Monday, 2018-04-23

*** dpawlik has joined #openstack-cinder00:03
*** dpawlik has quit IRC00:08
*** dave-mccowan has joined #openstack-cinder01:00
*** dave-mcc_ has joined #openstack-cinder01:02
*** dave-mccowan has quit IRC01:05
*** namnh has joined #openstack-cinder01:17
*** salv-orlando has joined #openstack-cinder01:28
*** kaisers_ has joined #openstack-cinder01:30
*** salv-orl_ has quit IRC01:31
*** kaisers has quit IRC01:34
*** zhaochao has joined #openstack-cinder01:34
*** bkopilov has quit IRC01:37
openstackgerritjunbo.li proposed openstack/cinder master: Add missing VolumeAttachStatus enum field  https://review.openstack.org/56344401:37
*** sapd has joined #openstack-cinder01:43
*** masuberu has quit IRC02:06
*** masuberu has joined #openstack-cinder02:06
*** deng_ has joined #openstack-cinder02:13
*** derekjhyang has joined #openstack-cinder02:16
*** deng_ has quit IRC02:24
*** wanghao has joined #openstack-cinder02:24
*** deng_ has joined #openstack-cinder02:25
*** moshele has joined #openstack-cinder02:27
*** armaan has quit IRC02:29
*** wanghao has quit IRC02:29
*** armaan has joined #openstack-cinder02:29
*** wanghao has joined #openstack-cinder02:30
*** wanghao has quit IRC02:35
*** wanghao has joined #openstack-cinder02:35
*** Dinesh_Bhor has joined #openstack-cinder02:39
*** dklyle has joined #openstack-cinder02:41
*** wanghao has quit IRC02:41
*** wanghao has joined #openstack-cinder02:42
*** psachin has joined #openstack-cinder02:43
*** wanghao has quit IRC02:47
*** wanghao has joined #openstack-cinder02:49
*** bkopilov has joined #openstack-cinder02:49
*** wanghao has quit IRC02:56
*** wanghao has joined #openstack-cinder02:56
*** dklyle has quit IRC03:00
*** wanghao has quit IRC03:00
*** wanghao has joined #openstack-cinder03:01
*** masuberu has quit IRC03:06
*** wanghao has quit IRC03:07
*** wanghao has joined #openstack-cinder03:08
*** masuberu has joined #openstack-cinder03:08
*** wanghao has quit IRC03:12
*** wanghao has joined #openstack-cinder03:12
*** masber has joined #openstack-cinder03:15
*** wanghao has quit IRC03:17
*** wanghao has joined #openstack-cinder03:18
*** masuberu has quit IRC03:18
*** wanghao has quit IRC03:22
*** wanghao has joined #openstack-cinder03:23
*** rcernin has quit IRC03:31
*** rcernin has joined #openstack-cinder03:35
*** lpetrut has joined #openstack-cinder03:39
*** psachin has quit IRC03:39
*** wanghao has quit IRC03:39
*** wanghao has joined #openstack-cinder03:41
*** wanghao has quit IRC03:46
*** wanghao has joined #openstack-cinder03:46
*** wanghao has quit IRC03:52
*** wanghao has joined #openstack-cinder03:53
*** wanghao has quit IRC03:53
*** wanghao has joined #openstack-cinder03:53
*** wanghao has quit IRC03:54
*** wanghao has joined #openstack-cinder03:54
*** wanghao has quit IRC03:55
*** wanghao has joined #openstack-cinder03:55
*** edmondsw has joined #openstack-cinder03:55
*** wanghao has quit IRC03:55
*** wanghao has joined #openstack-cinder03:56
*** wanghao has quit IRC03:56
*** wanghao has joined #openstack-cinder03:57
*** wanghao has quit IRC03:57
*** wanghao has joined #openstack-cinder03:57
*** wanghao has quit IRC03:58
*** dave-mcc_ has quit IRC04:00
*** bhagyashris has joined #openstack-cinder04:10
*** lpetrut has quit IRC04:11
*** lpetrut has joined #openstack-cinder04:22
*** pgadiya has joined #openstack-cinder04:23
*** pgadiya has quit IRC04:23
*** pgadiya has joined #openstack-cinder04:26
*** pgadiya has quit IRC04:26
*** abhishekk has joined #openstack-cinder04:46
*** lpetrut has quit IRC04:46
*** pooja_jadhav has joined #openstack-cinder04:47
*** armaan has quit IRC04:54
*** armaan has joined #openstack-cinder04:54
*** moshele has quit IRC04:55
*** psachin has joined #openstack-cinder04:59
*** edmondsw has quit IRC05:03
*** wanghao has joined #openstack-cinder05:03
*** edmondsw has joined #openstack-cinder05:03
*** wanghao has quit IRC05:04
*** wanghao has joined #openstack-cinder05:04
*** edmondsw has quit IRC05:08
*** wanghao has quit IRC05:08
*** wanghao has joined #openstack-cinder05:09
*** wanghao has quit IRC05:14
*** wanghao has joined #openstack-cinder05:15
*** wanghao has quit IRC05:24
*** wanghao has joined #openstack-cinder05:24
*** vivsoni__ has quit IRC05:27
*** vivsoni has joined #openstack-cinder05:29
*** wanghao has quit IRC05:29
*** wanghao has joined #openstack-cinder05:30
*** vivsoni has quit IRC05:31
*** wanghao has quit IRC05:33
*** wanghao has joined #openstack-cinder05:33
*** lkuchlan has joined #openstack-cinder05:38
*** wanghao has quit IRC05:40
*** wanghao has joined #openstack-cinder05:41
*** stakeda has joined #openstack-cinder05:43
*** wanghao has quit IRC05:44
*** wanghao has joined #openstack-cinder05:44
*** wanghao has quit IRC05:48
*** wanghao has joined #openstack-cinder05:49
*** moshele has joined #openstack-cinder05:50
*** wanghao has quit IRC05:52
*** wanghao has joined #openstack-cinder05:53
*** wanghao has quit IRC05:59
*** wanghao has joined #openstack-cinder06:00
*** e0ne has joined #openstack-cinder06:02
*** wanghao has quit IRC06:03
*** wanghao has joined #openstack-cinder06:04
*** wanghao has quit IRC06:10
*** wanghao has joined #openstack-cinder06:10
*** sar has joined #openstack-cinder06:12
*** links has joined #openstack-cinder06:13
*** wanghao has quit IRC06:15
*** wanghao has joined #openstack-cinder06:16
*** lpetrut has joined #openstack-cinder06:21
*** wanghao has quit IRC06:21
*** wanghao has joined #openstack-cinder06:22
*** wanghao has quit IRC06:27
*** wanghao has joined #openstack-cinder06:27
*** kaisers1 has quit IRC06:31
*** wanghao has quit IRC06:31
*** wanghao has joined #openstack-cinder06:32
*** kaisers has joined #openstack-cinder06:33
*** wanghao has quit IRC06:36
*** wanghao has joined #openstack-cinder06:37
*** ircuser-1 has joined #openstack-cinder06:39
*** wanghao has quit IRC06:41
*** wanghao has joined #openstack-cinder06:42
*** wanghao has quit IRC06:46
openstackgerritChris Yang proposed openstack/cinder master: QNAP: driver fails to create volume and snapshot in some case  https://review.openstack.org/56349706:47
*** e0ne has quit IRC06:49
*** wanghao has joined #openstack-cinder06:50
*** lkuchlan has quit IRC06:50
*** lkuchlan has joined #openstack-cinder06:51
*** wanghao has quit IRC06:52
*** wanghao has joined #openstack-cinder06:53
*** wanghao has quit IRC06:58
*** wanghao has joined #openstack-cinder06:58
*** wanghao has quit IRC07:02
*** wanghao has joined #openstack-cinder07:03
*** tesseract has joined #openstack-cinder07:04
*** kaisers_ has quit IRC07:05
*** wanghao has quit IRC07:05
*** wanghao has joined #openstack-cinder07:06
*** hamdyk has joined #openstack-cinder07:06
*** wanghao has quit IRC07:11
*** wanghao has joined #openstack-cinder07:11
*** sapd has quit IRC07:13
*** wanghao has quit IRC07:16
*** wanghao has joined #openstack-cinder07:16
*** rcernin has quit IRC07:19
*** dpawlik has joined #openstack-cinder07:19
*** wanghao has quit IRC07:20
*** wanghao has joined #openstack-cinder07:21
*** dpawlik has quit IRC07:22
*** dpawlik has joined #openstack-cinder07:22
*** pcaruana has joined #openstack-cinder07:25
*** wanghao has quit IRC07:25
*** wanghao has joined #openstack-cinder07:26
*** wanghao has quit IRC07:29
*** wanghao has joined #openstack-cinder07:29
*** wanghao has quit IRC07:36
*** AlexeyAbashkin has joined #openstack-cinder07:36
*** lhx_ has joined #openstack-cinder07:37
*** wanghao has joined #openstack-cinder07:37
*** wanghao has quit IRC07:42
*** wanghao has joined #openstack-cinder07:42
*** wanghao has quit IRC07:47
*** wanghao has joined #openstack-cinder07:48
*** wanghao has quit IRC07:51
*** wanghao has joined #openstack-cinder07:52
*** alexchadin has joined #openstack-cinder07:56
*** wanghao has quit IRC07:56
*** wanghao has joined #openstack-cinder07:58
*** Dinesh__Bhor has joined #openstack-cinder08:03
*** Dinesh_Bhor has quit IRC08:05
*** threestrands has quit IRC08:17
*** wanghao has quit IRC08:19
*** wanghao has joined #openstack-cinder08:20
*** finucannot is now known as stephenfin08:20
*** armaan has quit IRC08:24
*** wanghao has quit IRC08:25
*** armaan has joined #openstack-cinder08:25
*** wanghao has joined #openstack-cinder08:25
*** wanghao_ has joined #openstack-cinder08:31
*** wanghao has quit IRC08:34
*** wanghao_ has quit IRC08:34
*** wanghao has joined #openstack-cinder08:34
*** wanghao has quit IRC08:39
*** wanghao has joined #openstack-cinder08:40
openstackgerritChris Yang proposed openstack/cinder master: QNAP: driver fails to create volume and snapshot in some case  https://review.openstack.org/56349708:43
*** wanghao has quit IRC08:44
*** wanghao has joined #openstack-cinder08:44
*** kaisers_ has joined #openstack-cinder08:45
*** wanghao has quit IRC08:48
*** Alexey_Abashkin has joined #openstack-cinder08:48
*** wanghao has joined #openstack-cinder08:49
*** Dinesh__Bhor has quit IRC08:49
*** Alexey_Abashkin has quit IRC08:50
*** Alexey_Abashkin has joined #openstack-cinder08:50
*** wanghao has quit IRC08:50
*** wanghao has joined #openstack-cinder08:51
*** AlexeyAbashkin has quit IRC08:51
*** sapd has joined #openstack-cinder08:53
*** Dinesh_Bhor has joined #openstack-cinder08:53
*** AlexeyAbashkin has joined #openstack-cinder08:54
*** Alexey_Abashkin has quit IRC08:54
*** sapd1 has joined #openstack-cinder08:56
*** wanghao has quit IRC08:58
openstackgerritTommyLike proposed openstack/cinder master: Collect resource id from RequestSpec parameter  https://review.openstack.org/56340508:58
*** wanghao has joined #openstack-cinder08:58
*** Alexey_Abashkin has joined #openstack-cinder09:02
*** wanghao has quit IRC09:02
*** e0ne has joined #openstack-cinder09:03
*** wanghao has joined #openstack-cinder09:03
*** dalgaaf has joined #openstack-cinder09:04
*** AlexeyAbashkin has quit IRC09:05
*** Alexey_Abashkin is now known as AlexeyAbashkin09:05
*** wanghao has quit IRC09:06
*** wanghao has joined #openstack-cinder09:06
*** dalgaaf has quit IRC09:09
*** alexchad_ has joined #openstack-cinder09:10
*** wanghao has quit IRC09:10
*** wanghao has joined #openstack-cinder09:11
*** alexchadin has quit IRC09:12
*** armaan has quit IRC09:13
*** dalgaaf has joined #openstack-cinder09:14
*** wanghao has quit IRC09:14
*** armaan has joined #openstack-cinder09:14
*** wanghao has joined #openstack-cinder09:14
*** wanghao has quit IRC09:19
*** wanghao has joined #openstack-cinder09:20
*** namnh has quit IRC09:22
*** namnh has joined #openstack-cinder09:23
*** wanghao has quit IRC09:24
*** wanghao has joined #openstack-cinder09:24
*** bkopilov has quit IRC09:27
*** liuzz has quit IRC09:29
*** wanghao has quit IRC09:29
*** wanghao has joined #openstack-cinder09:30
*** wanghao has quit IRC09:35
*** wanghao has joined #openstack-cinder09:36
*** Dinesh_Bhor has quit IRC09:37
*** liuzz has joined #openstack-cinder09:38
*** wanghao has quit IRC09:38
*** wanghao has joined #openstack-cinder09:38
*** liuzz_ has joined #openstack-cinder09:40
*** wanghao has quit IRC09:42
*** liuzz has quit IRC09:43
*** wanghao has joined #openstack-cinder09:43
*** wanghao has quit IRC09:49
*** wanghao has joined #openstack-cinder09:50
*** wanghao has quit IRC09:53
*** lhx_ has quit IRC09:54
*** lhx_ has joined #openstack-cinder09:55
*** wanghao has joined #openstack-cinder09:56
openstackgerritKumar Prashant proposed openstack/cinder master: VMAX Driver - Snapshot Time to Live  https://review.openstack.org/56355509:58
*** wanghao has quit IRC10:00
*** hamdykhader has joined #openstack-cinder10:01
*** hamdyk has quit IRC10:02
*** nicolasbock has joined #openstack-cinder10:09
*** vivsoni has joined #openstack-cinder10:16
vivsonihi10:16
vivsonigetting 2018-04-23 11:44:42.285 519110 ERROR oslo.service.wsgi [req-e7c428b6-21bb-4f27-9421-d5e96717fea0 - - - - -] Could not bind to 0.0.0.0:877610:16
vivsoni2018-04-23 11:44:42.286 519110 CRITICAL cinder [req-e7c428b6-21bb-4f27-9421-d5e96717fea0 - - - - -] error: [Errno 98] Address already in use10:16
vivsoniwhile starting 'cinder api' service10:17
*** namnh has quit IRC10:17
*** armaan has quit IRC10:17
openstackgerritKumar Prashant proposed openstack/cinder master: VMAX Driver - Snapshot Time to Live  https://review.openstack.org/56355510:17
*** armaan has joined #openstack-cinder10:18
*** armaan has quit IRC10:22
*** pcaruana has quit IRC10:26
*** lkuchlan has quit IRC10:32
*** pcaruana has joined #openstack-cinder10:33
*** armaan has joined #openstack-cinder10:35
*** e0ne has quit IRC10:35
*** e0ne_ has joined #openstack-cinder10:35
*** nicolasbock has quit IRC10:39
*** abhishekk has quit IRC10:41
*** dalgaaf has quit IRC10:45
*** AlexeyAbashkin has quit IRC10:47
*** nicolasbock has joined #openstack-cinder10:50
*** alexchad_ has quit IRC10:50
*** AlexeyAbashkin has joined #openstack-cinder10:51
*** sapd1 has quit IRC10:54
*** wanghao has joined #openstack-cinder11:03
*** hoonetorg has quit IRC11:03
*** e0ne_ is now known as e0ne11:07
*** wanghao has quit IRC11:08
*** yikun has quit IRC11:10
*** yikun_ has joined #openstack-cinder11:10
*** stakeda has quit IRC11:10
e0nevivsoni: it means that you already have cinder-api running11:11
openstackgerritMargarita Shakhova proposed openstack/cinder master: Storwize: get list of all volumes for ensure_export.  https://review.openstack.org/54648911:12
vivsonie0ne: there are 3 controller node.. cinder api was not running on any node11:13
*** yikun_ has quit IRC11:13
vivsonithen, i tried to start the cinder api and ends up with that error11:13
*** yikun_ has joined #openstack-cinder11:13
*** hoonetorg has joined #openstack-cinder11:16
*** yikun_ has quit IRC11:20
*** hoonetorg has quit IRC11:21
*** tpsilva has joined #openstack-cinder11:26
*** wanghao has joined #openstack-cinder11:28
*** hoonetorg has joined #openstack-cinder11:39
*** wanghao has quit IRC11:41
*** wanghao has joined #openstack-cinder11:42
*** deng_ has quit IRC11:43
*** edmondsw has joined #openstack-cinder11:47
*** abishop has joined #openstack-cinder11:49
*** ganso has joined #openstack-cinder11:50
*** pcaruana has quit IRC11:54
*** dalgaaf has joined #openstack-cinder12:07
*** dalgaaf has quit IRC12:08
*** dalgaaf has joined #openstack-cinder12:09
*** pcaruana has joined #openstack-cinder12:10
*** liverpooler has joined #openstack-cinder12:16
*** pchavva has joined #openstack-cinder12:19
*** lkuchlan has joined #openstack-cinder12:24
*** zhaochao has quit IRC12:29
*** zhaochao has joined #openstack-cinder12:30
*** edmondsw has quit IRC12:37
*** wanghao_ has joined #openstack-cinder12:40
*** wanghao has quit IRC12:44
*** lseki has joined #openstack-cinder12:47
*** dustins has joined #openstack-cinder12:49
*** dave-mccowan has joined #openstack-cinder12:49
lsekixyang: hi, could you +w please? https://review.openstack.org/#/c/561367/13:03
*** eharney has joined #openstack-cinder13:08
*** sar has quit IRC13:10
*** wanghao_ has quit IRC13:11
*** diablo_rojo has joined #openstack-cinder13:12
*** moshele has quit IRC13:16
*** psachin has quit IRC13:17
*** wanghao has joined #openstack-cinder13:17
*** wanghao has quit IRC13:22
*** diablo_rojo has quit IRC13:29
*** lvdombrkr has joined #openstack-cinder13:30
*** r-daneel has joined #openstack-cinder13:37
*** felipemonteiro_ has joined #openstack-cinder13:38
*** KeithMnemonic has joined #openstack-cinder13:43
*** felipemonteiro__ has joined #openstack-cinder13:44
*** wanghao has joined #openstack-cinder13:44
*** felipemonteiro_ has quit IRC13:47
*** wanghao has quit IRC13:48
*** r-daneel has quit IRC13:51
openstackgerritSean McGinnis proposed openstack/cinder stable/ocata: OVO: Fix VolumeType create not bound to a Session  https://review.openstack.org/56362813:51
*** liuzz_ has quit IRC13:54
*** liuzz has joined #openstack-cinder13:55
*** edmondsw has joined #openstack-cinder13:58
*** wanghao has joined #openstack-cinder13:59
*** felipemonteiro_ has joined #openstack-cinder14:00
*** felipemonteiro__ has quit IRC14:00
*** gnufied has joined #openstack-cinder14:04
*** mriedem has joined #openstack-cinder14:08
*** kaisers_ has quit IRC14:11
*** felipemonteiro__ has joined #openstack-cinder14:20
*** liuzz has quit IRC14:23
*** felipemonteiro_ has quit IRC14:24
*** liuzz has joined #openstack-cinder14:24
*** links has quit IRC14:30
*** lpetrut has quit IRC14:31
xyanglseki: done14:34
*** namnh has joined #openstack-cinder14:34
*** lseki_ has joined #openstack-cinder14:35
lseki_xyang: thanks!14:35
*** arnewiebalck_ has joined #openstack-cinder14:36
*** lseki has quit IRC14:37
*** lvdombrkr89 has joined #openstack-cinder14:37
*** bkopilov has joined #openstack-cinder14:38
*** hemna_ has joined #openstack-cinder14:39
arnewiebalck_jungleboyj geguileo eharney jbernard smcginnis As suggested in the Cinder meeting some weeks ago, I now put up a spec for deferred deletion in rbd leveraging the rbd trash commands in Ceph Luminous: https://review.openstack.org/#/c/562311/14:39
arnewiebalck_I’ll put it also on the agenda for this week.14:39
*** lvdombrkr has quit IRC14:40
jungleboyjarnewiebalck_:  Cool.  Thanks.  Will try to take a look before the meeting.14:41
geguileoarnewiebalck_: didn't the issue that this spec tried to resolve get fixed with the tpool configuration option?14:41
arnewiebalck_geguileo: adding that option definitly helps, yes14:43
arnewiebalck_geguileo: it’s mentioned in the spec14:43
geguileoarnewiebalck_: I think this specs introduces some potentially big issues...14:45
arnewiebalck_geguileo: like … ?14:46
geguileobecause like you mention in your spec Cinder expects a specific behavior14:46
geguileoa delete request means that on completion:14:46
geguileo- Space will be available14:46
geguileo- No leftovers will be left on the backend referring to this volume once we remove the DB entry14:47
*** lvdombrkr89 has quit IRC14:47
geguileo- User used quota = provisioned space in backend14:47
arnewiebalck_geguileo: space: if you run that close to the limit, there is an issue anyway14:48
geguileoIt's not running that close to the limit14:48
arnewiebalck_geguileo: leftovers: how would Cinder be affected?14:48
arnewiebalck_geguileo: quota: again, that’s for the admin to decide14:49
geguileoWhat happens if you request to remove a lot of volumes all within one minute or so?14:49
arnewiebalck_geguileo: in the end, this is all optional14:49
arnewiebalck_geguileo: they would be moved to the trash14:49
geguileoyeah, but considering this optional behavior will go against how any other backend works means that users will NOT use it properly or understand it properly14:49
geguileoarnewiebalck_: but the space would not be freed14:50
arnewiebalck_geguileo: correct14:50
arnewiebalck_for the space14:50
arnewiebalck_not for the correct usage :)14:50
geguileoand that means that you won't be able to use that space14:50
*** Alexey_Abashkin has joined #openstack-cinder14:50
arnewiebalck_that’s true14:50
geguileoand that's the problem14:50
arnewiebalck_again: I don’t think anyone runs that close to the limit14:50
arnewiebalck_and if they do, they should not use this option14:51
geguileosuuuuuuuuure14:51
*** felipemonteiro__ has quit IRC14:51
geguileobecause that's how people work   ;-)14:51
geguileoThere are people sharing the same Ceph pool for Cinder and other workloads14:51
*** felipemonteiro__ has joined #openstack-cinder14:51
geguileoConsidering that, it is likely that people will use this without fully understanding it and then they'll run into a lot of troubles14:52
arnewiebalck_if they don’t understand this, they won’t understand other things either14:52
arnewiebalck_and run into trouble .)14:52
*** kaisers_ has joined #openstack-cinder14:53
geguileonot really, because this will be different than any other driver14:53
*** gcb has joined #openstack-cinder14:53
geguileoDon't get me wrong, I understand the feature and your use case14:53
arnewiebalck_I don’t think that people know for instance that the parent volume of clones will be kept if you delete it14:53
arnewiebalck_and take up space14:53
*** r-daneel has joined #openstack-cinder14:53
arnewiebalck_until the child is deleted14:54
*** zhaochao has quit IRC14:54
arnewiebalck_I ack there may be implications.14:54
geguileoMaybe not, but that's usually of less significance of deleting hundreds of volumes14:54
geguileoand not having that space available14:54
arnewiebalck_We had a case this week where it would have been very handy to be able to bring the volume back.14:54
*** AlexeyAbashkin has quit IRC14:54
*** Alexey_Abashkin is now known as AlexeyAbashkin14:54
geguileoAnyway, I won't oppose the feature, since you have a valid use case14:55
arnewiebalck_agreed14:55
geguileoI'm just not too convinced about it...14:55
arnewiebalck_Got it :-D14:55
geguileoThe spec looks good otherwise (except for a minor nit I'll comment in the spec)14:55
arnewiebalck_I appreciate you express your concerns and scrutinize the proposal.14:55
arnewiebalck_What about the regular purge?14:56
arnewiebalck_We’re in touch with the Ceph team to see if Ceph could clean up itself in case the trash gets too full.14:57
arnewiebalck_So, you’d have some high water mark and if more space is requested, the trash gets emptied.14:57
arnewiebalck_No promises though Ceph will accept this :)14:58
geguileoarnewiebalck_: even with that it wouldn't help that much, because in the meantime the free space provided by the stat collection would still report the actual free space and that would be used in the scheduler14:59
geguileobut that feature would be cool14:59
arnewiebalck_we have disabled the stat collection :-D15:00
arnewiebalck_I agree.15:00
geguileoalthough I don't think they'll accept it... because you may be using the trash with expiration dates with the idea of being able to recover the data if necessary15:00
arnewiebalck_actually, there is already a proposal how to do this :)15:00
geguileoI added a feature to partially disable the stat collection15:00
arnewiebalck_I know, we’re using this now instead.15:01
geguileoglad it's of use  :-)15:01
arnewiebalck_oh, it’s required :)15:01
*** guitarzan has quit IRC15:01
geguileoI can imagine given the amount of volumes you have   XD15:02
arnewiebalck_5680 right now15:03
arnewiebalck_for the spec, let’s re-discuss on wed, see what the others say15:03
geguileook15:03
*** arnewiebalck_ has quit IRC15:07
*** rajinir has joined #openstack-cinder15:08
*** felipemonteiro_ has joined #openstack-cinder15:09
*** arnewiebalck_ has joined #openstack-cinder15:10
*** hamdykhader has quit IRC15:10
arnewiebalck_geguileo: about the stats: how does that actually work for shared pools?15:11
geguileoyou mean on a pool that is shared between Cinder and something else?15:12
arnewiebalck_geguileo: Cinder does a df on the pool and compares with the size of all volumes?15:12
arnewiebalck_yes15:12
geguileoIn that case Cinder has to actually calculate the provisioned size15:12
geguileoso it can do the calculation with the used GB15:12
*** felipemonteiro__ has quit IRC15:13
arnewiebalck_provisioned means actually used?15:13
geguileowhen it's not shared it can just do a simple DB query to know the provisioned size15:13
geguileono, it's the virtual size15:13
arnewiebalck_ok15:13
geguileoand in the RBD case the DB query is a lot faster than querying the volumes15:13
arnewiebalck_right15:13
arnewiebalck_and Cinder queries the pool size I imagine?15:13
geguileobecause you have to N+1 queries to calculate it (where N is the number fo volumes)15:13
geguileoyup, Cinder queries the free size15:14
geguileoand the pool size15:14
geguileoand the quota size for the pool15:14
geguileoif there's no quota it uses the pool size as the total size15:14
geguileoif there's a quota it uses that as the total15:14
arnewiebalck_is there over-provisioning?15:14
arnewiebalck_or does Cinder refuse to create in that case?15:15
*** markstur has joined #openstack-cinder15:15
geguileothe over provisioning is calculated by the scheduler based on those parameters and the max over provisioning configured in the cinder volume service15:16
arnewiebalck_ok15:16
geguileothe default is 20 times15:16
geguileoalthough we now have the option of setting it to auto15:16
geguileowhich will adjust the over provisioning based on current ratio15:17
geguileosince this was something that was very hard to adjust when deploying and could change depending on the use15:17
arnewiebalck_ok, so by default, in a full pool if all volumes claimed their space, it would be 20 times too small?15:17
geguileoyup15:18
arnewiebalck_that does not seem to cause an issue15:18
geguileobut you can set the over provisioning ration to 1.015:18
geguileoI have never heard anybody having an issue using Ceph15:18
arnewiebalck_it’s thinly provisioned15:19
geguileobecause cloned volumes don't take up the space15:19
geguileoyup15:19
arnewiebalck_still15:19
arnewiebalck_it supports my point that operators work with some headroom15:19
*** pcaruana has quit IRC15:19
geguileobut the scenario you are talking about (deleting hundreds of big volumes) changes things  ;-)15:20
arnewiebalck_yeah, I guess15:20
*** felipemonteiro_ has quit IRC15:20
*** felipemonteiro_ has joined #openstack-cinder15:20
arnewiebalck_but our Ceph backend is also pretty large (>5PB)15:21
geguileoand it seems to be growing every time I attend a CERN talk!!!   :-)15:22
geguileoso cool!!!15:22
arnewiebalck_Cinder is really popular!15:22
*** lkuchlan has quit IRC15:22
arnewiebalck_ok, I got to go … thanks for your feedback!15:23
geguileoarnewiebalck_: np15:23
*** arnewiebalck_ has quit IRC15:24
*** sar has joined #openstack-cinder15:32
*** dklyle has joined #openstack-cinder15:32
*** lpetrut has joined #openstack-cinder15:37
*** moshele has joined #openstack-cinder15:38
*** moshele has quit IRC15:40
*** Swanson has joined #openstack-cinder15:43
*** abhishekk has joined #openstack-cinder15:44
zigoHi there! Trying to add Debian to puppet-openstack, I noticed an issue when calling openstack volume:15:45
zigohttp://paste.openstack.org/show/719756/15:45
zigoAny idea what's going on?15:45
openstackgerritTom Swanson proposed openstack/cinder master: Dell EMC SC: Removed space from debug statement  https://review.openstack.org/56367815:48
*** dklyle has quit IRC15:49
smcginniszigo: That doesn't make any sense to me. Maybe a cliff issue?15:49
zigosmcginnis: How can I investigate this further?15:50
smcginniszigo: Not really sure. Maybe check with the cliff team to see if they recognize that kind of error?15:50
zigodhellmann: ^15:51
zigoWho else in the cliff team?15:51
* smcginnis didn't know Doug was involved in cliff15:51
smcginnisNot sure15:51
zigoDean Troyer.15:52
smcginnisdtroyer15:52
smcginnisHah, yep15:52
zigoI believe Doug is the main author of cliff.15:52
dhellmannthat's not a cliff error, it's coming from the resource object returned by the volume command and it's triggered in some osc-lib code15:53
dhellmannmy guess is the resource object isn't the sort of thing osc-lib expects to see15:53
dhellmannzigo : I would modify osc_lib/utils/__init__.py to print out what resource is (or use pdb to look)15:54
dhellmannperhaps there's an incompatible version of something in there15:54
*** e0ne has quit IRC15:54
Swansonsmcginnis, you can workflow my patch without waiting for gate and or CI, I think.15:55
zigodhellmann: If I print(resource), I get:15:56
zigo<VolumeType: BACKEND_1>15:56
smcginnisSwanson: Was just doing that.15:56
zigodhellmann: What is resource.get() supposed to do?15:57
dhellmannzigo : it looks like it's supposed to be a way to get properties of the thing15:57
dhellmannzigo : what command are you actually running there?15:58
zigodhellmann: openstack volume type set --property volume_backend_name=BACKEND_1 BACKEND_115:58
zigoIt just spits: "get() takes 1 positional argument but 2 were given" as return ... :(15:58
Swansonsmcginnis, thanks! Irritation patch.16:01
dhellmannzigo : I don't know anything about how the volume commands work, so I'm not sure where that VolumeType class is defined16:02
zigo:(16:02
*** gcb has quit IRC16:03
imacdonnit's defined in https://github.com/openstack/python-cinderclient/blob/master/cinderclient/v2/volume_types.py16:03
*** armaan has quit IRC16:03
*** namnh has quit IRC16:04
*** armaan has joined #openstack-cinder16:04
smcginnisWould be good to know what object it is working on here. If it's a VolumeTypeManager, this may be the cause: https://github.com/openstack/python-cinderclient/blob/master/cinderclient/v2/volume_types.py#L9316:05
smcginnisBut I find it odd it says two arguments are given.16:05
dhellmannself is an argument16:05
dhellmannso we're looking for a get(self) method16:06
dhellmannwhich is likely to be in a model where "get" means "fetch from the remote source" rather than "act like a dict"16:06
smcginnisOh, I suppose that is an instance there.16:06
*** dpawlik has quit IRC16:06
imacdonnI think the actual get() method is inherited from https://github.com/openstack/python-cinderclient/blob/master/cinderclient/apiclient/base.py#L52016:06
dhellmannzigo : can you print out the __module__ property of the resource?16:07
dhellmannprint(resource.__module__)16:07
dhellmannor print(resource.__class__.__module__) more likely16:07
dhellmannimacdonn : that looks like a good candidate16:07
zigoSure, doing that.16:08
dhellmannimacdonn : although dtroyer has been working to remove the use of the client libraries from OSC16:08
dhellmannzigo : it would also be good to know if other volume type commands work properly16:08
dhellmann(list, etc.)16:08
*** armaan has quit IRC16:09
zigoThe value for resource.__class__.__module is: cinderclient.v2.volume_types16:09
*** armaan has joined #openstack-cinder16:09
dhellmannok, so it looks like imacdonn is correct16:09
imacdonnI wonder if this is code that rarely gets executed ..and has a latent bug ....16:10
dhellmannand the bug is in the "volume type set" command implementation16:10
dhellmannyeah, that seems likely16:10
zigodhellmann: "openstack volume type list" works perfectly and returns BACKEND_1, which is what puppet-openstack setups as LVM backend.16:10
dhellmannzigo : how about volume type show?16:10
openstackgerritMerged openstack/cinder master: NetApp ONTAP: Set new sub-lun clone limit for ONTAP driver  https://review.openstack.org/56136716:10
dhellmannor get or whatever they ended up calling it where you get the details of 1 thing16:10
zigodhellmann: Works as well.16:11
zigoopenstack volume type show 1d8667bc-d0cd-461f-aa66-8448d1747e1a16:11
dhellmannok, so it sounds like its just the set command then16:11
dhellmannI don't know the right way to fix it, but I'll bet looking at some of the other volume commands would give someone a hint16:12
dhellmannand zigo you could file the bug with all of the details from what we figured out16:12
zigodhellmann: Sure.16:12
zigodhellmann: You've noticed I'm using Py3 only, right?16:12
zigoNot sure if that is important ...16:12
dhellmannI did see that. I'll be interested to hear how it works out.16:13
*** pcaruana has joined #openstack-cinder16:13
dhellmannI expect it to mostly work, but I'm sure we'll find bugs.16:13
zigodhellmann: I think I failed to past the begining, which is a 404: http://paste.openstack.org/show/719758/16:18
zigoWeird that its saying 404 when searching for volume type BACKEND_1, no?16:18
zigoI mean, when I do "openstack volume type list" it really is listed ...16:19
dhellmannyou're well beyond my understanding of what the client is doing there, so I don't know if it's weird or not16:19
zigoIndeed.16:19
imacdonnzigo: It'll take what you provide and try to query for that as an ID... if that finds nothing, it'll try to find the resource by name .. at least that's how it's supposed to work ... so the 404, in itself, is OK16:20
imacdonnzigo: you could try your set command with the ID instead of name16:21
*** lhx_ has quit IRC16:24
*** edmondsw has quit IRC16:29
*** Alexey_Abashkin has joined #openstack-cinder16:31
*** AlexeyAbashkin has quit IRC16:34
*** Alexey_Abashkin is now known as AlexeyAbashkin16:34
*** dpawlik has joined #openstack-cinder16:46
*** dpawlik has quit IRC16:50
openstackgerritJon Bernard proposed openstack/cinder master: Add support for promoting a failed over backend  https://review.openstack.org/55889916:58
*** AlexeyAbashkin has quit IRC16:59
*** sar has quit IRC17:05
*** mvk has quit IRC17:22
*** tesseract has quit IRC17:23
*** e0ne has joined #openstack-cinder17:26
*** edmondsw has joined #openstack-cinder17:34
*** edmondsw has quit IRC17:37
*** edmondsw has joined #openstack-cinder17:37
*** lkuchlan has joined #openstack-cinder17:43
*** diablo_rojo has joined #openstack-cinder17:48
*** dpawlik has joined #openstack-cinder17:58
*** r-daneel has quit IRC18:00
*** r-daneel has joined #openstack-cinder18:01
*** dpawlik has quit IRC18:02
*** felipemonteiro__ has joined #openstack-cinder18:08
*** abhishekk has quit IRC18:11
*** felipemonteiro_ has quit IRC18:12
*** lkuchlan has quit IRC18:25
*** jhawk has joined #openstack-cinder18:31
*** jhawkinson has quit IRC18:33
*** jhawkinson has joined #openstack-cinder18:33
*** jhawk has quit IRC18:36
*** harlowja has joined #openstack-cinder18:44
openstackgerritPavlo Shchelokovskyy proposed openstack/cinder master: RBD: Handle ImageNotFound exception in _get_usage_info correctly  https://review.openstack.org/56331318:52
*** sar has joined #openstack-cinder18:53
openstackgerritJon Bernard proposed openstack/cinder master: RBD: add support for active/active replication  https://review.openstack.org/55665818:58
openstackgerritJon Bernard proposed openstack/cinder master: Add support for promoting a failed over backend  https://review.openstack.org/55889918:58
*** felipemonteiro__ has quit IRC19:06
*** felipemonteiro__ has joined #openstack-cinder19:07
*** lkuchlan has joined #openstack-cinder19:07
*** harlowja has quit IRC19:07
*** rajinir has quit IRC19:08
*** dklyle has joined #openstack-cinder19:11
*** lkuchlan has quit IRC19:11
*** imacdonn has quit IRC19:13
*** imacdonn has joined #openstack-cinder19:14
*** lpetrut has quit IRC19:41
openstackgerritMerged openstack/cinder master: Dell EMC SC: Removed space from debug statement  https://review.openstack.org/56367819:52
*** lkwan has joined #openstack-cinder19:57
zigodhellmann: imacdonn: It doesn't look like a client issue to me, I was using uwsgi, when I switch back to eventlet, then the command works.20:03
dhellmannzigo : *that* sounds like a client issue, if the client behaves differently in those cases20:03
zigodhellmann: When I switch stuff to uwsgi, I often get a 104, connection reset by peer.20:04
*** dpawlik has joined #openstack-cinder20:04
zigodhellmann: Would it be possible that somehow, there's issues with content-lenght ?20:04
zigolength20:04
dhellmannI have no idea :-(20:04
zigo:)20:04
zigodhellmann: At least, now I have the workaround to continue using eventlet, even though that's not ideal.20:05
*** dpawlik has quit IRC20:05
*** liverpooler has quit IRC20:05
*** dpawlik has joined #openstack-cinder20:05
imacdonnzigo dhellmann I still suspect that it's client code that only gets used in some corner case .. and using uwsgi is causing the corner case.... but haven't tried to prove it20:06
zigoimacdonn: I can provide some env with uwsgi if needed.20:06
imacdonnI've not looked at running cinder under uwsgi yet20:07
*** lkuchlan has joined #openstack-cinder20:09
*** dave-mccowan has quit IRC20:14
*** armaan has quit IRC20:14
*** lkuchlan has quit IRC20:14
SwansonGood luck to Cinder's Kevin Phillips-Bong candidate for TC, smcginnis.20:26
*** sar has quit IRC20:26
smcginnisNew project idea - a snopes.com like site to be able to look up all of Swanson's obscure references.20:27
smcginnisBut I will be sure to wear my slightly silly pants.20:28
Swanson:)20:29
*** felipemonteiro__ has quit IRC20:29
*** dpawlik has quit IRC20:34
*** dpawlik has joined #openstack-cinder20:37
*** rajinir has joined #openstack-cinder20:45
*** mvk has joined #openstack-cinder20:45
*** dpawlik has quit IRC20:49
*** hemna_ has quit IRC20:49
*** dpawlik has joined #openstack-cinder20:50
*** hemna_ has joined #openstack-cinder20:53
*** idlemind has quit IRC20:59
jungleboyjsmcginnis:  ++21:02
*** pchavva has quit IRC21:06
*** abishop has quit IRC21:21
*** dklyle has quit IRC21:31
*** r-daneel_ has joined #openstack-cinder21:35
*** dustins has quit IRC21:35
*** r-daneel has quit IRC21:36
*** r-daneel_ is now known as r-daneel21:36
_alastor_smcginnis jungleboyj: I'm implementing the snapshot manage feature and I need to store information about the parent volume.  Can I use a field on the Snapshot object to do that?  Maybe provider_location?21:40
*** eharney has quit IRC21:42
jungleboyj_alastor_:  If you are implementing a new feature is there a reason to not add a field for it?21:43
_alastor_jungleboyj: I mean specifically in my driver.  The feature already exists in Cinder21:43
_alastor_jungleboyj: I'm just looking for a place to store info about the managed snapshot's parent volume21:43
jungleboyjOh ... hmmm, good question.21:44
smcginnis_alastor_: Yeah, I think others are using provider_location for that type of metadata IIRC.21:44
jungleboyj_alastor_:  I am not the best person to ask as I haven't implemented something like that but I think that approach sounds familiar.21:45
jungleboyjsmcginnis:  :-)21:45
_alastor_smcginnis: Ok cool.  I just wanted to make sure I wasn't stepping on existing functionality for that field.21:45
smcginnisNope, that field is kind of for the driver implementors discretion, so should be good.21:46
_alastor_smcginnis: Thanks21:46
*** dpawlik has quit IRC21:47
*** diablo_rojo has quit IRC21:51
*** diablo_rojo has joined #openstack-cinder21:57
*** e0ne has quit IRC22:14
*** ganso has quit IRC22:17
*** lbragstad has quit IRC22:18
*** dklyle has joined #openstack-cinder22:26
*** rcernin has joined #openstack-cinder22:31
*** dklyle has quit IRC22:39
*** threestrands has joined #openstack-cinder22:41
*** threestrands has quit IRC22:41
*** threestrands has joined #openstack-cinder22:41
*** felipemonteiro__ has joined #openstack-cinder22:42
*** edmondsw has quit IRC22:43
*** lseki_ has quit IRC22:49
*** tpsilva has quit IRC23:05
*** felipemonteiro__ has quit IRC23:07
*** diablo_rojo_ has joined #openstack-cinder23:09
*** diablo_rojo has quit IRC23:12
*** _alastor_ has quit IRC23:15
*** r-daneel has quit IRC23:21
*** _alastor_ has joined #openstack-cinder23:43
*** _alastor_ has quit IRC23:48
*** wanghao has quit IRC23:59

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!