Tuesday, 2018-01-30

*** sapd_ has joined #openstack-cinder00:00
*** _pewp_ has joined #openstack-cinder00:01
*** namnh has quit IRC00:03
*** namnh has joined #openstack-cinder00:07
*** wanghao_ has joined #openstack-cinder00:08
*** salv-orlando has joined #openstack-cinder00:09
*** wanghao has quit IRC00:10
*** salv-orlando has quit IRC00:13
*** namnh has quit IRC00:19
*** namnh has joined #openstack-cinder00:20
*** namnh has quit IRC00:24
*** liverpooler has quit IRC00:26
*** namnh has joined #openstack-cinder00:28
*** namnh_ has joined #openstack-cinder00:32
*** felipemonteiro_ has quit IRC00:32
*** namnh_ has quit IRC00:32
*** dave-mccowan has quit IRC00:32
*** namnh has quit IRC00:33
*** wanghao_ has quit IRC00:44
*** wanghao has joined #openstack-cinder00:44
*** wanghao has quit IRC00:45
*** wanghao has joined #openstack-cinder00:46
*** yangyapeng has joined #openstack-cinder00:51
*** yangyapeng has quit IRC00:55
*** wanghao_ has joined #openstack-cinder00:57
*** wanghao has quit IRC00:57
*** Dinesh_Bhor has joined #openstack-cinder00:57
*** junboli has joined #openstack-cinder00:58
*** andreas_s has joined #openstack-cinder01:01
*** andreas_s has quit IRC01:05
*** melwitt has quit IRC01:10
*** zhaochao has joined #openstack-cinder01:11
*** yangyapeng has joined #openstack-cinder01:11
*** wanghao has joined #openstack-cinder01:13
*** yangyapeng has quit IRC01:16
*** wanghao_ has quit IRC01:16
*** yangyapeng has joined #openstack-cinder01:18
*** Dinesh_Bhor has quit IRC01:21
*** _dalgaaf has quit IRC01:29
*** Dinesh_Bhor has joined #openstack-cinder01:29
*** liverpooler has joined #openstack-cinder01:30
*** Dinesh_Bhor has quit IRC01:30
*** Dinesh_Bhor has joined #openstack-cinder01:30
*** dalgaaf has joined #openstack-cinder01:31
*** Dinesh_Bhor has quit IRC01:33
*** salv-orlando has joined #openstack-cinder01:33
*** derekjhyang has joined #openstack-cinder01:36
*** salv-orlando has quit IRC01:39
*** felipemonteiro_ has joined #openstack-cinder01:40
*** namnh has joined #openstack-cinder01:43
*** melwitt has joined #openstack-cinder01:56
*** melwitt is now known as Guest9430601:56
*** Guest94306 is now known as jgwentworth01:58
*** jgwentworth is now known as melwitt02:01
*** gcb has joined #openstack-cinder02:04
*** harlowja has quit IRC02:10
*** Dinesh_Bhor has joined #openstack-cinder02:11
*** moshele has joined #openstack-cinder02:18
*** felipemonteiro_ has quit IRC02:18
*** Dinesh_Bhor has quit IRC02:18
*** kaisers__ has joined #openstack-cinder02:19
*** Dinesh_Bhor has joined #openstack-cinder02:20
*** kaisers_ has quit IRC02:22
*** Dinesh_Bhor has quit IRC02:34
*** Dinesh_Bhor has joined #openstack-cinder02:36
*** Dinesh_Bhor has quit IRC02:39
*** Dinesh_Bhor has joined #openstack-cinder02:40
*** Dinesh_Bhor has quit IRC02:46
*** edmondsw has joined #openstack-cinder02:54
*** edmondsw has quit IRC02:58
*** dave-mccowan has joined #openstack-cinder02:58
*** Dinesh_Bhor has joined #openstack-cinder02:59
*** zhurong has joined #openstack-cinder03:03
*** moshele has quit IRC03:12
*** psachin has joined #openstack-cinder03:25
*** uberjay has quit IRC03:28
*** uberjay has joined #openstack-cinder03:29
*** moshele has joined #openstack-cinder03:31
*** sumitshatwara has quit IRC03:33
*** wanghao has quit IRC03:34
*** wanghao has joined #openstack-cinder03:35
*** salv-orlando has joined #openstack-cinder03:36
*** salv-orlando has quit IRC03:41
*** uberjay has quit IRC03:43
*** uberjay has joined #openstack-cinder03:45
*** uberjay has quit IRC03:47
*** uberjay has joined #openstack-cinder03:47
*** hoangcx has quit IRC03:50
*** daidv has quit IRC03:50
*** daidv has joined #openstack-cinder03:50
*** hoangcx has joined #openstack-cinder03:50
*** dave-mccowan has quit IRC03:52
*** moshele has quit IRC03:56
*** abhishekk has joined #openstack-cinder03:57
*** zhurong has quit IRC04:05
*** mylu has quit IRC04:09
*** links has joined #openstack-cinder04:10
*** markstur has quit IRC04:12
*** harlowja has joined #openstack-cinder04:21
*** mylu has joined #openstack-cinder04:21
*** harlowja has quit IRC04:26
*** harlowja has joined #openstack-cinder04:26
*** harlowja has quit IRC04:27
*** salv-orlando has joined #openstack-cinder04:37
*** pgadiya has joined #openstack-cinder04:38
*** salv-orlando has quit IRC04:42
*** edmondsw has joined #openstack-cinder04:42
*** andreas_s has joined #openstack-cinder04:45
*** edmondsw has quit IRC04:46
*** moshele has joined #openstack-cinder04:46
*** andreas_s has quit IRC04:49
*** Dinesh_Bhor has quit IRC04:54
*** Dinesh_Bhor has joined #openstack-cinder04:57
*** Dinesh_Bhor has quit IRC05:01
*** Dinesh_Bhor has joined #openstack-cinder05:03
*** mylu_ has joined #openstack-cinder05:06
*** mylu has quit IRC05:08
nikeshmerlon_ : hi05:36
*** salv-orlando has joined #openstack-cinder05:38
*** adisky__ has joined #openstack-cinder05:38
*** zhurong has joined #openstack-cinder05:41
*** salv-orlando has quit IRC05:43
*** kaisers__ has quit IRC05:45
*** caisan_ has joined #openstack-cinder05:47
*** gkadam has joined #openstack-cinder05:52
*** kaisers_ has joined #openstack-cinder05:56
*** salv-orlando has joined #openstack-cinder05:59
*** Dinesh_Bhor has quit IRC05:59
*** Dinesh_Bhor has joined #openstack-cinder06:00
*** jappleii__ has quit IRC06:01
*** caisan has joined #openstack-cinder06:02
*** caisan_ has quit IRC06:05
*** Dinesh_Bhor has quit IRC06:08
nikeshmany patch for deprecating options suggested by https://review.openstack.org/#/c/53485406:09
*** Dinesh_Bhor has joined #openstack-cinder06:11
*** mylu_ has quit IRC06:16
*** Dinesh_Bhor has quit IRC06:24
*** pgadiya has quit IRC06:27
*** caisan_ has joined #openstack-cinder06:28
*** edmondsw has joined #openstack-cinder06:30
*** caisan has quit IRC06:30
*** pgadiya has joined #openstack-cinder06:33
*** AlexeyAbashkin has joined #openstack-cinder06:33
*** izombie has joined #openstack-cinder06:34
*** edmondsw has quit IRC06:34
*** AlexeyAbashkin has quit IRC06:37
*** pgadiya has quit IRC06:41
*** pgadiya has joined #openstack-cinder06:41
*** Dinesh_Bhor has joined #openstack-cinder06:43
*** fnordahl has quit IRC06:44
amitocan one more core (other than smcginnis) please take a look at https://review.openstack.org/#/c/536778/ ? it's pending a +2 and workflow for some time now. thanks.06:44
*** fnordahl has joined #openstack-cinder06:46
*** caisan has joined #openstack-cinder06:56
*** zhurong has quit IRC06:58
*** caisan_ has quit IRC06:59
*** zhurong has joined #openstack-cinder07:02
*** wolsen has quit IRC07:09
*** pcaruana has joined #openstack-cinder07:10
*** wolsen has joined #openstack-cinder07:11
*** andreas_s has joined #openstack-cinder07:17
*** caisan_ has joined #openstack-cinder07:21
*** caisan has quit IRC07:24
*** kaisers_ has quit IRC07:28
*** sticker has joined #openstack-cinder07:30
*** salv-orlando has quit IRC07:36
*** e0ne has joined #openstack-cinder07:36
*** salv-orlando has joined #openstack-cinder07:36
openstackgerritJeremy Zhang proposed openstack/python-cinderclient master: Add api_version wraps for group snapshot list in v3/shell.py  https://review.openstack.org/53912407:38
*** salv-orlando has quit IRC07:41
*** AlexeyAbashkin has joined #openstack-cinder07:41
*** kaisers_ has joined #openstack-cinder07:42
*** AlexeyAbashkin has quit IRC07:46
*** caisan has joined #openstack-cinder07:47
*** alexchadin has joined #openstack-cinder07:49
*** AlexeyAbashkin has joined #openstack-cinder07:50
*** caisan_ has quit IRC07:51
*** wanghao_ has joined #openstack-cinder07:58
*** salv-orlando has joined #openstack-cinder07:59
*** josecastroleon has joined #openstack-cinder08:00
*** wanghao has quit IRC08:01
*** Dinesh_Bhor has quit IRC08:05
*** caisan_ has joined #openstack-cinder08:13
*** izombie has quit IRC08:14
*** kaisers_ has quit IRC08:15
*** caisan has quit IRC08:16
*** Dinesh_Bhor has joined #openstack-cinder08:16
*** stakeda has quit IRC08:16
*** tesseract has joined #openstack-cinder08:17
*** Dinesh_Bhor has quit IRC08:18
*** edmondsw has joined #openstack-cinder08:18
*** Dinesh_Bhor has joined #openstack-cinder08:19
*** armaan_ has quit IRC08:19
*** Dinesh_Bhor has quit IRC08:19
*** armaan has joined #openstack-cinder08:19
*** edmondsw has quit IRC08:23
*** lkuchlan has joined #openstack-cinder08:29
*** caisan has joined #openstack-cinder08:29
*** bkopilov has joined #openstack-cinder08:29
openstackgerritShunei Shiono proposed openstack/cinder master: Update NEC driver configuration reference for Queens  https://review.openstack.org/53914108:30
*** caisan_ has quit IRC08:33
*** AlexeyAbashkin has quit IRC08:35
*** Dinesh_Bhor has joined #openstack-cinder08:36
*** Dinesh_Bhor has quit IRC08:38
openstackgerritTommyLike proposed openstack/cinder master: Add 'list_volume' in Group show&list API Doc  https://review.openstack.org/53914508:38
*** AlexeyAbashkin has joined #openstack-cinder08:38
*** AlexeyAbashkin has quit IRC08:41
*** AlexeyAbashkin has joined #openstack-cinder08:42
*** dklyle has joined #openstack-cinder08:43
*** david-lyle has quit IRC08:44
*** Dinesh_Bhor has joined #openstack-cinder08:45
*** s-shiono has quit IRC08:45
*** andreas_s has quit IRC08:50
*** lennyb has quit IRC08:50
*** andreas_s has joined #openstack-cinder08:50
*** wanghao_ has quit IRC08:56
openstackgerritTommyLike proposed openstack/cinder master: Update document for worker cleanup API  https://review.openstack.org/53915208:57
*** wanghao has joined #openstack-cinder08:57
*** Dinesh_Bhor has quit IRC08:58
*** vivsoni has quit IRC09:00
*** vivsoni_ has joined #openstack-cinder09:00
*** Dinesh_Bhor has joined #openstack-cinder09:01
*** Miouge has joined #openstack-cinder09:01
MiougeDoes cinder RBD driver support EC pool with Luminous?09:02
MiougeCeph doc says “rbd create --size 1G --data-pool ec_pool replicated_pool/image_name”, but there is only a “rbd_pool” setting in cinder.conf, no rbd_data_pool setting :(09:03
*** Dinesh_Bhor has quit IRC09:04
*** kaisers_ has joined #openstack-cinder09:04
*** wanghao_ has joined #openstack-cinder09:04
*** Dinesh_Bhor has joined #openstack-cinder09:05
*** lkuchlan has quit IRC09:05
*** lkuchlan has joined #openstack-cinder09:07
*** wanghao has quit IRC09:08
*** lkuchlan has quit IRC09:11
openstackgerritTommyLike proposed openstack/cinder master: Add 'metadata' in backup detail&show API document  https://review.openstack.org/53915609:13
*** salv-orlando has quit IRC09:17
openstackgerritliuyamin proposed openstack/cinder master: Use constants for cinder-backup and cinder-schedule  https://review.openstack.org/53418109:18
*** salv-orlando has joined #openstack-cinder09:18
*** salv-orlando has quit IRC09:23
*** mvk has quit IRC09:29
*** andreas_s_ has joined #openstack-cinder09:31
*** caisan_ has joined #openstack-cinder09:33
*** lennyb has joined #openstack-cinder09:34
*** Dinesh_Bhor has quit IRC09:35
*** andreas_s has quit IRC09:35
*** caisan has quit IRC09:35
openstackgerritTommyLike proposed openstack/cinder master: Update document for worker cleanup API  https://review.openstack.org/53915209:40
*** lkuchlan has joined #openstack-cinder09:40
*** aydintd has joined #openstack-cinder09:43
*** aydintd has quit IRC09:44
*** itlinux has joined #openstack-cinder09:51
*** hoangcx has quit IRC09:51
*** namnh has quit IRC09:51
*** daidv has quit IRC09:51
*** daidv has joined #openstack-cinder09:51
*** hoangcx has joined #openstack-cinder09:51
*** namnh has joined #openstack-cinder09:51
*** mvk has joined #openstack-cinder09:58
*** josecastroleon has quit IRC09:58
*** armaan has quit IRC09:59
*** caisan has joined #openstack-cinder09:59
erlon_nikeshm, tpsilva is compiling a list with all drivers that are broken and should be fixed10:01
*** zhurong has quit IRC10:02
*** itlinux has quit IRC10:02
*** caisan_ has quit IRC10:02
*** caisan has quit IRC10:03
*** caisan has joined #openstack-cinder10:03
*** pgadiya has quit IRC10:03
*** wanghao_ has quit IRC10:04
nikeshmerlon_ : just pinged u for the email u sent10:05
nikeshmDeprecation of Kaminario and Pure automatic calculation for max_over_subscription_ratio10:06
*** edmondsw has joined #openstack-cinder10:06
*** itlinux has joined #openstack-cinder10:07
*** armaan has joined #openstack-cinder10:08
erlon_nikeshm, ow, right, thanks for reaching, so, you are the maintainer os kaminario right?10:09
*** edmondsw has quit IRC10:10
*** lpetrut has joined #openstack-cinder10:12
*** edmondsw has joined #openstack-cinder10:12
*** edmondsw has quit IRC10:12
erlon_nikeshm, so, you just need to set the deprecation flag on the auto_calc_max_oversubscription_ratio option, then when queens is closed, you can remove the option, and leave only this: https://github.com/openstack/cinder/blob/f98c9da944b46875cdec91cf4c0c28ce89e1ac6a/cinder/volume/drivers/kaminario/kaminario_common.py#L806-L80710:14
*** jessewu has quit IRC10:14
erlon_nikeshm, then, if the used set max_over_subscription_ratio = auto, the scheduler will use this formula: https://github.com/openstack/cinder/blob/f98c9da944b46875cdec91cf4c0c28ce89e1ac6a/cinder/utils.py#L1091-L109710:16
erlon_nikeshm, which has the same behaviour as the one in your driver10:17
*** pgadiya has joined #openstack-cinder10:21
*** caisan has quit IRC10:21
nikeshmerlon_ : this formula is used by our engineering team  provisioned_vol / float(cap.total - cap.free)10:24
nikeshmwhich looks different from 1 + (provisioned_capacity_gb/(  total_capacity_gb - free_capacity_gb + 1))10:25
nikeshmwhere 'provisioned_capacity_gb': provisioned_vol / units.Mi10:26
*** sumitshatwara has joined #openstack-cinder10:29
*** alexchadin has quit IRC10:30
erlon_nikeshm, yes, its a bit different, that's is to avoid critical points when 'total_capacity_gb - free_capacity_gb' == 0 and when (provisioned_capacity_gb/(  total_capacity_gb - free_capacity_gb + 1)) < 110:31
*** salv-orlando has joined #openstack-cinder10:31
*** alexchadin has joined #openstack-cinder10:31
*** dave-mccowan has joined #openstack-cinder10:32
nikeshmerlon_ : ok i will discuss with team once10:33
*** andreas_s has joined #openstack-cinder10:33
nikeshmis it mandatory to deprecate in queens10:33
*** pgadiya has quit IRC10:33
erlon_nikeshm, provisioned_capacity_gb is usually has at least 2 degrees of magnitude than 1, which is irrelevant10:33
*** bhagyashri_s has joined #openstack-cinder10:34
*** pooja-jadhav has joined #openstack-cinder10:34
erlon_nikeshm, I simulated the behaviour of kaminario's formula, pure's and other proposal, in the end, what will matter is, how long (considering the free space) will the schedule allow a volume to be schedule10:34
*** nsingh has joined #openstack-cinder10:34
erlon_and they all have the same bevaiour and similar limits10:35
nikeshmerlon_ : ok10:35
*** andreas_s_ has quit IRC10:36
*** niraj_singh has quit IRC10:36
*** pooja_jadhav has quit IRC10:37
*** bhagyashris has quit IRC10:37
erlon_nikeshm, https://docs.google.com/spreadsheets/d/1wpNg-80YkHyrQqSWk120znkKRMOg1xJB8va12-L_vso/edit?usp=sharing10:37
*** wanghao has joined #openstack-cinder10:37
erlon_nikeshm, simulations, see that yours and pure has a division by zero10:38
*** mvk has quit IRC10:39
*** itlinux has quit IRC10:41
*** wanghao has quit IRC10:41
*** ganso has joined #openstack-cinder10:43
nikeshmerlon_: total capacity in ur sheet is 1 GB10:43
*** tpsilva has joined #openstack-cinder10:43
nikeshmerlon_ : is it correct10:43
erlon_nikeshm, yep sorry, it uses brazillian conventions with takes '.' -> ','10:45
*** itlinux has joined #openstack-cinder10:46
erlon_geguileo, should see this as well ^10:46
geguileoerlon_: what's up?10:46
*** itlinux has quit IRC10:46
erlon_geguileo, hey, this is the simulations I did related the automatic mosr calculations10:46
*** alexchadin has quit IRC10:47
*** alexchadin has joined #openstack-cinder10:49
geguileoerlon_: nice!10:49
geguileoit even has nice colors   XD10:50
*** AlexeyAbashkin has quit IRC10:50
*** AlexeyAbashkin has joined #openstack-cinder10:51
*** mvk has joined #openstack-cinder10:51
*** wanghao has joined #openstack-cinder10:52
*** armaan has quit IRC10:52
nikeshmerlon_ : when is this case useful (provisioned_capacity_gb/(  total_capacity_gb - free_capacity_gb + 1)) < 110:52
*** armaan has joined #openstack-cinder10:53
nikeshmerlon_ : sorry i got the point10:53
erlon_geguileo, changing the page, I was debugging a error in our CI yesterday, and reach to os-brick connect_volume. I notice that the multipath is taking a way that does not connects to the portal before trying to show the current connections10:53
geguileoerlon_: that's always been the case10:54
erlon_geguileo, not sure that was something that changed recently or is the expected behaviour10:54
geguileoerlon_: you have to check first if you are already connected10:54
geguileoerlon_: it's always been like that10:54
geguileoerlon_: if you connect LUN 1 first and you are not connected that check will fail and you'll do the connection10:55
geguileoerlon_: if you connect to LUN 2 later you'll see that you already have the connection and just proceed to do the scanning10:55
geguileoerlon_: it's the same for single and multipathed connections10:55
*** andreas_s_ has joined #openstack-cinder10:55
erlon_but in this case brick is not retrying10:56
*** AlexeyAbashkin has quit IRC10:56
nikeshmerlon_ : is the simulation covered for (provisioned_capacity_gb/(  total_capacity_gb - free_capacity_gb + 1)) < 110:56
erlon_geguileo, these are the command flows, unyil the last test (from image) everything was ok10:56
erlon_nikeshm, you mean without the '1' in the formula?  1 + (prov ...)10:57
geguileoerlon_: there are a loooooot of ways connections could go wrong, and they are not necessarily os-brick issues10:57
geguileoerlon_: and the commands without DEBUG logs are not that usuful I'm afraid10:58
*** andreas_s has quit IRC10:58
*** junboli has quit IRC10:59
openstackgerritGorka Eguileor proposed openstack/cinder-specs master: Update Provisioning Improvements  https://review.openstack.org/53827110:59
*** abhishekk has quit IRC11:00
erlon_geguileo, hmm, the commands without debug are much easier to follow, at least for me. The thing is, when Cinder is trying to connect, in the 'from image' case, it does not log onto the portal neither scans11:01
erlon_geguileo, so, iscsiadm -m node --op show -p will obviously fail11:01
erlon_If I scan manually and run the same command on cinder it works11:01
*** alexchadin has quit IRC11:02
geguileoerlon_: you say the show obviously fails because you have no attached volume, right?11:02
*** alexchadin has joined #openstack-cinder11:03
*** alexchadin has quit IRC11:03
*** alexchadin has joined #openstack-cinder11:03
erlon_geguileo, yes, this is the first time cinder is trying to attach, and it should do something else before11:03
*** alexchadin has quit IRC11:04
*** alexchadin has joined #openstack-cinder11:04
*** alexchadin has quit IRC11:04
*** namnh has quit IRC11:05
*** alexchadin has joined #openstack-cinder11:05
*** alexchadin has quit IRC11:05
*** lkuchlan has quit IRC11:05
*** alexchadin has joined #openstack-cinder11:06
*** alexchadin has quit IRC11:06
geguileoerlon_: there's something missing from that sequence you sent me11:07
geguileoerlon_: that's why logs are better  XD11:07
*** bkopilov has quit IRC11:08
openstackgerritSam Wan proposed openstack/os-brick master: Sanity check before referring 'initiator_target_map'  https://review.openstack.org/53918411:12
*** andreas_s_ has quit IRC11:19
*** andreas_s has joined #openstack-cinder11:20
*** mvenesio has joined #openstack-cinder11:20
*** AlexeyAbashkin has joined #openstack-cinder11:22
erlon_erlon_, ahhrg, ok, ill try to reproduce that later and will gather the full logs11:22
openstackgerritLucian Petrut proposed openstack/cinder master: Update Windows docs  https://review.openstack.org/53831311:29
*** salv-orlando has quit IRC11:32
*** salv-orlando has joined #openstack-cinder11:32
*** armaan has quit IRC11:35
*** armaan has joined #openstack-cinder11:35
*** salv-orlando has quit IRC11:37
*** andreas_s has quit IRC11:47
*** andreas_s has joined #openstack-cinder11:47
*** alexchadin has joined #openstack-cinder11:49
*** AlexeyAbashkin has quit IRC11:51
*** AlexeyAbashkin has joined #openstack-cinder11:52
*** itlinux has joined #openstack-cinder11:54
*** itlinux has quit IRC12:00
*** itlinux has joined #openstack-cinder12:01
*** sticker has quit IRC12:06
*** edmondsw has joined #openstack-cinder12:08
*** zhurong has joined #openstack-cinder12:12
*** alexchadin has quit IRC12:17
*** itlinux has quit IRC12:18
*** alexchadin has joined #openstack-cinder12:18
*** itlinux has joined #openstack-cinder12:18
*** lkuchlan has joined #openstack-cinder12:19
*** alexchadin has quit IRC12:22
*** alexchadin has joined #openstack-cinder12:22
*** bkopilov has joined #openstack-cinder12:25
*** itlinux has quit IRC12:26
geguileoerlon_: don't do it for me, I mean, I'm not really interested in going through more logs (that's most of what I do lately), I only offer if it helps you   XD12:36
erlon_geguileo, well, sometime we ill need to solve that multipath problem, just needed some direction to see if my findings were correct. When we schedule a internal task here to thoroughly investigate we can talk again with more details about it12:44
geguileoerlon_: if you reproduce and have logs I can see if it's a bug or what is it, and if it's a bug I'll probably go and fix it  ;-)12:45
erlon_geguileo, great! thanks!12:47
*** tobasco has quit IRC12:49
*** links has quit IRC12:49
*** edmondsw_ has joined #openstack-cinder12:49
*** tobasco has joined #openstack-cinder12:50
*** edmondsw has quit IRC12:51
*** salv-orlando has joined #openstack-cinder12:52
openstackgerritErlon R. Cruz proposed openstack/cinder master: Updates overprovisioning dev-ref  https://review.openstack.org/53898912:52
*** gcb has quit IRC12:56
*** edmondsw_ has quit IRC12:58
*** edmondsw has joined #openstack-cinder13:00
*** abishop has joined #openstack-cinder13:01
*** zhurong_ has joined #openstack-cinder13:03
*** salv-orl_ has joined #openstack-cinder13:13
openstackgerritLucian Petrut proposed openstack/cinder master: Update Windows docs  https://review.openstack.org/53831313:13
*** caisan has joined #openstack-cinder13:16
*** salv-orlando has quit IRC13:16
*** zhurong has quit IRC13:17
*** hemna_ has joined #openstack-cinder13:18
*** zhurong_ has quit IRC13:19
_pewp_hemna_ ( *՞ਊ՞*)ノ13:19
*** zhurong has joined #openstack-cinder13:19
*** zhurong has quit IRC13:20
*** alexchadin has quit IRC13:21
*** chenying_ has quit IRC13:21
*** sshatwar_ has joined #openstack-cinder13:22
*** sumitshatwara has quit IRC13:25
*** lhx_ has joined #openstack-cinder13:25
openstackgerritVivek Soni proposed openstack/cinder master: 3PAR: Fixed setting of sync_period value  https://review.openstack.org/53920813:29
*** yangyapeng has quit IRC13:30
*** yangyapeng has joined #openstack-cinder13:30
openstackgerritErlon R. Cruz proposed openstack/cinder master: Updates overprovisioning dev-ref  https://review.openstack.org/53898913:34
*** moshele has quit IRC13:34
*** yangyapeng has quit IRC13:35
*** dave-mccowan has quit IRC13:37
amitoerlon_: hi, sup? any idea if zuul gate is still stuck / busy? we have a patch with +1 on workflow that has been stuck for a few hours now - https://review.openstack.org/#/c/536778/13:37
*** caisan has quit IRC13:39
*** izombie has joined #openstack-cinder13:39
*** caisan has joined #openstack-cinder13:40
-openstackstatus- NOTICE: Our ubuntu-xenial images (used for e.g. unit tests and devstack) are currently failing to install any packages, restrain from *recheck* or *approve* until the issue has been investigated and fixed.13:42
*** links has joined #openstack-cinder13:44
*** alexchadin has joined #openstack-cinder13:52
openstackgerritVivek Soni proposed openstack/cinder master: 3PAR: Fixed sync period updation in rcg  https://review.openstack.org/53920813:54
*** alexchadin has quit IRC13:55
*** izombie has quit IRC13:55
*** alexchadin has joined #openstack-cinder13:56
*** psachin has quit IRC13:57
*** alexchadin has quit IRC14:00
*** yangyapeng has joined #openstack-cinder14:01
*** abhishekk has joined #openstack-cinder14:06
jdillamanMiouge: you can use cinder multi-backend w/ different Ceph users per backend to achieve the desired effect14:08
Miougejdillaman: Yes, that is the plan: several cinder backends with different Ceph pools, but one of the pools is an Erasure Coded pool. RBD metadata needs to be stored on a replicated pool while the RBD data is in a EC pool. So there are 2 pool names to pass: metadata pool and data pool. I do not see how to configure that in cinder.14:11
jdillamanFor the "client.XYZ" user used for the EC backend, just add a section in your ceph.conf to override the config option for data pool14:12
*** dustins has joined #openstack-cinder14:12
jdillamanMiouge: a la https://pastebin.com/d5VuEwL814:13
Miougejdillaman: oh I see what you mean. Clever! This config in ceph.conf is required only on cinder-volume or also in nova-compute nodes?14:15
jdillamanMiouge: only on the cinder hosts14:16
Miougejdillaman: in your example “cinder_ex” is the rbd_user for the EC ciner backend right?14:18
jdillamanMiouge: yes -- a specific Cinder user for a specific multi-backend14:19
MiougeOK, then nova-compute/libvirt/librbd is able to find out the data-pool from the metadata in the rbd_pool. I’ll try this setup. Thanks jdillaman14:24
jdillamanMiouge: correct -- the data pool is only needed during the image creation14:24
jdillamanMiouge: once the image is created, its data pool is recorded in the image header14:25
*** derekjhyang has quit IRC14:26
*** AlexeyAbashkin has quit IRC14:26
*** AlexeyAbashkin has joined #openstack-cinder14:27
*** dklyle has quit IRC14:29
*** alexchadin has joined #openstack-cinder14:30
*** caisan has quit IRC14:37
Miougejdillaman: I see that python-librbd has support for data_pool on the create() method. Is there any interrest in adding that in the cinder driver? Maybe something I can try to add.14:38
*** abhishekk has quit IRC14:39
jdillamanCertainly wouldn't hurt. I think it would be a nice-to-have addition for the RBD drivers in OpenStack14:39
*** eharney has joined #openstack-cinder14:44
openstackgerritKushal Wathore proposed openstack/cinder master: 3PAR: fix cloning issue in create group from source.  https://review.openstack.org/53922514:45
*** csaikia has joined #openstack-cinder14:45
*** dave-mccowan has joined #openstack-cinder14:55
openstackgerritMayur Indalkar proposed openstack/cinder master: Implementation of Cinder driver over ISCSI for Veritas Access  https://review.openstack.org/53923014:55
*** r-daneel has joined #openstack-cinder14:57
erlon_amito, gate has been facing some slowliness these last week. My patch toke 26h to merge after workflow+114:59
*** kwathore has joined #openstack-cinder15:01
amitoerlon_: oh, bah... thanks.15:01
openstackgerritEric Harney proposed openstack/cinder master: bump pylint to recent version  https://review.openstack.org/50975515:02
amitoerlon_: isn't Zuul supposed to comment anyway after setting workflow +1 though?15:03
erlon_amito, it will but only after run the gate tests15:04
amitoerlon_, there is usually a comment stating that it is about to start gate tests or something of the sort...15:05
*** lhx_ has quit IRC15:05
erlon_amito, ow yes, that is. but the tests will take a looong time to start15:05
*** links has quit IRC15:05
*** lhx_ has joined #openstack-cinder15:05
erlon_amito, http://zuul.openstack.org/15:05
erlon_you can see you patch in the line here15:05
*** alexchadin has quit IRC15:05
erlon_waiting time is being about 20h15:06
*** andreas_s has quit IRC15:10
*** andreas_s has joined #openstack-cinder15:11
*** gkadam has quit IRC15:14
*** zhongjun has quit IRC15:18
openstackgerritErlon R. Cruz proposed openstack/cinder master: Updates overprovisioning dev-ref  https://review.openstack.org/53898915:18
openstackgerritErlon R. Cruz proposed openstack/cinder master: Updates overprovisioning dev-ref  https://review.openstack.org/53898915:19
*** esker has joined #openstack-cinder15:24
*** crushil has joined #openstack-cinder15:26
*** andreas_s has quit IRC15:29
*** andreas_s has joined #openstack-cinder15:30
*** gcb has joined #openstack-cinder15:31
*** eharney has quit IRC15:32
*** itlinux has joined #openstack-cinder15:37
*** izombie has joined #openstack-cinder15:37
*** andreas_s has quit IRC15:40
*** zhaochao has quit IRC15:41
*** itlinux has quit IRC15:42
*** andreas_s has joined #openstack-cinder15:45
*** pchavva has joined #openstack-cinder15:46
*** felipemonteiro__ has joined #openstack-cinder15:47
*** eharney has joined #openstack-cinder15:47
*** AlexeyAbashkin has quit IRC15:50
*** AlexeyAbashkin has joined #openstack-cinder15:50
*** esker has quit IRC15:52
*** AlexeyAbashkin has quit IRC15:55
*** AlexeyAbashkin has joined #openstack-cinder15:55
*** david-lyle has joined #openstack-cinder15:57
*** itlinux has joined #openstack-cinder15:57
*** felipemonteiro has joined #openstack-cinder15:58
*** andreas_s has quit IRC15:59
*** andreas_s has joined #openstack-cinder16:01
*** felipemonteiro__ has quit IRC16:01
*** alexchadin has joined #openstack-cinder16:03
*** markstur has joined #openstack-cinder16:03
*** armaan has quit IRC16:04
*** andreas_s has quit IRC16:05
*** markstur_ has joined #openstack-cinder16:05
*** andreas_s has joined #openstack-cinder16:06
*** markstur has quit IRC16:08
*** e0ne has quit IRC16:12
*** lkuchlan has quit IRC16:18
*** kaisers_ has quit IRC16:29
*** chandankumar is now known as chkumar|off16:30
*** david-lyle has quit IRC16:34
*** alexchadin has quit IRC16:35
*** AlexeyAbashkin has quit IRC16:38
*** andreas_s has quit IRC16:40
*** andreas_s has joined #openstack-cinder16:44
*** lhx_ has quit IRC16:45
*** david-lyle has joined #openstack-cinder16:47
*** david-lyle has quit IRC16:48
*** david-lyle has joined #openstack-cinder16:49
*** andreas_s has quit IRC16:53
*** andreas_s has joined #openstack-cinder16:54
*** gkadam has joined #openstack-cinder16:58
*** kaisers_ has joined #openstack-cinder16:59
*** gcb has quit IRC17:00
*** andreas_s has quit IRC17:00
*** erlon_ has quit IRC17:03
*** mvk has quit IRC17:10
*** crose has joined #openstack-cinder17:12
*** felipemonteiro has quit IRC17:14
*** felipemonteiro__ has joined #openstack-cinder17:14
SwansonWhere are timeouts for rest calls set? Dumb question, I'm sure, but I have a REST call that is taking 3 minutes to fail. Seems excessive.17:17
*** diablo_rojo has quit IRC17:23
*** csaikia has quit IRC17:26
*** csaikia has joined #openstack-cinder17:26
*** AlexeyAbashkin has joined #openstack-cinder17:29
smcginnisSwanson: I think it's a parameter to the requests call, IIRC.17:35
*** izombie has quit IRC17:35
Swansonsmcginnis, It doesn't appear to be part of the session. So I presume it is on each call...17:35
patrickeastsmcginnis: Swanson : yea its something you have to pass in on each one17:35
patrickeastmight be some global config to change the default17:36
*** markvoelker has joined #openstack-cinder17:36
Swansonpatrickeast, Thanks. Looking into that now.17:38
*** prateek has joined #openstack-cinder17:40
*** tesseract has quit IRC17:46
*** felipemonteiro has joined #openstack-cinder17:49
*** felipemonteiro__ has quit IRC17:51
*** alexchadin has joined #openstack-cinder17:58
SwansonNot sure why this has bumped up to 3 minute timeouts for this customer but I think I'm just going to add timeouts to the individual posts and whatnot..17:58
hemna_some calls are synch calls afaik18:01
hemna_vs. async18:01
*** david-lyle has quit IRC18:13
openstackgerritLucian Petrut proposed openstack/cinder master: Update Windows docs  https://review.openstack.org/53831318:15
*** diablo_rojo has joined #openstack-cinder18:16
*** lpetrut has quit IRC18:25
*** prateek has quit IRC18:26
*** alexchadin has quit IRC18:27
*** itlinux has quit IRC18:36
*** alexchadin has joined #openstack-cinder18:38
*** salv-orl_ has quit IRC18:39
*** salv-orlando has joined #openstack-cinder18:39
*** alexchadin has quit IRC18:41
*** armaan has joined #openstack-cinder18:43
*** salv-orlando has quit IRC18:44
*** AlexeyAbashkin has quit IRC18:46
*** moshele has joined #openstack-cinder18:53
*** armaan has quit IRC18:58
*** ircuser-1 has quit IRC19:00
*** armaan has joined #openstack-cinder19:00
*** kaisers_ has quit IRC19:09
*** kaisers_ has joined #openstack-cinder19:09
*** alexchadin has joined #openstack-cinder19:09
*** alexchadin has quit IRC19:11
*** adisky__ has quit IRC19:12
*** harlowja has joined #openstack-cinder19:17
*** ayoung has joined #openstack-cinder19:20
*** armaan has quit IRC19:21
ayoungsmcginnis, or anyone who might know:  if I try to make a volume bootable, and it fails (why is a different story for now) I can't quite figure out how to free up the volume.  It is still attached, but to an instance that does not exists, and I can't (as a normal user) change that state.19:22
ayoung b5d7ec69-772b-4969-ad6c-438e06ef202c | cfme_volume   | in-use |   80 | Attached to c54316f8-2b6f-4f33-bfcb-25b62cc7cbf3 on /dev/vda19:22
ayoungand openstack server remove volume 21b801c3-fab3-4845-b395-b61eb40ae643 279d37ed-16f1-42b9-8cc1-543be3b4133219:23
ayoungNo server with a name or ID of '21b801c3-fab3-4845-b395-b61eb40ae643' exists.19:23
*** armaan has joined #openstack-cinder19:23
ayoung openstack volume set --state error b5d7ec69-772b-4969-ad6c-438e06ef202c19:23
ayoungFailed to set volume state: Policy doesn't allow volume_extension:volume_admin_actions:reset_status to be performed. (HTTP 403) (Request-ID: req-c9706c71-12f3-4648-9084-a7cd3e41f36d)19:23
ayoungOne or more of the set operations failed19:23
ayoungis there any reason not to change that policy, and make it so a _member_ can perform reset_status?19:24
eharneythe primary reason is that it would allow users to set things into inconsistent states, we'd need a less dangerous/more limited reset command if we wanted to users to access it, i think19:26
ayoungeharney, yeah, but that can't work for existing deploys19:29
ayoungeharney, what would be the worst that could happen?  A user could corrupt a device already attached to a server?19:30
hemna_so....thoughts on lvmetad ?19:30
hemna_enable or not?19:30
eharneyayoung: not sure really, a lot of messy things could likely happen inside of cinder19:30
ayoungeharney, like, say a user changes the state of an attached drive, makes it available, and then another user attaches it to another server?19:32
eharneyayoung: yes19:32
ayoungnova would have two servers attached to the same device, and corruption of an epic scale?19:32
ayoungeharney, right now, there is a bunch of internet posts about fixing the problem I am seeing by hacking the CInder MySQL instance19:33
ayoungthat is arguably just as bad or badder19:34
ayoungmaybe badderest19:34
ayoungI tried to create a volume from a glance image.  The image requires a disk larger than the largest flavor provides, so I made it 80 GB.  And then booted the Nova server using that volume.  That boot failed, and I can't delete or re-attached the image.19:35
ayoungIts kimndof a sad failure mode19:35
eharneyIMO the answer is to redesign things to not have numerous "an error occurred and now you're stuck in bad state X" traps all over Cinder that require reset operations19:35
smcginnisayoung: Looks like that first command is failing due to using the wrong arguments of osc.19:35
eharneyi started sketching out a plan for how to do this a bit ago, maybe i should pick it up again...19:35
smcginnisayoung: For a boot from volume server, I beleive you will need to delete the server first.19:36
ayoungsmcginnis, yep, but once the server is deleted, the volume is wedged19:36
ayoungcan't change its state19:36
smcginnisayoung: So the server is deleted and gone, but the volume still shows attached?19:36
ayoungthis is an older deploy, probably OSP10 or 11, which maps to Newton or Ocata, I think19:37
smcginnisOK, then you need to reset status of the volume.19:38
smcginnisBut like eharney said, it's risky to allow that for all users.19:38
ayoungright...but this is a situation that the average user could inflict on himself...I am a pretty average user.  Maybe below average.19:38
smcginnisayoung: Well, _normally_ they shouldn't get in this state, but sounds like there is some other error that got you here.19:39
ayoungsmcginnis, right so when I booted the volume, I got this error: Build of instance 21b801c3-fab3-4845-b395-b61eb40ae643 aborted: Block Device Mapping is Invalid.19:40
ayoungI created the volume using the os_volume ansible module from an image.  It might be that the image was bad.19:40
ayoung  os_volume:19:41
ayoung    state: present19:41
ayoung    cloud: "{{ cloudname }}"19:41
ayoung    size: 8119:41
ayoung    display_name: cfme_volume_219:41
ayoung    image:  cfme-rhos-
ayoungdoesn't seem crazy.  and it seems like the only way to boot this image, as the image is pretty big:19:41
ayoungobviously, I could ask for a larger flavor to be created, too19:41
smcginnisYeah, that's all fine. You can boot an instance from a volume. But if this was done through nova, and nova failed to boot the instance and decided to delete the instance without detaching the volume, then that's the root of the problem.19:42
ayoungso I called "delete server"19:42
smcginnisJust saying, theoretically if things are working right, the need to administratively override the volume status should be pretty low.19:43
ayoungIt would not let me detach the volume first, as it was the root volume19:43
*** alexchadin has joined #openstack-cinder19:43
*** lpetrut has joined #openstack-cinder19:43
ayoungI'm trying to not just solve the problem, but come up with a general solution that can be applied when I am not here...or the only other admin for this cluster who is not available at the moment.19:44
hemna_anyone have a devstack up with an lvm backend configured?19:46
ayoungsmcginnis, you are saying that the problem is Nova did not set the volume status when deleting the server?19:46
hemna_I can't create a volume from an lvm snap19:46
hemna_device-mapper: reload ioctl on (254:23) failed: No data available19:46
smcginnisayoung: Just a guess, but I think the problem now is you deleted the server, but the osc command you are trying to run is through the server interface.19:47
hemna_I get that from the lvchange command19:47
smcginnishemna_: Is the snap activated?19:47
ayoungsmcginnis, but when I delete a server, normally you would expect it to change the volume state, regardless of what state the server is in, right?19:47
hemna_"activated" ?19:47
hemna_I created an lvm volumed, then snapped it, then trying to create a new vol from the snap19:48
smcginnisayoung: Yes, I would.19:48
*** david-lyle has joined #openstack-cinder19:51
*** pcaruana has quit IRC19:52
ayoungsmcginnis, https://bugs.launchpad.net/nova/+bug/140486719:53
openstackLaunchpad bug 1404867 in OpenStack Compute (nova) "Volume remains in-use status, if instance booted from volume is deleted in error state" [Medium,In progress] - Assigned to melanie witt (melwitt)19:53
smcginnisOh man, 2015.19:54
ayoungsmcginnis, and melwitt is on it, still.19:54
smcginnisLooks like every fix got reverted.19:55
melwittI just rebased that patch yesterday :P19:55
hemna_the lvchange had -a in it to activate it19:55
smcginnismelwitt: Third (or fourth) times a charm? :)19:55
melwittyeah, back in the day, mistakes in the patch kept breaking the gate, specifically only the postgres job19:55
hemna_err -ay19:56
ayoungmelwitt, its ok, I am once again on the hook for Bug 968696 and that was opened in 201219:56
openstackbug 968696 in OpenStack Identity (keystone) ""admin"-ness not properly scoped" [High,In progress] https://launchpad.net/bugs/968696 - Assigned to Adam Young (ayoung)19:56
ayoungmelwitt, In your humble opinion, is this backportable?19:57
melwittayoung: yes, it's backportable but not 100% straightforward because of differences in how volumes worked in the past. and I don't think it would be backported upstream because of the relatively low severity of the bug19:59
ayoungmelwitt, fair enough.  Its just annoying that a user can, easily, get into a situation which requires an admin to get them out of.20:02
ayoungAnd even if your fix goes it, I don't get to use it until we have a downstream version.20:03
*** alexchadin has quit IRC20:03
melwittayoung: yeah, agreed. I will ask if we can backport it upstream but regardless I'll be backporting it downstream20:03
smcginnisayoung: What version did you say you are running?20:03
ayoungsmcginnis, I'm not 100% certain.  I think RH OSP 11, which would  be Ocata20:04
ayounganyway I can tell as an end user?20:04
*** mvk has joined #openstack-cinder20:04
melwittmy bug is from liberty -_-20:05
smcginnisayoung: Probably could only backport to Pike at this point, and unless it merges in the next couple weeks, probably only can backport to stable/queens by that point.20:05
smcginnismelwitt: Speaking of sagas, what ever happened to the bug about not checking host storage availability for boot from volume instances? Did that one ever get resolved?20:06
melwittsmcginnis: you mean the root_gb=0 thing? not yet, the temporary stop-gap patches have been refused since resource providers work is supposed to solve the issue properly (with shared storage modeling). and the problem is super latent, it's always been that way. we're looking toward rocky for that now20:08
smcginnismelwitt: Yep, that one. Cool, hopefully that works out in Rocky.20:09
smcginnismelwitt: I know I've been pinged about that several times.20:09
openstackgerritEric Harney proposed openstack/cinder master: Rename lio job to lio-barbican  https://review.openstack.org/53930620:10
*** felipemonteiro has quit IRC20:15
*** felipemonteiro__ has joined #openstack-cinder20:15
ayoungsmcginnis, so...back to the dangerous thing...I realize that changing the status via that API really just changes the database table.  Thast in turn is used to determine what actions are legal for a given volume.  From the CInder side, it it possible to "yank" a volume from Nova, so that it is no longer available?20:25
ayounga "safe" way from the cinder perspective?20:29
*** lpetrut has quit IRC20:35
*** lpetrut has joined #openstack-cinder20:40
*** e0ne has joined #openstack-cinder20:42
SwansonGot one, a second +2 out there? https://review.openstack.org/#/c/538314/20:47
*** diablo_rojo has quit IRC20:48
*** KeithMnemonic has joined #openstack-cinder20:50
openstackgerritTom Swanson proposed openstack/cinder master: DellEMC SC: Added timeout to REST calls  https://review.openstack.org/53931820:54
*** e0ne has quit IRC20:55
smcginnisayoung: Sorry, was in a meeting.20:55
smcginnisayoung: So the storage is probably exported to the compute host.20:56
smcginnisayoung: Deleting the volume should get rid of that.20:56
ayoungsmcginnis, No need to apologize.  I am aware of how much an interruption and IRC question can be.20:56
smcginnisayoung: It should be safe as far as yanking it from the compute host.20:56
smcginnisayoung: Haha, yep, no worries. :)20:56
ayoungsmcginnis, but I can't delete the volume in the state...which is, I assume, a check on the user from pulling out their own plug.  However, would that be safer than allowing them to update the state?20:57
smcginnisayoung: Linux is pretty good about block devices going away, unlike some operating systems.20:57
smcginnisayoung: There's some cleanup that should be done, but I believe next time a volume is attached it should mostly clean itself up.20:57
smcginnisayoung: Yeah, you're going to have to update state, then delete.20:57
ayoungso let me put my keystone hat on for a moment20:57
ayoungsay I want to make my cluser as self maintaining as possible20:58
ayoungand I don't want people asking for cluster admins to clean up their mess20:58
ayoungI want a "smart storage person" on a project that can do more than just the basics20:58
ayoungI was thinking that we should have a role that is not-quite-admin20:58
smcginnisayoung: Unfortunately with failures like this, I think the only way to make it self-serve is to give the users more power than you probably want to.20:58
ayoungproject-manager or something.   So the question is what power to give to project-manager so that they can clean up their project, but not negatively affect other projects20:59
smcginnisayoung: Project admin I think has been brought up before. I like that idea, but not sure how many years it will take to get there. :]20:59
ayoungsmcginnis, its been 6 years thus far,  what are another 1/2 dozen?20:59
* smcginnis has a sad cry21:00
ayoungsmcginnis, one of my goals is to get some proof-of-concept deployments out there to iron out the real world issues with these21:00
ayoungI recently switched from Software Engineer to solutions architect in order to talk more with customers so that these discussions would have the weight of real world pain behind them21:00
smcginnisayoung: ++ awesome21:01
smcginnisayoung: A noble reason to go to the dark(ish) side. :)21:01
ayoungsmcginnis, so, our team has a cloud for running demos, and this is where I wedged things21:01
smcginnisayoung: I think we need more of that. Feedback on real pain points would definitely be welcome.21:01
ayoungI'm sure I will eventually get the "keys to the kingdom" for this cluster, but right now I am enjoying being a _member_ and keeping track of the ways I can mess things up21:02
smcginnisNot sure how to best address this pain point, other than getting that fix in nova, but if we can do something in Cinder to make that better without compromising other use cases, then I'm all for it.21:02
ayoungsmcginnis, question on the root error:  is there a timeout problem when booting from a volume?  I had an 80GB volume, and that seems to be why the server boot failed21:03
ayoungsome artivcles seem to say that I need to expand a timeout to let it succeed21:03
ayounghttps://ask.openstack.org/en/question/62636/instance-creation-fails-block-device-mapping/  for example21:03
smcginnisayoung: I wouldn't think so. It should discover the new block device like any other new volume, then the instance *should* be able to boot from it almost like a local ephemeral disk.21:04
smcginnisThere is a timeout issue when having the volume created through the nova API calls.21:04
smcginnisSo if you tell Nova to create a new instance and to create a volume for it, sometimes it takes too long for nova to proxy those calls to cinder and everything to complete in time for it to succeed within the default window.21:05
smcginnisSo in those cases it does need to be increased.21:05
ayoungnot what I am dfoing here though21:05
ayoungI am trying to use an existing volume:21:05
smcginnisayoung: Yeah, if you pre-create the volume, then tell Nova to boot from it, that shouldn't be an issue here.21:05
smcginnisAnd really is the preferable, though less streamlines, workflow.21:05
ayounggot the same error reported in one of the comments: Block Device Mapping is Invalid:21:05
smcginnisIt's kind of a catch all I think when the nova instance isn't able to find or use the block device I think.21:06
ayoungsmcginnis, what would happen if the original image were bogus?  The one I used to create the volume in the first call?21:07
ayoungis there any validation done?21:07
smcginnisIIRC, there is basic md5 hash verification.21:09
smcginnisayoung: There is a proposal to add support for image signing.21:09
ayoungMeh...that would not do much.  If someone uploads a bogus image, they would still end up here21:10
smcginnisayoung: But that's just the download step. There's nothing that verifies dd'ing the image to the volume didn't hit corruption as at that point I think it would have to reread the entire volume.21:10
smcginnisAnd yeah, nothing to say that the original image wasn't junk in the first place.21:10
*** felipemonteiro has joined #openstack-cinder21:10
smcginnisYou can create an image of a non-bootable volume and create a bootable volume from it.21:11
ayoungright, of course21:11
ayounga very important workflow, too21:11
*** diablo_rojo_phon has left #openstack-cinder21:12
*** pchavva has quit IRC21:13
*** felipemonteiro__ has quit IRC21:13
*** lpetrut has quit IRC21:19
*** eharney_ has joined #openstack-cinder21:21
SwansonDumb multiattach question. If I get a force-detach (nontype connector) I'm whacking ALL attachments, right?21:22
smcginnisSwanson: Correct.21:23
smcginnisSwanson: I believe we went through things, and the manager will not send a None connector down to the driver otherwise.21:23
*** eharney has quit IRC21:23
*** eharney_ is now known as eharney21:23
Swansonsmcginnis, This is bogus? https://git.openstack.org/cgit/openstack/cinder-specs/tree/specs/queens/enable-multiattach.rst#n14221:24
SwansonMaybe I should look at the impl and not the spec(l).21:24
smcginnisNo habla ingles.21:25
ildikovsmcginnis: I guess we could do a better job with docs as usual...21:25
smcginnisAlways seems to be the case.21:26
ildikovSwanson: the spec is always just an indication on what we might want to achieve :)21:26
SwansonYou spelled "is" wrong.21:26
*** moshele has quit IRC21:26
smcginnisildikov: Feel free to slap him.21:26
ildikovsmcginnis: I have a high threshold :)21:27
smcginnisildikov: Feel free to slap him anyway. :)21:27
SwansonAnyone have the review link for this in their pocket?21:27
SwansonAh, there it is.21:28
ildikovsmcginnis: ah, so you want me to do the dirty work for you?21:28
smcginnisThey know my face there. :)21:28
*** eharney has quit IRC21:30
SwansonSoooo.... set shared_targets to True and then test that multiple attachments actually work? Spiffy.21:32
ildikovI think shared_targets is True if not set specifically to False21:33
SwansonThat seems nuts.21:35
SwansonI get it.21:37
SwansonSo, don't set dick and just see if it works?21:37
*** eharney has joined #openstack-cinder21:39
*** threestrands has joined #openstack-cinder21:40
*** kaisers_ has quit IRC21:44
*** kaisers_ has joined #openstack-cinder21:44
*** armaan has quit IRC21:46
ildikovSwanson: set it to True, it can't go wrong21:46
*** armaan has joined #openstack-cinder21:47
*** mvenesio has quit IRC21:53
*** dustins has quit IRC21:54
*** mvenesio has joined #openstack-cinder21:54
*** felipemonteiro has quit IRC21:57
*** eharney has quit IRC21:57
*** felipemonteiro has joined #openstack-cinder21:57
*** mvenesio has quit IRC21:58
*** mylu has joined #openstack-cinder21:59
*** crose has quit IRC22:02
jgriffithSwanson: why do you insist on making all the simple things so difficult :)22:03
jgriffithSwanson: you sir are a grinder :)22:03
Swansonjgriffith, We'll see how simple this is. We'll see!22:04
jgriffithSwanson: you just need to lower your standards22:04
jgriffithSwanson: when you set the bar as low as I do things are *awesome* :)22:05
smcginnisI've seen what he eats for lunch. Trust me, his standards are pretty low.22:05
jgriffithPickled Pigs Feet?22:05
jgriffithGood ol PPF22:05
jgriffithsome greens on the side22:05
SwansonHad a sandwich from the machine today. I'm not sure what meat is gray (other than Arbys) but I'll say this, it wasn't great.22:06
jgriffithSwanson: Ok, I stand corrected; anybody eating a sandwich from a machine has their bar set low enough22:09
*** edmondsw has quit IRC22:11
*** edmondsw has joined #openstack-cinder22:12
openstackgerritAlan Bishop proposed openstack/cinder master: Fix how backups handle encryption key IDs  https://review.openstack.org/53746222:13
*** salv-orlando has joined #openstack-cinder22:14
*** edmondsw has quit IRC22:17
*** abishop has quit IRC22:17
Swansonjgriffith, I've had worse. These are actually better than they used to be. Back before the health department shut the vendor down for 6 months or so.22:17
*** rcernin has quit IRC22:18
*** tpsilva has quit IRC22:22
*** felipemonteiro__ has joined #openstack-cinder22:27
jgriffithSwanson: Ha!!22:28
*** csaikia has quit IRC22:29
*** csaikia has joined #openstack-cinder22:29
*** felipemonteiro has quit IRC22:31
*** diablo_rojo has joined #openstack-cinder22:36
*** hemna_ has quit IRC22:45
*** itlinux has joined #openstack-cinder22:48
ildikovI'm so glad it's not meal time here for a while :)22:49
ildikovjgriffith: well, if he messes things up now we still have a little time to fix, so why not?22:51
smcginnisGo to bed ildikov!22:51
*** rcernin has joined #openstack-cinder22:52
*** itlinux has quit IRC22:52
* ildikov has a sad kid face now22:52
*** sticker has joined #openstack-cinder22:53
ildikovsmcginnis: I won't be able to sleep anyway...22:53
smcginnisJet lagged again?22:54
*** csaikia has quit IRC22:54
ildikovsmcginnis: I would drop the 'again' part :)22:54
ildikovsmcginnis: I don't have a time zone anymore, but sadly the accountant and the chiropractor do...22:55
smcginnisildikov: You should be able to fall asleep at any time of the day now. :)22:56
ildikovjgriffith: BTW, did you get anywhere with debugging that swap volume issue?22:56
ildikovjgriffith: I thought to at least re-run it on the gate, but it's been dying ever since hence I switched to patience...22:57
jgriffithildikov: I did not22:57
*** andreas_s has joined #openstack-cinder22:58
ildikovsmcginnis: yeah, well, sleeping was never my strong suit, I'm way better at daydreaming which I'm practicing day and night now22:58
ildikovjgriffith: then we're on the same page at least :)22:59
ildikovjgriffith: I will look into it when I can, I just wanted to know whether you might have some new hints on that one22:59
jgriffithildikov: I couldn't get it to fail and moved on23:00
jgriffithadded the logging so the gate job that fails might give us some more insight23:00
*** mriedem is now known as mriedem_afk23:02
*** andreas_s has quit IRC23:02
SwansonHuh. Multipath on unbuntu seems less than awesome.23:02
SwansonYeah, that.23:03
*** dave-mccowan has quit IRC23:03
SwansonAlways best to wait until the end of the cycle to revisit everything.23:03
ildikovjgriffith: fair enough23:07
*** gnufied has quit IRC23:07
ildikovjgriffith: I saw the logging, hence I thought to rerun the tests, but the gate is under heavy load, so I thought to wait a little with that23:07
ildikovjgriffith: but wasn't sure whether you ever got this to fail or not, that's why I asked23:08
*** salv-orlando has quit IRC23:21
*** salv-orlando has joined #openstack-cinder23:22
*** salv-orlando has quit IRC23:26
*** armaan has quit IRC23:45
*** stakeda has joined #openstack-cinder23:49
*** ganso has quit IRC23:53
*** rcernin has quit IRC23:57

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