Monday, 2016-02-08

*** haomaiwang has quit IRC00:01
*** haomaiwang has joined #openstack-cinder00:01
*** alonma has quit IRC00:03
*** salv-orl_ has quit IRC00:19
*** chlong has joined #openstack-cinder00:26
*** mylu has quit IRC00:31
*** haomaiwang has quit IRC01:01
*** dims has quit IRC01:01
*** haomaiwang has joined #openstack-cinder01:01
*** pots has joined #openstack-cinder01:01
*** chlong has quit IRC01:03
*** mylu has joined #openstack-cinder01:05
*** haomaiw__ has joined #openstack-cinder01:06
*** haomaiwang has quit IRC01:07
*** salv-orlando has joined #openstack-cinder01:19
*** dims has joined #openstack-cinder01:24
*** gouthamr has joined #openstack-cinder01:27
*** gouthamr has quit IRC01:39
openstackgerritAngus Lees proposed openstack/os-brick: Trivial rootwrap -> privsep replacement
*** thurloat_isgone is now known as thurloat01:45
*** dims has quit IRC01:46
*** alonma has joined #openstack-cinder01:47
*** EinstCrazy has joined #openstack-cinder01:50
*** salv-orlando has quit IRC01:50
*** alonma has quit IRC01:51
*** alonma has joined #openstack-cinder01:53
*** EinstCrazy has quit IRC01:54
*** mylu has quit IRC01:57
*** alonma has quit IRC01:58
*** mylu has joined #openstack-cinder01:58
*** haomaiw__ has quit IRC02:01
*** haomaiwang has joined #openstack-cinder02:01
*** gouthamr has joined #openstack-cinder02:03
*** gouthamr has quit IRC02:05
*** chlong has joined #openstack-cinder02:06
openstackgerritChuck Fouts proposed openstack/cinder: Support for iSCSI CHAP Uni-directional Auth
openstackgerritOpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements
*** changbl has quit IRC02:42
*** changbl has joined #openstack-cinder02:42
openstackgerritOpenStack Proposal Bot proposed openstack/python-cinderclient: Updated from global requirements
*** akshai has joined #openstack-cinder02:48
*** alonma has joined #openstack-cinder02:53
*** akshai has quit IRC02:54
*** mragupat has joined #openstack-cinder02:55
*** alonma has quit IRC02:58
*** pots1 has joined #openstack-cinder02:59
*** haomaiwang has quit IRC03:01
*** haomaiwang has joined #openstack-cinder03:01
*** pots has quit IRC03:03
*** mragupat has quit IRC03:09
*** mragupat has joined #openstack-cinder03:10
*** salv-orlando has joined #openstack-cinder03:10
*** jwcroppe has joined #openstack-cinder03:11
*** jwcroppe has quit IRC03:11
*** jwcroppe has joined #openstack-cinder03:12
*** salv-orlando has quit IRC03:13
*** dims has joined #openstack-cinder03:13
*** thurloat is now known as thurloat_isgone03:17
*** mylu has quit IRC03:17
*** mylu has joined #openstack-cinder03:17
*** pots has joined #openstack-cinder03:18
*** pots1 has quit IRC03:21
*** mylu has quit IRC03:27
*** mylu has joined #openstack-cinder03:28
*** mylu has quit IRC03:29
*** mylu has joined #openstack-cinder03:31
*** mylu has quit IRC03:32
*** mylu has joined #openstack-cinder03:34
*** mylu has quit IRC03:35
*** dims has quit IRC03:35
*** IlyaG has quit IRC03:35
*** mylu has joined #openstack-cinder03:36
*** dims has joined #openstack-cinder03:42
*** IlyaG has joined #openstack-cinder03:45
*** EinstCrazy has joined #openstack-cinder03:46
*** EinstCrazy has quit IRC03:51
*** alonma has joined #openstack-cinder03:54
*** alonma has quit IRC03:58
*** cknight has joined #openstack-cinder03:59
*** haomaiwang has quit IRC04:01
*** alonma has joined #openstack-cinder04:01
*** haomaiwang has joined #openstack-cinder04:01
*** alonma has quit IRC04:05
*** alonma has joined #openstack-cinder04:07
*** dims has quit IRC04:08
*** itzdilip has joined #openstack-cinder04:09
*** alonma has quit IRC04:12
*** links has joined #openstack-cinder04:14
*** mylu has quit IRC04:15
*** mylu has joined #openstack-cinder04:16
*** mragupat has quit IRC04:22
*** david-lyle has quit IRC04:30
*** sheel_ has joined #openstack-cinder04:33
*** sheel_ is now known as sheel04:34
*** ByPasS has quit IRC04:35
*** winston-1 has joined #openstack-cinder04:36
*** jidar_ has joined #openstack-cinder04:38
*** ByPasS has joined #openstack-cinder04:38
*** winston-d has quit IRC04:38
*** jidar has quit IRC04:38
*** lennyb_ has quit IRC04:38
*** lennyb_ has joined #openstack-cinder04:38
*** jidar_ is now known as jidar04:38
openstackgerritAnthony Lee proposed openstack/cinder: LeftHand: Updating minimum client version
*** salv-orlando has joined #openstack-cinder04:44
*** david-lyle has joined #openstack-cinder04:45
*** akerr is now known as akerr_away04:46
sheeloops...why m talking to my self...04:53
*** shyama has joined #openstack-cinder04:55
*** salv-orlando has quit IRC04:56
*** mudassirlatif has joined #openstack-cinder04:58
*** sgotliv has joined #openstack-cinder05:00
*** haomaiwang has quit IRC05:01
*** haomaiwang has joined #openstack-cinder05:01
*** bkumar has joined #openstack-cinder05:03
*** mudassirlatif has quit IRC05:07
*** mudassirlatif has joined #openstack-cinder05:08
*** alonma has joined #openstack-cinder05:08
*** chhavi has joined #openstack-cinder05:11
*** alonma has quit IRC05:13
*** alonma has joined #openstack-cinder05:15
*** alonma has quit IRC05:19
*** ankit_ag has joined #openstack-cinder05:21
*** alonma has joined #openstack-cinder05:24
*** mudassirlatif has quit IRC05:28
*** alonma has quit IRC05:28
*** alonma has joined #openstack-cinder05:30
*** alonma has quit IRC05:35
*** alonma has joined #openstack-cinder05:37
openstackgerritAnkit Agrawal proposed openstack/cinder: NFS snapshots
*** alonma has quit IRC05:41
*** cknight has quit IRC05:42
*** alonma has joined #openstack-cinder05:43
ankit_agHi all, can someone please help reviewing few of cinderclient patches
*** alonma has quit IRC05:47
*** IlyaG has quit IRC05:48
*** alonma has joined #openstack-cinder05:50
*** alonma has quit IRC05:54
bkumarHi all, anyone here using hpe 3par as a cinder backend ?05:54
*** alonma has joined #openstack-cinder05:57
*** mylu has quit IRC05:58
*** mudassirlatif has joined #openstack-cinder05:58
*** haomaiwang has quit IRC06:01
*** haomaiwa_ has joined #openstack-cinder06:01
*** alonma has quit IRC06:02
*** alonma has joined #openstack-cinder06:08
*** alonma has quit IRC06:13
*** mudassirlatif has quit IRC06:19
*** mudassirlatif has joined #openstack-cinder06:20
*** sgotliv has quit IRC06:28
*** mudassirlatif has quit IRC06:40
*** salv-orlando has joined #openstack-cinder06:48
*** IlyaG has joined #openstack-cinder06:49
*** savihou has joined #openstack-cinder06:50
*** IlyaG has quit IRC06:54
*** salv-orlando has quit IRC06:55
*** EinstCrazy has joined #openstack-cinder06:56
*** haomaiwa_ has quit IRC07:01
*** haomaiwang has joined #openstack-cinder07:01
*** EinstCrazy has quit IRC07:02
*** chlong has quit IRC07:03
*** anshul has joined #openstack-cinder07:03
*** arecknag has joined #openstack-cinder07:04
*** smoriya has quit IRC07:07
*** alonma has joined #openstack-cinder07:09
openstackgerritAngus Lees proposed openstack/os-brick: Trivial rootwrap -> privsep replacement
*** alonma has quit IRC07:14
*** belmoreira has joined #openstack-cinder07:14
*** alonma has joined #openstack-cinder07:15
openstackgerritAnkit Agrawal proposed openstack/cinder: NFS snapshots
*** alonma has quit IRC07:20
*** alonma has joined #openstack-cinder07:21
*** alonma has quit IRC07:25
*** alonma has joined #openstack-cinder07:27
*** alonma has quit IRC07:32
*** alonma has joined #openstack-cinder07:33
*** alonma has quit IRC07:38
*** alonma has joined #openstack-cinder07:39
*** mudassirlatif has joined #openstack-cinder07:40
*** alonma has quit IRC07:44
*** alonma has joined #openstack-cinder07:46
*** alonma has quit IRC07:50
*** alonma has joined #openstack-cinder07:52
*** smoriya_afk has joined #openstack-cinder07:53
*** smoriya_afk is now known as smoriya07:53
*** alonma has quit IRC07:56
*** alonma has joined #openstack-cinder07:58
*** haomaiwang has quit IRC08:01
*** haomaiwang has joined #openstack-cinder08:01
*** alonma has quit IRC08:02
*** sinese has joined #openstack-cinder08:04
*** salv-orlando has joined #openstack-cinder08:07
*** rcernin has joined #openstack-cinder08:09
*** haomaiwang has quit IRC08:10
*** haomaiwa_ has joined #openstack-cinder08:11
*** salv-orlando has quit IRC08:12
*** dulek has joined #openstack-cinder08:15
*** alonma has joined #openstack-cinder08:20
*** alonma has quit IRC08:25
*** alonma has joined #openstack-cinder08:31
*** jistr has joined #openstack-cinder08:33
*** jistr is now known as jistr|sick08:33
*** alonma has quit IRC08:35
*** alonma has joined #openstack-cinder08:37
openstackgerritPhilipp Marek proposed openstack/cinder: DRBD: Fix arguments for resize_volume DBus API call.
*** alonma has quit IRC08:42
*** liverpooler has joined #openstack-cinder08:43
*** alonma has joined #openstack-cinder08:44
*** chlong has joined #openstack-cinder08:48
*** alonma has quit IRC08:49
*** alonma has joined #openstack-cinder08:52
*** nkrinner has joined #openstack-cinder08:53
*** bvivek has joined #openstack-cinder08:56
*** bvivek_ has joined #openstack-cinder08:56
*** alonma has quit IRC08:56
*** alonma has joined #openstack-cinder08:58
*** haomaiwa_ has quit IRC09:01
*** EinstCrazy has joined #openstack-cinder09:01
*** haomaiwa_ has joined #openstack-cinder09:01
*** openstackgerrit has quit IRC09:02
*** openstackgerrit has joined #openstack-cinder09:03
openstackgerritMatan Sabag proposed openstack/cinder: Manage/unmanage volume in ScaleIO driver
*** alonma has quit IRC09:03
*** EinstCrazy has quit IRC09:06
*** alonma has joined #openstack-cinder09:09
*** alonma has quit IRC09:14
*** lpetrut has joined #openstack-cinder09:14
*** alonma has joined #openstack-cinder09:15
*** alonma has quit IRC09:20
*** alonma has joined #openstack-cinder09:22
*** alonma has quit IRC09:26
*** isaacb has joined #openstack-cinder09:28
*** alonma has joined #openstack-cinder09:28
*** salv-orlando has joined #openstack-cinder09:31
*** aarefiev_ has quit IRC09:31
openstackgerritAnkit Agrawal proposed openstack/cinder: NFS snapshots
*** aarefiev has joined #openstack-cinder09:32
*** alonma has quit IRC09:33
*** alonma has joined #openstack-cinder09:34
*** ildikov has quit IRC09:37
*** alonma has quit IRC09:38
*** alonma has joined #openstack-cinder09:49
openstackgerritVictor Stinner proposed openstack/cinder: Port objects unit tests to Python 3
*** mvk has joined #openstack-cinder09:50
*** alonma has quit IRC09:53
*** haomaiwa_ has quit IRC10:01
*** alonma has joined #openstack-cinder10:01
*** haomaiwang has joined #openstack-cinder10:01
*** alonma has quit IRC10:05
*** bvivek_ has quit IRC10:07
*** bvivek has quit IRC10:07
*** ildikov has joined #openstack-cinder10:10
*** alonma has joined #openstack-cinder10:10
*** alonma has quit IRC10:15
*** alonma has joined #openstack-cinder10:16
*** alonma has quit IRC10:20
*** alonma has joined #openstack-cinder10:23
openstackgerritMatan Sabag proposed openstack/cinder: Manage/unmanage volume in ScaleIO driver
*** isaacb has quit IRC10:24
*** alonma has quit IRC10:27
*** ildikov has quit IRC10:28
*** ociuhandu has quit IRC10:30
*** alonma has joined #openstack-cinder10:36
*** ildikov has joined #openstack-cinder10:36
openstackgerritPhilipp Marek proposed openstack/cinder: DRBD: Fix arguments for resize_volume DBus API call.
*** salv-orl_ has joined #openstack-cinder10:40
*** alonma has quit IRC10:41
*** alonma has joined #openstack-cinder10:43
*** salv-orlando has quit IRC10:43
*** alonma has quit IRC10:47
*** tpsilva has joined #openstack-cinder10:49
*** alonma has joined #openstack-cinder10:49
*** e0ne has joined #openstack-cinder10:53
*** alonma has quit IRC10:54
openstackgerritLiucheng Jiang proposed openstack/cinder: Huawei: Implement v2 replication (managed)
*** alonma has joined #openstack-cinder10:55
*** alonma has quit IRC10:59
*** haomaiwang has quit IRC11:01
*** 21WAA0U3N has joined #openstack-cinder11:01
*** alonma has joined #openstack-cinder11:01
*** alonma has quit IRC11:06
*** alonma has joined #openstack-cinder11:11
*** alonma has quit IRC11:15
*** akerr has joined #openstack-cinder11:17
*** alonma has joined #openstack-cinder11:17
*** akerr_away has quit IRC11:18
*** rcernin has quit IRC11:19
*** alonma has quit IRC11:21
*** rcernin has joined #openstack-cinder11:22
*** zhonghua has quit IRC11:22
*** zhonghua-lee has joined #openstack-cinder11:23
*** alonma has joined #openstack-cinder11:24
*** rcernin has quit IRC11:25
*** rcernin has joined #openstack-cinder11:26
*** alonma has quit IRC11:31
*** anshul has quit IRC11:32
*** alonma has joined #openstack-cinder11:32
*** jistr|sick has quit IRC11:37
*** alonma has quit IRC11:39
*** alonma has joined #openstack-cinder11:41
*** alonma has quit IRC11:45
*** alonma has joined #openstack-cinder11:47
*** anshul has joined #openstack-cinder11:47
*** alonma has quit IRC11:51
*** alonma has joined #openstack-cinder11:53
*** liverpooler has quit IRC11:53
*** vgridnev has joined #openstack-cinder11:55
*** alonma has quit IRC11:57
*** thurloat_isgone is now known as thurloat11:58
*** alonma has joined #openstack-cinder11:59
*** mvk has quit IRC12:00
*** 21WAA0U3N has quit IRC12:01
*** erlon has joined #openstack-cinder12:01
*** haomaiwang has joined #openstack-cinder12:01
*** liverpooler has joined #openstack-cinder12:02
*** vgridnev has quit IRC12:07
*** dims has joined #openstack-cinder12:07
*** alyson_ has joined #openstack-cinder12:09
*** vgridnev has joined #openstack-cinder12:13
*** liverpooler has quit IRC12:13
*** liverpooler has joined #openstack-cinder12:15
*** salv-orl_ has quit IRC12:18
*** isaacb has joined #openstack-cinder12:18
*** adrianofr has joined #openstack-cinder12:20
*** dims has quit IRC12:20
*** dims has joined #openstack-cinder12:22
*** timcl has joined #openstack-cinder12:23
*** ociuhandu has joined #openstack-cinder12:25
*** mvk has joined #openstack-cinder12:33
*** akerr has quit IRC12:34
*** smoriya has quit IRC12:36
*** smoriya_afk has joined #openstack-cinder12:37
*** smoriya_afk is now known as smoriya12:38
*** yhayashi has quit IRC12:39
*** krotscheck_dcm is now known as krotscheck12:39
*** smoriya has quit IRC12:41
*** nikeshm has joined #openstack-cinder12:41
nikeshmi am back from marriage leaves :)12:41
nikeshmfeel free to ping me if any need12:42
*** salv-orlando has joined #openstack-cinder12:42
*** smoriya_afk has joined #openstack-cinder12:42
*** smoriya_afk is now known as smoriya12:42
*** mdavidson has joined #openstack-cinder12:43
*** bkumar has quit IRC12:45
*** alonma has quit IRC12:52
openstackgerritPetrut Lucian proposed openstack/os-brick: [WIP] Add Windows connectors
openstackgerritPetrut Lucian proposed openstack/os-brick: WIP os-brick refactor get_connector_properties
openstackgerritRonen Mesonzhnik proposed openstack/cinder: Support backup import on another Storage database
*** haomaiwang has quit IRC13:01
*** haomaiwang has joined #openstack-cinder13:01
*** EinstCrazy has joined #openstack-cinder13:05
*** krotscheck has quit IRC13:06
*** EinstCrazy has quit IRC13:09
*** xyang1 has joined #openstack-cinder13:16
*** krotscheck has joined #openstack-cinder13:18
*** links has quit IRC13:20
*** smoriya has quit IRC13:23
*** smoriya_afk has joined #openstack-cinder13:24
*** smoriya_afk is now known as smoriya13:24
*** laughterwym has joined #openstack-cinder13:25
*** smoriya has quit IRC13:28
scottdae0ne: If you are around, the microversion patches are passing Jenkins now:
scottdae0ne: Thanks for all your help on that.13:29
*** smoriya_afk has joined #openstack-cinder13:30
scottdaI'm going to address Patrick's comments later this morning.13:30
*** smoriya_afk is now known as smoriya13:30
*** sgotliv has joined #openstack-cinder13:30
e0nescottda: hi. thanks for the update13:30
scottdaSure. I'll keep working on this. I'm optimistic that we can still get this in in Mitaka.13:31
*** smoriya has quit IRC13:31
*** edmondsw has joined #openstack-cinder13:32
ankit_agHi all, can someone please help reviewing few of cinderclient patches
*** smoriya_afk has joined #openstack-cinder13:33
*** smoriya_afk is now known as smoriya13:33
*** smoriya has quit IRC13:33
*** smoriya_afk has joined #openstack-cinder13:34
*** ygl has joined #openstack-cinder13:35
*** smoriya_afk is now known as smoriya13:35
yglHi ALl13:35
yglHi All13:35
scottdae0ne: If you have time to look, please tell me what you think of Patrick's comments in / cinder/api/v3/router.py13:35
ygli am having issues while creating bootable volumes13:35
yglcan anyone help me please13:35
scottdae0ne: It would be great to avoid copying all the v2 code to v3/ and the tests as well. I'm wondering if you or anyone knows of any issues with that .13:36
*** smoriya has quit IRC13:36
yglcan anyone help me please13:37
*** smoriya_afk has joined #openstack-cinder13:37
*** smoriya_afk is now known as smoriya13:37
*** tongli has joined #openstack-cinder13:38
e0nescottda: I'll take a look later today13:38
*** smoriya has quit IRC13:38
e0nescottda: added 'recheck' to run rally again.13:38
yglcan anyone help me please13:39
scottdaygl: Please be patient. We see your message, and more people will come online as the day goes on. Sorry, I cannot help at the moment.13:39
*** smoriya_afk has joined #openstack-cinder13:39
*** smoriya_afk is now known as smoriya13:39
*** smoriya has quit IRC13:40
e0nescottda: did you see error in apache job?
scottdae0ne: Nope. Looking now...13:40
*** smoriya_afk has joined #openstack-cinder13:40
e0nescottda: I don't know if it is related or not13:41
*** smoriya_afk is now known as smoriya13:41
*** smoriya has quit IRC13:42
*** ygl has quit IRC13:42
scottdaLooks likely to be related:13:42
*** smoriya_afk has joined #openstack-cinder13:42
*** smoriya_afk is now known as smoriya13:43
*** permalac has joined #openstack-cinder13:44
*** sgotliv has quit IRC13:45
openstackgerritSzymon Borkowski proposed openstack/cinder: Fix for glance_metadata during volume migration
*** porrua has joined #openstack-cinder13:50
openstackgerritPetrut Lucian proposed openstack/os-brick: [WIP] Add Windows connectors
*** akerr has joined #openstack-cinder13:52
openstackgerritAlexey Khodos proposed openstack/cinder: NexentaStor5 iSCSI driver unit tests
*** sheel has quit IRC13:57
*** baumann has joined #openstack-cinder14:00
*** haomaiwang has quit IRC14:01
*** haomaiwang has joined #openstack-cinder14:01
*** boichev has joined #openstack-cinder14:02
*** ankit_ag has quit IRC14:05
*** gouthamr has joined #openstack-cinder14:05
openstackgerritIvan Kolodyazhny proposed openstack/cinder: Check for service existance in capabilities API
*** rlrossit has joined #openstack-cinder14:06
*** baumann has quit IRC14:07
*** baumann has joined #openstack-cinder14:08
*** merooney has joined #openstack-cinder14:20
*** cfriesen has joined #openstack-cinder14:20
*** edtubill has joined #openstack-cinder14:23
*** anshul has quit IRC14:27
*** cknight has joined #openstack-cinder14:31
*** jungleboyj has joined #openstack-cinder14:31
*** esker has quit IRC14:31
*** knikolla has joined #openstack-cinder14:32
openstackgerritSzymon Borkowski proposed openstack/cinder: Fix for glance_metadata during volume migration
*** diablo_rojo has joined #openstack-cinder14:37
*** laughterwym has quit IRC14:37
*** jgregor has joined #openstack-cinder14:38
*** salv-orlando has quit IRC14:41
*** superdan is now known as dansmith14:45
*** anshul has joined #openstack-cinder14:47
*** dustins has joined #openstack-cinder14:49
*** lprice1 has joined #openstack-cinder14:50
*** lprice has quit IRC14:51
smcginnisnikeshm: Welcome back and congratulations!14:54
*** salv-orlando has joined #openstack-cinder14:55
*** cknight has quit IRC14:55
*** pots is now known as Guest283614:56
*** pots2 is now known as pots14:56
*** EinstCrazy has joined #openstack-cinder14:57
*** pots1 has joined #openstack-cinder14:57
*** vgridnev has quit IRC14:59
*** haomaiwang has quit IRC15:01
*** krtaylor has joined #openstack-cinder15:01
*** haomaiwa_ has joined #openstack-cinder15:01
*** EinstCrazy has quit IRC15:04
*** lcurtis has joined #openstack-cinder15:05
*** Sunkum has joined #openstack-cinder15:10
*** laughterwym has joined #openstack-cinder15:11
*** eharney has joined #openstack-cinder15:13
*** baumann has quit IRC15:15
*** ildikov has quit IRC15:16
*** mragupat has joined #openstack-cinder15:21
*** cknight has joined #openstack-cinder15:21
*** sgotliv has joined #openstack-cinder15:22
*** vgridnev has joined #openstack-cinder15:22
*** anshul has quit IRC15:26
*** baumann has joined #openstack-cinder15:27
*** laughterwym has quit IRC15:29
*** dulek has quit IRC15:31
*** sheel has joined #openstack-cinder15:35
*** mtanino has joined #openstack-cinder15:37
*** esker has joined #openstack-cinder15:38
*** esker has quit IRC15:39
openstackgerritAlex Meade proposed openstack/cinder-specs: User facing error messages
*** esker has joined #openstack-cinder15:39
*** adrianofr has quit IRC15:40
*** rcernin has quit IRC15:42
sheele0ne:hi there15:43
e0nesheel: hi15:43
sheelcould you please review - needs final cinderclient15:43
sheele0ne: its cinder counterpart is already released15:44
*** arecknag has quit IRC15:44
diablo_rojopatrickeast: A while back you offered to test this patch ( ) on Pure's stuff, is that still an option ? :)15:45
*** arch-nemesis has joined #openstack-cinder15:46
patrickeastdiablo_rojo: yea I can test it out this afternoon15:46
*** adrianofr has joined #openstack-cinder15:48
*** vincent_hou has joined #openstack-cinder15:48
sheele0ne:Thank you :)15:48
e0nesheel: np15:49
*** ildikov has joined #openstack-cinder15:49
diablo_rojopatrickeast: Cool! that would be incredibly helpful.15:49
*** links has joined #openstack-cinder15:49
*** crose has joined #openstack-cinder15:50
sheelscottda: hi15:50
scottdasheel: Hi15:50
sheelscottda: I was looking at cli counterpart of microversioning..15:51
sheelscottda: are you done with the implementation or need some help in it?15:51
scottdaI have a patch up for the cinderclient:15:51
*** lprice1 has quit IRC15:51
*** lprice has joined #openstack-cinder15:52
scottdaBut I need to update it for the new /v3 endpoint. I might not get to that until tomorrow.15:52
*** lprice1 has joined #openstack-cinder15:55
*** lprice has quit IRC15:56
*** links has quit IRC15:57
*** vgridnev has quit IRC15:59
*** haomaiwa_ has quit IRC16:01
*** haomaiwang has joined #openstack-cinder16:01
ildikovscottda: hi16:01
scottdaildikov: Hi16:02
ildikovscottda: I just wanted to continue the discussion we started about the detach issue with multiattach16:02
*** vgridnev has joined #openstack-cinder16:03
ildikovscottda: I basicly wonder whether the db direction is viable or not16:03
scottdaOK, can we wait for hemna to come online?16:03
diablo_rojoDuncanT: smcginnis: When you get a sec I know you mentioned in the Cinder meeting having opinions about this cross project spec :)
smcginnisdiablo_rojo: Thanks, I haven't looked at it in a while. I probably should.16:05
smcginnisI initially liked the idea of it, but I would be surprised if we were ever able to actually accomplish the goal.16:05
openstackgerritEric Harney proposed openstack/cinder: Improve logging to debug invalid "extra_specs" entries
*** IlyaG has joined #openstack-cinder16:06
diablo_rojosmcginnis: No problem :) it does seem like a lofty idea, but a good one if it could actually get accomplished.16:06
ildikovscottda: sure16:06
smcginnisYeah, I think so too. I'm behind the idea, just have a lot of concern about how we would actually do it.16:07
openstackgerritPetrut Lucian proposed openstack/os-brick: [WIP] Add Windows connectors
ildikovscottda: I just thought to understand what you were suggesting before discussing whether we actually do this or not, but I think you're right it's better to have hemna from the beginning :)16:08
*** Sunkum has quit IRC16:08
jungleboyjdiablo_rojo: Just a heads up that there is an oslo.config release coming out with a fix for options being out of order.16:08
jungleboyjdiablo_rojo: Not sure that that will impact any of your code but wanted to let you know.16:09
scottdaildikov: Yes, and I'll have to remember what I was suggesting :)16:09
smcginnisjungleboyj: Speaking of oslo...16:09
scottdaildikov: and jgriffith was in on that conversation IIRC16:09
smcginnisjungleboyj: Just a reminder I think you were going to look in to what the plan is for common.16:09
jungleboyjsmcginnis: Yes.  You are correct sir.16:09
diablo_rojojungleboyj:  It shouldn't since the list is just presorted by the script that generates the file and that gets sent to oslo config to do its thing with.16:10
smcginnisjungleboyj: :)16:10
diablo_rojojungleboyj: Can you link me the review though?16:10
jungleboyjdiablo_rojo: They are actually talking about all the different ways that the configs are handled.16:10
ildikovscottda: jgriffith_away had the suggestion earlier to try to export a target per attachment IIRC16:10
jungleboyjYou could stop in openstack-meeting-alt and pimp your topic proposal.16:10
*** IlyaG has quit IRC16:10
scottdaildikov: Yes, but I don't think every driver can do that.16:11
ildikovscottda: yeap, I'm afraid so too16:11
ildikovthis is why I wanted to see what other options we have16:11
diablo_rojojungleboyj:  Interesting. My topic proposal?16:11
*** gouthamr has quit IRC16:11
scottdaildikov: I think it's worth another discussion, here in IRC, with whomever is interested. Then we probably should capture all the ideas with pros and cons and sent it out to the ML, or some other way to carry the design forward.16:12
jungleboyjdiablo_rojo: Not specifically, if you come over to that meeting you could mention it though.16:12
*** sinese has quit IRC16:12
*** belmoreira has quit IRC16:12
diablo_rojojungleboyj: Thanks for the heads up.16:12
jungleboyjdiablo_rojo: Of course.16:13
*** isaacb has quit IRC16:15
ildikovscottda: yeah, we really need to figure out something and I don't think pushing everything to Nova will work out16:15
ildikovscottda: so it would be great if we could elaborate all the options we have16:15
scottdaildikov: Yes, I think that's the way to go. Gather what we can in IRC, put them in a ML post and ask for any other ideas and feedback. Mabye that will help us get to consensus.16:16
ildikovscottda: hopefully16:16
ildikovscottda: I also think we need to disable multiattach support on the LVM driver until we don't have a solution implemented16:17
scottdaildikov: Yes, we need to figure out all the logistics and timing/release issues as well.16:18
ildikovscottda: yes, we need to be smart now otherwise we will not have this in Newton either16:18
*** gouthamr has joined #openstack-cinder16:20
ildikovhemna: hi :)16:27
hemnaildikov, heyas.   how goes it ?16:27
ildikovhemna: with scottda we are about to discuss the detach issue with multiattach16:27
ildikovhemna: what to solve where and how we are just waiting for everyone to gather :)16:28
openstackgerritMatan Sabag proposed openstack/cinder: Support for consistency groups in ScaleIO driver
hemnaso, the volume itself has a list of attachments16:28
hemnawith instance_uuids16:29
hemnanova can go through those instance_uuids and find out if there are >1 on the same host where the detach is happening.16:29
smcginnisscottda: Go Broncos.16:29
scottdasmcginnis: Thanks!16:29
hemnaif so, then don't call os-brick disconnect_volume ?16:29
hemnaeharney, :)16:29
hemnaildikov, that should work even w/o the host being set in the attachment16:30
*** esp has joined #openstack-cinder16:30
hemnathe problem is16:30
hemnasome Cinder drivers create a new target for every attach16:30
scottdahemna: But doesn't that depend on the driver? On whether or not the driver creates 1 export for each attach to the same host, or multiplexes them and keeps the same 1 export even if a new attach is created on that host?16:30
ildikovhemna: the two attachments on the same host is the problem16:30
hemnain which case you want to call os-brick's disconnect_volume.16:30
scottdahemna: Yeah, what you said!16:30
hemnascottda, correct.16:30
*** vgridnev has quit IRC16:31
hemnasome create new targets16:31
hemnaothers don't16:31
*** vgridnev has joined #openstack-cinder16:31
*** vgridnev has quit IRC16:31
scottdaIt seems there is a basic question: Who (Nova or Cinder) should do the accounting and figure out when to disconnect the volume?16:31
hemnathe connection_info for the volume might enable us to figure that out though16:31
*** mragupat has quit IRC16:32
hemnathat's the information that comes back from initialize_connection.16:32
*** bvivek_ has joined #openstack-cinder16:32
*** bvivek has joined #openstack-cinder16:32
hemna(thinking out loud here)16:32
*** mragupat has joined #openstack-cinder16:32
hemnaI think brick can give you a list of volume paths for the connection_info16:32
hemnaas well as the wwn of the volume (for SCSI based volumes, iSCSI, FC, ?)16:33
*** gouthamr has quit IRC16:33
scottdaI guess I could have included brick in the possibilities of whom does the accounting...16:33
ildikovI wanted to ask about that exactly whether we can figure out from os-brick who's connected to what16:34
*** garthb has joined #openstack-cinder16:34
hemnawell all brick can do is give you a list of paths for a given connection_info16:34
*** gouthamr has joined #openstack-cinder16:34
hemnait can't do 'accounting'16:34
hemnaI'm not sure this helps16:35
hemnadunno, just trying to chew through it16:35
scottdaMaybe accounting is the wrong word. Perhaps it just takes logic to determine "Is this the last connection using this target"16:35
*** esp has quit IRC16:37
hemnais it true that if the connection_info for each attachment is the same, then that the target is always the same?16:37
hemnaI think so16:37
scottdaIt might help to think about these questions: 1) What is the best place to determine when to disconnect the volume from an architectural POV? 2) What's the best place from the POV of landing code (i.e. it might be 10x as hard in Nova)? 3) What's the best place given current implementation (Can we do this now?)16:37
ildikovbasicly the question is that how we can check how many attchments are using the same target16:37
hemnain the case of 3PAR, we'd have a different LUN id for each target16:37
hemnatarget portal + iqn + LUN ID16:38
hemnais basically the iSCSI target16:38
scottdahemna: and 3par has a different export for each target, so that makes sense.16:38
hemnaif those are all the same, then it's the same target16:38
*** savihou has quit IRC16:38
hemnaFC is even simpler, I believe16:38
hemnaso, I wonder if nova could get all the attachments on a given host16:39
hemnathen loop through those and call cinder's initialize_connection16:39
hemnaand compare those16:39
hemnaif there is > 1 that's equal, then it's a shared target.16:39
hemnaelse, it's a unique target16:39
hemnaI'm not sure how else to track it16:39
hemnaespecially when live migration comes into play16:40
*** salv-orl_ has joined #openstack-cinder16:40
ildikovhmm, it does not good to call out to Cinder that much16:40
hemnait would only have to be done in the multi-attach enabled case.16:40
hemnaand only where you have more than 1 attachment on the same host.16:41
hemnaI'd guess it would only take 2 calls to find out16:41
ildikovyou mean to check on Nova side how many attachments we have on that host and then call out to Cinder?16:41
hemnaas a Cinder driver will either export the same target N times, or 1 times.16:41
hemnait wouldn't be a mix16:41
hemnaso I think it's safe to say that you would only have to call initialize_connection twice16:42
hemnaeven if there are 10 attachments on that same host16:42
hemnaildikov, yes16:42
hemnaam I making any sense?16:42
hemnalet me back up16:42
scottdaOK, so we should determine if nova can get all the attachements on a given host, and then capture this as Possibility #116:42
ildikovtwice you mean we would need to store what driver the volume is on?16:42
hemnasay we have a volume that's attached 10 times to nova instances.16:43
*** salv-orlando has quit IRC16:43
hemna4 of those attachments are on the host we are currently trying to detach.16:43
hemnanova loops through the attachments on the volume.16:43
*** bill_az has joined #openstack-cinder16:43
hemnafinds those 4 attachments on the host it cares about16:43
hemnathen for each of those, it calls Cinder's initialize_connection16:43
hemnato fetch the connection_info16:43
hemnaeach loop it compares the connection_info with the last target16:44
*** mudassirlatif has quit IRC16:44
hemnaif they are equal, you know it's a shared target, and you are done.16:44
hemnaif it's not a shared target, then you are done.16:44
hemnaI don't think we have cinder drivers that do a mixed mode for every attachment16:44
hemnamake sense ?16:44
scottdahemna: That makes sense16:44
hemnait's either a shared target, or it's not16:45
hemnaand comparing the connection_info should tell you.16:45
*** jgriffith_away is now known as jgriffith16:45
smcginnisReminder that we have through the end of next week for os-brick changes:
hemnaas the target_portal(s), iqn, LUN id uniquely identify it.16:45
hemnayah :(!16:45
*** IlyaG has joined #openstack-cinder16:45
hemnaI'm not sure how else we can do this16:45
hemnawithout creating a new Cinder API16:46
scottdahemna: What about the possibility of Cinder storing the connection_info when it is originally called with initialize_connection at first attach time?16:47
hemnathat information will change during live migration.16:47
hemnaand it would need to get updated16:47
scottdaYes, but we need a Cinder API to change that info during live migration anyway.16:47
*** chris_morrell has quit IRC16:47
scottdaI think we decided that at the mid-cycle? To add that API to fix live migration?16:48
hemnawe covered so many things, I can't keep it straight now16:48
*** vivekd has joined #openstack-cinder16:48
hemnaI remember talking about storing the initiator info16:49
hemnabut not the target info16:49
openstackgerritMerged openstack/python-cinderclient: Bootable filter for listening volumes from CLI
johnthetubaguyscottda: storing the connection_info makes good sense to me, making sure you know whats needed on the backend if the compute is dead, etc16:49
hemnajohnthetubaguy, yah I think we need that in the case of force detach16:50
scottdajohnthetubaguy: Yes, we wanted that for a 'cinder force-detach' in case the Nova host disappears16:50
hemnawhere nova is dead and/or nova doesn't know anything about the attachment or the instance anymore.16:50
johnthetubaguyyeah, it feels the best way to go16:51
scottdaSo, given that we could make changes for Cinder to store more info, i.e. intiator, target, everything, it is possible for Cinder to have enough knowledge to make the decision as to when to detach. Could be done in Nova, or in Cinder16:51
*** mragupat_ has joined #openstack-cinder16:51
*** jdurgin1 has joined #openstack-cinder16:52
hemnascottda, in the case of multiattach detach16:52
hemnanova has to decide up front to call disconnect_volume16:52
johnthetubaguythink two nova's sharing a single cinder16:52
hemnaCinder's terminate_connection is always after that.16:52
hemnanova needs to decide to call it or not.16:52
hemnaand I think the only way for Nova to determine that is my algorithm described above.16:53
johnthetubaguyhemna: we can call it, and it be a no op though, right?16:53
scottdahemna: For certain? I'm just thinking of the possibilities here, not trying to make a decsion.16:53
hemnajohnthetubaguy, what specifically do you mean by 'it' in that case ?16:53
hemnaI just want to be clear.16:53
*** sinese has joined #openstack-cinder16:54
*** mragupat has quit IRC16:54
johnthetubaguyhemna: I guess I mean terminate_connection?16:55
*** bvivek_ has quit IRC16:55
*** bvivek has quit IRC16:55
hemnayah that depends on the driver.16:55
hemnathe important part is to call disconnect_volume or not.16:55
johnthetubaguyhemna: thats fine, if we know it might be needed, cinder can deal with the if anything needs to happen, I think?16:55
johnthetubaguyit feels bad that the API user needs to know what driver is being used, in general16:56
*** xyang1 has quit IRC16:57
hemnaI think it's safe for nova to call terminate_connection16:57
hemnathe cinder driver will have to cope with it.16:58
hemnabut if the target is gone from the host (disconnect_volume), there isn't really a way to recover that.16:58
hemnaa scsi rescan might recover an iSCSI/FC volume16:58
ildikovI would leave it on Cinder side to decide about detach if possible16:58
hemnabut by then, it's too late, as i/o will have been borked.16:58
ildikovotherwise we need to store too much info in Nova to decide16:59
ildikovhemna: disconnect_volume removes the initiator, right?17:00
hemnais there a BDM/DB API in nova to find out which hosts instances are on ?17:00
scottdahemna: I think the idea we're trying to flesh out is, could Cinder store enough info to choose when to remove the target, regardless of when/how_many_times Nova calls disconnect_volume ?17:00
hemnaI could probably throw together a WIP on nova to hack this17:00
hemnathis = my algorithm I just came up with.17:00
johnthetubaguyI don't getting nova to store more data is a good idea17:00
*** jgriffith is now known as jgriffith_away17:00
*** haomaiwang has quit IRC17:01
hemnascottda, well, if we had the host in the attachment on cinder, Cinder has what it needs.17:01
hemnajohnthetubaguy, I'm not asking nova to store any more than it already has17:01
*** haomaiwang has joined #openstack-cinder17:01
scottdaI think we should list the possible solutions, pros and cons, and then start an ML list to try to get some consensus. It seems we can easily get stuck on "Nova should figure it out" vs. "Cinder should figure it out"17:01
ildikovjohnthetubaguy: I'm on the same page as you, I wouldn't want to put this on Nova either17:02
johnthetubaguyscottda: yeah, I can't see the full picture, thats the issue here17:02
*** permalac has quit IRC17:02
hemnaI'm pretty sure what I have in mind would work17:02
ildikovscottda: +117:02
scottdahemna: I believe it will work. But there are other possibilities.17:02
johnthetubaguyits not about it working right, its about having something thats going to be easy to maintain over time, across upgrades, etc17:02
ildikovhemna: it's not a question we can hack this out :)17:02
scottdaAnd if we don't discuss all the possibilities, they will come up in review.17:03
ildikovhemna: but it would be good to find something let's say future proof, where Im not saying that this algorithm you're describing is not that, but at least myself I cannot decide at this point17:03
hemnaildikov, sure.  I'm not sure this wouldn't be17:03
hemnabut without trying it, I'm not sure what else to do.17:04
*** jgriffith_away is now known as jgriffith17:04
*** apoorvad has joined #openstack-cinder17:04
*** salv-orl_ has quit IRC17:05
johnthetubaguyfor me, with a nova hat on, and also with my operator hat, I want simple interfaces that generally don't change much, between the components, that are easy to debug, i.e. no matter what the driver is, I really want the same set of API calls made.17:05
*** salv-orlando has joined #openstack-cinder17:05
hemnayah, agreed17:05
ildikovhemna: isn't there any way to figure out how many attachments are using the same target?17:05
openstackgerritAdriano Freires Rosso proposed openstack/cinder: HNAS driver: Fix SSH and cluster_admin_ip0 bug
hemnaI think what I have in mind provides that17:06
ildikovhemna: same for initiators I guess17:06
hemnait's just a matter of hacking it out and seeing what you guys think.17:06
scottdaSo, are 2 of the possibilities that we have now: #1 hemna idea #2 Change Nova to pass host + attachement to Cinder (recently backed out and needs fixing) and have Cinder make the decision17:06
hemnaildikov, that's exactly what my algorithm will do (pretty sure)17:06
hemnaildikov, it compares the connection_info between 2 attaches17:06
hemnaif they are the same, then it's a shared target17:06
ildikovbut his happens on Nova side, or?17:07
hemnascottda, I don't think #2 helps multiattach detach case17:07
hemnait helps other things17:07
hemnaand is still needed.17:07
ildikovscottda: when we were talking about storing target/attachment info in the db was that #2?17:07
scottdaildikov: No, I think that was some additional info that Cinder would need to start keeping.17:08
ildikovhemna: so when you say this compare it is Nova, who would do this, right?17:08
hemnanova would do:17:08
hemna1) if multiattach volume17:08
hemna2) get list of attachments on current host17:08
hemna3) call cinder's initialize_connection on each of those17:08
ildikovscottda: so if that's possible that would be option #3, where we still need to figure out how not to remove the initiator from the host17:09
hemna4) compare connection_info17:09
hemnaif connection_info is same, then it's a shared target.17:09
hemnaelse, unique target17:09
ildikovwhat is initialize_connection for exactly?17:09
hemnaif it's a shared target, don't call disconnect_volume.17:09
hemnaif it's unique, call disconnect_volume17:09
hemnaildikov, initialize_connection returns the target information from a cinder backend17:10
ildikovas reading this list if I wouldn't be in context it would be pretty much misleading why we call initialize_connection for a detach17:10
hemnathe target portal, iqn, LUN id of a target (for iSCSI)17:10
ildikovyeah, got it17:10
hemnathose 3 things make up the unique target17:10
ildikovalthough the name is misleading in this case17:10
hemnaildikov, we call initialize_connection in detach for live migration :)17:10
hemnawe = nova17:10
hemnato make sure we have the correct information to detach on the source side, after the volume has been attached on the destination host.17:11
ildikovNova manages the volume attachment stuff in a pre_live_migration function and it is exactly what I had in mind17:11
*** esker has quit IRC17:11
ildikovhemna: remember when we were wondering how Nova manages live_migration?17:12
ildikovhemna: if not necessary I wouldn't want to make anyone having the same circles :)17:12
hemnathis is similar17:12
hemnait's just a decision nova has to make to call disconnect_volume or not.17:12
ildikovyeah, but when someone tries to figure out the code and change something things are getting messy :(17:13
scottdaIt seems the sticky issue is that Nova host needs to remove the initiator, and only Nova can do that. So Nova will have to figure out when this needs to be done, right?17:14
*** dims has quit IRC17:14
hemnaildikov, comments ?17:14
scottdaildikov: johnthetubaguy ^^^17:14
hemna# we do this in the case of multiattach to make sure we don't remove a volume that's being used by other instances.17:14
*** edtubill has quit IRC17:14
*** asselin has joined #openstack-cinder17:14
hemna# change this or die.17:15
ildikovscottda: for the initiator that's unfortunately right :(17:15
johnthetubaguyscottda: cinder could tell us if that initiator is needed, if we tell you which host the volume is on, I guess?17:15
*** asselin__ has joined #openstack-cinder17:15
hemnascottda, not the initiator, but the attached volume devices.17:15
scottdajohnthetubaguy: That is possible, but currently volume_api.detach and volume_api.terminate_connection are asynchronouss17:15
ildikovjohnthetubaguy: would that work with the order of the calls?17:15
ildikovscottda: we have a disconnect_volume call that goes to os-brick17:16
*** xyang1 has joined #openstack-cinder17:16
ildikovscottda: IIRC that's actually a third call here17:16
scottdaildikov: Yes, but Nova calls disconnect_volume. hemna Could Cinder call Brick.disconnect_volume?17:17
scottdajohnthetubaguy: I guess we could add a Cinder API that could tell whether an initiator is needed, given the proper info....17:18
hemnascottda, no17:18
hemnaildikov, that call to disconnect_volume is the main problem, of which I think I have a solution.17:18
*** csky has joined #openstack-cinder17:18
ildikovscottda: yeah, I meant that way that Nova calls it, but I'm not sure it calls it late enough to wait for Cinder to confirm it can be called17:18
hemnaildikov, nova calls disconnect_volume first17:19
hemnathen calls terminate_connection, and detach17:19
scottdahemna: I think we can agree that you have a solution. I think it does not hurt to take some time to consider the alternatives, if there are any.17:19
hemnaheh sure17:19
scottdaI don't believe any Nova changes can land in Mitaka at this point. Am I right, ildikov ?17:19
ildikovscottda: no more code in Nova for Mitaka17:20
*** leeantho has joined #openstack-cinder17:20
johnthetubaguybug fixes are fine17:20
johnthetubaguybut thats a fine line17:21
ildikovjohnthetubaguy: pre-bug fixes also? :)17:21
scottdaWell, since multi-attach won't be turned on in the Nova API until Newton, there is need to think about any of this for Mitaka, bug fix or not.17:21
hemnaisn't multi-attach disable though in Nova for M?17:21
ildikovit is not enable in Nova yet17:23
scottdaThe only reason to have any haste is if there were some changes on Cinder that were simple and safe, such as storing more info in the DB. Not possible if there's an API change (such as the reverted patch to pass host+instance_id) but if Cinder already has some important info, like connector info during initialize_connection call...17:23
openstackgerritPetrut Lucian proposed openstack/os-brick: [WIP] Add Windows connectors
ildikovthe question is what need to do on Cinder side to fix detach and will need to be done on Nova side now or at the beginning of Newton17:23
*** mudassirlatif has joined #openstack-cinder17:24
ildikovscottda: right and even the host info if we can have microversion support for it17:24
hemnascottda, I think the stumbling block I had with storing the connector was finding the right attachment at initialize_connection time.  we don't have the instance_uuid at initialize_connection time.17:24
*** sinese has quit IRC17:24
hemnaand also there is no cinder API for updating the attachment, which is needed due to live migration.17:25
hemnathis is what prevented me from adding this at the start of M17:25
openstackgerritCory Stone proposed openstack/cinder: Fix dynamic import of CONF.volume_api_class
scottdaOK, so we need that update_attachment API and it's almost certainly not going to get in until Newton, yes?17:26
scottdaWhat's the next step? We've hemna proposal (which sounds viable and may be the right solution). We could stand to list alternatives, pros and cons, etc. Are we ready to take this to the ML, or should we find a time and venue to discuss further?17:27
*** lpetrut has quit IRC17:28
hemnaI can write up a ML post about my proposal for nova trying to figure out if it should call disconnect_volume17:28
hemnaI think that's 1 issue to hash out.17:28
ildikovyeah, that's definitely half of the issues17:29
scottdaOK, thanks. We can add any other options to the post. Sound good ildikov ?17:29
ildikovI'm not sure how hard that is to figure out the attachments on the same host though...17:30
*** sinese has joined #openstack-cinder17:30
hemnaildikov, on the surface I don't think it's difficult17:30
ildikovscottda: if hemna writes up his either of us can chime in to cover the whole chain17:30
*** mylu has joined #openstack-cinder17:30
hemnaI need a way to find out if an instance is on a host.17:30
hemnain nova17:31
ildikovyeah, so we need to go through the list of attachments the volume has in the volume info and check17:31
hemnathat's the only piece i don't know how to do in nova17:31
hemnaI'll have an instance_uuid17:31
ildikovI don't know how much this will be racy or not17:31
hemnais that on this host or not.17:31
*** dims has joined #openstack-cinder17:32
ildikovhemna: I don't know that either17:32
hemnaok off to a beating....bbiab.17:32
ildikovjohnthetubaguy: we have a basic question ^^ :)17:32
johnthetubaguyit could be very racey I suspect17:33
johnthetubaguybut not 100% sure17:33
ildikovjohnthetubaguy: you mean the process to identify whether we have two attachments from a list on the same host, right?17:33
*** dulek has joined #openstack-cinder17:34
johnthetubaguyildikov: yeah, not sure there is a good non-racey way of getting that our side17:34
ildikovjohnthetubaguy: yeah, seems tricky17:35
*** vgridnev has joined #openstack-cinder17:35
ildikovjohnthetubaguy: although waiting for info from Cinder regarding call disconnect_volume or not does not seem any better either17:35
*** porrua has quit IRC17:38
sheelDear All,17:39
sheelPlease find some time to review
*** mvk has quit IRC17:39
jgriffithpatrickeast: ping17:39
patrickeastjgriffith: hey17:40
jgriffithpatrickeast: hey17:40
*** IlyaG has quit IRC17:40
jgriffithpatrickeast: so wanted to talk through some of the comments on the rep patch if you have a minute?17:40
patrickeastjgriffith: sounds good, I've got time but am on mobile so slow to type responses ;)17:41
jgriffithpatrickeast: ahh... wait, I think I just found the answer to my own question :)17:41
*** sgotliv has quit IRC17:41
jgriffithpatrickeast: haha17:41
jgriffithpatrickeast: don't torture yourself :)  I think I found my answer in your response to jungleboyj 's comment anyway17:41
jgriffithpatrickeast: I'll get another (maybe final) patch up today and we'll go from there17:41
jgriffithSwanson: You on the other hand :)17:42
Swansonjgriffith, Me?17:42
jgriffithSwanson: yes.. *you*17:42
jgriffithSwanson: To answer your question, *yes* the failover will need to return the backend-id we failed over to in the case of the driver picking17:42
jgriffithSwanson: I just hadn't gotten that far yet17:43
Swansonjgriffith, does that come back to the driver?  (And I just coded that moments ago.)17:43
jgriffithSwanson: so we want the driver to return that info... because then the manager will need to update that DB record17:43
*** boris-42 has quit IRC17:43
jgriffithSwanson: I'll get that code updated today17:44
Swansonjgriffith, Yep.  So after I've failed over (which is just me going "Is that one alive? Yep, return that id.") I need that ID back.  Do I have to poke the DB or is that available on init or something?17:44
Swansonjgriffith, and once I recognized this is post meteor failover this became way way simpler.  :)17:45
*** e0ne has quit IRC17:46
patrickeastSwanson: you may wanna look at the wip change I put up for pures driver.. I took some guesses on those based on speculation17:46
patrickeastLike assuming on init we get the reply status and I'd for the backend17:46
Swansonpatrickeast, oh, I see that up at the top of the review.17:47
patrickeastSome of that stuff we talked about at the mid cycle, but it wasn't implemted yet in the code17:47
*** edtubill has joined #openstack-cinder17:47
akerrsmcginnis: when you get a free moment could you take a look at
Swansonpatrickeast, Looks like you implemented what I was expecting.  So, cool.  Works for me.17:48
patrickeastHaha, just noticed android helped "correct" repl to reply and id to  I'd17:50
*** vincent_hou has quit IRC17:51
Swansonpatrickeast, I saw that.  I'm not sure we need the status.  If an ID is set I would assume a failover has occurred.17:51
patrickeastYea, I use it to know if I should try to setup replication on init, if its failed over I won't try since the primary is assumed borked17:53
patrickeastNot 100% sure if we need to though17:53
*** asselin__ has left #openstack-cinder17:56
*** IlyaG has joined #openstack-cinder17:58
*** haomaiwang has quit IRC18:01
*** haomaiwang has joined #openstack-cinder18:01
*** vivekd_ has joined #openstack-cinder18:03
*** chhavi has quit IRC18:04
*** vivekd has quit IRC18:05
*** jwcroppe has quit IRC18:05
*** vivekd_ is now known as vivekd18:05
*** dulek has quit IRC18:07
jgriffithpatrickeast: so you're assumption WRT return from replication_failover was 100% correct, Swanson see in the doc-string the returns param18:07
jgriffithpatrickeast: also, YES we need to have the backend_id passed in to the driver on init.18:07
jgriffithpatrickeast: otherwise driver won't know "what" to do18:07
*** funeutron has joined #openstack-cinder18:08
*** shyama has quit IRC18:08
*** krtaylor has quit IRC18:08
*** jdurgin1 has quit IRC18:12
*** porrua has joined #openstack-cinder18:13
openstackgerritAnthony Lee proposed openstack/os-brick: Fix output returned from get_all_available_volumes
*** ociuhandu has quit IRC18:16
*** ircuser-1 has joined #openstack-cinder18:16
*** ociuhandu has joined #openstack-cinder18:17
*** jwang_ has quit IRC18:17
*** jwang has joined #openstack-cinder18:18
*** krtaylor has joined #openstack-cinder18:20
*** diablo_rojo has quit IRC18:21
sheeljungleboyj: hi there...18:23
sheeljungleboyj: need some eyes on for workflow18:23
*** mylu has quit IRC18:24
openstackgerritAlex O'Rourke proposed openstack/cinder: 3PAR: Create consistency group from source CG
*** vincent_hou has joined #openstack-cinder18:26
*** vincent_hou has quit IRC18:26
*** liverpooler has quit IRC18:28
openstackgerritVivek Dhayaal proposed openstack/cinder: Support ZeroMQ messaging in cinder multibackend
jungleboyjsheel: Ok, will take a look.18:29
sheeljungleboyj: thank you ..18:30
openstackgerritAlex O'Rourke proposed openstack/cinder: Remove old client version checks from 3PAR driver
openstackgerritAlex O'Rourke proposed openstack/cinder: 3PAR: Create consistency group from source CG
jungleboyjsheel: Welcome.18:31
*** mylu has joined #openstack-cinder18:32
*** mylu has quit IRC18:33
*** savihou has joined #openstack-cinder18:34
*** diablo_rojo has joined #openstack-cinder18:39
*** bardia has joined #openstack-cinder18:39
krotscheckAnyone around to refresh a +A on ?18:42
krotschecksmcginnis ^^ ? :)18:42
smcginniskrotscheck: Got it.18:43
krotschecksmcginnis: Awesome, thanks!18:43
*** e0ne has joined #openstack-cinder18:44
e0nehemna: hi. are you around?18:47
*** daneyon has joined #openstack-cinder18:47
hemnawhat's up18:47
openstackgerritAdriano Freires Rosso proposed openstack/cinder: HNAS driver: Fix SSH and cluster_admin_ip0 bug
e0nehemna: few questions according to
e0nehemna: I'm agree that your comments are reasonable18:48
hemnahrmm, that crap18:48
mtaninojgriffith: Hi,18:48
*** lpetrut has joined #openstack-cinder18:48
*** mylu has joined #openstack-cinder18:48
hemnaso, I don't know why they need to do that TBH18:48
e0neso, we've got 2 options (at least)18:48
jgriffithmtanino: hey ya18:48
hemnaI'm not an XP expert18:48
mtaninojgriffith: I got a feedback from Gorka for this.
e0nehemna: erlon confirmed that they are not needed for his driver too18:49
e0neif we want to leave these methods, we need to:18:50
e0nea) move code back from manager to driver:(18:50
jgriffithmtanino: yeah, so the only problem I have with Gorka's approach is the backport problem18:50
mtaninojgriffith: use DB data or use external saved data. Which one do you recommend?18:50
e0neor b) create something like def _before_data_copy and def _after_data_copy in drivers18:50
mtaninojgriffith: I think so, and from my glance of the code, we need to fix for Nova side to store initiator on DB.18:51
hemnae0ne, both of those kinda suck18:51
erlone0ne: are those the only drivers with the code left?18:51
hemnabut I tend to like b) better.18:51
e0nehemna: +118:51
e0neerlon: yes, only 2 drivers are affected18:52
jgriffithmtanino: not sure I follow the problem on the initiator side of it?18:52
openstackgerritAdriano Freires Rosso proposed openstack/cinder: HNAS driver: Fix SSH and cluster_admin_ip0 bug
erlone0ne: for the Hitachi drivers I have double checked and the code is not touched.  CI was passed as well18:52
e0nethe same for HP - this code is not executed at all18:53
jgriffithmtanino: so the most correct answer WRT targets is IMO to use the DB18:53
jgriffithmtanino: and then just do ensure_exports18:54
e0nethe questio is: is that code needed?18:54
jgriffithmtanino: BUT18:54
jgriffithI hate trying to backport db updates :)18:54
mtaninojgriffith: ok, so you recommend to use DB data to recreate iSCSI target.18:54
mtaninojgriffith: but?18:54
jgriffithmtanino: I'm looking... :)18:54
jgriffithNeed to refresh my memory a bit here :)18:54
mtaninojgriffith: sure. my concern is how to store initiator info to DB...18:55
erlone0ne: though CI results are not relevant in these patches,  IIRC, the copy_volume_data was used during migration, which does not have tests in tempest18:55
jgriffithmtanino: well, right now we can't :)18:55
*** daneyon_ has joined #openstack-cinder18:55
mtaninojgriffith: It's passed from Nova's attach call.18:55
e0neerlon, hemna: ok, let's try "option b"18:56
jgriffithmtanino: unless there's a hook in LIO that does it that I don't remember/know about18:56
*** chris_morrell has joined #openstack-cinder18:56
jgriffithmtanino: in "older" versions as well?18:56
jgriffithyes, that's right... that was a requirement for LIO18:56
erlone0ne: I didnt get the point of this option18:56
jgriffithin order to build the access list18:57
mtaninojgriffith: I think so.18:57
jgriffithmtanino: yes, you're correct.18:57
e0neerlon: make this code executable again18:57
jgriffithmtanino: so I'm not familiar with rtstool's "restore" command18:58
jgriffithmtanino: how does that work exactly?18:58
mtaninojgriffith: That's option restore iSCSI target from saved file.18:58
*** daneyon has quit IRC18:58
*** salv-orl_ has joined #openstack-cinder18:58
jgriffithmtanino: ok, so just like tgtadm... it has it's own persistence file?18:59
erlone0ne: hmm, that looks bad, you have a code that for some reason, stopped being called, the driver keeps, without anyone noticing that, for 1/2 releases, then you start calling that again18:59
mtaninojgriffith: Current LIO target save current configuration to persistenfile when we change someting to target configuration.18:59
mtaninojgriffith: yes, I think it's same as current tgtd behaviour.19:00
jgriffithI think that's fine to be honest19:00
*** haomaiwang has quit IRC19:01
erlone0ne: the code lost the meaning after the changes19:01
mtaninojgriffith: like this.
*** nkrinner has quit IRC19:01
jgriffithmtanino: yeah, so that seems fine to me19:01
erlone0ne:  hemna: i think vincent_hou did that in the migration refactor patches19:01
*** haomaiwa_ has joined #openstack-cinder19:01
*** salv-orlando has quit IRC19:01
jgriffithmtanino: I'm not completely clear why Gorka prefers a DB entry on this one19:01
e0neerlon, hemna: ok. now I see only one solution: get CI working to test  my patch19:02
erlone0ne: hemna: may be he could better explay why he did so19:02
e0neit was jbernard's patch:
mtaninojgriffith: So I guess there is a concern that someone changes the external saved file or the file is not correctly saved or etc?19:03
eharneythe external LIO configuration being saved or not depends on how target.service is setup on the system... it may reload things at reboot or may not19:03
*** EinstCrazy has joined #openstack-cinder19:04
jgriffitheharney: ok, that's fair.. but that sort of thing seems like a problem with "any" config :)19:04
jgriffitheharney: mtanino My vote was that the patch proposed is fine IMO, and good for backports, with the caveat that there might be a better long-term fix by adding the db entry19:05
jgriffitheharney: mtanino I updated in gerritt19:06
mtaninojgriffith: Thank you. I am going to talk Gorka again.19:06
jgriffithMaybe I'm completely wrong, but seems reasonable to make fwd progress19:06
*** EinstCrazy has quit IRC19:09
e0nehemna, smcginnis: who is maintainer of HPE XP driver?19:10
e0neDuncanT: ^^19:10
e0neI didn' see CI for it and it looks broken19:10
*** dramakri has joined #openstack-cinder19:10
erlone0ne: hard to know if that (jbernard) broke migration on those drivers or others. We needed migration tests running.19:10
smcginnis^^ That says asselin_19:11
asselinwhat's up?19:12
erlonsmcginnis: the CI is running, but there's no migration tests19:12
e0nesmcginnis: thanks19:12
* asselin updates19:12
smcginnisasselin: Questions on the XP CI. Are you still maintaining that?19:12
erlonsmcginnis: the migration tests are -2, becouse tempest guys want to see that running in CIs before they approve19:12
e0neasselin: IMO, HPEXPFCDriver is broken and CI is not reporting:(19:13
smcginnisasselin: This needs to be updated then:
asselinyup updating now19:13
jbernarde0ne: eh?19:13
smcginniserlon: We've had tempest issues like that before.19:13
e0ne - it imports class which is not exists19:14
hemnae0ne, nestor fernandez is the mx for the XP driver19:15
hemnae0ne, I emailed him earlier asking him to join the cinder channel19:15
hemnahe's supposed to be in here daily.....fwiw.19:15
e0nehemna: thanks19:15
erlonsmcginnis: you mean tests that tempest team does not approve?19:15
smcginniserlon: Yeah.19:15
asselinyeah, you should be able to contact him via that e-mail address19:16
smcginniserlon: After discussion at the midcycle I started looking into adding in-tree tempest tests to get around this.19:16
erlonsmcginnis: it looks like the chicken/egg problem  :/19:16
smcginniserlon: But they have been even less supportive on that as well.19:16
e0nefernnest: hi19:16
*** jgriffith is now known as jgriffith_away19:16
fernneste0ne, hi19:17
e0neI've got a question about xp-cinder-ci19:17
erlonjbernard: do you recall this patch?
fernneste0ne, yes?19:17
asselinfernnest, hi, please add your irc handle here in Contact Information:
erlonjbernard: do you remember what backends you have tested the migration?19:17
jbernardof course, i wrote it19:18
e0nefernnest: I didn't see any results from that CI19:18
e0nefernnest: and driver seems to be broken19:18
jbernarde0ne: hmm, give me a sec19:18
fernnestasselin, will do19:18
jbernarderlon: i made notes at some point19:18
jbernarderlon: ah yes, i put those notes in the commit message19:19
jbernarderlon: what is the problem?19:20
erlonjbernard: hmm, ok, the problem is that you remove the call for driver.copy_volume_data(), but, today still are some drivers that uses that19:21
e0nefernnest: I don't understan how it can work19:21
jbernardoh, we should discuss this19:22
*** mylu has quit IRC19:22
fernnest__e0ne, lemme look19:22
e0neok, waiting19:22
openstackgerritxing-yang proposed openstack/cinder: Update db in CGSnapshot create
erlonjbernard: what seems to me is that, your code does the same thing, copy the volume data, but it removes the ability of a driver to do any optiomization19:24
jbernarderlon: how so exactly?"19:24
erlonjbernard: ??19:25
fernnest__e0ne, there are files that are pulled from github that implement that module.19:26
jbernardthe patch is not particularly intrusive, the existing flow of execution remains for existing driver, and drivers the return an file object fall over to file-based copy19:26
e0nefernnest__: could you please give me a link?19:26
jbernarddriver-optimized migration is untouched19:26
fernnest__e0ne, one sec19:26
erlonjbernard: which is not done in case of Hitachi driver, it only calls, the default function defined in baseVD driver19:27
hemnaPure FC CI failed in 7 seconds19:27
e0nefaill fast:)19:27
hemnafail fast, and fail often.19:27
fernnest__e0ne, git clone
hemnapatrickeast, ping19:28
hemnapatrickeast, error: internal error: unable to execute QEMU command 'device_add': Device initialization failed.19:28
hemnaHBA is probably Offline is my guess19:29
patrickeastTry a recheck, I'll take a look in a bit19:29
e0nefernnest__: why it is not in cinder tree or at least in pypi?19:29
hemnaI'm just baby sitting a bunch of os-brick stuff right now to try and get them to land this week19:29
fernnest__e0ne, the developer for this is Hitachi, I've been trying to get them to submit the hidden files into the cinder tree since day 1.19:31
erlonfernnest__: you mean the code in github?19:32
*** akshai has joined #openstack-cinder19:32
e0nefernnest__: it's crazy19:35
*** esker has joined #openstack-cinder19:37
e0nefernnest__, smcginnis: in such case, we've got a driver in cinder which has not stable ci and kind of "hidden" code19:37
*** esker has quit IRC19:37
e0nefernnest__: how can I find that github repo? where is it documented?19:37
*** esker has joined #openstack-cinder19:38
fernnest__e0ne, here's the story, XP is OEM'ed Hitachi array.  Hitachi provides the driver.  We asked them to get all the files in the cinder tree. They are doing first the HSBD versions and then copy the files to the HPE_XP dir19:39
*** mylu has joined #openstack-cinder19:39
smcginnise0ne, fernnest__: The code issue isn't ideal, but the biggest thing is CI.19:41
smcginnise0ne, fernnest__: Is we don't have a good CI reporting results, we will need to pull the driver.19:41
smcginnis*If we19:41
fernnest__smcginnis,e0ne: I see the CI running correctly19:41
*** mylu_ has joined #openstack-cinder19:41
fernnest__smcginnis,e0ne: wheree are you looking for failed CI?19:42
smcginnisOh, that "refactor" patch. :|19:42
fernnest__smcginnis,e0ne: yup19:42
*** mylu has quit IRC19:42
e0nefernnest__: I've found you CI results now19:42
fernnest__smcginnis,e0ne: they want to get that in so that they can then turn around and do another submit of *nearly* identical changes19:43
fernnest__smcginnis,e0ne: the only changes would be the parameter names and file names.19:43
fernnest__e0ne, where are you looking?19:43
smcginnisfernnest__: If they are just wholesale copying the code and changing names, I would recommend doing what dothill did.19:43
fernnest__smcginnis:I'm all ears!19:44
* smcginnis looks for links...19:44
e0ne - only 5 results for the last 24 hours19:44
smcginnisfernnest__: They have a few different OEMs.19:45
smcginnisfernnest__: So there is one full "real" driver.19:45
smcginnisfernnest__: Then a OEMd versions that inherit from the base one.19:45
smcginnisfernnest__: Makes the additional drivers really small:
fernnest__e0ne: it is slow19:45
e0nefernnest__: does it support migration tests?19:46
fernnest__e0ne: I'm working on a fix to CI env to speed things up.19:46
*** ociuhandu has quit IRC19:46
smcginnisfernnest__: The OEM thing caused some discussion in channel and during a meeting as to whether to allow that.19:46
smcginnisfernnest__: But in the end, the one common code base is better than a bunch of copy/pasted code.19:47
e0nesmcginnis: +119:47
fernnest__smcginnis: code shared by two drivers but not generic enough for all/most drivers would go where?19:47
smcginnisfernnest__: In this case I would probably have the "base" code in the hitachi folder, then have the XP driver just inherit from that driver.19:48
smcginnisfernnest__: Then override or add hooks for "branding" or any version specific changes for the HPE XP driver.19:48
smcginnisfernnest__: With the example above, DotHill is the base driver that has all functionality.19:49
smcginnisfernnest__: Then the Lenovo and other OEM'd versions of it just are very small ones that inherit from it.19:50
smcginnisfernnest__: So Lenovo and HP MSA are just OEMs of DotHill it appears:
fernnest__smcginnis: so HPE driver would import from HBSD directory the lower levels of the code.  Then just override whatever needs to be overridden (config parms, that sort of thing)19:51
smcginnisfernnest__: Yep, basically at a high level that should be it.19:51
smcginnisfernnest__: Always some finer details to work through, but then at least it's not just a full copy of all the code.19:52
smcginnisfernnest__: With the need to know to fix in one place if you fix in the other.19:52
smcginnis*With avoiding the need19:52
fernnest__smcginnis: big difference ;)19:52
*** sinese has quit IRC19:53
*** sinese_ has joined #openstack-cinder19:53
jbernarde0ne: pre/post routines would work19:53
*** salv-orl_ has quit IRC19:53
e0nejbernard: I'll try it in a next patch19:53
openstackgerritabhiram moturi proposed openstack/cinder: Zfssaiscsi driver should not use 'default' initiator group
fernnest__smcginnis: let me study the MSA/OEM/Lenovo code and convince Hitachi...Can I get back to you if/when I have more questions?19:53
*** kfarr has joined #openstack-cinder19:54
jbernarde0ne: it looks like hitachi needs only discard, so just post might be enough19:54
smcginnisfernnest__: Definitely!19:54
smcginnisfernnest__: Thanks for looking into it.19:54
fernnest__smcginnis: thanks, really appreciated.  Can we make the review of that patch incumbent on what we've discussed?  I need the leverage.19:54
smcginnisfernnest__: Whatever I can do to help, just let me know.19:55
smcginnisfernnest__: I'm behind anything to get rid of code duplication.19:55
fernnest__smcginnis: Concur, I'll go get busy then.19:55
smcginnisfernnest__: Awesome, thanks!19:56
*** tongli has quit IRC19:57
*** akshai has quit IRC19:57
*** haomaiwa_ has quit IRC20:01
*** haomaiwang has joined #openstack-cinder20:01
*** mylu_ has quit IRC20:02
*** akshai has joined #openstack-cinder20:02
*** vgridnev has quit IRC20:05
*** ociuhandu has joined #openstack-cinder20:05
*** jgriffith_away is now known as jgriffith20:13
*** cknight has quit IRC20:13
*** esp has joined #openstack-cinder20:14
*** cknight has joined #openstack-cinder20:15
*** mylu has joined #openstack-cinder20:20
jgriffithbswartz: ping20:21
*** esp has quit IRC20:21
*** akshai has quit IRC20:24
*** mylu has quit IRC20:26
*** salv-orlando has joined #openstack-cinder20:27
*** edtubill_ has joined #openstack-cinder20:28
mtaninosmcginnis: sorry for several times, could you revisit this metadata support?
smcginnismtanino: Sure, should be able to in a few.20:28
*** mylu has joined #openstack-cinder20:28
mtaninosmcginnis: I'd appreciate it.20:28
*** edtubill has quit IRC20:30
*** dramakri has quit IRC20:34
*** salv-orlando has quit IRC20:36
*** akshai has joined #openstack-cinder20:36
*** vivekd_ has joined #openstack-cinder20:37
*** mylu has quit IRC20:38
mtaninosmcginnis: Thank you. I am going to talk with DuncanT too, but usually I can't find him...20:38
mtaninotime zone ;)20:39
DuncanTmtanino: I'm here now for a short while, as it happens :-)20:39
mtaninoDuncanT: oh, you there!20:39
*** vivekd has quit IRC20:40
mtaninoDuncanT: so about a glance metadata support patch for Cinder,
*** vivekd_ is now known as vivekd20:40
mtaninoDuncanT: I approached all comments from you and other reviewers, so please revisit if you have a time.20:41
*** leeantho has quit IRC20:42
mtaninoDuncanT: And then, I'll push glance side patch20:43
*** leeantho has joined #openstack-cinder20:43
DuncanTmtanino: Sure, looking now20:43
DuncanTmtanino: I was rather worried that most of the CIs failed20:44
*** leeantho has quit IRC20:44
*** leeantho_ has joined #openstack-cinder20:44
mtaninoDuncanT: I see. Let me recheck vendor CIs20:45
*** lpetrut has quit IRC20:46
DuncanTmtanino: Lots seem to fail with "aborted: Block Device Mapping is Invalid" which looks suspicious20:46
mtaninoDuncanT: Which one are you reffering now?20:47
DuncanTsolidfire and two others at random20:49
*** leeantho_ has quit IRC20:49
DuncanTSome failed to build with mirror problems though, so it might be unrelated20:49
*** chris_morrell has quit IRC20:49
DuncanTSolidfire is usually pretty good though20:49
*** chris_morrell has joined #openstack-cinder20:49
DuncanTI'll leave the review window open anyway, and check again in the morning, see how the rechecks have gone20:50
jgriffithDuncanT: mtanino I just hit recheck on SF; there were some issues with boot from volume that week20:50
jgriffithDuncanT: mtanino in general20:50
*** dims_ has joined #openstack-cinder20:51
mtaninoDuncanT: I pushed rechck. I'm going to check CIs result again too. Thank you for the advice.20:51
mtaninojgriffith: thank you.20:51
DuncanTmtanino: If there's a reasonable number of passes in the morning, the patch itself looks good20:51
jgriffithDuncanT: mtanino You'll note if you look at the SF logs under n-cpu....20:52
jgriffithDuncanT: mtanino it's the problem with that incompatable change that was merged to Cinder and Nova for multi-attach20:52
*** dims has quit IRC20:52
jgriffithDuncanT: mtanino it's possible that you will need to rebase the cinder patch though20:52
mtaninoDuncanT: jgriffith Was the patch reverted?20:53
jgriffithDuncanT: mtanino BTW, here's the signature for that in the n-cpu logs:
DuncanTjgriffith: thanks. I didn't investigate beyond the console log20:54
mtaninojgriffith: ok Thank you20:54
*** leeantho has joined #openstack-cinder20:54
jgriffithmtanino: yeah... sadly some CI's are running a rebase during check phase, while others (sos-ci) are not20:54
*** dramakri has joined #openstack-cinder20:54
jgriffithmtanino: so those that do a rebase will pass, and don't actually test compatability of things20:55
jgriffithmtanino: I don't know which is better or worse... but at least this pointed out that we were breaking compatbility for old-cinders and new Nova's20:55
mtaninojgriffith: DuncanT I see. If my patch still have some CIs issue, I'm gotin to rebase and try again.20:56
jgriffithmtanino: sounds good20:56
mtaninojgriffith: hmm, I'm not sure to.. which one is good..20:56
jgriffithmtanino: :)20:56
patrickeastjgriffith: isn't it only compatability between those tip of tree points as opposed to the whole version though? the one that does a merge is closer to what would actually be released... i think20:56
jgriffithmtanino: so I think it's good that it failed personally.  but bad that we allowed a change like that :)20:57
jgriffithpatrickeast: no, so that's the problem.20:57
jgriffithpatrickeast: if you have an "old" Cinder it will flat out fail20:57
jgriffithpatrickeast: this is the danger of the cross-project dependency model IMO20:57
openstackgerritMatan Sabag proposed openstack/cinder: Support for consistency groups in ScaleIO driver
patrickeastjgriffith: oh, like *any* old cinder and a M or newer nova20:57
patrickeastjgriffith: i see20:57
jgriffithpatrickeast: exactly20:57
jgriffithpatrickeast: so like, lord help you if you're upgrading in stages :)20:58
patrickeastjgriffith: stages? psh, who does that ;)20:58
DuncanTUpgrades? The old stuff works, nobody touch a thing!20:59
patrickeastseems like something we should have in our ci testing21:00
jgriffithpatrickeast: so that's the thing21:00
jgriffithpatrickeast: we used to NOT rebase in check I don't think21:00
*** haomaiwang has quit IRC21:01
patrickeastjgriffith: yea i think a while ago zuul didn't do it21:01
*** salv-orlando has joined #openstack-cinder21:01
patrickeastjgriffith: back when like recheck vs reverify was a thing, iiuc21:01
*** haomaiwang has joined #openstack-cinder21:01
openstackgerritabhiram moturi proposed openstack/cinder: Zfssaiscsi driver should not use 'default' initiator group
patrickeastwe kinda want both i guess21:02
jgriffithpatrickeast: exactly21:02
jgriffithpatrickeast: IMHO no rebase on check, and rebase of course on verify21:02
jgriffithbut whatever21:02
patrickeastyea, but thats a pain to find out that you would be blocked in gate so late21:03
jgriffitheven the guy who manages the "common CI" code doesn't know how it works :)21:03
patrickeastmakes sense why it does it ne check now21:03
*** mylu has joined #openstack-cinder21:03
patrickeastwe kind of want both in check and then merge on verify21:03
jgriffithpatrickeast: I'll stick with my SOS thank you very much :)21:03
*** mylu has quit IRC21:03
jgriffithpatrickeast: I can continue to be the canary in the coal mine :)21:04
patrickeasti dunno, the zuul code isn't too bad21:04
*** pots1 has quit IRC21:04
patrickeastits just the insane way things get configured that makes things mysterious21:04
jgriffithpatrickeast: nahh... I agree with you21:04
smcginnisjgriffith: Yeah, we've been trying to get common going here.21:04
smcginnisI'm happy with my scripts for now. :)21:04
jgriffithpatrickeast: the config and puppet sides are what drive me to drink21:04
patrickeastjgriffith: ditto21:04
openstackgerritMerged openstack/cinder: Allow for eradicating Pure volumes on Cinder delete
*** savihou has quit IRC21:09
*** jgregor has quit IRC21:10
*** pots has quit IRC21:11
*** angela-s has joined #openstack-cinder21:13
*** jgregor has joined #openstack-cinder21:13
openstackgerritScott DAngelo proposed openstack/cinder: cinder-api-microversions code
*** merooney has quit IRC21:13
openstackgerritMatan Sabag proposed openstack/cinder: Manage/unmanage volume in ScaleIO driver
*** sheel has quit IRC21:17
*** jgregor has quit IRC21:18
*** lpetrut has joined #openstack-cinder21:22
openstackgerritMatan Sabag proposed openstack/cinder: ScaleIO QoS Support
*** akerr has quit IRC21:23
*** jgregor has joined #openstack-cinder21:23
*** dramakri has quit IRC21:24
openstackgerritRyan McNair proposed openstack/cinder: WIP Fixup nested quotas to enforce child defaults
*** dustins has quit IRC21:24
*** jgregor has left #openstack-cinder21:25
*** jgregor has joined #openstack-cinder21:25
*** dims has joined #openstack-cinder21:26
*** dims_ has quit IRC21:29
*** dustins has joined #openstack-cinder21:30
openstackgerritHelen Walsh proposed openstack/cinder: VMAX-Replacing deprecated API EMCGetTargetEndpoints
angela-ssmcginnis, jgriffith: could you please take a look at this review?
angela-ssmcginnis, jgriffith: i had to abandon the original review as i could not resolve the conflicts after the parent review was merged21:32
smcginnisangela-s: Sure, I'll take a look.21:32
angela-sthank you!21:32
openstackgerritVincent Hou proposed openstack/cinder: Storwize: Implement v2 replication
*** mtanino has quit IRC21:34
*** baumann has left #openstack-cinder21:40
*** porrua has quit IRC21:46
openstackgerritAngela Smith proposed openstack/cinder: Adds support for configuring zoning in a virtual fabric
angela-ssmcginnis: thanks for catching that, argh!21:49
smcginnisangela-s: ;)21:49
*** sinese_ has quit IRC21:51
*** dustins has quit IRC21:53
*** krtaylor has quit IRC21:55
openstackgerritVincent Hou proposed openstack/cinder: Storwize: Implement v2 replication
*** dulek has joined #openstack-cinder21:55
*** boris-42 has joined #openstack-cinder21:58
*** haomaiwang has quit IRC22:01
*** haomaiwang has joined #openstack-cinder22:01
*** dramakri has joined #openstack-cinder22:02
*** timcl has quit IRC22:03
*** e0ne has quit IRC22:03
*** knikolla has quit IRC22:04
*** akshai has quit IRC22:05
*** krtaylor has joined #openstack-cinder22:07
*** cknight has quit IRC22:08
*** jgriffith is now known as jgriffith_away22:10
openstackgerritVivek Dhayaal proposed openstack/cinder: Removed potential races from volume update method
*** cknight has joined #openstack-cinder22:12
*** akshai has joined #openstack-cinder22:13
Swansoncgsnapshot question for anyone.  xyang1 maybe.  I'm calling objects.SnapshotList().get_all_for_cgsnapshot(context, snapshotid).  It looks like I no longer need to do this as those snapshots appear to be on the parameter list.  Correct?22:14
*** jgriffith_away is now known as jgriffith22:17
*** mylu has joined #openstack-cinder22:19
aorourkeSwanson, that is correct. The driver should not need to access the db at all now. Snapshots and volumes are passed into to cg functions as parameters now22:20
SwansonNevermind.  Actual docstrings farther down in
Swansonaorourke, Thanks for the confirmation.22:21
*** mylu has quit IRC22:21
*** esker has quit IRC22:22
*** diablo_rojo has quit IRC22:23
*** esker has joined #openstack-cinder22:23
*** krtaylor has quit IRC22:24
xyang1Swanson: all set?22:24
*** chris_morrell has quit IRC22:25
Swansonxyang1, yep!22:25
*** esp has joined #openstack-cinder22:25
*** dulek has quit IRC22:26
*** jungleboyj has quit IRC22:29
*** dims has quit IRC22:31
*** esp has quit IRC22:32
*** mylu has joined #openstack-cinder22:34
*** krtaylor has joined #openstack-cinder22:36
*** edtubill_ has quit IRC22:39
*** crose has quit IRC22:40
Swansonjgriffith, What exception for a replication_failover fail?22:40
jgriffithSwanson: depends :)22:40
jgriffithSwanson: so for the case of "replicaton not configured" I'm raising "invalidInput"22:41
jgriffithSwanson: in the case of the driver being hosed, it should return a "backenddriverexception"22:41
jgriffithwith an appropriate message22:41
Swansonjgriffith, I can roll with that.  Thanks!22:42
*** mylu has quit IRC22:42
*** mylu has joined #openstack-cinder22:44
*** jgregor has quit IRC22:46
*** dulek has joined #openstack-cinder22:48
*** rlrossit has quit IRC22:49
*** dulek has quit IRC22:49
*** merooney has joined #openstack-cinder22:49
angela-swho runs EMC XIO CI?  It's posted 'merge failed' comment 5 times in last 1 1/2 hrs to this review
*** merooney has quit IRC22:50
*** dims_ has joined #openstack-cinder22:51
aorourkeangela-s, it has also done this on and
eharneyangela-s: it did that on too.  not very useful.22:52
angela-sok, so it's gone crazy.  anyone know who the owner is?22:53
*** kfarr has quit IRC22:53
angela-seharney: yeah, seems like a bogus failure message to me22:53
eharneyIMO (again) CIs that post this are just making noise and should stop posting these results.22:54
eharneybut i've brought this up a few times and nobody seems to have a good idea for what to do about it22:54
xyang1angela-s: that's me22:54
xyang1angela-s: could be caused by log server full22:54
*** funeutron has quit IRC22:54
xyang1angela-s: we'll take a look22:55
eharneyxyang1: the problem is, the message it posts implies there's something wrong with the patchset22:55
angela-sxyang1: thanks for speaking up.22:55
eharneyor, not implies, but actually tells you to fix it22:55
xyang1eharney: I'll take a look22:55
eharneyxyang1: can you take a look at just not reporting when it hits that state?22:56
angela-sxyang1: and why does it keep posting?  is it doing automatic recheck?22:56
xyang1eharney: sure22:56
*** EinstCrazy has joined #openstack-cinder22:58
*** diablo_rojo has joined #openstack-cinder22:59
*** haomaiwang has quit IRC23:01
*** haomaiwang has joined #openstack-cinder23:01
*** mragupat_ has quit IRC23:02
*** esp has joined #openstack-cinder23:03
*** eharney has quit IRC23:04
patrickeasteharney: unfortunately I don't think zuul has that as an option (zuul-merger being the culpret behind those errors), although could probably be extended to do it23:04
jgriffithpatrickeast: you could use sos :)23:05
* patrickeast goes digging in
*** EinstCrazy has quit IRC23:05
jgriffithpatrickeast: seriously though, I can't imagine that not being configurable23:05
patrickeastjgriffith: nah, i'm way to committed to the infra tools23:05
jgriffithpatrickeast: seems that the CI's are maybe running the wrong config ?23:05
jgriffithpatrickeast: hehe... I hear ya23:06
patrickeastjgriffith: gotta keep driving business for my blossoming CI consulting23:06
*** mtanino has joined #openstack-cinder23:06
*** dan_nguyen has joined #openstack-cinder23:06
jgriffithrule #1 in OpenStack... "don't make it too easy, that's how you make money" :)23:06
patrickeastjgriffith: so the problem is that for regular jenkins, a merge failing is like a real problem23:06
patrickeastjgriffith: since it means the patchset would not be able to get committed and what not23:06
jgriffithpatrickeast: hmmm... I guess my issue is that it shouldn't be trying to merge anyway23:06
patrickeastjgriffith: haha yea23:07
patrickeastjgriffith: that may or may not be configurable23:07
jgriffithpatrickeast: in other words, that's a job for the gate-verify NOT the ci/check23:07
*** esp has quit IRC23:07
*** lpetrut has quit IRC23:08
xyang1patrickeast: how much are you charging for CI consulting?:)23:10
patrickeastxyang1: nothing yet23:10
*** vivekd has quit IRC23:12
openstackgerritKendall Nelson proposed openstack/os-brick: WIP: Lun id's > 255 should be converted to hex
*** dims_ has quit IRC23:14
*** salv-orlando has quit IRC23:14
*** gouthamr has quit IRC23:15
*** gouthamr has joined #openstack-cinder23:15
*** gouthamr has quit IRC23:15
*** salv-orlando has joined #openstack-cinder23:15
*** dramakri has quit IRC23:15
*** dims_ has joined #openstack-cinder23:19
*** diablo_rojo has quit IRC23:26
*** IlyaG has quit IRC23:27
*** dramakri has joined #openstack-cinder23:27
*** IlyaG has joined #openstack-cinder23:28
*** can8dnSix has joined #openstack-cinder23:36
*** liewegas has quit IRC23:39
*** lcurtis has quit IRC23:39
*** liewegas has joined #openstack-cinder23:40
*** smoriya_ has joined #openstack-cinder23:43
*** esker has quit IRC23:43
*** yhayashi has joined #openstack-cinder23:48
*** can8dnSix has quit IRC23:48
*** akshai has quit IRC23:58

Generated by 2.14.0 by Marius Gedminas - find it at!