| *** knikolla has quit IRC | 00:06 | |
| *** salv-orlando has joined #openstack-cinder | 00:09 | |
| *** diogogmt has joined #openstack-cinder | 00:09 | |
| *** knikolla has joined #openstack-cinder | 00:09 | |
| *** netapp has quit IRC | 00:10 | |
| *** netapp has joined #openstack-cinder | 00:11 | |
| *** salv-orlando has quit IRC | 00:12 | |
| *** laughterwym has joined #openstack-cinder | 00:17 | |
| *** jamielennox|away is now known as jamielennox | 00:20 | |
| *** raunak has joined #openstack-cinder | 00:22 | |
| *** laughter_ has joined #openstack-cinder | 00:23 | |
| *** laughterwym has quit IRC | 00:24 | |
| *** laughter_ has quit IRC | 00:24 | |
| *** ducttape_ has joined #openstack-cinder | 00:32 | |
| *** gouthamr has quit IRC | 00:34 | |
| *** darosale has joined #openstack-cinder | 00:34 | |
| *** laughterwym has joined #openstack-cinder | 00:35 | |
| *** gouthamr has joined #openstack-cinder | 00:35 | |
| *** takedakn has joined #openstack-cinder | 00:36 | |
| *** takedakn has quit IRC | 00:36 | |
| *** Julien-zte has joined #openstack-cinder | 00:49 | |
| *** takedakn has joined #openstack-cinder | 00:54 | |
| *** takedakn has quit IRC | 00:54 | |
| *** Julien-zte has quit IRC | 00:56 | |
| *** cheneydc has joined #openstack-cinder | 01:00 | |
| *** caoshufeng has joined #openstack-cinder | 01:08 | |
| caoshufeng | Good morning, all. | 01:08 |
|---|---|---|
| *** bardia has joined #openstack-cinder | 01:10 | |
| *** salv-orlando has joined #openstack-cinder | 01:12 | |
| *** takedakn has joined #openstack-cinder | 01:14 | |
| *** takedakn has quit IRC | 01:15 | |
| *** takedakn has joined #openstack-cinder | 01:16 | |
| openstackgerrit | Merged openstack/os-brick: Updated from global requirements https://review.openstack.org/322545 | 01:16 |
| *** salv-orlando has quit IRC | 01:17 | |
| *** takedakn has quit IRC | 01:17 | |
| *** takedakn has joined #openstack-cinder | 01:17 | |
| *** takedakn has quit IRC | 01:18 | |
| *** takedakn has joined #openstack-cinder | 01:18 | |
| openstackgerrit | Merged openstack/cinder: Updated from global requirements https://review.openstack.org/322982 | 01:19 |
| *** takedakn has quit IRC | 01:20 | |
| openstackgerrit | Merged openstack/os-brick: Mock time.sleep in ISCSIConnectorTestCase https://review.openstack.org/322821 | 01:21 |
| *** EinstCrazy has joined #openstack-cinder | 01:25 | |
| openstackgerrit | yangweiwei proposed openstack/cinder: Update the status of backup-restore https://review.openstack.org/275542 | 01:31 |
| *** Wenjun has joined #openstack-cinder | 01:34 | |
| *** gouthamr_ has joined #openstack-cinder | 01:35 | |
| *** gouthamr has quit IRC | 01:36 | |
| *** julim has quit IRC | 01:44 | |
| *** mtanino has quit IRC | 01:49 | |
| openstackgerrit | Deepti Ramakrishna proposed openstack/cinder-specs: New public core API to expose system capabilities https://review.openstack.org/306930 | 01:50 |
| *** markvoelker has joined #openstack-cinder | 01:51 | |
| *** Lee1092 has joined #openstack-cinder | 01:55 | |
| *** markvoelker has quit IRC | 01:55 | |
| openstackgerrit | Wilson Liu proposed openstack/cinder: Add unit tests for Huawei drivers https://review.openstack.org/283029 | 01:57 |
| *** chip__ has joined #openstack-cinder | 01:58 | |
| *** julim has joined #openstack-cinder | 02:00 | |
| *** zul has joined #openstack-cinder | 02:01 | |
| *** zhangjn has quit IRC | 02:01 | |
| *** julim has quit IRC | 02:04 | |
| *** zhangjn has joined #openstack-cinder | 02:06 | |
| *** Guest30993 has joined #openstack-cinder | 02:12 | |
| *** Guest30993 is now known as wilson-l | 02:12 | |
| *** takedakn has joined #openstack-cinder | 02:18 | |
| *** gouthamr_ has quit IRC | 02:18 | |
| *** gouthamr has joined #openstack-cinder | 02:20 | |
| *** takedakn has quit IRC | 02:22 | |
| *** ociuhandu has quit IRC | 02:28 | |
| *** wilson-l has quit IRC | 02:32 | |
| *** salv-orlando has joined #openstack-cinder | 02:33 | |
| openstackgerrit | wanghao proposed openstack/cinder: Fix snapshot stuck in 'deleting' after reboot c-v https://review.openstack.org/322418 | 02:34 |
| openstackgerrit | Merged openstack/cinder: Make IBM drivers return snapshot object for cg ops https://review.openstack.org/268283 | 02:36 |
| *** salv-orlando has quit IRC | 02:38 | |
| *** chip__ has quit IRC | 02:43 | |
| *** chip__ has joined #openstack-cinder | 02:44 | |
| *** gouthamr has quit IRC | 02:45 | |
| *** gcb has quit IRC | 02:46 | |
| *** sdake_ has joined #openstack-cinder | 02:46 | |
| *** sheel has joined #openstack-cinder | 02:47 | |
| *** chip__ has quit IRC | 02:49 | |
| *** sdake has quit IRC | 02:50 | |
| *** chip__ has joined #openstack-cinder | 02:51 | |
| *** adisky has joined #openstack-cinder | 02:59 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements https://review.openstack.org/323063 | 02:59 |
| *** EinstCrazy has quit IRC | 03:00 | |
| *** edmondsw has joined #openstack-cinder | 03:02 | |
| *** gcb has joined #openstack-cinder | 03:03 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/os-brick: Updated from global requirements https://review.openstack.org/323076 | 03:04 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-cinderclient: Updated from global requirements https://review.openstack.org/321715 | 03:05 |
| *** sdake_ has quit IRC | 03:07 | |
| openstackgerrit | YanLin Ren proposed openstack/cinder: IBM FlashSystem: Add host check for iSCSI driver https://review.openstack.org/323091 | 03:10 |
| *** Julien-zte has joined #openstack-cinder | 03:13 | |
| openstackgerrit | Cao ShuFeng proposed openstack/python-cinderclient: Add "--name" filter to get-pools https://review.openstack.org/322011 | 03:15 |
| *** EinstCrazy has joined #openstack-cinder | 03:30 | |
| *** darosale has quit IRC | 03:33 | |
| *** Julien-zte has quit IRC | 03:42 | |
| *** sgotliv has joined #openstack-cinder | 03:42 | |
| *** rcernin has joined #openstack-cinder | 03:49 | |
| *** links has joined #openstack-cinder | 03:50 | |
| *** markvoelker has joined #openstack-cinder | 03:52 | |
| *** neophy has joined #openstack-cinder | 03:54 | |
| *** markvoelker has quit IRC | 03:56 | |
| openstackgerrit | yuyafei proposed openstack/cinder: EMC driver should get ceiling of volume size https://review.openstack.org/322388 | 03:56 |
| *** arch-nemesis has joined #openstack-cinder | 04:00 | |
| *** takedakn has joined #openstack-cinder | 04:04 | |
| *** lpetrut has joined #openstack-cinder | 04:15 | |
| openstackgerrit | Cao ShuFeng proposed openstack/python-cinderclient: Add "--name" filter to get-pools https://review.openstack.org/322011 | 04:17 |
| *** arch-nemesis has quit IRC | 04:18 | |
| openstackgerrit | Cao ShuFeng proposed openstack/python-cinderclient: Add "--name" filter to get-pools https://review.openstack.org/322011 | 04:22 |
| *** ircuser-1 has quit IRC | 04:25 | |
| openstackgerrit | xing-yang proposed openstack/cinder: WIP: Add generic volume groups https://review.openstack.org/322459 | 04:37 |
| *** salv-orlando has joined #openstack-cinder | 04:39 | |
| openstackgerrit | xing-yang proposed openstack/python-cinderclient: WIP: Add generic volume groups https://review.openstack.org/322627 | 04:40 |
| *** salv-orlando has quit IRC | 04:44 | |
| *** itzdilip has joined #openstack-cinder | 04:45 | |
| *** avishay has joined #openstack-cinder | 04:48 | |
| *** rcernin has quit IRC | 04:52 | |
| *** laughter_ has joined #openstack-cinder | 04:53 | |
| *** chlong has quit IRC | 04:54 | |
| *** laughterwym has quit IRC | 04:56 | |
| *** lpetrut has quit IRC | 05:04 | |
| *** ChubYann has quit IRC | 05:08 | |
| *** jwcroppe has quit IRC | 05:12 | |
| *** chlong has joined #openstack-cinder | 05:12 | |
| *** Poornima has joined #openstack-cinder | 05:14 | |
| *** laughter_ has quit IRC | 05:24 | |
| *** jamielennox is now known as jamielennox|away | 05:24 | |
| *** laughterwym has joined #openstack-cinder | 05:25 | |
| openstackgerrit | Ryan Liang proposed openstack/cinder: Pass volume_type_id in when deleting snapshot. https://review.openstack.org/323125 | 05:27 |
| *** jamielennox|away is now known as jamielennox | 05:28 | |
| *** lpetrut has joined #openstack-cinder | 05:29 | |
| *** jgregor has joined #openstack-cinder | 05:29 | |
| *** salv-orlando has joined #openstack-cinder | 05:31 | |
| *** diogogmt has quit IRC | 05:31 | |
| *** pgadiya has joined #openstack-cinder | 05:32 | |
| *** diogogmt has joined #openstack-cinder | 05:32 | |
| openstackgerrit | Falk Reimann proposed openstack/cinder: Fix Pep8 https://review.openstack.org/323129 | 05:39 |
| *** Julien-zte has joined #openstack-cinder | 05:40 | |
| *** neophy has quit IRC | 05:46 | |
| *** lpetrut has quit IRC | 05:47 | |
| *** bardia has quit IRC | 05:49 | |
| *** GB21 has joined #openstack-cinder | 05:50 | |
| *** markvoelker has joined #openstack-cinder | 05:52 | |
| *** nkrinner_afk is now known as nkrinner | 05:55 | |
| *** markvoelker has quit IRC | 05:57 | |
| *** nkrinner has quit IRC | 06:06 | |
| *** nkrinner has joined #openstack-cinder | 06:07 | |
| *** lpetrut has joined #openstack-cinder | 06:08 | |
| *** mjura has joined #openstack-cinder | 06:10 | |
| *** bardia has joined #openstack-cinder | 06:10 | |
| *** lpetrut has quit IRC | 06:13 | |
| *** mjura has quit IRC | 06:15 | |
| *** jamielennox is now known as jamielennox|away | 06:15 | |
| *** mjura has joined #openstack-cinder | 06:15 | |
| *** liverpooler has joined #openstack-cinder | 06:16 | |
| *** liverpooler has quit IRC | 06:21 | |
| *** liverpooler has joined #openstack-cinder | 06:21 | |
| *** takedakn has quit IRC | 06:29 | |
| *** GB21 has quit IRC | 06:34 | |
| openstackgerrit | Falk Reimann proposed openstack/cinder: SwiftBackupDriver: Dont add project_id to authurl https://review.openstack.org/247964 | 06:39 |
| *** belmoreira has joined #openstack-cinder | 06:46 | |
| *** arecknag has joined #openstack-cinder | 06:49 | |
| *** salv-orlando has quit IRC | 06:49 | |
| *** Julien-zte has quit IRC | 06:49 | |
| openstackgerrit | yuyafei proposed openstack/os-brick: Disconnect multipath iscsi may logout session https://review.openstack.org/323154 | 06:50 |
| *** iceyao has joined #openstack-cinder | 06:53 | |
| *** iceyao has quit IRC | 06:53 | |
| *** neophy has joined #openstack-cinder | 06:53 | |
| *** GB21 has joined #openstack-cinder | 06:54 | |
| *** ircuser-1 has joined #openstack-cinder | 06:55 | |
| *** raunak has quit IRC | 06:56 | |
| *** e0ne has joined #openstack-cinder | 06:56 | |
| *** tesseract has joined #openstack-cinder | 06:57 | |
| *** salv-orlando has joined #openstack-cinder | 06:59 | |
| *** rcernin has joined #openstack-cinder | 07:02 | |
| *** neophy has quit IRC | 07:04 | |
| *** jamielennox|away is now known as jamielennox | 07:18 | |
| openstackgerrit | yuyafei proposed openstack/os-brick: Disconnect multipath iscsi may logout session https://review.openstack.org/323154 | 07:27 |
| *** bardia has quit IRC | 07:27 | |
| *** salv-orlando has quit IRC | 07:29 | |
| *** sdake has joined #openstack-cinder | 07:34 | |
| openstackgerrit | Falk Reimann proposed openstack/cinder: SwiftBackupDriver: Dont add project_id to authurl https://review.openstack.org/247964 | 07:35 |
| openstackgerrit | yuyafei proposed openstack/os-brick: Disconnect multipath iscsi may logout session https://review.openstack.org/323174 | 07:36 |
| *** bardia has joined #openstack-cinder | 07:45 | |
| openstackgerrit | Helen Walsh proposed openstack/cinder: VMAX Driver - SnapVX licensing checks for VMAX3 https://review.openstack.org/322972 | 07:46 |
| *** bardia has quit IRC | 07:50 | |
| *** markvoelker has joined #openstack-cinder | 07:53 | |
| openstackgerrit | Keiichi KII proposed openstack/os-brick: remove unneeded in_use flag checking for multipath https://review.openstack.org/264678 | 07:55 |
| *** markvoelker has quit IRC | 07:58 | |
| *** zzzeek has quit IRC | 08:00 | |
| *** zzzeek has joined #openstack-cinder | 08:01 | |
| *** avishay has quit IRC | 08:06 | |
| lixiaoy1 | danpb: Hi Daniel, could you kindly have a look at this patch https://review.openstack.org/#/c/247372/ , which is about cinder encrypts / decrypts volumes. you ever left comments | 08:07 |
| *** jordanP has joined #openstack-cinder | 08:08 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/cinder: [WIP] Testing latest u-c https://review.openstack.org/303766 | 08:10 |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/cinder: [WIP] Testing latest u-c https://review.openstack.org/303766 | 08:10 |
| openstackgerrit | yuyafei proposed openstack/cinder: SolidFire driver should get ceiling of volume size https://review.openstack.org/323191 | 08:11 |
| openstackgerrit | YanLin Ren proposed openstack/cinder: IBM FlashSystem: Add host check for iSCSI driver https://review.openstack.org/323091 | 08:17 |
| openstackgerrit | LisaLi proposed openstack/cinder: Handle quota exceed exception https://review.openstack.org/281666 | 08:20 |
| *** avishay has joined #openstack-cinder | 08:21 | |
| *** mkrcmari__ has quit IRC | 08:22 | |
| *** ircuser-1 has quit IRC | 08:24 | |
| openstackgerrit | LisaLi proposed openstack/cinder-specs: Retype encrypted volumes https://review.openstack.org/248593 | 08:32 |
| *** sdake_ has joined #openstack-cinder | 08:32 | |
| *** sdake has quit IRC | 08:35 | |
| *** jsheeren has joined #openstack-cinder | 08:44 | |
| jsheeren | hi all | 08:44 |
| jsheeren | i have a question about cinder using an nfs share for shared storage | 08:44 |
| jsheeren | i can create and extend volumes; attach and detach them | 08:44 |
| jsheeren | but when i want to extend a detached volume, i get permission denied | 08:44 |
| jsheeren | i see the ownership on the volume file on the share changed from cinder:cinder to nobody:nogroup | 08:45 |
| jsheeren | cinder cannot do any actions on the volume because of this | 08:45 |
| jsheeren | force extending a volume which is attached works as expected (because the libvirt-qemu;kvm ownership when attached) | 08:46 |
| jsheeren | scratch this: "force extending a volume which is attached works as expected (because the libvirt-qemu;kvm ownership when attached)" -> my brain fffd | 08:47 |
| jsheeren | has anyone seen this behaviour, is it normal expected behaviour? | 08:48 |
| *** laughterwym has quit IRC | 08:50 | |
| *** laughterwym has joined #openstack-cinder | 08:50 | |
| lixiaoy1 | jsheeren: no such problem in my env. I use devstack to install my openstack env, and with nfs as backend | 08:51 |
| jsheeren | lixiaoy1: what are your exports options? | 08:52 |
| lixiaoy1 | jsheeren: let me check | 08:52 |
| jsheeren | have you enabled rootsquash? | 08:52 |
| lixiaoy1 | jsheeren: I use devstack to install and with nfs as backend | 08:52 |
| jsheeren | lixiaoy1: ah devstack handles the nfs exports then? | 08:53 |
| lixiaoy1 | jsheeren: yes. | 08:53 |
| lixiaoy1 | jsheeren: rw, no_root_squash exports options | 08:54 |
| jsheeren | lixiaoy1: aha; thanks for the options; i'll double check with my storage guy to see which options he set. | 08:54 |
| *** mkrcmari__ has joined #openstack-cinder | 08:54 | |
| *** takedakn has joined #openstack-cinder | 08:54 | |
| lixiaoy1 | jsheeren: enable_plugin devstack-plugin-nfs <devstack-plugin-nfs github address> | 08:54 |
| *** takedakn has quit IRC | 08:54 | |
| *** laughterwym has quit IRC | 08:55 | |
| lixiaoy1 | jsheeren: have to leave now. Good luck! | 08:55 |
| jsheeren | lixiaoy1: thank you! | 08:55 |
| *** jordanP has quit IRC | 08:59 | |
| *** yhayashi has quit IRC | 09:00 | |
| openstackgerrit | yuyafei proposed openstack/os-brick: Disconnect multipath iscsi may logout session https://review.openstack.org/323174 | 09:05 |
| *** avishay has quit IRC | 09:10 | |
| *** permalac has quit IRC | 09:16 | |
| *** jsheeren has quit IRC | 09:16 | |
| *** jsheeren has joined #openstack-cinder | 09:21 | |
| *** avishay has joined #openstack-cinder | 09:22 | |
| *** andymaier_ has joined #openstack-cinder | 09:23 | |
| *** ircuser-1 has joined #openstack-cinder | 09:26 | |
| *** chlong has quit IRC | 09:29 | |
| openstackgerrit | wanghao proposed openstack/cinder: Fix snapshot stuck in 'deleting' after reboot c-v https://review.openstack.org/322418 | 09:31 |
| *** yangyapeng has joined #openstack-cinder | 09:34 | |
| *** takedakn has joined #openstack-cinder | 09:38 | |
| dulek | geguileo: Hi! Am I getting it right that with manifest we will have some more sophisticated rules of when to do version bump and when not to? | 09:41 |
| geguileo | Not really, it will be the simplest rule of all | 09:41 |
| geguileo | If you don't change a field, you don't bump a version | 09:42 |
| dulek | geguileo: I mean - on child change we won't need to bump parent's version, but hash will still change, right? | 09:42 |
| geguileo | Yes | 09:42 |
| geguileo | So we'll still need to change the test | 09:42 |
| geguileo | Unless we change the way we do the hashing for the test | 09:42 |
| geguileo | Which now that you mention it, we should | 09:42 |
| openstackgerrit | Merged openstack/cinder: Add test for checking object compatibilities https://review.openstack.org/259123 | 09:43 |
| *** mkrcmari__ has quit IRC | 09:43 | |
| dulek | geguileo: We're using fixtures from o.vo for that. | 09:44 |
| geguileo | I'll have a look now and see if there's an easy way to change it | 09:44 |
| dulek | geguileo: Maybe it's easier to just get more specific in the comment on top of test_objects.py instead? | 09:45 |
| geguileo | dulek: I would like to avoid having to make unnecesary changes to the tests hashes whenever we have a related object change | 09:46 |
| dulek | geguileo: This definitely makes live easier for anyone. :) | 09:46 |
| geguileo | That way it will not be misleading and you will know you really require a version bump if the test is complaining | 09:46 |
| *** takedakn has quit IRC | 09:46 | |
| geguileo | If it's easy I'll do it | 09:47 |
| geguileo | If it's not, I'll probably look the other way ;-) | 09:47 |
| *** chlong has joined #openstack-cinder | 09:47 | |
| *** mvk has joined #openstack-cinder | 09:47 | |
| dulek | geguileo: I've just checked. Looks like with your patch modification to the child does nothing with parent. Now when I recall it - wasn't it like that all the time? That's why we've missed so much required parent version bumps. | 09:52 |
| geguileo | dulek: I thought the hashing took into account child versions | 09:52 |
| geguileo | I was just about to check it | 09:52 |
| dulek | Yeah, there are child_versions taken into account when hashing, but as you're removing it… :) | 09:53 |
| *** markvoelker has joined #openstack-cinder | 09:54 | |
| *** ducttape_ has quit IRC | 09:57 | |
| *** sdake_ is now known as sdake | 09:57 | |
| *** markvoelker has quit IRC | 09:58 | |
| geguileo | dulek: Ok, I see now why it's not complaining with ovo field changes | 10:00 |
| geguileo | dulek: Because the repr is not reporting versions | 10:00 |
| geguileo | dulek: Ok, then we won't get problems in the test when we bump a related field version | 10:00 |
| dulek | geguileo: But should it? | 10:01 |
| geguileo | dulek: It depends on what you want to do... | 10:02 |
| geguileo | dulek: But I believe it should in general | 10:02 |
| geguileo | dulek: But we don't need it to and nova doesn't either | 10:02 |
| dulek | geguileo: You're talking about __repr__ in o.vo.base? That's for objects, not fields I think. | 10:02 |
| *** Julien-zte has joined #openstack-cinder | 10:03 | |
| *** cheneydc has quit IRC | 10:03 | |
| geguileo | dulek: Correct, but if a field is an object won't it get called? | 10:03 |
| geguileo | Give me a sec | 10:03 |
| dulek | geguileo: I think hashing is processing object *class*, not object *object. | 10:03 |
| * dulek knows this is bad explanation. | 10:04 | |
| geguileo | But the class has fields that link to another class | 10:04 |
| geguileo | Here https://github.com/openstack/oslo.versionedobjects/blob/788e3d08f7d073acb9a9958b4e474c9d51da934d/oslo_versionedobjects/fields.py#L143 | 10:05 |
| *** caoshufeng has quit IRC | 10:05 | |
| geguileo | It should be reporting the version when a field is a related object imo | 10:05 |
| geguileo | But we don't need it, nova doesn't need it... | 10:06 |
| geguileo | So I don't care all that much };-) | 10:06 |
| dulek | fields.Object doesn't implement __repr__. | 10:06 |
| dulek | Yeah, let's cut it. :P | 10:06 |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder: Refactor create, save, and destroy OVO methods https://review.openstack.org/315541 | 10:07 |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder: Add debug level logs on OVO backports https://review.openstack.org/321143 | 10:07 |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder: Implement OVO class init mechanism https://review.openstack.org/321144 | 10:07 |
| *** laughterwym has joined #openstack-cinder | 10:08 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/cinder: [WIP] Testing latest u-c https://review.openstack.org/303766 | 10:10 |
| *** laughterwym has quit IRC | 10:12 | |
| *** chlong has quit IRC | 10:15 | |
| *** laughterwym has joined #openstack-cinder | 10:15 | |
| *** laughterwym has quit IRC | 10:18 | |
| *** laughterwym has joined #openstack-cinder | 10:19 | |
| *** laughterwym has quit IRC | 10:23 | |
| *** chlong has joined #openstack-cinder | 10:28 | |
| *** Julien-zte has quit IRC | 10:28 | |
| *** pdeore has joined #openstack-cinder | 10:30 | |
| *** EinstCrazy has quit IRC | 10:32 | |
| *** yangyapeng has quit IRC | 10:35 | |
| *** sdague has joined #openstack-cinder | 10:36 | |
| *** laughterwym has joined #openstack-cinder | 10:47 | |
| *** permalac has joined #openstack-cinder | 10:48 | |
| *** liverpooler has quit IRC | 10:48 | |
| *** laughterwym has quit IRC | 10:50 | |
| *** merooney has joined #openstack-cinder | 10:53 | |
| *** Julien-zte has joined #openstack-cinder | 10:53 | |
| *** ducttape_ has joined #openstack-cinder | 10:57 | |
| *** pots2 has joined #openstack-cinder | 10:58 | |
| openstackgerrit | wanghao proposed openstack/cinder: Fix snapshot stuck in 'deleting' after reboot c-v https://review.openstack.org/322418 | 10:58 |
| *** pots1 has quit IRC | 11:00 | |
| *** liverpooler has joined #openstack-cinder | 11:01 | |
| *** ducttape_ has quit IRC | 11:02 | |
| *** merooney has quit IRC | 11:04 | |
| *** merooney has joined #openstack-cinder | 11:04 | |
| *** merooney has quit IRC | 11:06 | |
| *** raildo-afk is now known as raildo | 11:13 | |
| *** merooney has joined #openstack-cinder | 11:13 | |
| openstackgerrit | yuyafei proposed openstack/cinder: EMC driver should return ceiling of volume size https://review.openstack.org/322388 | 11:19 |
| openstackgerrit | yuyafei proposed openstack/cinder: SolidFire driver should return ceiling of volume size https://review.openstack.org/323191 | 11:21 |
| openstackgerrit | yuyafei proposed openstack/os-brick: Disconnect multipath iscsi may logout session https://review.openstack.org/323174 | 11:23 |
| openstackgerrit | yuyafei proposed openstack/python-cinderclient: unit.test_base.BaseTest.test_eq does not work https://review.openstack.org/322408 | 11:24 |
| *** ducttape_ has joined #openstack-cinder | 11:33 | |
| openstackgerrit | Wilson Liu proposed openstack/cinder-specs: Migrate volume between backends in an async way https://review.openstack.org/312853 | 11:35 |
| *** zhangjn has quit IRC | 11:36 | |
| *** pgadiya has quit IRC | 11:38 | |
| *** zhangjn has joined #openstack-cinder | 11:38 | |
| openstackgerrit | yuyafei proposed openstack/cinder: SolidFire driver should return ceiling of volume size https://review.openstack.org/323191 | 11:40 |
| openstackgerrit | yuyafei proposed openstack/cinder: SolidFire driver should return ceiling of volume size https://review.openstack.org/323191 | 11:41 |
| *** diablo_rojo has joined #openstack-cinder | 11:41 | |
| *** laughterwym has joined #openstack-cinder | 11:47 | |
| *** pgadiya has joined #openstack-cinder | 11:47 | |
| *** Julien-zte has quit IRC | 11:52 | |
| *** markvoelker has joined #openstack-cinder | 11:55 | |
| *** markvoelker has quit IRC | 11:57 | |
| *** markvoelker has joined #openstack-cinder | 11:57 | |
| *** ducttape_ has quit IRC | 11:58 | |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder: Move OVO methods from CinderObject to Persistent https://review.openstack.org/312930 | 11:58 |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder: Remove remotable from OVOs https://review.openstack.org/312929 | 11:58 |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder: Refactor create, save, and destroy OVO methods https://review.openstack.org/315541 | 11:58 |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder: Add debug level logs on OVO backports https://review.openstack.org/321143 | 11:58 |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder: Stop assuming persistent OVOs are also DictCompat https://review.openstack.org/321142 | 11:58 |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder: Implement OVO class init mechanism https://review.openstack.org/321144 | 11:58 |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder: Use manifest to backport OVOs during upgrades https://review.openstack.org/307075 | 11:58 |
| *** merooney has quit IRC | 11:59 | |
| *** avishay has quit IRC | 12:01 | |
| *** takedakn has joined #openstack-cinder | 12:02 | |
| *** takedakn has quit IRC | 12:02 | |
| *** merooney has joined #openstack-cinder | 12:03 | |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder: Refactor create, save, and destroy OVO methods https://review.openstack.org/315541 | 12:05 |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder: Implement OVO class init mechanism https://review.openstack.org/321144 | 12:05 |
| * geguileo forgot to update one commit message :-( | 12:05 | |
| *** pdeore has quit IRC | 12:08 | |
| *** diablo_rojo has quit IRC | 12:11 | |
| *** takedakn has joined #openstack-cinder | 12:12 | |
| *** diablo_rojo has joined #openstack-cinder | 12:12 | |
| *** JoseMello has joined #openstack-cinder | 12:12 | |
| *** Lee1092 has quit IRC | 12:13 | |
| *** martyturner has joined #openstack-cinder | 12:14 | |
| *** savihou has joined #openstack-cinder | 12:17 | |
| *** takedakn has quit IRC | 12:18 | |
| *** arch-nemesis has joined #openstack-cinder | 12:22 | |
| *** arch-nemesis has quit IRC | 12:22 | |
| *** arch-nemesis has joined #openstack-cinder | 12:23 | |
| *** akshai has joined #openstack-cinder | 12:23 | |
| *** tpsilva has joined #openstack-cinder | 12:24 | |
| *** ducttape_ has joined #openstack-cinder | 12:24 | |
| *** GB21 has quit IRC | 12:24 | |
| *** dave-mccowan has joined #openstack-cinder | 12:25 | |
| *** sdake has quit IRC | 12:27 | |
| *** merooney has quit IRC | 12:27 | |
| *** merooney has joined #openstack-cinder | 12:28 | |
| *** ducttape_ has quit IRC | 12:29 | |
| *** arch-nemesis has quit IRC | 12:29 | |
| *** ducttape_ has joined #openstack-cinder | 12:29 | |
| *** kfarr has joined #openstack-cinder | 12:30 | |
| *** ducttape_ has quit IRC | 12:34 | |
| *** jungleboyj_ has joined #openstack-cinder | 12:35 | |
| *** pgadiya has quit IRC | 12:36 | |
| openstackgerrit | Edwin Wang proposed openstack/cinder: ABC Metaclass for IBM FlashSystem https://review.openstack.org/323343 | 12:42 |
| *** arch-nemesis has joined #openstack-cinder | 12:42 | |
| *** timcl has joined #openstack-cinder | 12:43 | |
| openstackgerrit | Wilson Liu proposed openstack/cinder-specs: Migrate volume between backends in an async way https://review.openstack.org/312853 | 12:44 |
| *** ducttape_ has joined #openstack-cinder | 12:47 | |
| *** gouthamr has joined #openstack-cinder | 12:48 | |
| *** julim has joined #openstack-cinder | 12:53 | |
| *** diablo_rojo has left #openstack-cinder | 12:53 | |
| *** diogogmt has quit IRC | 12:56 | |
| *** diogogmt has joined #openstack-cinder | 12:57 | |
| *** kfarr has quit IRC | 13:01 | |
| *** eharney has joined #openstack-cinder | 13:03 | |
| openstackgerrit | Kaitlin Farr proposed openstack/cinder: Specify key algorithm and size for create_key https://review.openstack.org/251503 | 13:04 |
| *** akerr has joined #openstack-cinder | 13:04 | |
| *** Poornima has quit IRC | 13:05 | |
| *** cknight has joined #openstack-cinder | 13:05 | |
| *** merooney has quit IRC | 13:06 | |
| *** rlrossit has joined #openstack-cinder | 13:09 | |
| *** erlon_ has joined #openstack-cinder | 13:10 | |
| *** enriquetaso has joined #openstack-cinder | 13:15 | |
| *** merooney has joined #openstack-cinder | 13:15 | |
| *** xyang1 has joined #openstack-cinder | 13:18 | |
| *** bill_az_ has joined #openstack-cinder | 13:20 | |
| *** akshai has quit IRC | 13:24 | |
| *** diogogmt has quit IRC | 13:25 | |
| *** akshai has joined #openstack-cinder | 13:29 | |
| *** porrua has joined #openstack-cinder | 13:30 | |
| * dulek needs more coffee to finish reviewing geguileo's patches with API race conditions. | 13:31 | |
| smcginnis | Gallons! :) | 13:33 |
| smcginnis | Or is that liters for you? :D | 13:34 |
| *** bill_az_ has quit IRC | 13:36 | |
| *** diogogmt has joined #openstack-cinder | 13:36 | |
| *** lgreg has joined #openstack-cinder | 13:38 | |
| *** mtanino has joined #openstack-cinder | 13:39 | |
| dulek | smcginnis: Liters are converted to gallons linearly, so it's easy. ;) | 13:40 |
| dulek | BTW - ever heard that Fahrenheit was living in Gdansk, where I'm based? :D | 13:40 |
| smcginnis | dulek: Oh really? Hah, interesting. | 13:41 |
| *** jungleboyj_ has quit IRC | 13:43 | |
| merooney | could I get this blueprint approved when someone gets a chance? https://blueprints.launchpad.net/cinder/+spec/netapp-cheesecake-replication-support | 13:47 |
| *** baumann has joined #openstack-cinder | 13:47 | |
| *** rlrossit has quit IRC | 13:49 | |
| *** erlon_ is now known as erlon | 13:50 | |
| *** mtanino has quit IRC | 13:52 | |
| *** baumann1 has joined #openstack-cinder | 13:55 | |
| *** ametts has joined #openstack-cinder | 13:56 | |
| *** baumann has quit IRC | 13:56 | |
| *** diogogmt has quit IRC | 14:01 | |
| sheel | I can offer some coffee if someone can review my patch also :) | 14:03 |
| *** arch-nemesis has quit IRC | 14:03 | |
| sheel | https://review.openstack.org/#/c/316587/ | 14:04 |
| *** akshai has quit IRC | 14:10 | |
| dulek | geguileo: Can you take a look if I'm making sense in https://review.openstack.org/#/c/221442/ ? | 14:11 |
| *** jgregor has joined #openstack-cinder | 14:12 | |
| *** yangyapeng has joined #openstack-cinder | 14:12 | |
| *** Julien-zte has joined #openstack-cinder | 14:15 | |
| *** ducttape_ has quit IRC | 14:18 | |
| openstackgerrit | Merged openstack/cinder: Fix typos in Cinder files https://review.openstack.org/298078 | 14:18 |
| *** tomh_ has joined #openstack-cinder | 14:18 | |
| *** EinstCrazy has joined #openstack-cinder | 14:19 | |
| *** andymaier_ has quit IRC | 14:21 | |
| *** akshai has joined #openstack-cinder | 14:22 | |
| *** MrHeavy has joined #openstack-cinder | 14:24 | |
| *** ducttape_ has joined #openstack-cinder | 14:26 | |
| *** lprice has joined #openstack-cinder | 14:26 | |
| *** darosale has joined #openstack-cinder | 14:28 | |
| *** jungleboyj_ has joined #openstack-cinder | 14:32 | |
| *** lprice1 has joined #openstack-cinder | 14:34 | |
| *** lprice has quit IRC | 14:35 | |
| *** ducttape_ has quit IRC | 14:40 | |
| *** diogogmt has joined #openstack-cinder | 14:43 | |
| geguileo | dulek: You make total sense | 14:44 |
| geguileo | dulek: I just can't give you a good reason | 14:44 |
| geguileo | dulek: I wrote that 8 months ago :-( | 14:44 |
| dulek | geguileo: It's either simple mistake, copied from migrate, constraint got removed meanwhile, or - there's sense behind it. | 14:45 |
| geguileo | dulek: I'm afraid of the seonc option | 14:45 |
| geguileo | second | 14:45 |
| dulek | jgriffith: Hey, do you know if it should be possible to retype a volume with enabled replication? | 14:46 |
| * dulek finished the coffee with the review. And a second one awaits… | 14:47 | |
| *** jgregor has quit IRC | 14:49 | |
| *** timcl has quit IRC | 14:50 | |
| *** Lee1092 has joined #openstack-cinder | 14:50 | |
| *** jgregor has joined #openstack-cinder | 14:50 | |
| tobasco | any ubuntu maintainers sneaking around here, hit me up ;) | 14:50 |
| *** xinli has joined #openstack-cinder | 14:51 | |
| *** jungleboyj_ has quit IRC | 14:53 | |
| *** jgregor has quit IRC | 14:54 | |
| *** mtanino has joined #openstack-cinder | 14:55 | |
| *** cdelatte has joined #openstack-cinder | 14:56 | |
| *** jgregor has joined #openstack-cinder | 14:59 | |
| *** timcl has joined #openstack-cinder | 14:59 | |
| *** jwcroppe has joined #openstack-cinder | 15:00 | |
| jgriffith | dulek: Yes, but honestly I think there needs to be some thought there still | 15:07 |
| *** rcernin has quit IRC | 15:07 | |
| jgriffith | dulek: it's a pretty special case that requires some other things to happen (at least to do it correctly) | 15:07 |
| *** mjura has quit IRC | 15:09 | |
| dulek | jgriffith: geguileo accidentally blocked that in his patchset. I started to wonder if that wasn't really an accidentally bugfix. | 15:09 |
| *** ducttape_ has joined #openstack-cinder | 15:09 | |
| dulek | jgriffith: That's the patch: https://review.openstack.org/#/c/221442/ | 15:10 |
| *** timcl has quit IRC | 15:10 | |
| *** merooney has quit IRC | 15:11 | |
| *** Swanson has quit IRC | 15:11 | |
| *** jungleboyj has joined #openstack-cinder | 15:11 | |
| *** chris_morrell has joined #openstack-cinder | 15:12 | |
| *** Swanson has joined #openstack-cinder | 15:12 | |
| *** links has quit IRC | 15:15 | |
| Swanson | Internal REST API change for dell driver. Needs 1 more +2 and a +A. https://review.openstack.org/#/c/315226/ | 15:16 |
| jgriffith | dulek: looking | 15:16 |
| Swanson | ^^^ daily beg | 15:16 |
| smcginnis | Swanson: http://giphy.com/gifs/mpQkJpBMbSxMs | 15:17 |
| Swanson | Hmm.... BP stock is down half a buck today. They being forced to scrub ducks on a beach somewhere again? | 15:20 |
| Swanson | It's all just going as planned today. | 15:20 |
| Swanson | Hostility level at 11. | 15:20 |
| *** belmoreira has quit IRC | 15:22 | |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder: Remove API races from consistency groups https://review.openstack.org/259429 | 15:23 |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder: Remove API races from migrate and retype https://review.openstack.org/221442 | 15:23 |
| *** arch-nemesis has joined #openstack-cinder | 15:24 | |
| *** arch-nemesis has quit IRC | 15:24 | |
| *** arch-nemesis has joined #openstack-cinder | 15:24 | |
| *** diablo_rojo has joined #openstack-cinder | 15:24 | |
| *** EinstCrazy has quit IRC | 15:28 | |
| *** EinstCrazy has joined #openstack-cinder | 15:29 | |
| *** EinstCrazy has quit IRC | 15:29 | |
| *** arecknag has quit IRC | 15:29 | |
| *** lprice1 has quit IRC | 15:29 | |
| *** karthikp has joined #openstack-cinder | 15:32 | |
| *** openstackgerrit has quit IRC | 15:33 | |
| *** openstackgerrit has joined #openstack-cinder | 15:34 | |
| openstackgerrit | xiexs proposed openstack/cinder: Improve the retrieval feature of message https://review.openstack.org/323450 | 15:35 |
| openstackgerrit | xiexs proposed openstack/cinder: Destroy the error message after volume was deleted https://review.openstack.org/323451 | 15:35 |
| openstackgerrit | zhangsong proposed openstack/cinder: Add lock when create tgt target https://review.openstack.org/323452 | 15:36 |
| *** xyang has joined #openstack-cinder | 15:38 | |
| *** Apoorva has joined #openstack-cinder | 15:39 | |
| *** lgreg has quit IRC | 15:40 | |
| xyang | DuncanT, tbarron: hi, can you take a look of this? https://review.openstack.org/#/c/321943/ | 15:42 |
| DuncanT | xyang: I'm not massively sold on the backporting story - we changed the code an great deal, for good reasons. I'll review the patch though | 15:44 |
| xyang | DuncanT: but it broke some code that was working in Liberty | 15:44 |
| xyang | DuncanT: I thought you were ok with finding the local volume node when I asked you last time? | 15:45 |
| DuncanT | xyang: We need to fix the testing so it doesn't get broken again. I don't mind an ugly fix on -stable, but making a mess of master without a plan for getting it done properly makes me nervous. I think I'd rather just turn off the broken optimised path in one patch, backport that, then fix it properly. Maybe it isn't as bad as I think though - I'll take a | 15:47 |
| DuncanT | look at the patch | 15:47 |
| xyang | DuncanT: thanks for looking, but without fixing the optimal path, both non-disruptive backup and backup snapshot are broken in Mitaka | 15:49 |
| *** jungleboyj has quit IRC | 15:49 | |
| xyang | DuncanT: so the problem is there is no slow path for driver that uses snapshot to backup | 15:49 |
| *** yangyapeng has quit IRC | 15:50 | |
| DuncanT | We can just remove the call to the driver and make every driver go via the slow path - that looks like a small, clean patch | 15:50 |
| *** yangyapeng has joined #openstack-cinder | 15:50 | |
| xyang | DuncanT: but that was working in Liberty | 15:51 |
| eharney | xyang: did we lose an optimization or some user functionality? | 15:52 |
| DuncanT | xyang: And it got broken. I'd rather fix it cleanly than with a hacky mess. If we break offloaded backup while fixing this, then we've still got a broken feature. I need to spend some time on your patch to see | 15:52 |
| xyang | eharney: the functionality is currently broken | 15:53 |
| eharney | xyang: which functionality? | 15:53 |
| xyang | eharney: non-disruptive backup and backup snapshot | 15:53 |
| eharney | xyang: i don't know what "backup snapshot" means for end-user functionality | 15:54 |
| *** tesseract has quit IRC | 15:54 | |
| xyang | eharney: that is a feature we added in Mitaka | 15:54 |
| eharney | xyang: what is the feature? | 15:54 |
| xyang | eharney: backup a snapshot | 15:54 |
| eharney | ok | 15:54 |
| xyang | DuncanT: appreciate if you could take a closer look. It is an ugly patch as we discussed last week, but it gave us back what was supported before. | 15:55 |
| *** Julien-zte has quit IRC | 15:56 | |
| DuncanT | xyang: It looks like it breaks the offloaded backup though | 15:56 |
| xyang | DuncanT: does it? | 15:56 |
| xyang | DuncanT: this fix only affects the code path that uses snapshot, which does not work in the current code | 15:56 |
| xyang | DuncanT: I believe this should not affect the code that uses a temp volume | 15:57 |
| DuncanT | xyang: Not sure until I can spin up a multinode test environment, but it looks like it only fixes it if you co-host c-vol and c-bak | 15:57 |
| xyang | DuncanT: that is right | 15:57 |
| xyang | DuncanT: to fix multinode for snapshot, we need a bigger fix, which I also have a patch for | 15:57 |
| DuncanT | xyang: But it doesn't look like the fallback to the none-optimised path works | 15:58 |
| xyang | DuncanT: but that complete fix can’t be backported | 15:58 |
| jgriffith | smcginnis: dulek can I get your opinions on something? | 15:58 |
| DuncanT | xyang: I might be wrong though | 15:58 |
| smcginnis | jgriffith: Sure | 15:58 |
| *** nkrinner has quit IRC | 15:58 | |
| jgriffith | smcginnis: dulek https://review.openstack.org/#/c/321095/1 | 15:58 |
| xyang | DuncanT: if you are using the snapshot path, it does not fall back to non-optimized path. it never did that | 15:58 |
| jgriffith | smcginnis: dulek so the bug IMO is that attach_get_by_* does a first() call | 15:59 |
| xyang | DuncanT: so if we want to fall back, that will be a new thing we can add later | 15:59 |
| DuncanT | xyang: It needs to now though, that's my point | 15:59 |
| jgriffith | smcginnis: dulek the names of the methods suck and aren't very clear on what you get | 15:59 |
| xyang | DuncanT: I can think about that | 15:59 |
| DuncanT | xyang: Running c-vol and c-bak on separate nodes is now supported, so we need to make that work | 15:59 |
| jgriffith | smcginnis: dulek BUT, the object code for example interpreted things as I did... that these were a "get all" | 15:59 |
| smcginnis | jgriffith: Yeah, I don't see why we would only want to get the first one. | 15:59 |
| smcginnis | jgriffith: That would have been my expectation. | 16:00 |
| jgriffith | smcginnis: dulek Gorka would like to see me write new methods, I'd rather just fix the ones that are there | 16:00 |
| DuncanT | xyang: I think it's going to be easier to disable the optimisation (and so make it work non-optimised everywhere) then look at adding your fix on top of that, rather than the other way round - the disable patch should be very small | 16:00 |
| xyang | DuncanT: the problem is still that it was working in Liberty | 16:01 |
| jgriffith | smcginnis: dulek I suppose I can just change the names, but then I have to go through all the bull shit of object version updates etc etc and a bunch of crappy unit tests :( | 16:01 |
| dulek | jgriffith: Diving. A break from conditional updates will be nice. | 16:01 |
| jgriffith | I guess I'm being lazy and should get over it | 16:01 |
| jgriffith | dulek: haha | 16:01 |
| xyang | DuncanT: so the code that I added are copied from Liberty | 16:01 |
| xyang | DuncanT: I mean the code that looks for enabled backend on the backup node | 16:01 |
| DuncanT | xyang: The liberty approach is no longer good, we changed the deployment model | 16:02 |
| smcginnis | jgriffith: So it looks like you've updated every reference expect a mock in test_volume_attachment. | 16:02 |
| xyang | DuncanT: If they are on the same node, it works | 16:02 |
| dulek | jgriffith: Gorka is removing remotable decorator from objects methods in one of his commits. This would save you burden on bumping the versions. :) | 16:02 |
| DuncanT | xyang: That's my point - things are different now and we have to fix it in that context | 16:02 |
| smcginnis | jgriffith: Looking at that now... | 16:02 |
| smcginnis | dulek: Oh, there may be a reason. :) | 16:03 |
| xyang | DuncanT: so with this fix, it will work for the same node, which is the same as Liberty | 16:03 |
| DuncanT | xyang: But it will be broken for anybody deploying using the new model | 16:03 |
| jgriffith | dulek: oh that would be nice | 16:03 |
| xyang | DuncanT: it will not break | 16:03 |
| smcginnis | jgriffith: I guess I don't see a strong reason to add a new once since you seem to have covered all references. | 16:03 |
| xyang | DuncanT: this only affects driver that uses snapshot | 16:03 |
| smcginnis | jgriffith: But dulek has a fair point. :) | 16:04 |
| DuncanT | xyang: Disabling the optimisations gets it working for every supported config, in a very small patch | 16:04 |
| xyang | DuncanT: currently it does not work for them at all | 16:04 |
| jgriffith | smcginnis: ok, I'll just rename them I gues | 16:04 |
| jgriffith | guess | 16:04 |
| dulek | I'm not yet leaning for renaming. ;) | 16:04 |
| xyang | DuncanT: this change does not affect others | 16:04 |
| dulek | I've just thrown a data point. | 16:04 |
| *** jsheeren has quit IRC | 16:05 | |
| DuncanT | xyang: It's adding a mess of ugly code, and I think it's the wrong approach | 16:05 |
| DuncanT | xyang: Give me a day or so to get an env up to test it though, maybe I'll change my mind | 16:05 |
| xyang | DuncanT: sure. appreciated! | 16:05 |
| DuncanT | xyang: I'm just going by gut reaction at the moment | 16:05 |
| xyang | DuncanT: I asked you last week before making this “ugly” patch:) and you were okay at that time | 16:06 |
| jgriffith | it doesn't fix the fact that the object code is expecting a list and gets a single item though IMO | 16:06 |
| DuncanT | xyang: I didn't realise quite how ugly, nor that it is still broken in the new config with this patch | 16:06 |
| jgriffith | but whatever | 16:06 |
| *** yangyapeng has quit IRC | 16:07 | |
| xyang | DuncanT: I can only fix the local node config with this fix, otherwise I have another fix that also fix the remote config | 16:07 |
| dulek | jgriffith: Are we even using them? Looks like these should explode when used in current status. | 16:07 |
| DuncanT | xyang: I get the feeling it would be better to do the disable first, then the complex fix, but we're going round in circles at this point, I'll try to get back to you tomorrow | 16:08 |
| *** raunak has joined #openstack-cinder | 16:08 | |
| *** raunak has quit IRC | 16:08 | |
| *** asselin has joined #openstack-cinder | 16:08 | |
| openstackgerrit | Kaitlin Farr proposed openstack/cinder: Replace key manager with Castellan https://review.openstack.org/280492 | 16:08 |
| xyang | DuncanT: sure. I realize that we are talking in circles:). Talk to you tomorrow about it | 16:08 |
| *** jungleboyj has joined #openstack-cinder | 16:09 | |
| dulek | jgriffith: Okay, looks like geguileo wants to rename both DB methods in the manner of volume_attachment_get_by_host to volume_attachment_get_all_by_host | 16:11 |
| dulek | jgriffith: And this is totally reasonable. | 16:11 |
| jgriffith | dulek: I'm not sure I'm being clear.. just a sec :) | 16:11 |
| dulek | jgriffith: It would require you just fixing usages of them. No need to modify object's method names as these are named like returning a list. | 16:12 |
| jgriffith | dulek: https://github.com/openstack/cinder/blob/master/cinder/objects/volume_attachment.py#L84 | 16:12 |
| jgriffith | dulek: note the call in the object code is "get_all_by_host" | 16:12 |
| jgriffith | dulek: BUT, it calls a db method that just returns "first()" | 16:12 |
| jgriffith | dulek: That's a bug | 16:12 |
| jgriffith | dulek: so I tried to just fix it | 16:13 |
| jgriffith | dulek: the result is everybody throwing in their 2 cents about "new names", "why is this that way" etc etc | 16:13 |
| jgriffith | dulek: I thought it was a pretty simple fix, but now I'm getting the impression I'm living in a different universe :) | 16:13 |
| jgriffith | dulek: so you and smcginnis seem to be suggesting that I write new db methods and leave the broken usage in object/volume_attachments.py... or that I rewrite those as well. | 16:14 |
| smcginnis | jgriffith: Yeah, I don't really see the need to rename. I think it should just be fixed to work as its been assumed to. | 16:14 |
| jgriffith | smcginnis: yeah... don't get me wrong, I don't mind renaming them at all. But I'm saying those calls in object/volume_attachments have to be fixed | 16:15 |
| jgriffith | which means object version change etc | 16:15 |
| smcginnis | jgriffith: I agree. | 16:15 |
| dulek | jgriffith: Suuure, object's code is screwed up here. But it seems to me that if we're making db.volume_attachment_get_by_host return list now, we should rename it to db.volume_attachment_get_*all*_by_host and switch all the occurrences to use new name. | 16:16 |
| dulek | Doesn't it make sense? | 16:16 |
| *** kfarr has joined #openstack-cinder | 16:16 | |
| smcginnis | dulek: I don't know. get_by_host I would assume would return multiple. | 16:16 |
| smcginnis | We don't have a 1:1 attachment to host relationship. | 16:16 |
| jgriffith | dulek: yeah, that's cool. I'll just add some new calls and update those parts and get on with my day :) | 16:16 |
| smcginnis | So if I'm getting by host, I would expect to get everything for that host. | 16:16 |
| jgriffith | dulek: personally I think the existing ones should probably be removed | 16:17 |
| jgriffith | smcginnis: +1000 That's what i was trying to point out :) | 16:17 |
| dulek | Gimme a sec. | 16:17 |
| jgriffith | smcginnis: and it turns out other than the author who wrote the unit tests an the code that consumes it that's what everybody assumed | 16:17 |
| *** xyang has quit IRC | 16:17 | |
| smcginnis | It should just be fixed IMO. | 16:18 |
| jgriffith | smcginnis: I didn't even bring up the other ones... like volume_attachment_get_used_by_volume_id | 16:18 |
| smcginnis | jgriffith: :/ | 16:18 |
| *** xyang has joined #openstack-cinder | 16:19 | |
| jgriffith | smcginnis: which uses "all()" as we would expect, but by gorka and dulek 's thoughts I should change those too... which is ok, I'm doing it now :) | 16:19 |
| *** rueben_ has joined #openstack-cinder | 16:19 | |
| dulek | smcginnis, jgriffith: Looking through db.sqlalchemy.api it looks like we have a convention of naming methods with get_all when it returns list and with just get when it returns a single item. | 16:20 |
| *** lprice has joined #openstack-cinder | 16:20 | |
| jgriffith | dulek: yeah, I'll change them to match that convention :) | 16:20 |
| jgriffith | dulek: smcginnis thanks for the input | 16:20 |
| smcginnis | dulek: Fair enough. I guess I'd rather have consistency. | 16:20 |
| dulek | I don't really care that much, but I think it makes devs lives a little easier. | 16:20 |
| jgriffith | dulek: agreed | 16:21 |
| *** raunak has joined #openstack-cinder | 16:22 | |
| dulek | Also note that geguileo is trying to make some automated bindings between DB model and objects, so consistency in naming makes his life easier. :) | 16:22 |
| smcginnis | dulek: OK, there's a stronger reason IMO. ;) | 16:23 |
| *** savihou has quit IRC | 16:25 | |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder: Minimize cost of DB exists decorator checks https://review.openstack.org/319177 | 16:28 |
| *** chris_morrell has quit IRC | 16:28 | |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder: Improve DB volume check in LVM manage volume https://review.openstack.org/319323 | 16:28 |
| *** merooney has joined #openstack-cinder | 16:28 | |
| *** timcl has joined #openstack-cinder | 16:30 | |
| *** garthb has joined #openstack-cinder | 16:35 | |
| *** jungleboyj has quit IRC | 16:36 | |
| *** e0ne has quit IRC | 16:38 | |
| jgriffith | geguileo: hmm.. for some reason didn't think you were on line... otherwise would've just discussed that whole get_all thing with you :) | 16:38 |
| jgriffith | geguileo: sorry... I keep using "gor" for auto complete to find your name | 16:38 |
| geguileo | lol | 16:38 |
| jgriffith | geguileo: :) | 16:38 |
| geguileo | jgriffith: I was online but not on the computer | 16:40 |
| geguileo | jgriffith: So I missed the conversation | 16:40 |
| jgriffith | geguileo: haha.. well then I guess it all worked out :) | 16:41 |
| *** permalac has quit IRC | 16:42 | |
| openstackgerrit | John Griffith proposed openstack/cinder: Fix attachment_get_by_* to return entire list https://review.openstack.org/321095 | 16:46 |
| *** frakt has quit IRC | 16:46 | |
| *** frakt has joined #openstack-cinder | 16:47 | |
| *** laughterwym has quit IRC | 16:49 | |
| *** mvk has quit IRC | 16:51 | |
| *** jgregor has quit IRC | 16:52 | |
| *** xyang has quit IRC | 16:54 | |
| *** timcl has quit IRC | 16:55 | |
| *** diablo_rojo has quit IRC | 16:56 | |
| *** baumann has joined #openstack-cinder | 16:59 | |
| *** links has joined #openstack-cinder | 17:00 | |
| *** rcernin has joined #openstack-cinder | 17:02 | |
| *** diogogmt has quit IRC | 17:02 | |
| *** baumann2 has joined #openstack-cinder | 17:02 | |
| *** baumann1 has quit IRC | 17:02 | |
| *** diogogmt has joined #openstack-cinder | 17:03 | |
| *** baumann has quit IRC | 17:04 | |
| *** eharney has quit IRC | 17:04 | |
| *** ociuhandu has joined #openstack-cinder | 17:05 | |
| *** eharney has joined #openstack-cinder | 17:06 | |
| hemanthm | dulek: thanks for the cinder rolling upgrades link. Will ping you once read up on it | 17:06 |
| *** yangyapeng has joined #openstack-cinder | 17:07 | |
| *** links has quit IRC | 17:11 | |
| openstackgerrit | John Griffith proposed openstack/cinder: Fix attachment_get_by_* to return entire list https://review.openstack.org/321095 | 17:11 |
| *** yangyapeng has quit IRC | 17:12 | |
| *** diablo_rojo has joined #openstack-cinder | 17:12 | |
| *** jungleboyj has joined #openstack-cinder | 17:15 | |
| openstackgerrit | xiexs proposed openstack/cinder: Improve the retrieval feature of message https://review.openstack.org/323450 | 17:24 |
| *** lpetrut has joined #openstack-cinder | 17:30 | |
| *** raunak has quit IRC | 17:36 | |
| *** e0ne has joined #openstack-cinder | 17:39 | |
| *** xyang has joined #openstack-cinder | 17:44 | |
| *** raunak has joined #openstack-cinder | 17:45 | |
| *** suyash has joined #openstack-cinder | 17:46 | |
| *** raunak has quit IRC | 17:46 | |
| jgriffith | smcginnis: hemanthm ildikov so long term I think we should do something like this: https://gist.github.com/j-griffith/ce51ef48db38a4916d9b8f7ed20d9101 | 17:50 |
| jgriffith | smcginnis: hemanthm ildikov but that means a lot of change for callers, so may not be the right answer for now. | 17:50 |
| jgriffith | johnthetubaguy: ^^ I'd be curious what you think of something like that | 17:51 |
| ildikov | jgriffith: well, this seems like a big change indeed | 17:53 |
| *** Guest83344 has quit IRC | 17:53 | |
| *** jay-mehta has quit IRC | 17:53 | |
| *** bardia has joined #openstack-cinder | 17:53 | |
| jgriffith | ildikov: yeah it is a very big change for nova anyway | 17:53 |
| jgriffith | ildikov: I figure it won't be something folks are willing to invest in right now, but I did want to share my opinion | 17:54 |
| ildikov | jgriffith: exactly, I do like the way of simplifying things so if it works for Cinder I'm happy to investigate how we can use it in Nova | 17:54 |
| smcginnis | jgriffith: I like the granularity of it. One call to do one overall action. | 17:54 |
| jgriffith | smcginnis: yeah, the mlti-step calls are what have made things kinda ugly for us | 17:55 |
| ildikov | jgriffith: what do you think is there any way to do it as a step by step? | 17:55 |
| jgriffith | smcginnis: and I can't see a real reason why we have them that way still | 17:55 |
| smcginnis | jgriffith: More moving parts to break. More things to get confused about. More steps to possibly overlook. :) | 17:55 |
| ildikov | jgriffith: I think the reason of the current state is purely historical | 17:56 |
| *** timcl has joined #openstack-cinder | 17:56 | |
| jgriffith | ildikov: yeah, and rapid growth :) | 17:56 |
| ildikov | jgriffith: it just never got fixed, so maybe this is the time now :) | 17:56 |
| smcginnis | ildikov: +1 | 17:56 |
| ildikov | jgriffith: yeah, growth too, I agree | 17:56 |
| jgriffith | ildikov: yeah, it's really up to the folks on the Nova side and whether they'd be willing to do this large of a change | 17:57 |
| ildikov | jgriffith: I'm a great fan of refactoring and removing code we don't use/like anymore | 17:57 |
| jgriffith | ildikov: hehe me too | 17:57 |
| *** jay-mehta has joined #openstack-cinder | 17:57 | |
| smcginnis | I think it's going to have to happen at some point. Might as well get it all changed and right than piece meal our way there over 4-5 release cycles. | 17:58 |
| ildikov | jgriffith: smcginnis: johnthetubaguy: we will need to figure out the VM movement cases | 17:58 |
| jgriffith | ildikov: you mean Instance Migration? | 17:58 |
| ildikov | smcginnis: as we have API microversion I hope we don't have to spend 4-5 cycles on this | 17:58 |
| ildikov | jgriffith: that, both types and evacuation, shelve offload and I'm sure there's something I forget | 17:59 |
| smcginnis | ildikov: I'm hoping. | 17:59 |
| ildikov | jgriffith: as if we now switch to another granularity that might affect these operations | 18:00 |
| ildikov | jgriffith: shelve offload I think could be supported by Cinder as a term as well, I mean saving a volume to a shelved instance | 18:00 |
| jgriffith | ildikov: but with multi-attach those just fall out as disconnect/connect calls no? | 18:01 |
| ildikov | jgriffith: I think we already said that with the proper amount of info evacuation could be a force detach + attach combo | 18:01 |
| ildikov | jgriffith: migration is a question in my head from this perspective | 18:01 |
| jgriffith | ildikov: what I mean is like for migrate with attached volume, create new attachment on destination, then go back and disconnect the old one via attachment ID | 18:02 |
| jgriffith | ildikov: since each connection would get it's own unique attachment_id record we just use that everywhere | 18:02 |
| jgriffith | ildikov: you add attachment_id to the bdm for the volume | 18:02 |
| ildikov | jgriffith: with migrate the volume has to support multi-attach if you want to use that method | 18:02 |
| jgriffith | ildikov: well... we cheat and do it now anyway :) | 18:03 |
| jgriffith | anyway... just some thoughts, I'll work on the other method that isn't quite so intrusive | 18:03 |
| ildikov | jgriffith: with migrate the volume has to support multi-attach if you want to use that method | 18:04 |
| jgriffith | but long term I think we should consider something like the new flow | 18:04 |
| *** raunak has joined #openstack-cinder | 18:04 | |
| ildikov | I totally support you on the new flow, this is why I brought up these use cases as if we can address those with the new model we already took a step towards it | 18:05 |
| *** ChubYann has joined #openstack-cinder | 18:08 | |
| *** ChubYann has quit IRC | 18:09 | |
| *** ChubYann has joined #openstack-cinder | 18:09 | |
| ildikov | jgriffith: can your initialize_connection work be done in a way that can be reused for this new flow in any extent? | 18:09 |
| *** timcl has quit IRC | 18:10 | |
| ildikov | jgriffith: I'm just asking as it would be good to have some progress while we're discussing the bigger change although it would be better if we could work a bit proactively as well | 18:11 |
| jgriffith | ildikov: ya, actually what that would do is just call those new versions of initialize_connection etc | 18:15 |
| ildikov | jgriffith: ok, cool, so basically the new API would hide these functions in the future from the users, right? | 18:16 |
| jgriffith | ildikov: yup, that's one way to do it | 18:16 |
| *** xyang has quit IRC | 18:17 | |
| ildikov | jgriffith: ok, cool, got it | 18:17 |
| *** earlephilhower has joined #openstack-cinder | 18:18 | |
| *** timcl has joined #openstack-cinder | 18:19 | |
| *** Swanson has quit IRC | 18:20 | |
| hemna | yough | 18:24 |
| hemna | jgriffith, so yah that looks good. I think this is what scottda and I were hoping for in the refactor approach we started talking about a while back. | 18:27 |
| jgriffith | hemna: ahh | 18:28 |
| hemna | looks good | 18:28 |
| hemna | make it so | 18:28 |
| hemna | :P | 18:28 |
| jgriffith | hemna: well, I'm game if Nova folks are up to it | 18:28 |
| jgriffith | hemna: really need input from that side before doing much more there though I think | 18:28 |
| jgriffith | hemna: I BELIEVE I can make most things work with existing calls, but it doesn't adress some of the issues with multi-attach and race conditions etc | 18:29 |
| hemna | I'm still trying to hash out the removal of the check_attack stuff | 18:30 |
| hemna | nova is such a pita | 18:30 |
| hemna | overly complex, etc | 18:30 |
| hemna | they seem to do a lot of checking before they check, and then do it kinda stuff. | 18:31 |
| jgriffith | hemna: hehe... like meetings to plan meetings :) | 18:31 |
| hemna | yah pretty much :) | 18:31 |
| ildikov | hemna: jgriffith: yeah, those checks are overkill... | 18:32 |
| ildikov | hemna: jgriffith: I'm still on the side of trying to plan a few steps to get to this proposal | 18:34 |
| ildikov | hemna: jgriffith: that might keep Nova folks a bit more relaxed about it too | 18:35 |
| ildikov | hemna: jgriffith: I haven't sent out the reminder yet, but if you're available for the Thursday backup slot this week for our meeting I think we should bring this up there to get some initial thoughts from Nova people as well | 18:37 |
| jgriffith | ildikov: works for me | 18:37 |
| ildikov | jgriffith: cool, tnx | 18:39 |
| *** diablo_rojo has quit IRC | 18:39 | |
| *** merooney has quit IRC | 18:40 | |
| *** xyang has joined #openstack-cinder | 18:42 | |
| *** xyang has quit IRC | 18:45 | |
| *** xyang has joined #openstack-cinder | 18:47 | |
| *** jungleboyj has quit IRC | 18:47 | |
| *** baumann has joined #openstack-cinder | 18:47 | |
| *** baumann1 has joined #openstack-cinder | 18:48 | |
| *** baumann has quit IRC | 18:48 | |
| *** baumann2 has quit IRC | 18:49 | |
| *** xyang has quit IRC | 18:49 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements https://review.openstack.org/323063 | 18:50 |
| *** baumann has joined #openstack-cinder | 18:50 | |
| *** merooney has joined #openstack-cinder | 18:51 | |
| *** baumann1 has quit IRC | 18:53 | |
| *** diablo_rojo has joined #openstack-cinder | 18:56 | |
| *** xyang has joined #openstack-cinder | 18:57 | |
| *** baumann has left #openstack-cinder | 18:57 | |
| *** baumann has joined #openstack-cinder | 18:58 | |
| *** jungleboyj has joined #openstack-cinder | 19:00 | |
| *** diablo_rojo has quit IRC | 19:01 | |
| *** diablo_rojo1 has joined #openstack-cinder | 19:01 | |
| *** diablo_rojo1 has quit IRC | 19:02 | |
| *** diablo_rojo has joined #openstack-cinder | 19:02 | |
| *** timcl has left #openstack-cinder | 19:04 | |
| *** timcl has joined #openstack-cinder | 19:04 | |
| *** timcl has left #openstack-cinder | 19:05 | |
| *** timcl has joined #openstack-cinder | 19:06 | |
| *** jungleboyj has quit IRC | 19:06 | |
| *** e0ne has quit IRC | 19:07 | |
| *** karthikp has quit IRC | 19:08 | |
| *** jgregor has joined #openstack-cinder | 19:09 | |
| *** Swanson has joined #openstack-cinder | 19:09 | |
| *** akshai has quit IRC | 19:09 | |
| *** yangyapeng has joined #openstack-cinder | 19:09 | |
| *** akshai has joined #openstack-cinder | 19:11 | |
| *** raunak has quit IRC | 19:11 | |
| Swanson | Has pycharm ever successfully installed a pip package on its own? Or just locked up every single time? | 19:12 |
| *** yangyapeng has quit IRC | 19:14 | |
| *** e0ne has joined #openstack-cinder | 19:20 | |
| *** xyang has quit IRC | 19:21 | |
| *** jungleboyj has joined #openstack-cinder | 19:22 | |
| bardia | can any of the cores take a look at my review and finalize it please? https://review.openstack.org/#/c/318382/4 | 19:23 |
| *** xyang has joined #openstack-cinder | 19:25 | |
| *** sheel has quit IRC | 19:25 | |
| *** e0ne has quit IRC | 19:28 | |
| *** eharney has quit IRC | 19:29 | |
| *** jungleboyj has quit IRC | 19:31 | |
| *** xyang has quit IRC | 19:32 | |
| *** Lee1092 has quit IRC | 19:33 | |
| *** xyang has joined #openstack-cinder | 19:33 | |
| *** eharney has joined #openstack-cinder | 19:36 | |
| *** eharney has quit IRC | 19:39 | |
| *** eharney has joined #openstack-cinder | 19:41 | |
| *** mvk has joined #openstack-cinder | 19:42 | |
| *** jungleboyj has joined #openstack-cinder | 19:43 | |
| *** merooney has quit IRC | 19:44 | |
| bswartz | scottda: ping | 19:44 |
| *** sdake has joined #openstack-cinder | 19:45 | |
| bswartz | scottda: the midcycle etherpad says the hotel rooms have a "HP discount rate", but last time I tried to use that they wanted to know if I actually worked for HP | 19:45 |
| *** sdake_ has joined #openstack-cinder | 19:49 | |
| *** raunak has joined #openstack-cinder | 19:50 | |
| *** porrua has quit IRC | 19:50 | |
| *** pots2 has quit IRC | 19:51 | |
| *** sdake has quit IRC | 19:51 | |
| *** lpetrut has quit IRC | 19:53 | |
| *** lgreg has joined #openstack-cinder | 19:54 | |
| *** crose has joined #openstack-cinder | 19:55 | |
| *** crose has quit IRC | 19:55 | |
| *** crose has joined #openstack-cinder | 19:56 | |
| smcginnis | bswartz: That's just you. They have you on their bad list. | 19:56 |
| smcginnis | :) | 19:56 |
| smcginnis | bswartz: I was able to book it no problem. I didn't say I worked for HP, just that I was going to be there for meetings at the HP office. | 19:56 |
| * hemna wants to go to the midcycle | 19:57 | |
| * hemna has a sad | 19:57 | |
| *** jgregor has quit IRC | 19:57 | |
| smcginnis | hemna: No luck, huh? | 19:57 |
| openstackgerrit | Karthik Prabhu Vinod proposed openstack/cinder: Add ServiceBinary Enum https://review.openstack.org/322977 | 19:58 |
| *** crose has quit IRC | 19:59 | |
| Swanson | hemna, I bet reviewing this https://review.openstack.org/#/c/315226/ would cheer you up. | 19:59 |
| Swanson | (Worth a try.) | 19:59 |
| hemna | nothing yet | 19:59 |
| *** Apoorva has quit IRC | 19:59 | |
| *** lprice1 has joined #openstack-cinder | 20:01 | |
| *** lprice has quit IRC | 20:02 | |
| Swanson | hemna, I don't know if you're cheered up but I sure am! Thanks1 | 20:04 |
| *** Apoorva has joined #openstack-cinder | 20:06 | |
| *** jgregor has joined #openstack-cinder | 20:08 | |
| *** garthb_ has joined #openstack-cinder | 20:09 | |
| *** jsheeren has joined #openstack-cinder | 20:10 | |
| *** garthb has quit IRC | 20:10 | |
| *** jgregor has quit IRC | 20:14 | |
| bswartz | smcginnis: yeah I've been able to book in the past but when I actually showed up they wanted evidence that I was with HP | 20:16 |
| *** pots1 has joined #openstack-cinder | 20:16 | |
| bswartz | you're probably right that I'm just a shady looking guy and they put my on their bad list | 20:16 |
| smcginnis | bswartz: Just show then you NA business card and tell them the news isn't public yet. :D | 20:16 |
| bswartz | now that *would* be shady | 20:17 |
| smcginnis | hehe | 20:17 |
| ildikov | jgriffith: I added a few comments to the initialize_connection patch | 20:28 |
| *** timcl has quit IRC | 20:30 | |
| openstackgerrit | John Griffith proposed openstack/cinder: Introduce new calls for attaching volumes https://review.openstack.org/323571 | 20:31 |
| jgriffith | ildikov: smcginnis ^^ I'll update with more info/detail shortly on how this works | 20:31 |
| jgriffith | ildikov: smcginnis be back in like 10 minutes | 20:32 |
| *** cknight has quit IRC | 20:32 | |
| ildikov | jgriffith: regarding host_name, I think we should either change the API so that Nova can send it or remove the need for it in Cinder | 20:33 |
| *** akerr has quit IRC | 20:33 | |
| ildikov | jgriffith: cool, tnx, I'm reading it already :) | 20:33 |
| *** jgregor has joined #openstack-cinder | 20:36 | |
| *** jungleboyj has quit IRC | 20:37 | |
| *** julim has quit IRC | 20:38 | |
| ildikov | jgriffith: smcginnis: regarding naming I wonder how much it is confusing in the sense of having the connect term in os-brick | 20:39 |
| smcginnis | ildikov: Not sure I'm following whay you mean. | 20:40 |
| smcginnis | *what | 20:40 |
| *** eharney has quit IRC | 20:42 | |
| *** haypo has left #openstack-cinder | 20:43 | |
| ildikov | smcginnis: I meant the disconnect_volume Nova calls at detach time for instance, but maybe only I got confused because of this when I saw the renaming of attach to connect | 20:45 |
| smcginnis | ildikov: Ah, I see what you're saying. | 20:46 |
| *** raildo is now known as raildo-afk | 20:47 | |
| hemna | ildikov, we need the host name | 20:47 |
| ildikov | smcginnis: well, naming sucks, and as I'm still just learning how things are working or not that much working here I can get confused very easily, so I was just thinking out loud :) | 20:47 |
| hemna | well at least some of the drivers do, in order to export the volume from the array | 20:48 |
| ildikov | hemna: in attach you mean? | 20:48 |
| smcginnis | ildikov: No, I think half of our problems are due to naming not being clear enough to avoid confusion. ;) | 20:48 |
| hemna | ildikov, yup | 20:48 |
| *** grumpycatt has quit IRC | 20:48 | |
| ildikov | hemna: hmm, in attach_volume() shouldn't it come from Nova, when it calls attach? | 20:50 |
| hemna | *sigh* | 20:50 |
| hemna | trying to make sense of the nova client | 20:51 |
| ildikov | hemna: I'm playing with jgriffith's patch: https://review.openstack.org/#/c/320721/1/cinder/volume/manager.py in devstack and host_name here is None | 20:51 |
| *** Apoorva has quit IRC | 20:51 | |
| hemna | the nova boot process = pain | 20:51 |
| ildikov | hemna: but that patch does not really change host_name so I assume it's None originally as well | 20:51 |
| *** Apoorva has joined #openstack-cinder | 20:51 | |
| ildikov | hemna: but isn't this what's blocked in the Cinder API? | 20:52 |
| hemna | yuh | 20:52 |
| ildikov | hemna: I mean that you cannot specify host and instance in attach | 20:52 |
| hemna | correct | 20:53 |
| hemna | I'm trying to track down the BFV case | 20:53 |
| hemna | and find where reserve_volume is ever called | 20:53 |
| hemna | I can't find it | 20:53 |
| hemna | :( | 20:53 |
| ildikov | smcginnis: right, one of the most difficult tasks if you try to do it right... :) | 20:53 |
| ildikov | hemna: in BFV case they don't call reserve_volume | 20:54 |
| ildikov | hemna: I tested it in Devstack | 20:54 |
| ildikov | it's not called at all | 20:54 |
| ildikov | at least in case of attaching an already existing volume at boot time | 20:54 |
| *** lgreg has quit IRC | 20:55 | |
| ildikov | but I guess it's the same scenario when Nova creates the volume, although I haven't tested that one | 20:55 |
| hemna | so yah that's what I'm seeing as well | 20:55 |
| hemna | which is bad mmmkay | 20:55 |
| hemna | nova is so painful | 20:55 |
| ildikov | I think we talked about it last week | 20:56 |
| ildikov | I was wondering whether we could have it as the regular attach | 20:56 |
| hemna | I tried and tried and I can't reproduce the same issues in unit tests | 20:56 |
| ildikov | I don't really see why attach differs in case of BFV | 20:56 |
| ildikov | which issues you mean? | 20:56 |
| hemna | the accessing the bdm values | 20:56 |
| hemna | fails every time in the real case | 20:57 |
| hemna | and works in the unit tests | 20:57 |
| hemna | bdm.get('volume_id') | 20:57 |
| hemna | None | 20:57 |
| hemna | bdm.volume_id works | 20:57 |
| ildikov | isn't it because of some mocks? | 20:57 |
| hemna | the BDM objects are VERY different in the unit tess | 20:57 |
| hemna | tests | 20:57 |
| hemna | compared to the real runtime | 20:57 |
| hemna | can't quite comprehend how this is allowed | 20:57 |
| ildikov | I thought so, that was my guess as well | 20:58 |
| hemna | the fakes in the unit tests | 20:58 |
| hemna | are not even the same structure or type of object | 20:58 |
| hemna | smh | 20:58 |
| hemna | so they aren't even testing the real thing | 20:58 |
| ildikov | there are even fake methods, like for check_attach in Nova | 20:58 |
| *** diablo_rojo has quit IRC | 20:58 | |
| hemna | so I give up | 20:59 |
| ildikov | yeah, sometimes it's hard to find the border between unit and functional tests... :( | 20:59 |
| hemna | it makes no sense | 20:59 |
| jgriffith | hemna: ildikov smcginnis hmm... looks like I stirred up a hornets nest and left :) | 20:59 |
| hemna | oh well | 20:59 |
| hemna | I can reproduce the issue every time in the real BFV case | 20:59 |
| hemna | but johnthetubaguy wanted me to reproduce it in unit tests...which is impossible. | 20:59 |
| hemna | :( | 20:59 |
| hemna | not even the same objects | 20:59 |
| hemna | ultimately, I need to get nova to stop trying to check_attach in the BFV | 21:00 |
| ildikov | jgriffith: :) | 21:00 |
| hemna | as my change calls reserve instead of check_attach up front | 21:00 |
| jgriffith | ildikov: :) So I reworked things here: https://review.openstack.org/#/c/323571/ With the hope of making a clean transition on things | 21:00 |
| *** lprice has joined #openstack-cinder | 21:00 | |
| hemna | which causes failure later, because it tries to check_attack | 21:00 |
| * hemna head is spinning | 21:00 | |
| ildikov | hemna: don't give up, let's point out to Nova folks that this is the case and get a few eyes on it, they should know as well | 21:01 |
| ildikov | jgriffith: yeah, I recognized since then, host_name looks fine in it :) | 21:01 |
| ildikov | jgriffith: the rest I need to understand :) | 21:01 |
| *** lprice1 has quit IRC | 21:01 | |
| ildikov | hemna: where is the other check attach there? | 21:02 |
| johnthetubaguy | just push it back to us, I can try take a look, to see whats going on again | 21:02 |
| johnthetubaguy | well, in the morning | 21:02 |
| *** akshai has quit IRC | 21:02 | |
| ildikov | hemna: I guess you changed the first check_attach to reserve, right? | 21:02 |
| openstackgerrit | Merged openstack/cinder: Dell SC: Switched REST calls to async where available https://review.openstack.org/315226 | 21:02 |
| openstackgerrit | Jay Mehta proposed openstack/cinder: Fixed _create_3par_iscsi_host(), added unit test for 3PAR https://review.openstack.org/317763 | 21:03 |
| ildikov | johnthetubaguy: yeah, this is definitely a morning topic rather than quick before sleep one | 21:03 |
| hemna | ildikov, http://paste.openstack.org/show/506653/ | 21:03 |
| hemna | that nonsense | 21:03 |
| *** jgregor has quit IRC | 21:04 | |
| ildikov | hemna: :S | 21:05 |
| ildikov | hemna: the BFV case should be fine I think as I guess the check_attach is called from attach as do_check_attach is not set to False in BFV case as they currently don't call reserve at all | 21:06 |
| hemna | well, I'm trying to remove the check_attach call early on | 21:07 |
| hemna | race conditions, etc | 21:07 |
| hemna | it should reserve the volume up front | 21:07 |
| hemna | and then not check_attach | 21:07 |
| ildikov | it's called in validate_bdm or somewhere like this as well IIRC | 21:07 |
| jgriffith | hemna: you could just take a look at the patch I just proposed and use it's get_connection instead | 21:07 |
| jgriffith | hemna: do a simple search/replace in nova :) | 21:08 |
| hemna | bdm.attach calls really shouldn't check_attach | 21:08 |
| hemna | jgriffith, :) | 21:08 |
| *** martyturner has quit IRC | 21:08 | |
| jgriffith | hemna: IIRC some of that was there for things like getting everything put back together after a compute node reboot... but it's been a while and I could be wrong | 21:08 |
| ildikov | yeah, but if you call reserve when the first check_attach happens the second one should not be needed | 21:08 |
| ildikov | hemna: it's like in your attach diagram, the second check_attach is not called in reality | 21:09 |
| *** Wenjun has quit IRC | 21:10 | |
| hemna | well, if I replace the first check_attach call with reserve_volume | 21:10 |
| hemna | it fails | 21:10 |
| ildikov | hemna: so it's basically not needed, we need to ensure that reserve is called every time it has to be called | 21:10 |
| hemna | because nova does call check_attach later, and it's already in 'attaching' | 21:11 |
| hemna | so it pukes | 21:11 |
| ildikov | hemna: set do_check_attach False when calling attach | 21:11 |
| *** garthb_ has quit IRC | 21:12 | |
| johnthetubaguy | so most of this stuff was added because we didn't understand what we were doing, the reserve_volume makes it much clearer that we don't need to check_attach later on | 21:12 |
| openstackgerrit | Jay Mehta proposed openstack/cinder: Fixed _create_3par_iscsi_host(), added unit test for 3PAR https://review.openstack.org/317763 | 21:13 |
| *** eharney has joined #openstack-cinder | 21:13 | |
| ildikov | johnthetubaguy: +1 | 21:13 |
| hemna | I'm just hacking at this point | 21:13 |
| hemna | it's frustrating | 21:13 |
| johnthetubaguy | feel free to come back to us with what you think should happen, and make us check that it does what you want | 21:13 |
| hemna | I'm trying :) | 21:13 |
| hemna | but then I run into bdm.get('volume_id') always returning None | 21:14 |
| *** garthb_ has joined #openstack-cinder | 21:14 | |
| johnthetubaguy | I mean broken tests and random commented out code with TODOs would be fine, with a good commit message saying what the intent is | 21:14 |
| johnthetubaguy | yeah bdm involves all disks, not just volumes | 21:14 |
| *** jsheeren has quit IRC | 21:14 | |
| johnthetubaguy | oh, wait, that other thing | 21:14 |
| johnthetubaguy | yeah | 21:14 |
| johnthetubaguy | I wish I knew how that was happening | 21:15 |
| hemna | I tried to reproduce it in unit tests | 21:15 |
| johnthetubaguy | its worth trying to catch dansmith about that, he is the objects expert | 21:15 |
| hemna | but couldn't because the faked out bdm is a different object and doesn't act the same way :( | 21:15 |
| johnthetubaguy | hmm, thats odd | 21:16 |
| johnthetubaguy | anyways, I must stop, its late, just around for the TC meeting | 21:16 |
| *** gouthamr has quit IRC | 21:16 | |
| ildikov | johnthetubaguy: regarding check_attach we're trying to figure out whether we always call reserve when it's supposed to be called | 21:16 |
| ildikov | johnthetubaguy: I sent out the mail with a slot to Thursday as a back up slot for this week | 21:17 |
| ildikov | johnthetubaguy: we might touch these areas there if we cannot clarify these things earlier | 21:17 |
| hemna | http://paste.openstack.org/show/506654/ | 21:17 |
| hemna | that shows the fake bdm | 21:18 |
| hemna | and it acts different | 21:18 |
| johnthetubaguy | yeah, I should double check, I am off work thursday, driving around for some last min stuff at the moment | 21:18 |
| johnthetubaguy | yeah, I was really meaning just create a real bdm object in a test, then play with it to see if its the same | 21:19 |
| johnthetubaguy | you can do it in one of these functional ones: | 21:19 |
| ildikov | johnthetubaguy: ah, ok, sorry I didn't remember :( | 21:19 |
| johnthetubaguy | https://github.com/openstack/nova/blob/master/nova/tests/functional/db/test_compute_node.py | 21:19 |
| *** baumann has left #openstack-cinder | 21:19 | |
| johnthetubaguy | anyways, catch you later, must run | 21:20 |
| openstackgerrit | Jay Mehta proposed openstack/cinder: Implement match-set type 3PAR FC VLUN creation https://review.openstack.org/309613 | 21:24 |
| _alastor_ | Hey Folks, does anyone have an example of where the capabilities shown for a driver in "cinder get-capabilities host@driver" are populated? I assume the driver has some control over what shows up there. | 21:24 |
| *** ducttape_ has quit IRC | 21:25 | |
| ildikov | hemna: I need to go too, I have a pretty early morning meeting tomorrow | 21:26 |
| ildikov | hemna: I will look into the BFV check_attach too and ping you tomorrow | 21:28 |
| hemna | ildikov, ok | 21:28 |
| *** edmondsw has quit IRC | 21:29 | |
| ildikov | hemna: take care, chat tomorrow | 21:30 |
| *** ducttape_ has joined #openstack-cinder | 21:31 | |
| _alastor_ | nevermind, I think I found it | 21:34 |
| *** kfarr has quit IRC | 21:36 | |
| openstackgerrit | Szymon Wróblewski proposed openstack/cinder: Update initialize_connection to use versionedobjects https://review.openstack.org/238894 | 21:39 |
| *** sgotliv has quit IRC | 21:39 | |
| *** lprice has quit IRC | 21:42 | |
| *** andymaier_ has joined #openstack-cinder | 21:43 | |
| *** krtaylor has quit IRC | 21:47 | |
| *** karthikp has joined #openstack-cinder | 21:48 | |
| openstackgerrit | Tom Swanson proposed openstack/cinder: Dell SC: Added support for failover_host failback https://review.openstack.org/310064 | 22:06 |
| *** darosale has quit IRC | 22:06 | |
| openstackgerrit | Szymon Wróblewski proposed openstack/cinder: Replace locks in remotefs backend driver https://review.openstack.org/270240 | 22:07 |
| *** eharney has quit IRC | 22:09 | |
| *** Julien-zte has joined #openstack-cinder | 22:10 | |
| *** sdake has joined #openstack-cinder | 22:10 | |
| *** sdake_ has quit IRC | 22:12 | |
| *** ametts has quit IRC | 22:14 | |
| *** tpsilva has quit IRC | 22:17 | |
| *** JoseMello has quit IRC | 22:19 | |
| *** xyang1 has quit IRC | 22:26 | |
| *** rueben_ has quit IRC | 22:29 | |
| openstackgerrit | Tom Swanson proposed openstack/cinder: Dell SC: Added support for failover_host failback https://review.openstack.org/310064 | 22:30 |
| *** ducttape_ has quit IRC | 22:36 | |
| *** ducttape_ has joined #openstack-cinder | 22:36 | |
| *** bardia has quit IRC | 22:40 | |
| *** bardia has joined #openstack-cinder | 22:40 | |
| *** ducttape_ has quit IRC | 22:41 | |
| *** enriquetaso has quit IRC | 22:46 | |
| *** julim has joined #openstack-cinder | 22:56 | |
| *** julim has quit IRC | 22:59 | |
| *** andymaier_ has quit IRC | 23:15 | |
| *** sdague has quit IRC | 23:16 | |
| *** diogogmt has quit IRC | 23:19 | |
| openstackgerrit | Vasanthi Thirumalai proposed openstack/cinder: Violin Memory ISCSI storage for 7000 series AFA https://review.openstack.org/269796 | 23:20 |
| *** eharney has joined #openstack-cinder | 23:21 | |
| *** xinli has quit IRC | 23:26 | |
| *** cknight has joined #openstack-cinder | 23:29 | |
| *** cknight1 has joined #openstack-cinder | 23:31 | |
| *** eharney has quit IRC | 23:33 | |
| *** cknight has quit IRC | 23:34 | |
| bardia | smcginnis: https://review.openstack.org/#/c/318382/ :) | 23:35 |
| *** arch-nemesis has quit IRC | 23:35 | |
| *** ducttape_ has joined #openstack-cinder | 23:41 | |
| *** cknight1 has quit IRC | 23:43 | |
| *** yhayashi has joined #openstack-cinder | 23:49 | |
| *** Julien-zte has quit IRC | 23:50 | |
| *** chlong has quit IRC | 23:53 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!