Monday, 2015-07-27

*** alexpilotti has quit IRC00:00
*** chlong has joined #openstack-cinder00:00
*** harlowja_at_home has quit IRC00:05
*** merooney has quit IRC00:26
*** Apoorva has joined #openstack-cinder00:26
*** Apoorva has quit IRC00:31
*** heyun has joined #openstack-cinder00:33
*** r-daneel has joined #openstack-cinder00:38
*** laughterwym has joined #openstack-cinder00:43
*** r-daneel has quit IRC00:44
*** salv-orlando has joined #openstack-cinder00:45
*** salv-orl_ has joined #openstack-cinder00:49
*** salv-orlando has quit IRC00:50
*** salv-orl_ has quit IRC00:54
*** dims has joined #openstack-cinder00:58
*** lixiaoy1 has joined #openstack-cinder01:04
*** annashen has joined #openstack-cinder01:11
*** ericksonsantos has quit IRC01:11
*** harlowja_at_home has joined #openstack-cinder01:12
*** lixiaoy11 has joined #openstack-cinder01:15
*** ericksonsantos has joined #openstack-cinder01:15
*** lixiaoy1 has quit IRC01:17
*** lixiaoy1 has joined #openstack-cinder01:18
openstackgerritxing-yang proposed openstack/cinder: Clone CG
*** lixiaoy11 has quit IRC01:21
*** jasondotstar has joined #openstack-cinder01:26
*** harlowja_at_home has quit IRC01:26
openstackgerritDave Chen proposed openstack/python-cinderclient: Add commands for modifying image metadata
*** davechen has joined #openstack-cinder01:31
*** laughterwym_ has joined #openstack-cinder01:32
*** laughterwym has quit IRC01:35
*** Lee1092 has joined #openstack-cinder01:37
openstackgerritRaunak Kumar proposed openstack/cinder: Add extra spec capability for Nimble Cinder Driver
openstackgerritwanghao proposed openstack/cinder: Add pagination to backups
*** salv-orlando has joined #openstack-cinder01:53
*** daneyon has joined #openstack-cinder01:54
*** daneyon has quit IRC01:59
*** salv-orl_ has joined #openstack-cinder02:01
openstackgerritwanghao proposed openstack/cinder: Implement function to manage/unmanage snapshots
*** salv-orlando has quit IRC02:03
*** salv-orl_ has quit IRC02:06
*** smoriya has joined #openstack-cinder02:07
*** zhenguo has joined #openstack-cinder02:07
openstackgerritwanghao proposed openstack/cinder: Support volume backup of qcow2 format volume
openstackgerritwanghao proposed openstack/cinder: Incremental backup improvements for L
*** annashen has quit IRC02:35
*** annashen has joined #openstack-cinder02:37
*** annashen_ has joined #openstack-cinder02:40
*** annashen_ has quit IRC02:41
*** annashen has quit IRC02:42
*** annashen has joined #openstack-cinder02:42
*** annashen has quit IRC02:45
*** annashen has joined #openstack-cinder02:46
*** zhonghua-lee has joined #openstack-cinder02:49
*** annashen has quit IRC02:50
*** annashen has joined #openstack-cinder02:50
*** hakimo has joined #openstack-cinder02:52
*** hakimo_ has quit IRC02:54
*** markvoelker has joined #openstack-cinder02:58
openstackgerritMitsuhiro Tanino proposed openstack/cinder: Adds framework for get_capabilities() feature
*** Liu has joined #openstack-cinder02:58
openstackgerritxiaoqin proposed openstack/cinder: Port StorwizeSVCDriver to use ABCMeta driver model
*** dims has quit IRC02:59
openstackgerritliuxinguo proposed openstack/cinder: Add iSCSI multipath support for Huawei driver
*** markvoelker has quit IRC03:02
*** kcaj has quit IRC03:07
*** salv-orlando has joined #openstack-cinder03:08
*** annashen has quit IRC03:09
*** smoriya has quit IRC03:12
*** annashen has joined #openstack-cinder03:12
*** salv-orlando has quit IRC03:12
*** salv-orlando has joined #openstack-cinder03:13
*** kcaj has joined #openstack-cinder03:14
openstackgerritMitsuhiro Tanino proposed openstack/cinder: Efficient volume copy for generic volume migration
*** salv-orlando has quit IRC03:17
*** changbl has quit IRC03:19
*** jasondotstar has quit IRC03:20
*** smoriya has joined #openstack-cinder03:22
*** nzrf has left #openstack-cinder03:26
openstackgerritVincent Hou proposed openstack/cinder: Volume status management during migration
openstackgerritVincent Hou proposed openstack/cinder: Volume status management during migration
*** Apoorva has joined #openstack-cinder03:29
*** harlowja_at_home has joined #openstack-cinder03:29
*** coolsvap|away is now known as coolsvap03:30
*** changbl has joined #openstack-cinder03:32
*** Apoorva has quit IRC03:33
*** daneyon has joined #openstack-cinder03:43
*** zhonghua-li has joined #openstack-cinder03:43
openstackgerritYuanHui Xu proposed openstack/cinder: Add mock cases for IBM FlashSystem
*** zhonghua-lee has quit IRC03:46
*** daneyon has quit IRC03:48
*** annashen has quit IRC03:48
*** annashen has joined #openstack-cinder03:48
*** zul has quit IRC03:52
openstackgerritEdwin Wang proposed openstack/cinder: Test for dependency
*** zul has joined #openstack-cinder03:53
openstackgerritxing-yang proposed openstack/cinder: Clone CG
*** dims has joined #openstack-cinder04:00
*** haomaiwang has quit IRC04:02
*** haomaiwang has joined #openstack-cinder04:04
*** dims has quit IRC04:06
openstackgerritxing-yang proposed openstack/cinder: Add multiple pools support to VMAX driver
*** annashen has quit IRC04:16
*** annashen has joined #openstack-cinder04:16
*** annashen has quit IRC04:19
*** daneyon has joined #openstack-cinder04:20
openstackgerritxing-yang proposed openstack/cinder: EMC VMAX Create CG from CG Snapshot
*** daneyon_ has joined #openstack-cinder04:22
*** daneyon has quit IRC04:22
*** daneyon_ has quit IRC04:23
openstackgerritwanghao proposed openstack/cinder: Refactoring of manager's manage_existing flow
*** DericHorn-HP has joined #openstack-cinder04:26
*** zhonghua-li has quit IRC04:26
openstackgerritYuanHui Xu proposed openstack/cinder: Add mock cases for IBM FlashSystem
*** annashen has joined #openstack-cinder04:28
openstackgerritxing-yang proposed openstack/python-cinderclient: CLI: Clone CG
*** annashen has quit IRC04:35
*** zhonghua-li has joined #openstack-cinder04:35
*** krypton has joined #openstack-cinder04:40
openstackgerritVincent Hou proposed openstack/cinder: Adds the migration progress support for migration
*** markvoelker has joined #openstack-cinder04:58
*** salv-orlando has joined #openstack-cinder05:01
*** Zhongjun has joined #openstack-cinder05:02
*** markvoelker has quit IRC05:03
*** salv-orl_ has joined #openstack-cinder05:03
*** salv-orlando has quit IRC05:05
openstackgerritAndrey Pavlov proposed openstack/cinder: Fix doc string definitions
*** ub has quit IRC05:09
*** salv-orl_ has quit IRC05:11
*** salv-orlando has joined #openstack-cinder05:12
thingeewinston-d: ping05:13
*** pradipta has joined #openstack-cinder05:13
*** salv-orlando has quit IRC05:16
*** salv-orlando has joined #openstack-cinder05:19
*** lixiaoy1 has quit IRC05:23
*** avishay has joined #openstack-cinder05:24
*** lixiaoy1 has joined #openstack-cinder05:24
*** Apoorva has joined #openstack-cinder05:26
*** vlaza has joined #openstack-cinder05:33
*** BharatK has joined #openstack-cinder05:33
*** lpetrut has joined #openstack-cinder05:34
*** DericHorn-HP has quit IRC05:35
*** annashen has joined #openstack-cinder05:35
*** annashen has quit IRC05:40
*** ankit_ag has joined #openstack-cinder05:44
*** deepakcs has joined #openstack-cinder05:47
*** krypton has quit IRC05:50
*** nkrinner has joined #openstack-cinder05:50
*** krypton has joined #openstack-cinder05:51
*** avishay has quit IRC05:52
*** avishay_ has joined #openstack-cinder05:52
*** krypton_ has joined #openstack-cinder05:56
*** krypton has quit IRC05:59
*** jwcroppe has joined #openstack-cinder06:00
*** fifieldt has joined #openstack-cinder06:00
*** krypton_ has quit IRC06:01
*** dims has joined #openstack-cinder06:03
openstackgerritMike Perez proposed openstack/cinder: Remove deprecated SimpleScheduler
*** Apoorva has quit IRC06:04
*** harlowja_at_home has quit IRC06:04
*** agarciam has joined #openstack-cinder06:07
*** dims has quit IRC06:09
*** zhonghua-li has quit IRC06:09
*** markus_z has joined #openstack-cinder06:10
*** daneyon has joined #openstack-cinder06:12
openstackgerritMike Perez proposed openstack/cinder-specs: Updating Get Volume Driver Capabilities Spec
*** annashen has joined #openstack-cinder06:14
openstackgerritAbhijeet Malawade proposed openstack/cinder: Remove unused context parameter
*** daneyon has quit IRC06:17
*** markus_z has quit IRC06:17
*** vilobhmm has joined #openstack-cinder06:19
*** annashen has quit IRC06:23
thingeedavechen: ping06:26
*** alexpilotti has joined #openstack-cinder06:26
*** lan has quit IRC06:27
*** annashen has joined #openstack-cinder06:29
thingeedavechen: commented on ... just curious how one would see their volume glance metdata.06:29
thingeedavechen: obvious I could just look in the db, but there appears to be other cinderclient commands to see metadata, so maybe we need one for this as well?06:30
*** vilobhmm1 has joined #openstack-cinder06:30
*** annashen has quit IRC06:31
*** vilobhmm has quit IRC06:33
*** ub has joined #openstack-cinder06:33
*** annashen has joined #openstack-cinder06:33
*** sgotliv_ has quit IRC06:33
thingeejgriffith: by chance would you be able to push an update of this week, so we can push that along for ya based on consensus in the last meeting of auto-failover not being necessary this time around?06:39
*** annashen has quit IRC06:39
*** krypton has joined #openstack-cinder06:40
*** vilobhmm1 has quit IRC06:43
*** avishay_ has quit IRC06:44
*** avishay__ has joined #openstack-cinder06:44
*** annashen has joined #openstack-cinder06:46
*** annashen has quit IRC06:49
*** annashen has joined #openstack-cinder06:51
*** rushiagr_away is now known as rushiagr06:53
*** annashen has quit IRC06:53
*** vilobhmm has joined #openstack-cinder06:59
*** markvoelker has joined #openstack-cinder06:59
openstackgerritVipin Balachandran proposed openstack/cinder: VMware: Set virtual disk UUID to volume ID
openstackgerritVipin Balachandran proposed openstack/cinder: VMware: Add volume ID in vCenter's volume config
*** krypton has quit IRC07:03
*** krypton has joined #openstack-cinder07:04
*** markvoelker has quit IRC07:04
*** lpetrut has quit IRC07:09
*** markus_z has joined #openstack-cinder07:11
*** ronis has joined #openstack-cinder07:13
*** lpetrut has joined #openstack-cinder07:16
*** LiuXinguo has joined #openstack-cinder07:23
*** Liu has quit IRC07:26
*** zhonghua-li has joined #openstack-cinder07:28
*** chlong has quit IRC07:28
*** jwcroppe has quit IRC07:30
*** jwcroppe has joined #openstack-cinder07:31
*** vilobhmm has quit IRC07:38
*** salv-orlando has quit IRC07:39
*** fifieldt has quit IRC07:42
*** zhonghua-li has quit IRC07:43
*** zhonghua-lee has joined #openstack-cinder07:43
*** jistr has joined #openstack-cinder07:44
*** salv-orlando has joined #openstack-cinder07:45
*** bluex-pl has joined #openstack-cinder07:47
*** bluex-pl has quit IRC07:48
*** bluex-pl has joined #openstack-cinder07:48
*** bluex-pl has quit IRC07:49
*** chenying has joined #openstack-cinder07:49
*** bluex-pl has joined #openstack-cinder07:49
*** chenying has quit IRC07:49
*** lpetrut has quit IRC07:53
*** daneyon has joined #openstack-cinder08:01
*** jwcroppe has quit IRC08:01
*** jwcroppe has joined #openstack-cinder08:02
*** vincent_hou has joined #openstack-cinder08:02
*** daneyon has quit IRC08:06
*** jwcroppe has quit IRC08:06
*** dtadrzak_ has quit IRC08:06
*** aarefiev has joined #openstack-cinder08:07
*** ndipanov has joined #openstack-cinder08:08
*** git-harry_ is now known as git-harry08:09
*** sgotliv has joined #openstack-cinder08:09
*** Zhongjun has quit IRC08:12
*** jwcroppe has joined #openstack-cinder08:15
*** coolsvap is now known as coolsvap|away08:21
openstackgerritLei Zhang proposed openstack/cinder: Make concurrent backups for same volume impossible
*** coolsvap|away is now known as coolsvap08:29
*** coolsvap is now known as coolsvap|away08:31
*** aix has joined #openstack-cinder08:31
*** e0ne has joined #openstack-cinder08:32
*** gaurangt has joined #openstack-cinder08:35
*** annashen has joined #openstack-cinder08:38
*** laughterwym_ has quit IRC08:38
davechenthingee: sorry, Mike, just notice your message.08:39
davechenthingee: are you still around?08:39
openstackgerritAvishay Traeger proposed openstack/os-brick: Fix race in check and access of /dev/disk/by-path/
*** lpetrut has joined #openstack-cinder08:43
*** coolsvap|away is now known as coolsvap08:43
*** dalgaaf has joined #openstack-cinder08:45
davechenthingee: I will reply to you on that patch, thanks!08:46
openstackgerritGaurang Tapase proposed openstack/cinder: Adding NFS support to the GPFS Driver
*** subscope has joined #openstack-cinder08:55
*** vincent_hou has quit IRC08:57
*** markvoelker has joined #openstack-cinder09:00
*** markvoelker has quit IRC09:05
openstackgerritAbhijeet Malawade proposed openstack/cinder: Remove unused context parameter
*** dims has joined #openstack-cinder09:06
*** dims has quit IRC09:11
*** annashen has quit IRC09:18
*** lpetrut has quit IRC09:26
*** lpetrut has joined #openstack-cinder09:27
*** afazekas has joined #openstack-cinder09:28
*** laughterwym has joined #openstack-cinder09:28
*** jordanP has joined #openstack-cinder09:32
*** jasondotstar has joined #openstack-cinder09:32
*** kodoku has joined #openstack-cinder09:49
*** daneyon has joined #openstack-cinder09:50
kodokuHi, I upgraded juno to kilo but I have issue when I manage old volumes (created on JUNO) ==>
kodokuVolumeBackendAPIException: Bad or unexpected response from the storage volume backend API: Unable to fetch connection information from backend: 'NoneType' object has no attribute 'service_content'09:50
*** lixiaoy1 has quit IRC09:51
*** jwcroppe has quit IRC09:54
*** davechen has left #openstack-cinder09:54
*** kodoku has quit IRC09:55
*** daneyon has quit IRC09:55
*** jwcroppe has joined #openstack-cinder09:56
openstackgerritliuxinguo proposed openstack/cinder: Add hypermetro support for Huawei driver
deepakcsjordanP: u r not on nova channel ? :)09:59
openstackgerritliuxinguo proposed openstack/cinder: Enhance FC zone support for Huawei driver
*** dims has joined #openstack-cinder10:04
openstackgerritliuxinguo proposed openstack/cinder: Add iSCSI multipath support for Huawei driver
openstackgerritliuxinguo proposed openstack/cinder: Add smartX support for Huawei driver
*** dims_ has joined #openstack-cinder10:06
openstackgerritliuxinguo proposed openstack/cinder: Add volume migration support for Huawei driver
BharatKDuncanT, Please review the patch, (Modified as per the comments)10:09
*** dims has quit IRC10:10
LiuXinguoe0ne: I saw you leave a -1 on patch
LiuXinguoe0ne: But I can't see any comments outside the patch or inline10:14
e0neLiuXinguo: i posted my commenst with vote in the same time as you did a new patch10:15
LiuXinguoe0ne: :)10:15
LiuXinguoe0ne: I guessed that10:15
*** smoriya has quit IRC10:15
e0neLiuXinguo: that's why gerrit leave -1 and marked my comments as draft for the previous patch:(10:16
LiuXinguoe0ne: Maybe it's a bug, haha10:16
openstackgerritDaniel Tadrzak proposed openstack/cinder: ConsistencyGroup Object
LiuXinguoe0ne: So you need to submit the comments again?10:18
e0neLiuXinguo: done10:18
e0neLiuXinguo: thanks for pinging me for it10:19
LiuXinguoe0ne: No pro10:19
*** Lee1092_ has joined #openstack-cinder10:20
LiuXinguoe0ne: Thanks for you review, I will address your comments as soon as possible10:20
e0neLiuXinguo: you're welcome! do not hesitate to ping if my help is needed10:20
*** cppforlife__ has joined #openstack-cinder10:21
LiuXinguoe0ne: That's great10:21
LiuXinguoe0ne: thanks e0ne!10:21
*** Lee1092 has quit IRC10:21
*** cppforlife_ has quit IRC10:21
*** dobson` has quit IRC10:21
*** albertom has quit IRC10:21
*** rushiagr has quit IRC10:21
*** git-harry has quit IRC10:21
*** smcginnis has quit IRC10:21
*** smcginni1 has joined #openstack-cinder10:22
*** powerbsd has joined #openstack-cinder10:22
*** Lee1092_ is now known as Lee109210:22
*** powerbsd is now known as albertom10:22
*** cppforlife__ is now known as cppforlife_10:22
*** salv-orlando has quit IRC10:24
*** dobson has joined #openstack-cinder10:25
*** heyun has quit IRC10:29
*** coolsvap is now known as coolsvap|away10:35
*** gaurangt1 has joined #openstack-cinder10:37
*** gaurangt has quit IRC10:38
*** gaurangt1 has left #openstack-cinder10:39
*** gaurangt1 has joined #openstack-cinder10:39
*** git-harry has joined #openstack-cinder10:40
*** zhonghua-lee has quit IRC10:42
DuncanTBharatK: looking now10:45
BharatKDuncanT, Thank you very much.10:45
DuncanTBharatK: +2, but can you have a look at the sheepdog CI failure please? It seems to be the only one that is backup related10:50
BharatKThanks DuncanT,10:53
BharatKDuncanT, Backup testcases in sheepdog are failing for every patch set.10:53
DuncanTBharatK: Ok, thanks for checking10:53
openstackgerritVipin Balachandran proposed openstack/cinder: VMware: Set virtual disk UUID to volume ID
BharatKDuncanT, Thanks once again.10:54
*** DTadrzak has joined #openstack-cinder10:55
*** markvoelker has joined #openstack-cinder11:01
*** zhonghua-lee has joined #openstack-cinder11:03
openstackgerritVipin Balachandran proposed openstack/cinder: VMware: Set virtual disk UUID to volume ID
*** markvoelker has quit IRC11:05
*** avishay__ has quit IRC11:08
openstackgerritliuxinguo proposed openstack/cinder: Add smartX support for Huawei driver
*** avishay__ has joined #openstack-cinder11:11
*** pradipta has quit IRC11:12
LiuXinguoe0ne: Have addressed all of your comments, please check again11:15
LiuXinguoe0ne: just when you have timek, thanks!11:15
e0neLiuXinguo: thanks. will look on it a bit later11:15
LiuXinguoe0ne: thanks!11:15
openstackgerritliuxinguo proposed openstack/cinder: Add volume migration support for Huawei driver
*** ganso_ has joined #openstack-cinder11:21
*** simondodsley has joined #openstack-cinder11:24
*** salv-orlando has joined #openstack-cinder11:26
*** jwcroppe_ has joined #openstack-cinder11:30
*** salv-orlando has quit IRC11:30
*** marzif has joined #openstack-cinder11:31
*** salv-orlando has joined #openstack-cinder11:32
*** jwcroppe has quit IRC11:33
*** markus_z has quit IRC11:35
*** daneyon has joined #openstack-cinder11:39
*** avishay__ has quit IRC11:40
*** avishay__ has joined #openstack-cinder11:40
*** daneyon has quit IRC11:44
*** jwcroppe_ has quit IRC11:46
*** deepakcs has quit IRC11:46
*** markus_z has joined #openstack-cinder11:48
*** avishay__ has quit IRC11:50
*** gaurangt1 has left #openstack-cinder11:50
*** davechen has joined #openstack-cinder11:51
davechenthingee: Just replied to you on that patch (, pls kindly let me know if there is anything more needed.11:52
*** avishay__ has joined #openstack-cinder11:53
davechenthingee: I need to logoff and have supper, thanks!11:54
*** zhonghua-lee has quit IRC11:54
*** davechen is now known as davechen_afk11:54
*** BharatK has quit IRC11:57
*** salv-orlando has quit IRC11:58
openstackgerritzhongjun proposed openstack/cinder: Add volume retype support for Huawei driver
*** avishay__ has quit IRC12:00
*** shyama has joined #openstack-cinder12:01
*** marcusvrn has joined #openstack-cinder12:02
*** zhenguo has quit IRC12:02
*** thangp has joined #openstack-cinder12:06
*** jasondotstar has quit IRC12:10
*** bswartz has quit IRC12:13
openstackgerritDaniel Tadrzak proposed openstack/cinder: CGSnapshot Object
*** annashen has joined #openstack-cinder12:19
*** nzrf has joined #openstack-cinder12:19
DuncanTe0ne: You beat me to the db scaling comments on - I was busy spinning up devstack to check I was correct before making the comments12:23
*** annashen has quit IRC12:23
DuncanTe0ne: Good spot12:23
e0neDuncanT: thanks12:25
openstackgerritwanghao proposed openstack/cinder: Add pagination to backups
e0neDuncanT: i didn't test it, so i asked to get some performance results12:26
DuncanTe0ne: I'm trying to write a large-ops CI job, so I had some scripts to generate fake backups handy.12:27
openstackgerritThang Pham proposed openstack/cinder: Create service_versions table
e0neDuncanT: i didn't test cinder well with large numbers ob backups, but we've found several issues when we've got >30k volumes and snapshots in the db12:28
*** kaisers has quit IRC12:28
*** avishay__ has joined #openstack-cinder12:28
DuncanTe0ne: There have been quite a few fixed over the last 2 years, usually resulting from the same pattern (build a list then query the db for every item on the list before returning) - I try to keep an eye out for new ones but no doubt miss some12:30
*** edmondsw has joined #openstack-cinder12:30
*** timcl has joined #openstack-cinder12:30
DuncanTe0ne: Have you root-caused any of the issues you found to see if there's any other base issues?12:30
*** chlong has joined #openstack-cinder12:30
e0neDuncanT: not yet:(. i hope to have time for it after L-3 will be released12:31
*** jasondotstar has joined #openstack-cinder12:32
DuncanTe0ne: I'll carry on my slow progress too, and ping you if I find anything12:32
*** rushiagr_away has joined #openstack-cinder12:33
DuncanTe0ne: I should be able to automate issue detection to some degree - insert 1000+ dummy records into the db for each entity type (vol, snap, cg, bak, etc) and call all of the list interfaces - any call that causes 1000+ db queries is broken12:33
e0neDuncanT: thanks. we've got a reported bug for our distro i'll ask to check it with devstack too12:34
openstackLaunchpad bug 1471273 in Mirantis OpenStack 6.0-updates " 'cinder list' ERROR: Gateway Timeout (HTTP 504) when we have more than 50k volumes" [High,Confirmed] - Assigned to MOS Sustaining (mos-sustaining)12:34
e0neDuncanT: we tested cinder with fake driver12:34
DuncanTe0ne: I configure the fake driver, but just inserting into the db is much quicker... should give me something I can sensibly run on every commit. Need a fake backup driver to do that with backups, plus the fake driver needs some fixup for cgs etc12:35
DuncanTe0ne: Actually, I might go write a fake backup driver, shouldn't be hard... assuming we don't already have one12:36
e0neDuncanT: you're right. but with fake driver we could also see to create something works when we've got a lot of data in the DB12:37
*** eharney has quit IRC12:37
DuncanTe0ne: Yeah, definitely. Maybe insert a bunch of fake data then call create a few times?12:38
e0neDuncanT: sound like a good idea:). i like it12:38
*** kodoku has joined #openstack-cinder12:43
*** kaisers has joined #openstack-cinder12:43
*** zhonghua-lee has joined #openstack-cinder12:44
openstackgerritwanghao proposed openstack/cinder: Support volume backup of qcow2 format volume
kodokuHi, I upgraded juno to kilo but I have issue when I manage old volumes (created on JUNO) ==>
kodokuMaybe somebody can help me ?12:47
*** kaisers has quit IRC12:50
*** kaisers1 has joined #openstack-cinder12:50
*** kjnelson has joined #openstack-cinder12:52
*** bswartz has joined #openstack-cinder12:53
*** annegentle has joined #openstack-cinder12:54
*** akerr has joined #openstack-cinder12:56
*** markus_z has quit IRC12:59
*** kaisers1 has quit IRC13:00
DuncanTkodoku: It looks like that's an error within the VMWare driver itself13:01
*** merooney has joined #openstack-cinder13:01
*** zhenguo has joined #openstack-cinder13:01
*** cbader has joined #openstack-cinder13:02
*** markus_z has joined #openstack-cinder13:03
*** rushiagr_away is now known as rushiagr13:03
DuncanTkodoku: I've never looked at that driver, but I'll see if I can spot a possible cause13:03
*** julim has joined #openstack-cinder13:05
kodokuDuncanT hum do you think it's like this bug ==>
openstackLaunchpad bug 1380675 in VMwareAPI-Team "VMware: retype fails with AttributeError 'NoneType' object has no attribute 'service_content'" [Undecided,New]13:07
*** edmondsw has quit IRC13:08
*** diogogmt has quit IRC13:09
*** annegentle has quit IRC13:10
*** annegentle has joined #openstack-cinder13:11
*** diogogmt has joined #openstack-cinder13:11
*** diogogmt has quit IRC13:15
openstackgerritThang Pham proposed openstack/cinder: Create service_versions table
*** akshai has joined #openstack-cinder13:16
DuncanTkodoku: Doesn't look like it is that exact bug, but it could well be related13:17
DuncanTkodoku: I've not got a vmware system to poke around I'm afraid, so I can't dig any deeper13:17
*** eharney has joined #openstack-cinder13:22
*** kaisers has joined #openstack-cinder13:23
*** daneyon has joined #openstack-cinder13:28
*** daneyon has quit IRC13:32
*** markvoelker has joined #openstack-cinder13:34
*** gaurangt has joined #openstack-cinder13:37
*** annegent_ has joined #openstack-cinder13:38
*** Yogi1 has joined #openstack-cinder13:40
*** annegentle has quit IRC13:41
*** chlong has quit IRC13:42
*** jungleboyj has quit IRC13:44
*** alexpilotti has quit IRC13:46
*** coolsvap|away is now known as coolsvap13:47
*** jasondotstar has quit IRC13:48
e0nethingee: fyi, we'll have rally meeting in 10 minutes13:50
*** jistr has quit IRC13:52
*** jistr has joined #openstack-cinder13:54
openstackgerritOpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements
*** julim_ has joined #openstack-cinder13:57
*** edmondsw has joined #openstack-cinder13:57
*** thangp_ has joined #openstack-cinder13:57
*** julim has quit IRC14:00
*** mylu has joined #openstack-cinder14:00
*** Guest10__ has joined #openstack-cinder14:01
*** thangp has quit IRC14:01
*** boris-42 has joined #openstack-cinder14:02
*** jungleboyj has joined #openstack-cinder14:03
*** zhonghua-lee has quit IRC14:05
*** erlon has joined #openstack-cinder14:08
*** xyang has joined #openstack-cinder14:13
*** davechen_afk has left #openstack-cinder14:13
erlontbarron: ping14:13
tbarronerlon: wazzup?14:14
erlontbarron: hey, Tom, can you have a look on that patch when you have a change?14:14
erlontbarron: I have implemented you suggestion, with the @retry14:15
erlontbarron: its much cleaner! :)14:15
erlontbarron: thanks!14:15
*** Yogi1 has quit IRC14:16
tbarronerlon: will look in a half hour or so (meeting).  I like that retry decorator too.14:16
*** ankit_ag has quit IRC14:18
openstackgerritMatthew Edmonds proposed openstack/cinder: move oslo.vmware to test-requirements
*** salv-orlando has joined #openstack-cinder14:21
*** edtubill has joined #openstack-cinder14:23
ameadesomeone want to put a nail in this simple patch?
*** mylu has quit IRC14:27
smcginni1ameade: Would if I could. ;)14:27
*** smcginni1 is now known as smcginnis14:27
*** jasondotstar has joined #openstack-cinder14:28
openstackgerritSean McGinnis proposed openstack/cinder: Dell SC: Add check of current value on retype
*** mylu has joined #openstack-cinder14:28
openstackgerritliuxinguo proposed openstack/cinder: Add volume migration support for Huawei driver
*** changbl has quit IRC14:32
*** vilobhmm has joined #openstack-cinder14:35
ameadesmcginnis: maybe if you just click that +2 extra hard? ;)14:35
*** Guest14296 is now known as dansmith14:35
*** salv-orlando has quit IRC14:35
ameadethat may have come off weird with the winky face >.>14:35
smcginnisameade: Hah! ;)14:36
jungleboyjameade: What are you suggesting?14:36
ameadejungleboyj: ;)14:36
*** vilobhmm1 has joined #openstack-cinder14:36
*** bluex-pl has quit IRC14:36
* ameade is suggesting that jungleboyj review that patch14:37
smcginnisameade: Nice one!14:37
jungleboyjameade: Get the big old +2 out.  Ok.  Got it.14:37
*** bluex-pl has joined #openstack-cinder14:37
*** bluex-pl has quit IRC14:38
*** merooney has quit IRC14:38
*** bluex-pl has joined #openstack-cinder14:38
*** merooney has joined #openstack-cinder14:38
*** vilobhmm has quit IRC14:39
*** salv-orlando has joined #openstack-cinder14:44
jungleboyjameade: You are welcome.  ;-)14:45
ameadejungleboyj: ty sir!14:45
jungleboyjameade: Welcome.14:46
*** r-daneel has joined #openstack-cinder14:46
*** mtanino has joined #openstack-cinder14:48
*** jaypipes has joined #openstack-cinder14:48
openstackgerritMatthew Edmonds proposed openstack/cinder: move oslo.vmware to test-requirements
LiuXinguoWhy lots of patch not have bug link or bp link in the commit message?14:51
LiuXinguoI am a little confused about this14:51
smcginnisLiuXinguo: A lot of them are not bugs or major changes to cinder. Most don't need them.14:53
*** annegent_ has quit IRC14:54
LiuXinguosmcginnis: So I can just commit code to gerrit and needn't register a bug or bp?14:55
*** vlaza has left #openstack-cinder14:55
smcginnisIn many cases, yes.14:55
*** Yogi1 has joined #openstack-cinder14:55
smcginnisLiuXinguo: If it's a legitimage bug that might affect someone where it would be useful to have a way for them to search and find the bug, then file a bug.14:55
smcginnisLiuXinguo: If it's a change to how Cinder behaves or significant enough work that it should be tracked and made visible then file a blueprint.14:56
smcginnisLiuXinguo: Otherwise just make a change and submit it.14:56
*** kbyrne has quit IRC14:56
openstackgerritJohn Griffith proposed openstack/cinder: Use a single base class for the VolumeDriver
LiuXinguosmcginnis: Thanks for your explain, smcginnis14:57
LiuXinguosmcginnis: I've got that :)14:58
*** blinky_ghost has joined #openstack-cinder15:01
aarefievjungleboyj: hi, could look at please , it blocks this one
*** Swanson has joined #openstack-cinder15:01
*** ronis has quit IRC15:01
jungleboyjaarefiev: Sure, can take a look.15:02
*** alexpilotti has joined #openstack-cinder15:02
aarefievjungleboyj: thanks15:02
blinky_ghosthi all, is it possible to use cinder-backup service with an attached volume to a VM? I get this error: cinder backup-create cirrosiscibackup "ERROR: Invalid volume: Volume to be backed up must be available (HTTP 400) (Request-ID: req-7e23875c-4663-4147-91e1-974375e36070)" Is this normal or a bug? Thanks15:03
*** kodoku has quit IRC15:03
*** jdurgin1 has joined #openstack-cinder15:05
*** gaurangt has quit IRC15:05
*** diogogmt has joined #openstack-cinder15:06
*** nkrinner has quit IRC15:09
guitarzanblinky_ghost: that's currently being worked on15:10
*** kbyrne has joined #openstack-cinder15:10
*** crose has joined #openstack-cinder15:10
guitarzanblinky_ghost: right now it's not possible, no15:10
thangp_guitarzan: actually, it recently made it into master -
guitarzanthangp_: even better15:11
thangp_so depends on what version blinky_ghost has15:11
erlonjungleboyj: hi Jay, can you give a look in this change again when you have a change?15:14
*** salv-orlando has quit IRC15:16
*** daneyon has joined #openstack-cinder15:16
*** kmartin has joined #openstack-cinder15:18
*** dannywilson has joined #openstack-cinder15:19
*** eharney has quit IRC15:20
*** dims_ has quit IRC15:21
*** daneyon has quit IRC15:21
*** tsekiyama has joined #openstack-cinder15:21
*** dims has joined #openstack-cinder15:21
*** aix has quit IRC15:22
jungleboyjerlon: Can try.  :-)15:22
*** rushiagr is now known as rushiagr_away15:22
*** zhenguo has quit IRC15:22
*** edtubill has quit IRC15:22
*** edtubill has joined #openstack-cinder15:23
*** dobson has quit IRC15:29
openstackgerritCurt Bruns proposed openstack/cinder: Add deactivate step to extend_lv
*** dobson has joined #openstack-cinder15:34
*** eharney has joined #openstack-cinder15:35
*** afazekas has quit IRC15:36
*** ganso_ has quit IRC15:40
anteayathingee: when you are about? can we go over the systems you would like disabled? the third-party-announce list tends to be broadcasting actions to operators that have already happened or decisions that have been made, it isn't really a discussion forum and honestly it isn't the best way to tell me what systems you would like disabled15:41
anteayathingee: I get confused15:41
openstackgerritliuxinguo proposed openstack/cinder: Add volume migration support for Huawei driver
*** rushiagr_away is now known as rushiagr15:45
openstackgerritDanny Wilson proposed openstack/cinder-specs: Add ability for Cinder backend to report discard/unmap/trim
*** kfox1111_ has left #openstack-cinder15:46
hemnaaaaand it's Monday.15:48
*** daneyon has joined #openstack-cinder15:48
*** nkrinner has joined #openstack-cinder15:48
*** daneyon_ has joined #openstack-cinder15:51
LiuXinguohah, yeah Monday again15:51
*** agarciam has quit IRC15:52
*** daneyon has quit IRC15:53
*** harlowja_at_home has joined #openstack-cinder15:54
LiuXinguohemna: have you saw my leave words, I'm so sorry to disturb you just from Monday ^-^15:55
*** xyang has quit IRC15:55
*** mgarza_ has joined #openstack-cinder16:00
thingeeanteaya: in the future how would you like me to request CI's to be disabled. This was the previous way people were making requests.16:03
thingeeah an etherpad, ok16:03
anteayaor irc16:04
openstackgerritTom Barron proposed openstack/cinder: Fix cleanup_temp_volume_snapshots for missing vol
*** jdurgin1 has quit IRC16:07
*** e0ne has quit IRC16:10
*** avishay__ has quit IRC16:14
thingeesmcginnis: I would like to see WIP not touched as well
*** bluex-pl has quit IRC16:16
thingeemkoderer: what were your reasons for not wanting to go all in on jgriffith's approach here?
openstackgerritTomoki Sekiyama proposed openstack/cinder: Efficient image download and upload for Glance Cinder backend
*** mragupat has joined #openstack-cinder16:18
*** leeantho has joined #openstack-cinder16:19
openstackgerritSean McGinnis proposed openstack/cinder: Add abandon script to clean up orphaned reviews
*** thangp_ has quit IRC16:21
thingeesmcginnis: thanks16:21
openstackgerritAnton Arefiev proposed openstack/cinder: Add entry create and cast tasks to manage workflow
openstackgerritAnton Arefiev proposed openstack/cinder: Fix lvm manage existing volume
*** ganso_ has joined #openstack-cinder16:23
*** Apoorva has joined #openstack-cinder16:23
aarefievthingee: hi, is there any chance that it lands in liberty ?16:24
*** salv-orlando has joined #openstack-cinder16:24
*** patrickeast has joined #openstack-cinder16:25
*** salv-orl_ has joined #openstack-cinder16:26
*** vilobhmm1 has quit IRC16:27
anteayathingee: thanks I have to run errands now, will check the etherpad in a few hours when I return16:28
anteayathank you16:28
*** vokt has joined #openstack-cinder16:28
*** salv-orlando has quit IRC16:29
*** laughterwym has quit IRC16:30
DuncanTthingee: jgriffith's ABC replacement patch I didn't like.... I'm not sure what it achieves. The problem (highlighted by the current approach) is that a bunch of 'mandatory' features are not universally implemented. IMO jgriffith's patch is only needed because we haven't faced up to / fixed that problem. It goes back to 'if things are broken, that's the16:30
DuncanTdriver author's fault and they should have done better', rather than making the problem clear to everybody16:30
thingeeDuncanT: I think it's the idea of having a single class as oppose to all these broken out classes.16:31
DuncanTthingee: All of the  features that are universally implemented can go in one class, sure, but there aren't that many of those...16:31
thingeeDuncanT: There was a good point raised on the list though
DuncanTthingee: snapshots etc are currently defacto optional... being able to put them in the base class proves your problem detection mechanism is broken16:32
thingeeI would agree with that post on the ML that THAT'S broken16:32
DuncanTthingee: The ABC approach to that would be to introduce a second ABC that is a superset of the first, then migrate all of the existing drivers to the superset over time. The *point* of the ABC work, to a certain degree, was to make situations like that obvious (both in the code and the testing)16:34
eharneydoesn't putting things back in the base class remove the ability to use abc to ensure we know which features are implemented in which drivers, and that the interface for those features is correctly shaped?16:34
DuncanTthingee: Introspection means it doesn't blow up until run time, and our test coverage is not exactly high16:34
eharneylike Duncan (i think), i'm still kind of lost on the motivation16:34
*** jasondotstar has quit IRC16:34
thingeeI'm confused by both your guys' comments. are you for or against the current approach? :P16:35
thingeerather current merged approach16:35
*** afazekas has joined #openstack-cinder16:36
eharneysorry, i'm generally for the current approach, and don't really understand the aim with the new patch16:36
*** mylu has quit IRC16:37
thingeeeharney: my main take away from jgriffith's email was there are too many classes. lets put everything into one and one by one start decorating methods we mark as required features.16:37
thingeeeharney: the whole goal is to not know what drivers do what features. the main goal was to eventually graduate features into being required.16:37
thingeeeharney: avoid a matrix16:37
DuncanTthingee: There can be one class for every required feature, fine, but we have lots of defacto optional features, the classes make that clear16:37
*** _cjones_ has joined #openstack-cinder16:38
*** mylu has joined #openstack-cinder16:38
DuncanTthingee: We *have* a matrix, pretending we don't is a fantasy16:38
eharneyyeah, there's a reason we haven't made more features required in the code16:38
thingeeDuncanT: do we always see these features being optional?16:38
thingeeDuncanT: there's no hiding here. The point of this is a movement away from that.16:38
DuncanTthingee: CGs for example? Yes, absolutely16:38
eharneythe only one i see moving any time soon is basic snapshot support, but even that's not ready yet16:38
DuncanTthingee: Snapshots I'd love to see going away from being optional, but people don't want to kill the block driver16:39
jgriffithDuncanT: thingee eharney so all those points aside...16:39
eharneymore people don't want to kill the NFS driver16:39
thingeeI guess that goes back to people being unhappy that some features aren't capable of being fulfilled in the reference implementation16:39
*** thangp has joined #openstack-cinder16:39
jgriffithis there a reason having 12 classes is "better" preferred to having "one"16:39
DuncanTthingee: replication - always optional16:39
jgriffith(in terms of the abstract classes)16:40
thingeejgriffith: from what I got from DuncanT, yes. Just to be clear of what features a driver supports16:40
eharneyjgriffith: well, i think it depends on what your personal goals are, and it seems that we have different ones16:40
thingeejgriffith: since we'll never escape a matrix with the directiont hings have gone16:40
*** jasondotstar has joined #openstack-cinder16:40
thingeejgriffith: that's it though.16:40
DuncanTSo we have BaseVD for everything that is mandatory16:41
jgriffithDuncanT: but we don't really.  It' s not complete, it's not used, and I don't see any value in breaking everything out by method.16:41
*** harlowja_at_home has quit IRC16:41
jgriffithThat being said, I'm also not going to argue this any longer.  I noted my opinion and provided an alternative.  If it's not good enough... well then that's life16:42
thingeejgriffith: by feature... but I knew what you meant.16:42
jgriffiththingee: Ooops.. :)  yeah, sorry16:42
DuncanTjgriffith: Only optional features was the original idea... it just turned out many feature are optional16:42
jgriffithFTR, I think the diverging features is a bad thing; but probably inevitable as mentioned.  In which case I think it should be the vendors problem not mine or anybody elses16:43
thingeejgriffith: I think your idea is fine, if we decide to just get rid of folks who can't do snapshots and remove features like CG and replication16:43
DuncanTjgriffith: SnapVD, TransferVD MigrateVD, ExtendVD can all go into the base class once all of the drivers actually support the features16:44
DuncanTjgriffith: Putting them in a base class now is just broken since not all drivers support them16:44
thingeeat least mkoderer started some patch in moving some to the base16:44
jgriffithDuncanT: ok.  My point was I just don't see any real value in that; or continuing to modify the object heirarchy16:44
*** krypton_ has joined #openstack-cinder16:44
thingeejgriffith: did you have any reasons of disliking it besides it just being (my words here) unusual to have all these classes broken out?16:45
eharneythis was one of the value-adds, just for context:
jgriffiththingee: it's poor object design16:46
DuncanTjgriffith: Well, part of the problem is that you consider having (relatively) crap drivers in cinder and letting the vendors suffer for that ok, I don't - I get hassle whenever there's a 'cinder' problem, whether it is ultimately a vendor issue or not, so I want every driver in cinder to get better (or get out)16:46
jgriffiththingee: inheritance can be bad enough, let alone multi-way inheritance16:46
DuncanTjgriffith: It is interfaces, which is not poor OO IMO16:46
jgriffiththingee: it's not implemented fully anywhere16:46
jgriffiththingee: it's lead to a mix of both the "old" inheritance structure and the new16:47
*** sghanekar_ has joined #openstack-cinder16:47
jgriffiththingee: There are a number of drivers that have done a pick and choose of which classes they use and which ones they don't16:47
DuncanTjgriffith: In part because cleaning it up is damn tricky because out drivers are a weird mess16:47
jgriffiththingee: inheritance is the least flexible form of code design and object usage16:47
thingeeDuncanT: I have to say though wrt I don't think creating yet more classes is a good idea =/16:48
DuncanTjgriffith: Interfaces are *not* classical inheritence, though in python they look like it16:48
jgriffithDuncanT: yes, and my patch in part cleaned some of that up IMHO.  And also, that's not going to change by adding more layers until something is actually done to enforce the changes16:48
*** krypton has quit IRC16:48
jgriffithDuncanT: The way we've used them, they kinda are though... but regardless.  I really don't want to argue this any more.16:49
*** krypton_ has quit IRC16:49
*** jwcroppe has joined #openstack-cinder16:49
jgriffithif eharney doesn't agree in particular then "ok"16:49
*** sgotliv has quit IRC16:49
jgriffithand given that nobody else has really raised any issue other than me, and all the other folks writing drivers seem good with it then maybe i'm just being paranoid and overly picky16:50
eharneyi'm not really terribly vested either way on this, i'm just trying to figure out if there's a route here that's a clear win or not16:50
jgriffithI'm not changing my opinion, and I think it is going to come back and bight us in the butt16:50
*** markus_z has quit IRC16:50
jgriffitheharney: Well if there's no "clear win" then why introduce the extra complexity?16:51
DuncanTjgriffith: I thing you're *right* to think that what we have is ugly, but where we disagree is the cause of the ugly... I think the ABC work highlights a bunch of ugly that is and was there all along, and it will look better once we've fixed the underlying mess16:51
jgriffithwinston-d: as in "chomp chomp"16:51
jgriffithDuncanT: ok16:51
thingeeto sum up this discussion, it appears people value seeing a clear idea of what optional features are supported by which drivers, and fine with the cost being a bunch of classes16:52
eharneyjgriffith: i think the win is that we are making our mess more understandable and documenting it -- it's a mess either way, but that sounds like a win to me16:52
eharneyyeah what thingee just said16:52
jgriffithDuncanT: but I'd argue that given we've had drivers merge in Liberty that ignored or half implemented the abc semantics that were added in Kilo it's likely not going to work out16:52
jgriffitheharney: but I've been saying it's not "more understandable" to most16:52
DuncanTthingee: jgriffith: As for the case discussed on the mailing list, having a new interface that is implemented in only a few drivers is the root ugly, and ABCs make that ugly obvious - it all goes away once everybody has migrated to the new structure. I call that a feature of the ABC system, not a bug16:52
jgriffiththingee: yes, that seems to be the concensus16:53
eharneyjgriffith: i think that's because of the inherent complexity that would be there for any way that we handle this, not due to abc, though16:53
eharneybut i'm kinda guessing there16:53
jgriffithDuncanT: and by the way, I argue that the reason for such a wide divergence in the implementations in all of our drivers is the lack of a clear well defined base class16:53
jgriffithDuncanT: and having 12 base classes isn't going to help that, it's going to make it worse16:54
*** nkrinner has quit IRC16:54
jgriffitheharney: DuncanT yes, to be clear I'm not opposed to ABC at ALL.  If you looked at my patch you will notice that I utilized it16:54
jgriffithabstract base class definitions is not the issue IMO16:54
jgriffithi.e. the issue to debate16:55
*** jistr has quit IRC16:55
thingeejgriffith: I'd like to understand more on your point of it not being understandable. 1) For example, people wrote this to see a clear list of what's supported by which drivers ... 2) You can just look at the driver's class of long list of inherited classes16:55
jgriffiththingee: how does that help a person writing a driver?16:56
eharneyjgriffith: i'll admit that i need to study your patch more to really figure out what all is going on there, i think i'm still missing some of the detail16:56
Swansonjgriffith:  Two driver implementation things.16:57
thingeejgriffith: Well we have a BaseVD class. That's the base. As a new volume driver, you should start there. I agree though that by us diverging from just that idea makes things more interesting in Cinder.16:57
jgriffiththingee: DuncanT so my biggest concern is the goal seems to be "build an automated feature matrix" but my contention is still that we shouldn't do that16:57
Swansonjgriffith: I add consistency groups so I can just see what is in ConsistencyGroupVD and add those functions.  Easy.  No searching.16:57
*** afazekas has quit IRC16:58
*** ub has quit IRC16:58
Swansonjgriffith: I also need to change existing functions that are in basevd.  Not so easy.  Lots of searching.16:58
DuncanTjgriffith: So kick out all of the drivers that don't meet minimums, move transfer, snap and migrate into baseVD, and the answer is 'implement all of baseVB'16:58
eharneyi don't think avoiding a feature matrix is a feasible option any time soon16:58
jgriffithSwanson: sure.. but what's wrong with ### Consistency Group Interface Begin #####    ..... END16:58
thingeejgriffith: I agree. I think though that before we can say this idea of optional feature classes is bad, we should use it as a point that we just don't like the idea of replication and consistency group in CInder.16:58
jgriffitheharney: fair enough (I guess)16:58
thingeeDuncanT: that's fair, I think that should be a goal of next release.16:58
jgriffitheharney:  but that means I've been doing things wrong for the last 4 years16:58
DuncanTjgriffith: That was kind of where I hoped we'd be by now, TBH16:59
*** annashen has joined #openstack-cinder16:59
thingeeDuncanT: I didn't focus on this release of graduating features. IMO I need to do early communication to maintainers.16:59
jgriffitheharney: I should be submitting all the "special" features that I have been hiding in extensions or in volume-types16:59
smcginnisI think for M a goal should be to have all drivers switched to ABC.16:59
smcginnisThat will be a move to identifying what is really common and what is extra.16:59
smcginnisWith the long term goal of getting as much as possible into base.16:59
Swansonjgriffith: Either way works for me.  What we had was a pain.  xxxVD is better.  Your patch is fine, too.17:00
DuncanTjgriffith: Nova said no extensions, only core API, because otherwise your workload portability is shot.... not an entirely bad point17:00
*** afazekas has joined #openstack-cinder17:00
jgriffithDuncanT: Cinder != Nova last time I checked17:00
jgriffithDuncanT: and you're the most famous for saying "I don't care what project X does"17:00
DuncanTjgriffith: I didn't say they were the same, I said they had a not entirely bad point in this case17:00
DuncanTjgriffith: Just because they're full of crazy doesn't mean they always make bad points17:01
thingeeok lets bring things back...17:01
jgriffithDuncanT: well.. don't take this the wrong way please, but that's awfully convenient.  Use the argument when it suits you?17:01
thingeeDo we all agree that avoiding a matrix at this point is not feasible?17:01
DuncanTjgriffith: Huh? The argument that sometimes nova make good points?17:02
smcginnisthingee: +117:02
jgriffiththingee: -117:02
DuncanTthingee: Agreed. +1.17:02
*** jordanP has quit IRC17:02
thingeejgriffith: should we start considering on removing some of these features we have in Cinder?17:02
jgriffiththingee: I'd personally be fine with that (more than fine actually)17:02
jgriffithMost of them aren't used by any real customers anyway17:03
jgriffithnor are they tested anywhere17:03
smcginnisOptions I see are 1) remove functionality from cinder (that may be used by someone), 2) remove drivers that can't be fully functional, or 3) have a matrix.17:03
jgriffiththingee: but I also realize that's a tough thing to do, because as soon as we do remove something somebody is going to hollar17:04
thingeesmcginnis: that would remove our reference implementation at this point :)17:04
* DuncanT would say kill replication, keep cgs, and improve testing and discoverability17:04
smcginnisthingee: Exactly!17:04
*** akshai has quit IRC17:04
DuncanTCGs are cross vendor, replication is currently single vendor17:04
thingeeDuncanT: might change once things are accessible by jgriffith's efforts.17:04
jgriffithHow many of the drivers do CG's now?17:04
jgriffithSo maybe there should be at least a 50% rule or something :)17:05
smcginnisIf they were all using ABC we could know pretty quick. :P17:05
*** akshai has joined #openstack-cinder17:05
* smcginnis hides17:05
thingeeall of emc, if not a patch up for review by now, pure, dell17:05
jgriffithsmcginnis: not true actually17:05
SwansonAs a semirelated aside TransferVD seems to actually be implemented by solidfire only.  Haven't finished my migratevd search.17:05
smcginnisjgriffith: Well, if they were all switched over we could see which ones implement the consistency group interface.17:06
jgriffithSwanson: FWIW, transfer is implemented by quite a few drivers.  Somebody updated the SF driver to use the class though17:06
*** coolsvap is now known as coolsvap|away17:06
smcginnisIf they don't support it, they don't implement the interface.17:06
DuncanTjgriffith: If the ABCs were done properly, we should know easily and in code... if that isn't possible then I'd liketo know about it17:06
thingeejgriffith: so lets say we did remove CG and replication all together...17:06
DuncanTSwanson: Most drivers don't need to do anything for transfer, it just works17:06
*** kjnelson has quit IRC17:06
thingeejgriffith: I've been sort of curious how drivers would do this today without having some knowledge in the manager to handle the requests. it's not enough for their to be an api extension17:06
DuncanTSwanson: ONly backends that store tenant ids etc need to care17:07
jgriffithDuncanT: Id' say go look at the code, pick some "features' and start looking at drivers.  It doesn't work currently at all17:07
blinky_ghostguitarzan: I use icehouse so guess not :) thanks anyway17:07
thingeejgriffith: and maybe your answer is it just shouldn't be possible in Cinder.17:07
Swansonjgriffith: really?  When I did a search on accept_transfer that was the only driver I found.  I found more than a few that inherited transfervd but only the one that implemented its function.  Maybe the scope of my search is off.17:07
jgriffiththingee: yeah, well you can do direct calls to db and roc layer from the extension (not saying that's good... just saying you can)17:07
jgriffithSwanson: interesting17:08
DuncanTjgriffith: If you're going to do that, why bother with cinder as a standard at all?17:08
jgriffithSwanson: did you by chance notice that it's implemented at a general layer?17:08
DuncanTjgriffith: For most drivers, transfer is a no-op17:08
jgriffithSwanson: it's only implemented as a method in my driver so I can update the attributes on the cluster device17:08
winston-dSwanson: jgriffith was right, most drivers don't need driver level accept_transfer 'cos they don't use volid internally.17:09
jgriffithSwanson: so it's a good example of a "feature" abc-class that gets picked up by the base driver and nobody has to implement17:09
jgriffithbut that doesn't mean it will actually "work"17:09
DuncanTSorry, got to run. I'll look again at the class hierachy, see if I can suggest some cleanup that doesn't involve throwing out the baby with the bath water17:09
jgriffithSwanson: my driver for example can do that; but if I didn't update the volume object on the cluster the "new" owner would never be able to find the volume or access it17:10
jgriffithSwanson: I wouldn't be surprised if there are some drivers that have this problem TBH and don't even know it17:10
jgriffithSwanson: which IMO shows that ABC doesn't actually "do" what some people think it does17:10
winston-dDuncanT: btw, before you leave, quick question about decouple backup from volume service17:10
winston-dDuncanT: anything in progress?17:11
jgriffithSwanson: and being that transfer has a general implementation in the base driver and the reference works, why is it a "separate" feature in the matrix?17:11
smcginnisIf things are done right, we should be able to do isinstance(driver, transferableVD).17:11
jgriffithsmcginnis: but that will "pass"17:11
smcginnisThen they aren't inheriting correctly.17:11
jgriffithsmcginnis: because the method is implemented in the next layer of base driver17:12
smcginnisThey should only inherit if they implement that interface.17:12
jgriffithsmcginnis: which the vendor driver inherits from17:12
smcginnisjgriffith: That's a problem.17:12
jgriffithsmcginnis: but that doesn't mean the feature works17:12
smcginnisNone of these should be in base driver implementations unless it's clear what it means to inherit from that base.17:12
jgriffithsmcginnis: that's what I've been trying to say here17:12
thingeeI think this will be excellent to discuss with ya'll at the midcycle. My decision is still going to be that unless we're really interested in doing a hard push of removing these features and telling a small number of customers (according to sources in this room) tough, that's not what Cinder does, what we're doing is fine.17:12
Swansonjgriffith: i do use the volume id for about everything.  I wonder if I'm hosed.17:12
tbarronwinston-d: i've got a bp spec and POC up17:12
smcginnisjgriffith: But moving them ALL into a base volumedriver is the wrong direction then.17:13
thingeeespecially since we don't have an answer for drivers to do this outside of CInder easily.17:13
jgriffiththingee: that's cool17:13
*** krypton has joined #openstack-cinder17:13
jgriffithsmcginnis: volume-id doesn't change17:13
jgriffitherrr.. Swanson17:13
smcginnisthingee: +2, this should be a midcycle discussion.17:13
tbarronwinston-d: want me to add you as reviewer to these?17:13
jgriffithSwanson: in my case I have multi-tenancy on the cluster that maps to tenants in OpenStack, so I need to make that update17:13
thingeejgriffith: I have to say though, I think it's funny that you're both wanting to remove replication and fix it. :P17:14
winston-dtbarron: scalable backup?17:14
tbarronwinston-d: yes, really just a pre-conditiono for scalable backup :-)17:14
Swansonjgriffith: I'm thinking it doesn't matter because of that.  We had plans for some backend weirdness but smcginnis fixed that..17:14
thingeethat's the kind of guy jgriffith is... "well if it has to be here, I'm going to fix it."17:14
tbarronwinston-d: I'm re-working some of the POC code in light of the recent merge of Xing's non-disruptive backuip17:15
tbarronbackup :-)17:15
*** krypton_ has joined #openstack-cinder17:15
winston-dtbarron: thx, i was expecting sth like 'decouple backup from vol', sth more straightforward, but it makes sense.17:15
jgriffiththingee: well... I guess.  But if it's going to be "in" might as well try and help make it better no?17:15
tbarronwinston-d: -1 and have me change the title if you want :-)17:15
tbarronwinston-d: actually, I agree.17:15
winston-dtbarron: nah, i don't mind. i'm just too lazy to read the code17:16
thingeejgriffith: I was trying to communicate your last sentence, yes. :)17:16
Swansonjgriffith: (Still would be the same tenant so even the weirdness I was planning wouldn't have mattered.)17:17
winston-dtbarron: one question regarding to lisa's change on adding 'volume_name' to backup_restore?17:17
jgriffiththingee: LOL.. yeah, I see that line now (missed it)17:17
winston-dtbarron: you said you understand the motivation, could you elabrate a bit, 'cos I don't.17:18
*** krypton has quit IRC17:18
winston-dbut i am not familiar with backup, so i must have missed sth obvious17:18
*** angela-s has joined #openstack-cinder17:19
*** krypton_ has quit IRC17:20
*** vilobhmm has joined #openstack-cinder17:20
erlonsmcginnis: hey Sean, you have just reviewed my patch 'concurrent attaches on HNAS iSCSI driver', there's still another comment I did, about adding a new exception. Can you answer my question so I can address both points in one commit?17:21
smcginniserlon: Sure. Was that the question about whether to add a new exception or not?17:22
*** afazekas has quit IRC17:22
*** e0ne has joined #openstack-cinder17:22
erlonsmcginnis: yes, currenty every driver addis its own exception17:22
smcginniserlon: Not every driver.17:23
erlonsmcginnis: so I just I just did the same, as I couldn't  reuse the one there17:23
erlonsmcginnis: not not all, but several17:23
smcginniserlon: OK, I see how you are using it now. I'm fine if you leave it as is.17:24
*** xyang has joined #openstack-cinder17:24
smcginniserlon: But the text passed to does need to have _().17:24
tbarronwinston-d: just got back, got pulled AFK for a bit.  Catching up.17:25
openstackgerritabhiram moturi proposed openstack/cinder: Enhance unit tests for zfssa drivers
erlonsmcginnis: ok then, ill fix that, I understood what you meant.17:25
smcginniserlon: OK, great. Thanks erlon!17:26
tbarronwinston-d: the dependent change wants to display volume_name along with the other info.  Do you agree that this could be a useful thing to see?17:27
tbarronwinston-d: her use case is where the same restore cmd is run several times in a row17:28
tbarronwinston-d: showing the name would make it apparent that one is creating a bunch of volumes with the same name17:28
tbarronwinston-d: maybe the use case isn't that compelling.  I just figured someone had gotten into the situation  and showing the name would have been helpful.17:29
winston-dtbarron: i just want to understand how is that helpful.17:32
winston-dif you restore backup using same name multiple times, you'd need a pen to track down vol uuid instead of name, right?17:33
*** krypton has joined #openstack-cinder17:33
tbarronwinston-d: right.17:34
tbarronwinston-d: you should press the point "how is this really useful".  Maybe I was too hasty.17:35
tbarronwinston-d: I was figuring it provided a way of grouping these volumes and making their origin clear.  But Lisa should explain rather than me making up an explanation.17:36
*** avishay__ has joined #openstack-cinder17:36
winston-dtbarron: yeah, i should ask here, but i did in previous comments.17:37
winston-dmaybe i should call her instead.17:37
tbarronwinston-d: you can probably make a local call?17:37
*** lpetrut has quit IRC17:37
winston-dtbarron: yeah, sometime this week.17:38
dannywilsonwinston-d: with regards to and reusing the scsi controller, do you know how to detect it was already added?17:39
winston-dstill stuck at airport, my flight has been delayed for almost 8hrs.17:39
winston-ddannywilson: tbh, no, not yet. :)17:40
tbarronwinston-d: lisa is reviewing the decouple backup stuff too, so maybe encourage her to hang out on IRC.  She doesn't have to be on US time of course.17:40
smcginniswinston-d: Oh man. Good luck.17:40
dannywilsonwinston-d: ouch17:40
tbarronwinston-d: which airport?17:40
*** boris-42 has quit IRC17:40
dannywilsonwinston-d: okay, I will keep looking, good luck17:40
winston-dtbarron: beijing, it's almost 2am. ;017:40
tbarronwinston-d: I was going to say ...  Where are you going?17:41
winston-dtbarron: heading back to hometown, somewhere in south china.17:42
tbarronwinston-d: ah, well enjoy being home then.17:42
tbarronwinston-d: when you eventually get there :(17:42
*** raunak has joined #openstack-cinder17:42
winston-ddannywilson: we piggyback the image properties for discard support.17:43
winston-dno pretty, but works.17:43
winston-ddannywilson: but i really would like your change to land, especially Nova's part.17:43
*** akshai has quit IRC17:44
*** mgarza_ has quit IRC17:44
winston-ddannywilson: we should talk to Daniel P. Berrangé, he must know how17:47
dannywilsonwinston-d: okay, I'll check with him on nova irc, thanks17:47
winston-dtbarron: thx, hopefully i can be there, eventually.17:48
*** afazekas has joined #openstack-cinder17:49
tbarronwinston-d: yes, we'll "keep our fingers crossed" for you17:50
*** akshai has joined #openstack-cinder17:50
winston-dtbarron: thx, my notebook battery is dead, need to get off now.17:51
*** winston-d is now known as winston-d_zZZ17:51
winston-d_zZZswitching to zZZ mode, but not sleeping actually.17:51
winston-d_zZZyou guys have a good day.17:52
dannywilsonwinston-d_zZZ: safe travels17:52
*** harlowja has joined #openstack-cinder17:52
openstackgerritOpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements
*** annegentle has joined #openstack-cinder17:58
*** e0ne has quit IRC18:01
*** kjnelson has joined #openstack-cinder18:03
openstackgerritThang Pham proposed openstack/cinder: Conversion to volume object
*** mylu has quit IRC18:07
*** mylu has joined #openstack-cinder18:08
*** blinky_ghost has quit IRC18:08
openstackgerritErickson Filipe Guedes dos Santos proposed openstack/cinder: Fix get default quota values for subprojects
openstackgerritErickson Filipe Guedes dos Santos proposed openstack/cinder: Nested Quota Driver: Get Project Hierarchy
*** changbl has joined #openstack-cinder18:12
*** salv-orl_ has quit IRC18:14
*** annegentle has quit IRC18:15
*** annashen has quit IRC18:16
*** avishay__ has quit IRC18:17
*** avishay has joined #openstack-cinder18:17
*** annashen has joined #openstack-cinder18:18
*** afazekas has quit IRC18:19
*** jasondot_ has joined #openstack-cinder18:25
openstackgerritJon Bernard proposed openstack/cinder: RBD: use user-configured value for max_clone_depth
*** annegentle has joined #openstack-cinder18:27
*** jasondotstar has quit IRC18:27
*** casusbelli has joined #openstack-cinder18:31
*** boris-42 has joined #openstack-cinder18:42
*** annegentle has quit IRC18:43
*** sgotliv has joined #openstack-cinder18:46
*** mylu has quit IRC18:47
jbernardi'll be at the glance minisummit (since it's in the town where is live), if anyone would like a topic to be raised or anything, im happy to help18:49
*** andrey-mp has joined #openstack-cinder18:49
*** mragupat_ has joined #openstack-cinder18:50
*** rushiagr is now known as rushiagr_away18:50
*** mylu has joined #openstack-cinder18:51
*** casusbelli has quit IRC18:52
*** lpetrut has joined #openstack-cinder18:52
*** mragupat has quit IRC18:54
openstackgerritGorka Eguileor proposed openstack/cinder: WIP: Cleanup stuck jobs
*** raunak has quit IRC18:57
andrey-mphi all. can anyone tell me when such type of review close - ? I've re-based it ten times and i bored ^) that very simple change requires very much efforts...18:57
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder: Add connector object to create_export
*** zhonghua-lee has joined #openstack-cinder19:00
*** annegentle has joined #openstack-cinder19:02
thingeejbernard: is the glance ptl going to be present?19:04
thingeejbernard: last I heard the ptl wasn't going to be present at the tokyo summit.19:05
jbernardthingee: i expect so, it's at the rackspace office19:05
jbernardthingee: (which is where nikhil is located)19:05
thingeehas anyone tried the cinder glance store?
thingeeI cc'd Nikhil since I haven't heard anything from this and it has some Cinder folks approval19:06
jbernardthingee: im happy to read through and raise it in the adjenda19:07
thingeejbernard: I think someone just needs to test it out.19:07
thingeepatrickeast: ^19:08
jbernardthingee: i can do that too19:08
thingeepatrickeast: not sure if you tried it out yet. I'm bring up an instance now with the patch19:08
patrickeastthingee: i haven’t tried out the glance side yet, just the cinder changes for it19:08
patrickeastthingee: i’ll try it out later today too19:08
thingeepatrickeast: there are cinder changes for it?19:08
openstackgerritSean Chen proposed openstack/cinder: Tintri snapshot id
patrickeastthingee: yea there is a change… one sec lemmie find a link19:09
patrickeastthingee: it can kind of work standalone without the glance side, which is what i have tried out19:10
*** avishay has quit IRC19:10
andrey-mpthingee: could you tell me when such type of review close - ? I've re-based it ten times and i bored :) so very simple change requires very much efforts...19:11
*** salv-orlando has joined #openstack-cinder19:11
*** annegentle has quit IRC19:11
*** annegentle has joined #openstack-cinder19:12
tbarronandrey-mp: I'm tired of +1-ing your change :-)19:15
*** zhonghua-lee has quit IRC19:16
*** ub has joined #openstack-cinder19:17
*** lpetrut has quit IRC19:18
*** xyang has quit IRC19:20
*** edtubill1 has joined #openstack-cinder19:21
*** xyang has joined #openstack-cinder19:22
thingeepatrickeast: thanks19:23
thingeemy goal today is to be further along with the volume migration patches and tested out the cinder glance store.19:24
*** edtubill has quit IRC19:25
andrey-mpthingee: thank you19:27
*** shyama has quit IRC19:28
*** andrey-mp has left #openstack-cinder19:32
*** lpetrut has joined #openstack-cinder19:36
openstackgerritErickson Filipe Guedes dos Santos proposed openstack/cinder: Fix get default quota values for subprojects
openstackgerritErickson Filipe Guedes dos Santos proposed openstack/cinder: Nested Quota Driver: Get Project Hierarchy
*** thangp has quit IRC19:44
tsekiyamathingee: thanks for testing cinder glance store. Just FYI, you also need glance rootwrap patch to make it work:
*** harlowja has quit IRC19:45
*** vilobhmm has quit IRC19:45
*** raunak has joined #openstack-cinder19:49
*** vilobhmm has joined #openstack-cinder19:56
j_kingdoes provider_location provide any information about the driver/backend? Noticing that the tests in assume provider location can be something in the ballpark of r'[a-zA-Z0-9-.]+:/{1}.*$' and wants to check the volume in ensure_export() to see that the volume was created by the same provider as is presently configured...19:58
openstackLaunchpad bug 1287134 in Cinder "C-vol crashes when driver class is changed after some volumes exist on that backend" [Low,In progress] - Assigned to j_king (james-agentultra)19:58
j_kingcode doesn't say much and presently trying to scrape through documentation but turning up short19:59
*** harlowja has joined #openstack-cinder20:01
*** Apoorva has quit IRC20:01
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder-specs: Add Connector to create_export
*** andreykurilin__ has joined #openstack-cinder20:07
*** diogogmt_ has joined #openstack-cinder20:19
*** breitz has quit IRC20:19
*** breitz has joined #openstack-cinder20:19
*** Apoorva has joined #openstack-cinder20:21
j_kingunless I'm totally lost it seems like there's no way for the driver or the manager to know what backend a volume came from20:21
*** diogogmt has quit IRC20:22
*** diogogmt_ is now known as diogogmt20:22
Swansonj_king: it doesn't in the dell driver.  I do not return provider_location.20:22
hemnaj_king, the volume table has the host field, which is what backend and pool the volume came from.20:22
*** merooney has quit IRC20:22
patrickeasthemna: beat me to it :p20:23
j_kinghemna: host might be good; in the nfs case I would I be right assuming that's a parseable url?20:23
j_kingpatrickeast: I have a buffer open on that file atm, ty20:23
hemnaj_king, the format of the host field should always be foo@bar#pool20:24
hemnaregardless of the backend.20:24
*** sghanekar_ has quit IRC20:26
*** lpetrut has quit IRC20:31
*** mragupat_ has quit IRC20:36
jungleboyjhemna: smcginnis xyang Would appreciate some eyes on these simple patches to help get our drivers back to proper functionality: and
*** mragupat has joined #openstack-cinder20:37
*** mragupat has quit IRC20:37
*** annegentle has quit IRC20:37
*** mragupat has joined #openstack-cinder20:37
*** julim_ has quit IRC20:38
*** mragupat has quit IRC20:38
hemnajungleboyj, Is there CI for the XIV  driver?20:39
*** mragupat has joined #openstack-cinder20:39
hemnamaybe I'm not seeing it20:39
jungleboyjNot at the moment because it has been disabled partially because of this bug.20:39
jungleboyjhemna: I am working that issue internally.  It has been verified that that patch works internally.20:40
hemnabut shouldn't it be reporting for that patch to prove it fixes it ?20:40
jungleboyjhemna: How would that happen if they have disabled it?20:40
SwansonI want to mark a patch as also being a problem in Kilo.  I don't remember how to do this.  Anyone?20:40
hemnaenable the CI and it tests against that patch and reports success ?20:41
jungleboyjSwanson: Nominate for Series20:41
jungleboyjRight, but they aren't allowing us to report results anymore.20:41
*** Yogi1 has quit IRC20:42
openstackgerritEric Harney proposed openstack/cinder: Move update_migrated_volume to correct class
openstackgerritEric Harney proposed openstack/cinder: Log which service is down
*** raildo has joined #openstack-cinder20:46
patrickeasttsekiyama: hey, so i’m looking into the v2 glance api that has virtual_size, but i’m not seeing it ever be anything other than ‘None’… is there some glance configuration required to get those filled it? or is it a manual step after creating the image?20:46
*** harlowja_ has joined #openstack-cinder20:47
openstackgerritMike Perez proposed openstack/cinder: Remove deprecated config options for Liberty
tsekiyamapatrickeast: hmm, as far as I tried so far, it is not yet completed. Currently it is set only when you use the import flow to register an image.
patrickeasttsekiyama: ah ok cool20:49
*** vilobhmm has quit IRC20:50
patrickeastpatrickeast: so I think what i’ll do is check for it, if its there we use it, otherwise look for a property with the right name (like it does right now)20:51
patrickeasttsekiyama: ^20:51
*** harlowja has quit IRC20:51
patrickeasttsekiyama: and for the ones we upload from cinder we can set the value for them based on the volume size20:51
tsekiyamapatrickeast: yeah I think that is reasonable20:52
*** eharney has quit IRC20:53
*** bswartz has quit IRC20:54
openstackgerritErickson Filipe Guedes dos Santos proposed openstack/cinder: Fix get default quota values for subprojects
openstackgerritErickson Filipe Guedes dos Santos proposed openstack/cinder: Nested Quota Driver: Get Project Hierarchy
thingeetsekiyama: thanks20:54
*** raunak has quit IRC20:58
*** raildo has quit IRC21:00
*** edtubill1 has left #openstack-cinder21:00
Swansonjungleboyj: I'm not familiar with nominate for series.  I did find Cherry Pick to Branch which I think I've done before but that doesn't seem to be working for me.  Where do I find the nominate for series bit?  Or is it some kind of typey typey thing.21:00
*** raunak has joined #openstack-cinder21:01
*** sghanekar_ has joined #openstack-cinder21:02
*** annegentle has joined #openstack-cinder21:06
*** akerr has quit IRC21:11
*** annashen has quit IRC21:14
openstackgerritJon Bernard proposed openstack/cinder: Add support for file I/O volume migration
*** xyang has quit IRC21:21
*** crose has quit IRC21:31
thingeeSwanson: should be under the bug information of assignee and priority21:31
*** vilobhmm has joined #openstack-cinder21:38
*** annashen has joined #openstack-cinder21:39
*** salv-orlando has quit IRC21:39
*** mylu has quit IRC21:41
jungleboyjSwanson: So, you can cherry-pick the change to stable/kilo and then Launchpad takes care of the deails of showing a fix has been proposed.21:44
jungleboyjSwanson: Otherwise, below the table that shows Affect, Status, etc. There should be a little clock button that says 'Nominate for Series' there.21:45
jungleboyjThat is another way to indicate that something should be cherry-picked back.21:45
thingeegit cherry-pick -x21:45
jungleboyjYeah.  :-)21:46
*** julim has joined #openstack-cinder21:46
jungleboyjSwanson: I am wondering if that button is only something that stable maintainers have.21:51
jungleboyjSwanson: Anyway, if you want to propose a patch for backport, do the cherry pick.  I will get an e-mail and review it.21:51
smcginnisThanks jungleboyj, I don't have that option on my screen, so probably the same for Swanson.21:52
Swansonjungleboyj:  Thanks!  I think I've gone blind trying to find that.21:52
smcginnisSwanson: No comment.21:53
Swansonthingee: thanks!21:53
jungleboyjsmcginnis: I am biting my tongue as well.21:53
Swansonsmcginnis: quiet you21:53
jungleboyjSecond time things have gone wrong out here today.21:53
smcginnisGotta have some fun on a Monday. :)21:54
jungleboyjTell me about it.  I need that today.21:54
jungleboyjAnyway, I gotta drop off.  I will watch for the backport and take a look at it.  That is a big code change, but it is limited to your driver.  So, I can look.21:54
Swansonjungleboyj: Thanks!21:55
*** sgotliv has quit IRC21:55
jungleboyjSwanson: Welcome sir.  Don't go blind.21:55
*** mragupat has quit IRC21:57
*** ganso_ has quit IRC22:00
*** jungleboyj has quit IRC22:01
*** Guest10__ has quit IRC22:07
*** Lee1092 has quit IRC22:07
SwansonOh, what an ugly merge.22:07
*** chlong has joined #openstack-cinder22:09
*** salv-orlando has joined #openstack-cinder22:09
*** xyang has joined #openstack-cinder22:11
patrickeasttsekiyama: i just noticed after the latest rebase there is a new error looks like there was some cleanup to remove extra parameters in create volume22:13
*** edmondsw has quit IRC22:14
openstackgerritSonia Ghanekar proposed openstack/cinder: Nimble driver version shows the correct version instead of N/A in the volume.log
*** jasondot_ has quit IRC22:15
*** harlowja_ has quit IRC22:21
*** harlowja has joined #openstack-cinder22:21
openstackgerritDanny Wilson proposed openstack/cinder: Add discard to connection properties
*** raunak has quit IRC22:24
*** xyang has quit IRC22:24
*** raunak has joined #openstack-cinder22:24
*** marcusvrn has quit IRC22:25
tsekiyamapatrickeast: thanks, I'm looking into it.22:28
openstackgerritSonia Ghanekar proposed openstack/cinder: Nimble driver version shows the correct version instead of N/A in the volume.log
*** robotiX has joined #openstack-cinder22:42
thingeetsekiyama: which group does show_multiple_locations=True get set to? This not obvious from the commit message or /etc/glance/glance-api.conf22:44
*** robotiX has left #openstack-cinder22:45
tsekiyamathingee: It's [DEFAULT] section in /etc/glance/glance-api.conf.22:47
thingeeOk, it would probably be good to have that in the glance sample conf or something22:47
tsekiyamathingee: I agree22:47
*** annegentle has quit IRC22:50
*** DericHorn-HP has joined #openstack-cinder22:51
openstackgerritMitsuhiro Tanino proposed openstack/cinder: Adds framework for get_capabilities() feature
thingeetsekiyama: ok, about to test things22:54
*** alexpilotti has quit IRC22:54
*** akshai has quit IRC22:55
openstackgerritSonia Ghanekar proposed openstack/cinder: Nimble driver version shows the correct version instead of N/A in the volume.log
*** Apoorva has quit IRC22:57
*** DericHorn-HP has quit IRC22:58
thingeetsekiyama: things exploded :(
thingeetsekiyama: this was uploading to image22:59
tsekiyamathingee: hmm23:00
tsekiyamathingee: let me see what is wrong...23:00
patrickeastoh sweet, i’m seeing that too… i thought for sure my cache stuff i’m playing with was breaking it23:00
*** r-daneel has quit IRC23:02
patrickeastthingee: tsekiyama: fwiw i only see that error uploading a volume to image with the v2 glance api enabled23:02
*** jwcroppe has quit IRC23:02
patrickeastenabled in cinder.conf*23:03
*** jwcroppe has joined #openstack-cinder23:03
*** DericHorn-HP has joined #openstack-cinder23:05
thingeepatrickeast: the patch noted it has to be v2, did it succeed for you with v1?23:06
* thingee is happy patrickeast is also trying this23:06
tsekiyamapatrickeast: thingee: ah...  so it is a cinder bug, not related to this patch23:06
*** arahal has joined #openstack-cinder23:06
*** DericHorn-HP has quit IRC23:06
patrickeastthingee: i don’t even have the cinder:// url stuff enabled and just creating an image from volume with v2 enabled does it23:06
tsekiyamathingee: can you try just "glance image-create --file ... ..." to create a new image?23:07
thingeedid you all know about btw?23:08
*** jasondotstar has joined #openstack-cinder23:08
thingeebam! glance image-create command already ready
patrickeastoh wow23:08
patrickeastthats pretty sweet23:08
thingeeYeah I'm really, really happy to see some of this surface. While I agree the foundation needs work first and foremost, this useful bits are quite appreciated!!23:09
tsekiyamaIt's nice. if you add --store cinder (only supported by glance v1 API), you may put the image into cinder volume23:09
thingeetsekiyama: so far it seems to be working. the image is in a saving state23:10
thingeetsekiyama: fwiw, cinder upload-to-image was working fine for me previously. If it's not this patch, it might be v2 enabled what is hitting this issue23:12
openstackgerritMitsuhiro Tanino proposed openstack/cinder: Adds framework for get_capabilities() feature
*** annashen has quit IRC23:12
*** jasondotstar has quit IRC23:13
tsekiyamathingee: Have you also applied my cinder patch for "efficient upload/download for glance cinder store"?23:16
thingeetsekiyama: I applied the glance patch for rootwrap, made sure things are setup in /etc/glance/rootwrap.d, applied the Cinder patch, applied the glance_store patch23:17
*** annashen has joined #openstack-cinder23:17
*** Apoorva has joined #openstack-cinder23:18
thingeetsekiyama: the only config changes I made is what's listed here
thingeeexcept for image_upload_use_cinder_backend and image_upload_use_internal_tenant23:18
openstackgerritMitsuhiro Tanino proposed openstack/cinder: Efficient volume copy for generic volume migration
tsekiyamathingee: ah ok. then my cinder patch is doing something bad23:18
*** jaypipes has quit IRC23:19
thingeetsekiyama: have you tried applying the patches to a clean environment?23:19
tsekiyamathingee: yeah, but it looks like there were rebase error before proposing the last revision23:20
patrickeastso that error about the keys looks like it is due to glance giving back a 204 and thus giving us None here
openstackgerritVilobh Meshram proposed openstack/cinder: ConsistencyGroup Object
thingeepatrickeast: ah, yeah I thought it might be related to v1 and v2 differences23:22
patrickeastwhich is different than the rebase error which will look like
*** annashen has quit IRC23:23
patrickeastthingee: yea, i wonder if this is a glance bug or if we are expecting the wrong thing23:24
patrickeastdoh, glance is behaving as expected
patrickeastthis call is wrong
patrickeastthere is not image_meta coming back from that upload call23:28
*** jasondot_ has joined #openstack-cinder23:28
thingeepatrickeast: was it for v1?23:29
patrickeastthingee: for v1 we call a different method, which does appear to return the new image data23:29
thingeeok so v2 glance in cinder is just broken regardless23:30
patrickeasti think so23:30
tsekiyamalooks like the returned value from v2 is really different from v123:31
thingeetsekiyama: yeah that's what patrickeast was saying23:31
*** salv-orlando has quit IRC23:32
*** arahal has quit IRC23:37
*** arahal has joined #openstack-cinder23:37
*** arahal has quit IRC23:37
tsekiyamathingee: I'm wondering why it is trying to upload the image data. It should just register the location to Glance and if that succeeds, it won't upload the data. Do you have some other error logs before ?23:38
*** markvoelker has quit IRC23:40
*** erlon has quit IRC23:43
*** jasondotstar has joined #openstack-cinder23:44
*** DericHorn-HP has joined #openstack-cinder23:44
*** diogogmt has quit IRC23:44
*** jasondot_ has quit IRC23:47
*** vokt has quit IRC23:48
*** bswartz has joined #openstack-cinder23:51
tsekiyamathingee: okey I confirmed "upload-to-image --disk-format=qcow2 --container-format=bare vol1 img1" fails when glance_api_version is set to 2 even without any patches.23:53
*** Yogi1 has joined #openstack-cinder23:56
*** dannywilson has quit IRC23:59

Generated by 2.14.0 by Marius Gedminas - find it at!