithtbarron: and you should be able to have that IMO14:37
tbarronjgriffith: kumaiah14:38
jgriffithPeace Love and Twinkies14:38
*** Tross has quit IRC14:38
tbarronjgriffith: sounds like Colorado snack food14:38
tbarronthough a little early in the day14:39
*** IanGovett has joined #openstack-cinder14:40
jgriffithDuncanT: you around?14:41
*** belmoreira has quit IRC14:42
*** e0ne has joined #openstack-cinder14:45
ameadewinston-d:  would you be so kind as to sponsor the netapp driver?
*** rajinir_r has quit IRC14:45
*** rajinir_r has joined #openstack-cinder14:46
avishayjgriffith: guitarzan: it's working!!!14:53
jgriffithavishay: sweeeeet!!!!14:53
avishayjgriffith: guitarzan: i found my old VM image, went into ipython and looked at the history :)14:54
avishayjgriffith: guitarzan: don't know how to get it to work from the command line, but works from python import14:54
e0nejgriffith: hi. i would like to get you advice14:55
jgriffithe0ne: you sure about that?  :)14:55
e0nejgriffith: should we merge new drivers ( and allow to make code cleanup in the following patches or not?14:55
jgriffithe0ne: what's up?14:55
guitarzanavishay: victory!14:55
openstackgerritDmitry Guryanov proposed openstack/os-brick: remotefs: add virtuozzo storage support
jgriffithe0ne: looking14:56
jgriffithe0ne: so you're primarily referring to the last comment L#435?14:57
e0nejgriffith: personally, i prefer to not merge patches which require cleanip14:57
jgriffithe0ne: IMO I would agree with you14:57
jgriffithe0ne: the fact that you need a "cleanup" BP before you even merge the first version is a bad sign IMHO14:58
e0nejgriffith: L#435 is only one case. i'm going to answer to other comments14:58
*** hemnafk is now known as hemna14:58
e0nejgriffith: great comment! thanks14:58
jgriffithe0ne: there, I also added my 2 cents to the review15:03
*** jungleboyj has joined #openstack-cinder15:03
e0nejgriffith: thanks for the help and advice. i very appreciate it15:04
jgriffithe0ne: certainly15:05
*** Longgeek_ has quit IRC15:07
*** Longgeek has joined #openstack-cinder15:07
jungleboyjhemna: Morning.  Happy Friday!15:08
*** Miouge has quit IRC15:08
mtaninojungleboyj: thanks :)15:09
avishayjgriffith: guitarzan:
hemnayes it is!15:09
avishayjungleboyj: hemna: hello hello15:09
openstackgerritEdwin Wang proposed openstack/cinder: Add iSCSI protocol support for IBM FlashSystem
jungleboyjmtanino: My pleasure.  Thank you.15:10
jungleboyjavishay: Howdy sir!15:10
e0nehemna: evening:)15:11
guitarzanavishay: great, cinder still works with noauth! :)15:13
avishayguitarzan: create and list worked ;)15:14
guitarzanthat's great news imo15:14
avishayguitarzan: yep15:14
*** cknight has quit IRC15:15
*** dannywilson has joined #openstack-cinder15:15
*** dannywilson has quit IRC15:17
openstackgerritEdwin Wang proposed openstack/cinder: Add iSCSI protocol support for IBM FlashSystem
*** dannywilson has joined #openstack-cinder15:17
lpetrutjgriffith: Hi. Could you please spare some time to take a look over this patch? It already got a +2, Merging this bug fix and the one depending on it is quite important for us as at the moment we have to apply them on the CI before running tests15:19
*** rushil has quit IRC15:24
*** rushil has joined #openstack-cinder15:24
*** Miouge has joined #openstack-cinder15:25
*** e0ne is now known as e0ne_15:29
*** e0ne_ is now known as e0ne15:30
*** markvoelker has joined #openstack-cinder15:30
*** e0ne is now known as e0ne_15:30
*** e0ne_ is now known as e0ne15:30
*** patrickeast has joined #openstack-cinder15:31
*** markvoelker has quit IRC15:34
*** tsekiyama has joined #openstack-cinder15:36
*** daneyon has joined #openstack-cinder15:36
*** daneyon has quit IRC15:38
*** daneyon has joined #openstack-cinder15:38
hemnajgriffith, hey man, what happened with the replication spec.  I noticed it was abandoned this morning.  I was going to take another look at it.15:38
patrickeast: ( yea i just saw that15:39
jgriffithhemna: patrickeast I abandoned it15:40
jgriffithwell... obviously :)15:40
patrickeastjgriffith: so… like forever?15:40
hemnajust curious why ?15:40
jgriffithI dunno15:40
hemnaIt looked like we were getting close to finishing it up no ?15:41
jgriffithWell it's been almost six months and no real progress, seems like it's not really going to work15:41
patrickeasti don’t want that thing to die… i’m willing to take it up and argue with the ibm folks to get it through15:41
jgriffithhemna: I thought so, until that last round of -1's15:41
dannywilsonpatrickeast: +115:41
hemnaoh wrt to the failover ?15:41
hemnaI forgot, but what did we agree to wrt failover in Vancouver ?15:41
hemnawhat's described in the spec?15:42
patrickeasti think if we scale back how much automatic failover happens, so its just that cinder replication-failover <volume-d> command and not any volume operation that hits the scheduler it would alieve a lot of their concerns15:42
patrickeastand as far as the enable/disable15:43
patrickeastwell who gives a shit if v1 had it15:43
jgriffithpatrickeast: yeah, so what I think I'll try is maybe a POC15:43
jgriffithpatrickeast: see if that helps with some of this15:43
patrickeastjgriffith: that seems pretty reasonable15:43
jgriffithfrankly the spec process just burns me out :)15:44
hemnaso it looks like there are some good questions in the review, but there are also some comments that just don't seem reasonable to me15:44
patrickeastbut its soooo close15:44
*** Tross has joined #openstack-cinder15:44
jgriffithI'll try a poc, then build a spec off of it15:45
jgriffithsee if that helps15:45
hemnaI'd rather see us get something in that works for more than 1 vendor, rather than trying to solve every replication gotcha and waiting another 5 years.15:45
jgriffithshould be able to start working on it today, maybe have something ready by Wednesdays Cinder meeting15:45
jgriffithhemna: +115:45
jgriffithhemna: patrickeast honestly there's another approach as well15:46
dannywilsonjgriffith: maybe we can help on spec while you work on POC?15:46
*** david-lyle has quit IRC15:46
jgriffithdannywilson: sure15:46
*** david-lyle has joined #openstack-cinder15:46
jgriffithpart of the reason I've abandoned it now is that it's gone in circles on a couple things, and frankly I'm not even sure it's "right" anymore :)15:47
dannywilsonjgriffith: what do you think about the suggestion from patrickeast above about scaling back on how much automatic failover happens?15:47
*** rushil has quit IRC15:47
hemnawhy not make it an admin thing for now?15:48
jgriffithdannywilson: that's actually the *easy* part15:48
hemnadunno, I'm not a replication expert15:48
jgriffithhemna: well that's what it is currently15:48
jgriffithhemna: so the thing is, the comments that are coming in now are mostly just pushing it back to what the current implementation already is15:48
jgriffithso let me work on it this afternoon15:48
hemnawhich never worked for anyone other than 1 vendor15:49
jgriffithsee if we can get something finished up15:49
jgriffithhemna: yeah, but remember they have *20* customers just last week alone that started using it15:49
hemnaI'd hate to see this fail again, but I understand how complex it is and getting a gazillion vendors to buy off on something that makes them all happy.15:50
jgriffithhemna: patrickeast dannywilson Just give me a couple days to try something different15:50
jgriffithand see if if works15:50
hemnawe have to remember as a community that we should take small steps to get things right.15:50
*** markus_z has quit IRC15:50
dannywilsonjgriffith: okay, let us know if we can help out in any way, we can also try a POC on our driver if you think that would help15:50
smcginnishemna: +115:50
hemnaand sometimes that means putting something into place that doesn't make everyone happy/work the first time around.15:50
jgriffithhemna: TBH I think part of the problem is a misunderstanding of the flexibility I'm proposing15:50
*** leeantho has joined #openstack-cinder15:51
jgriffithdannywilson: that's what I was thinking, a POC might be better than having everybody interpret the "words" in a spec15:51
hemnaThis is probably one of the most complex features for Cinder15:51
jgriffithI honestly think people will be fairly happy with what I'm trying to put together15:51
hemnaso it's hard for me to wrap my head around it all in a spec.15:51
*** sgotliv has joined #openstack-cinder15:52
hemnabut my own issue.15:52
openstackgerritEdwin Wang proposed openstack/cinder: Add iSCSI protocol support for IBM FlashSystem
jgriffithhemna: I'd agree with you on that15:54
jgriffithhemna: and I think that's the problem all the way around15:54
jgriffithhemna: so maybe a spec WITH some POC code will help all of us15:54
*** _cjones_ has joined #openstack-cinder15:55
*** sgotliv has quit IRC15:55
*** sgotliv has joined #openstack-cinder15:55
hemnajgriffith, +115:56
*** Miouge has quit IRC16:00
lpetruthemma: Related to your comment regarding the Microsoft CI skipping attach test, where you reffering to test_attach_detach_volume ?16:00
*** jistr has quit IRC16:03
hemnalpetrut, yah, both the linux and windows smbfs skip the attach test16:05
*** shyama has quit IRC16:06
*** shyama has joined #openstack-cinder16:06
lpetruthemma: well, it seems that the validation.run_validation config option was not set, defaulting to False, in which case some of the tests are skipped. Thanks for bringing this up! We'll enable this.16:08
*** IanGovett has quit IRC16:08
openstackgerritTom Swanson proposed openstack/cinder: Dell: Added verify cert option for REST calls
*** avishay has quit IRC16:13
*** merooney has quit IRC16:14
thingeeDuncanT: casual review friday16:22
thingeeI'm guessing people are doing easy reviews today16:22
*** Miouge has joined #openstack-cinder16:23
*** e0ne is now known as e0ne_16:23
openstackgerritJohn Griffith proposed openstack/cinder-specs: Cinder move drivers/plugins to external libs
jgriffitheharney: maybe that's better, maybe it's not ^^16:24
*** e0ne_ has quit IRC16:24
jgriffithpatrickeast: smcginnis ^^ would alos be curious to get your thoughts on the alternative suggestion16:25
*** sgotliv has quit IRC16:26
smcginnisjgriffith: Will take a look now.16:26
* patrickeast looking16:26
smcginnisjgriffith: I actually lean somewhat toward alternative 2.16:27
smcginnisjgriffith: I think there's something to be said for public visibility of the driver code.16:28
smcginnisAnd public shaming. :)16:28
jgriffithsmcginnis: yeah, I figured most people would :)16:28
jgriffithsmcginnis: the only other question is still, "Am I the only one that cares about this at all"16:28
jgriffithsmcginnis: which is fine16:28
jgriffithif I am16:28
smcginnisjgriffith: I don't think that's the case.16:29
smcginnisjgriffith: I've heard a lot of people behind this general idea.16:29
*** jordanP has joined #openstack-cinder16:29
smcginnisjgriffith: I can just say from our experience, having our driver go through the standard review process has definitely made it better.16:29
patrickeastsmcginnis: +116:29
smcginnisjgriffith: We could have just implemented our driver in isolation, but we would have missed a lot.16:30
patrickeasti doesn’t need to be checked into the cinder repo, but having cinder reviewers look at our drivers has been hugely important16:30
smcginnisMy preference would be for a different repo with maybe some different requirements or even a different set of cores.16:30
openstackgerritSean Chen proposed openstack/cinder: Tintri driver to manage existing backend storage objects
smcginnisThe biggest concern I have is someone does their own driver in isolation and it has problems.16:31
smcginnisThen an end user tries to use it and hits some bugs.16:31
smcginnisYes, it will make that vendor look bad and it will fall on them to fix it.16:31
smcginnisBut it will also give somewhat an impression that Cinder/OpenStack quality is subpar.16:32
jgriffithsmcginnis: so myself, I'm not concerned about that at all... because it's that Vendors problem and not Cinders :)16:32
jgriffithsmcginnis: but yes, I see your point16:32
SwansonDrivers "in the box" is a plus for cinder.16:32
jgriffithpatrickeast: yeah, I get it16:32
jgriffithso I agree with everything you all are saying, that's why I modified to propose a compromise16:32
smcginnisjgriffith: Agreed, but I think it affects the overall impression of Cinder even if it isn't part of cinder proper.16:32
*** jdurgin1 has quit IRC16:32
smcginnisjgriffith: I appreciate that. Good for discussion.16:33
smcginnisjgriffith: I'm certainly open to being in the minority here too, but it's good to talk about.16:33
*** dmz has joined #openstack-cinder16:33
*** merooney has joined #openstack-cinder16:34
*** crose has joined #openstack-cinder16:35
*** anshul has quit IRC16:36
*** yamada-h has joined #openstack-cinder16:37
*** breitz has quit IRC16:38
*** breitz has joined #openstack-cinder16:39
*** e0ne has joined #openstack-cinder16:40
*** Daviey has joined #openstack-cinder16:41
*** yamada-h has quit IRC16:42
*** Miouge has quit IRC16:42
*** e0ne has quit IRC16:43
patrickeastjgriffith: i like the new option… but plz no subrepo16:43
patrickeasti worked on a project with a bunch of subrepos and i still have nightmares16:43
patrickeastor crazy sub-tree shenanigans16:44
*** Miouge has joined #openstack-cinder16:44
*** IanGovett has joined #openstack-cinder16:44
smcginnisNo shenanigans allowed. :)16:44
*** Miouge has quit IRC16:45
*** Miouge has joined #openstack-cinder16:46
*** annegentle has quit IRC16:46
patrickeastmy personal rule of thumb is that if you have to type git submodule or git subtree you’ve done something wrong16:47
patrickeastalthough i might be biased16:47
*** jordanP has quit IRC16:48
DavieyHi, I'm having a problem with XtremIO driver initialisation.  Anyone here familiar with EMC kit?16:48
patrickeastxyang: ^16:49
xyangDaviey: hi, what is issue?16:50
openstackgerritSzymon Wróblewski proposed openstack/cinder: Tooz locks
Davieyxyang: When XtremeIO loads it tries to contact http://x/api/json/types/xms , which is 40416:51
xyangDaviey: where did you get the driver and what xtremio version are you running16:52
Davieyxyang: current master, 3.0.316:52
xyangDaviey: can you send me logs?  I'll ask the xtremio team to take a look.  my email: xing.yang@emc.com16:53
*** annegentle has joined #openstack-cinder16:54
*** cknight has joined #openstack-cinder16:55
flip214thingee: ping?16:56
thingeeflip214: hi16:56
flip214from -infra:16:56
flip21418:55 < jeblair> flip214, asselin__, fungi: do we need a cinder sign-off on that? or is the cinder core team/ptl expecting this job to16:56
*** cknight has quit IRC16:56
flip214be added?"16:56
*** sasi has left #openstack-cinder16:56
*** cknight has joined #openstack-cinder16:56
flip214this is about, CI for the DRBD driver.16:56
*** breitz has quit IRC16:57
*** cknight has left #openstack-cinder16:57
*** raunak has joined #openstack-cinder17:01
flip214thingee: apart from watching the results, do I need to submit the DRBD driver in some way to have it in the list of pending reviews?17:02
flip214*results of CI tests against other commits, that is17:02
thingeepropose the patch, I will compare what was already in merge...should be a quick merge. I will only merge once the CI passes though.17:03
thingeeflip214: ^17:03
thingeewhat was already merged in Kilo*17:03
thingeeflip214: so wait to propose until that gate test is passing a few times would be my advice17:03
flip214yeah, okay.17:04
flip214well, what you reverted in April is quite a bit different from what I submitted, because of several review comments...17:04
thingeeflip214: ok, just being optimistic of it being quick :)17:05
thingeeI would like it to be quick. it's a friday afterall17:05
thingeeDuncanT promised me casual review friday and he's not delivering17:06
* thingee thinks he's probably already out dancing somewhere to care17:06
thingeeI'll bb17:06
*** annegentle has quit IRC17:08
*** IanGovett has quit IRC17:10
*** vilobhmm has joined #openstack-cinder17:10
flip214thingee: I'll be gone soon, too, as it's already past 7pm here, and I've got to bring the kids to bed.17:10
flip214I'll be back on Monday, though!17:11
flip214have a nice weekend.17:11
jgriffithsmcginnis: hemna DuncanT jungleboyj thingee
smcginnisjgriffith: Small (but important) typo.17:12
jgriffithsmcginnis: ruhroh17:12
jgriffithsmcginnis: DOH!!!!17:12
thingeejgriffith: seems fine. We'll be doing the push on Monday.17:13
thingeeAfter typo fix17:13
openstackgerritJohn Griffith proposed openstack/python-cinderclient: Prep for 1.2.2
jgriffiththingee: Oh17:13
jgriffiththingee: I was going to push it as soon as that landed17:14
thingeejgriffith: in the infra channel, I got a sense this isn't urgent to do on a Friday.17:14
jgriffiththingee: didn't know you had *plans*17:14
jgriffithOk, well then never mind17:14
thingeeInfra and I have plans :). Not just me17:14
jgriffiththingee: s'alright17:14
*** merooney has quit IRC17:15
*** rushil has joined #openstack-cinder17:16
vilobhmmjgriffith : ^^17:18
*** merooney has joined #openstack-cinder17:18
*** asselin__ has quit IRC17:19
DuncanTThingee: I've got to come up with a new plan for casual review Fridays... I forgot the very important detail that Friday isn't a work day here. I'll start preparing a list Thursday I guess17:19
*** markvoelker has joined #openstack-cinder17:19
DuncanTThingee: I was actually at the opera, not dancing, as it happens17:19
jgriffithvilobhmm: thanks17:20
jgriffithvilobhmm: apparantly it's a moot point, thingee "has plans"17:20
openstackgerritDmitry Guryanov proposed openstack/cinder: Add Virtuozzo Storage Volume Driver
vilobhmm"plans" about what :P17:20
jgriffithvilobhmm: releasing a client version17:21
jgriffithvilobhmm: so that update was so I could release something as soon as it landed17:21
jgriffithvilobhmm: but there's already a schedule in place17:21
*** crose has quit IRC17:22
*** markvoelker has quit IRC17:24
thingeejgriffith: I was ready to do the same commit an hour ago, but fungi requested this not happen on a Friday. Not my plans.17:25
openstackgerritEdwin Wang proposed openstack/cinder: Add iSCSI protocol support for IBM FlashSystem
*** harlowja has quit IRC17:27
jgriffiththingee: Ahhh... yeah, that makes sense17:28
jgriffiththingee: I forgot about the "no client pushes on friday" thing17:28
jgriffiththingee: sorry about that, totally makes sense17:28
jgriffiththingee: maybe I can sort out the Kilo compat issues between now and then too17:28
* smcginnis envisions DuncanT in a tux with opera glasses17:28
openstackgerritEdwin Wang proposed openstack/cinder: Add iSCSI protocol support for IBM FlashSystem
smcginnisjgriffith: thingee So go ahead with but wait to push client? Or just wait all together?17:30
*** xyang has quit IRC17:30
jgriffithsmcginnis: merging the bump in version num should be fine17:31
jgriffithsmcginnis: thingee was just reminding me that we DONT push to pypi on Fridays :)17:31
smcginnisjgriffith: Done. OK, no casual pypi Fridays then. :)17:32
*** harlowja has joined #openstack-cinder17:32
jgriffithsmcginnis: LOL17:33
*** rushiagr is now known as rushiagr_away17:40
*** lpetrut has quit IRC17:41
*** rushil has quit IRC17:42
*** krtaylor has quit IRC17:46
openstackgerritMikhail Khodos proposed openstack/cinder: Re-add Nexenta driver.
openstackgerritEarle F. Philhower, III proposed openstack/cinder: Implement Cinder Volume driver for HGST Solutions
*** e0ne has joined #openstack-cinder17:48
smcginnisjgriffith: thingee Before I do something wrong - is it OK to push through python-cinderclient global requirement updates. Just newer minimum version of keystoneclient.17:49
*** rushil has joined #openstack-cinder17:50
*** annegentle has joined #openstack-cinder17:50
*** merooney has quit IRC17:52
*** Tross has quit IRC17:52
e0nesmcginnis: why not if it passed CI?17:54
thingeesmcginnis: seems fine to me. otherwise we'd have to do another push to not be conflicting on that as well17:54
smcginnise0ne: That's what I assume, just didn't know if we had a policy on this.17:54
thingeeit's the pypi dance17:54
smcginnisthingee: In general, as long as the OpenStack Proposal Bot patches look OK should those just be ninja'd through?17:55
*** akerr has quit IRC17:56
* smcginnis off to last meeting of the day17:56
jgriffithsmcginnis: thingee wait17:56
jgriffithsmcginnis: so that will introduce another problem17:57
jgriffithsmcginnis: thingee that change says >=1.6.017:57
jgriffithsmcginnis: thingee but Cinder kilo has: keystonemiddleware>=1.5.0,<1.6.017:58
jgriffithsmcginnis: thingee not sure how those tie in?17:58
jgriffithsmcginnis: thingee maybe the anser is in fact just "you can't run cinderclient on node"17:58
e0nejgriffith: if i did't miss anything, smcginnis talks about keystoneclient, _not_ keyston middleware17:58
jgriffithe0ne: Yes, that's what I'm asking17:58
jgriffithe0ne: if that relationship matters or not17:59
e0negood point17:59
* e0ne verifying...17:59
*** thangp has quit IRC17:59
*** crose has joined #openstack-cinder17:59
e0nein kilo keystonemiddleware has: python-keystoneclient>=1.1.018:01
e0nejgriffith, thingee: afair, keystoneclient is deprecated. should we switch to openstackclient instead?18:02
*** nikeshm has joined #openstack-cinder18:02
jgriffithe0ne: I'm staying out of that one ;)18:03
*** rwsu has quit IRC18:04
*** rajinir_r has quit IRC18:12
*** akerr has joined #openstack-cinder18:12
*** earlephilhower has quit IRC18:14
openstackgerritAlex Meade proposed openstack/cinder: NetApp E-Series driver: Remove caching logic
openstackgerritAlex Meade proposed openstack/cinder: NetApp E-Series: Refactor class structure for FC
openstackgerritAlex Meade proposed openstack/cinder: NetApp E-Series: Add Fibre Channel Support
*** merooney has joined #openstack-cinder18:16
e0nejgriffith: :)18:18
e0neeharney: hi. are you around? i would like to discuss some issues with ceph driver and eventlet before sending it to openstack-dev ml18:20
*** krtaylor has joined #openstack-cinder18:20
*** earlephilhower has joined #openstack-cinder18:22
*** shyama has quit IRC18:23
*** kaisers1 has quit IRC18:35
*** mriedem is now known as mriedem_madmax18:35
*** mriedem_madmax has quit IRC18:35
*** kaisers1 has joined #openstack-cinder18:37
*** kaisers has quit IRC18:37
*** Tross has joined #openstack-cinder18:38
*** kaisers has joined #openstack-cinder18:39
*** xyang1 has joined #openstack-cinder18:41
thingeejgriffith: e0ne hahah18:46
*** harlowja has quit IRC18:46
jgriffiththingee: :)18:47
thingeejgriffith: the cinderclient installed on nodes is a particular version though right?18:47
jgriffiththingee: I just updated the bug, I can't repro18:47
jgriffiththingee: the problem was reported as: "I'm running Kilo, WTF... you said I could set private types"18:48
* jgriffith said.. you can just "pip install --upgrade python-cinderclient"18:48
thingeepeople should just be happy that it's friday18:48
eharneye0ne: i'm back18:48
jgriffiththey said:  "WTF, cinder --version PUUUKES now, oslo.cfg error"18:48
thingeeI mean come on18:48
jgriffiththingee: so I filed a bug... then I tried it in devstack and ... well18:50
jgriffith"it worked in devstack"18:50
thingeejgriffith: speaking of that, did you see the new shirt at the summit?18:50
* jgriffith sends them a T-shirt18:50
jgriffithI did not18:50
e0nefree T-shirts?18:50
openstackgerritJoshua Huber proposed openstack/cinder: Add cinder volume driver for Blockbridge EPS
e0neeharney: actually, i've got 2 questions18:51
e0neeharney: i want to try implement retries for rbd driver on connection timeouts18:52
e0neeharney: do you think it good idea? i just started testing work with timeouts so i don't know almost anything about realted errors18:52
*** harlowja has joined #openstack-cinder18:53
*** ociuhandu has quit IRC18:53
eharneye0ne: it seems like a reasonable idea to me18:53
*** ctracey has quit IRC18:54
*** serverascode has quit IRC18:54
*** thingee has quit IRC18:54
*** wanghao has quit IRC18:54
*** rhefner has quit IRC18:54
*** scottda has quit IRC18:54
*** ameade has quit IRC18:54
*** briancurtin has quit IRC18:54
*** zhiyan has quit IRC18:54
e0neeharney: do you have any expirience with timeouts on your envs? they are disabled by default18:54
eharneye0ne: no, i don't18:54
e0neeharney: thanks. one more questions18:56
*** lpetrut has joined #openstack-cinder18:57
e0neeharney: i don't have good enough steps to reproduce, but some times we've got an issue, when cinder tries to connect to ceph for delete volume and this thread (real python thread, eventlet.tpool) hangs18:57
*** rushil has quit IRC18:57
e0neeharney: did you see anything related?18:58
eharneye0ne: well, i think that's what came up in the Nova review of the threaded delete patch, that it could deadlock like that18:58
eharneye0ne: what's in Cinder now seems to be unsafe18:58
*** rushil has joined #openstack-cinder18:59
e0neeharney: do you mean this one?18:59
*** raunak has quit IRC18:59
eharneye0ne: yes19:00
*** angela-s has joined #openstack-cinder19:01
e0neeharney: in my env, only threads with rdb call are locked. main thread of c-vol feels good and working19:01
eharneye0ne: well... it sounds like that could still be related, though it's hard to say if that's the expected outcome19:03
uberjayhemna: hey there -- how would you feel about taking another look at for me? rebased due to conflict in, but otherwise the same.19:03
e0neeharney: agree, it could be related. i'm just confused that threads are blocked and python-rbd  doesn't delete volumes without any messaged in cinder and ceph logs19:06
e0neTBH, i didn't set debug logs for ceph19:07
*** markvoelker has joined #openstack-cinder19:08
asselindevstack patch to pull in dev version of os-brick19:08
asselindevstack-gate change is:
asselinwhich will update devstack jobs to setup the workspace with os-brick git cloned with zuul ref19:09
uberjayhemna: you had -1'd me before, but I think just for requests vs. httplib & lack of utilization reporting. if there's more to it than that, i'd like to get on it asap, since the fast-approaching deadline is making me nervous!19:10
openstackgerritMikhail Khodos proposed openstack/cinder: Re-add Nexenta driver.
*** Lee1092 has quit IRC19:11
*** markvoelker has quit IRC19:12
*** annegentle has quit IRC19:14
*** rushil has quit IRC19:14
*** dannywilson has quit IRC19:16
*** Longgeek has quit IRC19:16
*** rushil has joined #openstack-cinder19:16
*** pradipta has joined #openstack-cinder19:19
*** rushil has quit IRC19:20
*** rushil has joined #openstack-cinder19:23
*** jnicklin has joined #openstack-cinder19:23
*** merooney has quit IRC19:27
*** rushil has quit IRC19:31
*** leakypipes has quit IRC19:34
openstackgerritSean Chen proposed openstack/cinder: Remove duplicate volume.filters entry
*** e0ne has quit IRC19:39
*** merooney has joined #openstack-cinder19:41
*** earlephilhower has quit IRC19:48
*** akerr has quit IRC19:51
*** Yogi1 has quit IRC19:51
*** annegentle has joined #openstack-cinder19:54
*** xyang has joined #openstack-cinder19:55
*** dims__ has joined #openstack-cinder19:57
*** merooney has quit IRC19:58
xyangasselin: ping19:59
hemnauberjay, heyas20:00
hemnawhat's up20:00
*** dims_ has quit IRC20:00
*** boris-42 has quit IRC20:02
*** rex_lee has quit IRC20:02
*** sweston has quit IRC20:02
*** annegentle has quit IRC20:04
*** annegentle has joined #openstack-cinder20:04
*** dustins has quit IRC20:07
uberjayhemna: hey there20:08
*** asselin_ has joined #openstack-cinder20:10
*** dims__ has quit IRC20:11
*** rushil has joined #openstack-cinder20:12
uberjayhemna: I'm thinking our driver is ready for merge, and any additional fixes and features should go in as an additional patch. I certainly don't want to bother core devs if I don't have to, but the deadline is soon, which makes me nervous. Have any guidance? :)20:14
hemnauberjay, ok I'll look again20:14
uberjayhemna: thanks, I appreciate it!20:15
smcginnisuberjay: Blockbridge, right?20:15
uberjaysmcginnis: yep, that's us :)20:16
smcginnisuberjay: You have some pep8 errors to fix.20:16
*** e0ne has joined #openstack-cinder20:16
uberjayack, must have snuck in. i'll fix them.20:16
*** tellesnobrega_ has joined #openstack-cinder20:17
*** rushil has quit IRC20:17
hemnauberjay, what's the value in providing the blockbridge_volume_query cinder.conf entry ?20:17
hemnawhy do admins need that ?20:17
hemnathat seems like a driver internals thingy20:17
*** ociuhandu has joined #openstack-cinder20:19
uberjaythe volume query is used to select a subset of storage advertised via a particular blockbridge storage service -- essentially an attribute query string.20:20
smcginnisuberjay: Also, is there really a need for BlockBridgeAPIException instead of just using VolumeBackupAPIException?20:21
smcginnisI see the code actually does both.20:21
hemnauberjay, why would an admin need that ?20:22
uberjaythere's some overlap with the cinder scheduler, of course. e.g., consider some storage service that is used by cinder and non-cinder clients. the query gives the admin a way to select what resources should be consumed by cinder.20:22
hemnauberjay, like a pool for provisioning a volume ?20:23
uberjaysmcginnis: (i'll take a look, thanks)20:23
*** markvoelker has joined #openstack-cinder20:24
uberjayhemna: something like that, yes. the pool of backend datastores used for provisioning is selectable by attribute. the admin may configure this on the backend, or via the cinder driver configuration. for example, you could use it to expose two different pools of storage (as different volume types), both provisioned via the same backend blockbridge service.20:25
hemnauberjay, heh ok20:26
hemnawell, I'd suggest using Cinder's built in mechanism for 'pool's to do this then20:26
hemnainstead of cinder.conf or extra specs20:26
hemnayou'll be back dooring the scheduler's built in support for pools20:27
hemnawhich will cause you problems20:27
hemnaa better approach is to list the pools you want to support in cinder.conf20:27
hemnathen have your get_volume_stats report on those pools20:27
uberjaycan you point me at some more details about storage pools? we don't really have a storage pool concept  --  we just have datastores, which may or may not be tagged with a particular attribute does that make sense? I'm not sure if it maps well, but it may.20:27
hemnaand let the cinder scheduler pick which pool to use.20:28
*** markvoelker has quit IRC20:28
hemnauberjay, that's what we do for our driver.20:33
hemnawe have a list of 'pool's that volumes can be provisioned from and we report stats up for each of those.20:33
hemnathe scheduler actually picts which one is used.20:33
hemnacomes from that ^^^20:33
uberjaymm, gotcha. currently all we could do is directly expose each discrete datastore as a pool to support the pool-aware scheduler. it doesn't really express the full abilities of *our* provisioning query support, though. ah, interesting. i'm looking.20:34
*** annashen has joined #openstack-cinder20:34
uberjayah, that's actually a great way to go. that's much better than having to define multiple driver instances to select different pools.20:35
uberjayultimately, though, the way we select each "pool" of storage boils down to a query (for our system)20:36
hemnasure, just abstract into something you can render into your query that goes into the query on line 32220:37
hemnameh, that made no sense20:37
uberjayheh :)20:37
hemnaif you have some description or name of the store/pool, then just use that as the pool name list in cinder.conf20:37
hemnathen render that name into the query that is used as part of the _create_volume() call20:38
*** tellesnobrega_ has quit IRC20:38
*** tellesnobrega_ has joined #openstack-cinder20:39
uberjaysure, yes. that would work. Essentially the blockbridge_query_string *is* the pool name.20:42
hemnathere ya go20:42
hemnaso I'd probably rename that conf entry and call it something that makes sense, but make it a list of supported 'pool' names.20:43
hemnaand FWIW, this is a prime example of why keeping drivers in tree is helpful for driver writers.20:43
hemnaand forcing core devs to do driver reviews.20:44
uberjaycan conf options be dicts?20:44
hemnanot sure20:44
hemnacan you do a CSV ?20:44
hemnathat's what we do20:44
hemnafoo,bar,blah,something sucky, something less sucky, awesomeness20:44
uberjayyou use a list, presumably because your pool names map directly into some backend object name20:45
*** e0ne has quit IRC20:45
smcginnishemna: +1 on benefits of in tree as well as pool names. :D20:45
*** ctracey has joined #openstack-cinder20:46
hemnasmcginnis, but I love the idea of pip install <vendor driver>20:46
hemnaright now we can't really fix bugs against older shipping releases.20:46
smcginnisI wonder if we had it in a different repo if we could do something a la cinderclient to "publish" them.20:46
smcginnisThen have them pulled down on demand.20:46
smcginnisSo all in tree, but not all in source distro.20:47
hemna40 subprojects of cinder20:47
hemnaP A I N20:47
smcginnis50 shades of cinder?20:47
xyangsmcginnis: hi, the ScaleIO CI name is fixed now. you can filter it out now20:47
smcginnisNot each a subproject.20:47
smcginnisxyang: Cool!20:47
uberjayyeah, I was reading that discussion and thinking the same. I can see how getting the core team away from the large (and growing) 3rd party drivers would be a big relief, but... quality control and such. unfortunately, people are going to come to you when a driver reaches behind and corrupts the cinder database. :(20:47
hemnasmcginnis, so every driver is in 1 subproject repo cinder-drivers ?20:47
smcginnishemna: I was just thinking on repo "cinder-drivers" and have some way to pull out each driver as a module published to be able to pip install per driver.20:48
*** e0ne has joined #openstack-cinder20:48
smcginnishemna: No idea if that's at all feasible, just brainstorming on a Friday afternoon. :)20:49
*** pradipta has quit IRC20:49
*** lpetrut has quit IRC20:49
*** serverascode has joined #openstack-cinder20:49
uberjayrun 3rd party drivers in another process/sandbox, maybe? ;) *grin*20:49
*** ameade has joined #openstack-cinder20:51
*** gouthamr has quit IRC20:51
*** sweston has joined #openstack-cinder20:53
hemnasmcginnis, yah I'm not sure how you could separate each individual driver in that shared repo.20:54
*** zhiyan has joined #openstack-cinder20:54
smcginnishemna: Script that expects a certain folder structure maybe?20:54
hemnaI guess it's just a higher level dir that contains dirs of pip package dirs20:54
smcginnisfor each driver in /cinder-drivers/drivers/*20:55
smcginnisSomething along those lines.20:55
smcginnisI'm sure I'm vastly simplifying what would be a complex task though.20:55
hemnayah seems doable20:56
smcginnisHeading out. Have a good weekend everyone.20:58
mtaninosmcginnis: you too!20:58
hemnasmcginnis, l8rs!20:58
openstackgerritMikhail Khodos proposed openstack/cinder: Re-add Nexenta driver.
*** boris-42 has joined #openstack-cinder21:00
hemnaspeaking of moving drivers out of tree....21:00
*** briancurtin has joined #openstack-cinder21:05
*** e0ne has quit IRC21:06
*** krtaylor has quit IRC21:08
*** ociuhandu has quit IRC21:09
*** tellesnobrega_ has quit IRC21:11
*** raunak has joined #openstack-cinder21:11
*** thingee has joined #openstack-cinder21:12
*** bswartz has joined #openstack-cinder21:13
*** rhefner has joined #openstack-cinder21:13
*** scottda has joined #openstack-cinder21:14
uberjayhemna smcginnis -- thanks for your feedback. i'm taking care of the issues :)21:15
*** wanghao has joined #openstack-cinder21:17
*** akerr has joined #openstack-cinder21:20
*** akerr_ has joined #openstack-cinder21:21
Swansonxyang: jgriffith thanks for pushing through 182396.21:23
*** akerr has quit IRC21:25
jgriffithSwanson: sure21:28
*** esker has quit IRC21:29
xyangSwanson: welcome21:29
*** rhagarty has quit IRC21:32
*** cbader has quit IRC21:32
openstackgerritEarle F. Philhower, III proposed openstack/os-brick: Add HGST Solutions connector
*** Miouge has quit IRC21:37
*** akerr_ has quit IRC21:38
*** openstack has joined #openstack-cinder21:43
*** rmesta has quit IRC21:43
*** rex_lee has joined #openstack-cinder21:43
*** rmesta has joined #openstack-cinder21:44
*** rmesta has quit IRC21:44
*** jnicklin has quit IRC21:44
*** rmesta has joined #openstack-cinder21:44
*** rmesta has quit IRC21:44
*** rmesta has joined #openstack-cinder21:45
*** rmesta has quit IRC21:45
*** rmesta has joined #openstack-cinder21:45
*** krtaylor has joined #openstack-cinder21:54
*** earlephilhower has joined #openstack-cinder21:54
*** primechuck has quit IRC21:59
*** dannywilson has joined #openstack-cinder22:03
*** dannywilson has quit IRC22:03
*** earlephilhower has quit IRC22:06
openstackgerritEarle F. Philhower, III proposed openstack/os-brick: Add HGST Solutions connector
*** markvoelker has joined #openstack-cinder22:12
*** earlephilhower has joined #openstack-cinder22:15
*** markvoelker has quit IRC22:17
*** earlephilhower has quit IRC22:19
*** annashen has quit IRC22:19
*** earlephilhower has joined #openstack-cinder22:19
*** jungleboyj has quit IRC22:21
openstackgerritTom Swanson proposed openstack/cinder: Dell SC: Added support for alternate iscsi portals
*** xyang has quit IRC22:26
openstackgerritEarle F. Philhower, III proposed openstack/cinder: Implement Cinder Volume driver for HGST Solutions
*** xyang1 has quit IRC22:29
openstackgerritMikhail Khodos proposed openstack/cinder: Re-add Nexenta driver.
*** daneyon has quit IRC22:33
*** daneyon has joined #openstack-cinder22:33
*** daneyon has quit IRC22:33
*** daneyon has joined #openstack-cinder22:34
hemnatbarron, ping22:35
*** openstackgerrit has quit IRC22:37
*** openstackgerrit has joined #openstack-cinder22:37
earlephilhowerthingee:  Howdy,  Can I ask a 3rd party Cinder CI question?  Read your post on it 2015-May/064072.html, but unclear about CI posting its results to gerrit prior to approval of new driver.22:39
asselinearlephilhower, your ci shoudl post resutls prior to approval22:40
hemnaearlephilhower, do you have a driver you want to submit to Cinder ?22:41
*** lcurtis has quit IRC22:41
earlephilhowerasselin:  Thx.  So real Zuul-initiated gerrit comment addition, or just link to the logs from a manual started Jenkins run?  I'm running your setup (well, as it was 2 weeks ago before diskbuilder change).22:42
tbarronhemna: pong22:43
asselinearlephilhower, actually, you should test your ci with openstack-dev/ci-sandbox22:43
hemnatbarron, so the compression patch22:43
earlephilhowerhemna:  Yes, I've got the BP up along w/code.  Working through the rest of the formalities now (I hope).22:43
asselinsend me the link of the log files I can take a quick look. once you're happy witht the way it works in the sandbox you can enable in cinder.22:43
hemnatbarron, since this patch basically makes it possible to have blocks that aren't compressed, does the restore of the backup work correctly?22:44
asselinearlephilhower, be careful when posting to cinder. If you create too much noise (i.e. brokensystem not adding any value) your accound can get disabled22:44
hemnatbarron, dealing with compressed vs. uncompressed blocks?22:44
tbarronhemna: yes22:44
tbarroneach chunk has its own metadata22:44
tbarronwhich includes the type of compression, including 'none'22:44
tbarronrestore uses this22:44
earlephilhowerasselin:  Yes, it's been committing to the ci-sandbox for a while now.  Just hooked up to openstack/cinder on silent mode and chugging on Nexenta patch.  Logs at$LOG_PATH/ (LOG_PATH is fixed for real gerrit stuff)22:45
tbarronall that pre-existed this change22:45
hemnatbarron, ah ok cool.   that was my only concern with this since I was reading it and thinking about it.22:45
tbarronpls do look down at the restore a chunk method, or whatever that's called22:45
asselinearlephilhower, what do you mean 'fixed for real gerrit stuff?22:45
tbarronyou are asking the right question and shouldn't just take my word for it22:45
hemnatbarron, ah ok I see the 'none', data being returned and that's getting stuffed into the metadata22:45
earlephilhowerasselin, Got it!  I saw the IBM system that went loopy..  LOG_PATH "fixed" means it comes out as ""22:46
earlephilhowerJust the manually started Jenkins jobs don't define LOG_PATH22:46
earlephilhowerSo it ends up with ugly URL22:47
asselinok, yeah, do you have one of your driver?22:47
earlephilhowerYes, the "$LOG_PATH" one.  It really had dollarsign-LOGPATH in the URL:$LOG_PATH/22:48
earlephilhowerI didn't think "$" was legit in URLs, guess I was wrong...22:48
asselinearlephilhower, don't have one on silent mode?22:48
asselinearlephilhower, real patch being ci-ed?22:49
*** ozialien has joined #openstack-cinder22:49
earlephilhowerasselin, do you mean like one initiated by gerrit and uploaded to the logserver but no gerrit review sent back?22:50
tbarronhemna: in restore_v1 you'll see an iteration through metadata_objects.  For each, we look at the compression algo22:51
tbarron(which may be 'none')22:51
hemnatbarron, ok perfect.   thanks.  I was just curious and concerned22:51
tbarronor rather, at this point None22:51
tbarronand if not None, we decompress22:52
earlephilhowerJust kicked silent mode on this afternoon, and only 2 patches since have come through.  Looks like about 10 minutes left on the first one.22:52
tbarronIt was when I realized that the compression algo was saved for each chunk that I realized we could do this.22:52
earlephilhowerasselin:  logs uploading right now, sec for url22:53
hemnatbarron, so I presume the 'none' algorithm simply means jam this block behind the last one and continue ?22:53
tbarronmore or less.  it means write it at the designated offset and continue.22:54
hemnamakes sense.22:54
tbarronincrementals do a full restore (where all writes end up being continguous) then jump around on subsequent passes22:55
tbarronnot "around" - always forwards, but not contiguously22:56
earlephilhowerasselin:  Logs from silent mode checkin here:
* asselin looks22:57
tbarronhemna: thanks, walt!22:57
hemnatbarron, np22:58
asselinearlephilhower, overall very nice!22:59
asselinearlephilhower, a couple of this.22:59
asselin1. please increae your regex to 'volume'22:59
asselin2. doesn't show as html22:59
asselinare you using my log server script? that should set it up correctly....
asselinwhich will also setup log-analyze (optional, but a big plus)23:00
asselinin the mean time, I personally don't have issues with your ci commenting as -is23:01
earlephilhowerasselin, thanks.  #1 easy-peasy.  #2, I'm gonna have to break out my apache-fu.  We've got some corporate requirements for external servers so I couldn't use your script as-is.  I kind of hit the bare-minimum as a start.23:01
asselinearlephilhower, ok, fair enough....take a look at the apache conf for clues then23:02
*** ozialien has quit IRC23:02
thingeeearlephilhower: for companies with requirements poking holes in the firewall to allow access with some webserver, I get that. Just use some external web server.23:02
thingeethat's managed elsewhere23:02
thingeeas a last resort though23:03
hemnaasselin, jenkins is happy now.23:04
earlephilhowerasselin, I'll grok that link as best I can.  I'm actually running on an external server already, but it's also serving some other needs so I couldn't change things too greatly.23:04
hemnathingee, FYI
asselinthingee,  please...and all your ptl friends :)23:05
* thingee wishes he had friends23:05
*** ozialien has joined #openstack-cinder23:05
* asselin will friend for +123:06
earlephilhowerasselin, so I'm going to update the regex right now and see about the couple log-things.  Appreciate the feedback!23:06
asselinyou're welcome23:06
openstackgerritMikhail Khodos proposed openstack/cinder: Nexenta Edge iSCSI backend driver.
earlephilhowerasselin, so if over the weekend on silent I don't see anything bad it's OK to move to check on Mon?  Don't want to move to check right before weekend, and have to hack stuff at 3AM Sat. nite.23:08
asselinearlephilhower, yeah that's fine23:09
hemnahave a good weekend guys.23:09
*** hemna is now known as hemnafk23:09
*** dannywilson has joined #openstack-cinder23:15
*** angela-s has quit IRC23:16
openstackgerritMike Perez proposed openstack/cinder: Revert "Adds drivers for DotHill Storage Arrays."
*** annegentle has quit IRC23:18
*** annashen has joined #openstack-cinder23:20
*** dannywilson has quit IRC23:23
*** dannywilson has joined #openstack-cinder23:23
*** annashen has quit IRC23:26
*** ian_brown has quit IRC23:33
*** ozialien has quit IRC23:33
*** daneyon has quit IRC23:34
*** ian_brown has joined #openstack-cinder23:37
*** jwcroppe has joined #openstack-cinder23:41
*** ian_brown has quit IRC23:42
*** raunak has quit IRC23:45
*** ian_brown has joined #openstack-cinder23:47
mtaninojgriffith: Thanks for that comment :)23:50
*** _cjones_ has quit IRC23:58

Generated by 2.14.0 by Marius Gedminas - find it at!