Tuesday, 2015-08-18

*** Yogi1 has quit IRC00:09
*** diogogmt has quit IRC00:09
*** akerr_ has quit IRC00:10
*** angela-s has quit IRC00:10
*** Ctina_ has quit IRC00:11
*** dannywilson has quit IRC00:11
xyang1jgriffith: Maybe find something sounds better than meatball?:)00:12
xyang1jgriffith: I see what you are saying.  i just don't have a better solution to solve the ABC mess00:13
*** lixiaoy1 has joined #openstack-cinder00:18
*** takedakn has quit IRC00:23
*** garthb_ has quit IRC00:25
*** tsekiyama has quit IRC00:26
*** mtanino has quit IRC00:32
*** dims has joined #openstack-cinder00:34
*** merooney has joined #openstack-cinder00:36
*** daneyon has joined #openstack-cinder00:39
*** mriedem has joined #openstack-cinder00:43
*** leeantho_ has quit IRC00:44
*** mriedem1 has quit IRC00:45
*** merooney has quit IRC00:45
*** daneyon_ has joined #openstack-cinder00:46
openstackgerritLisaLi proposed openstack/cinder: Prevent uploading encrypted volume to images  https://review.openstack.org/21361600:46
openstackgerritLisaLi proposed openstack/cinder: Prevent uploading encrypted volume to images  https://review.openstack.org/21361600:48
*** mriedem has quit IRC00:48
*** daneyon has quit IRC00:49
*** mriedem has joined #openstack-cinder00:50
*** zhonghua-lee has joined #openstack-cinder00:52
mriedemjungleboyj: in https://review.openstack.org/#/c/174783/ - where does the test actually mock out the port_status?00:58
mriedemi think i see it now01:01
mriedemdef _cmd_lsnode(self, **kwargs): overrides the ssh call, and node2 has an active and inactive port01:01
*** dflorea has quit IRC01:02
mriedemseems unfortunate that the storwize fc ci failed and the change is only for storwize fc :(01:03
openstackgerritYusuke Hayashi proposed openstack/cinder: Update deprecated version of novaclient  https://review.openstack.org/21073701:03
*** vilobhmm has quit IRC01:05
*** skylerberg has quit IRC01:05
*** xyang1 has quit IRC01:08
openstackgerritxing-yang proposed openstack/cinder: Allow CG without snapshot to be deleted  https://review.openstack.org/21389201:13
*** mriedem has quit IRC01:13
*** mtanino has joined #openstack-cinder01:14
*** tiny-hands has quit IRC01:20
*** tiny-hands has joined #openstack-cinder01:21
*** zhenguo has joined #openstack-cinder01:22
openstackgerritOpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements  https://review.openstack.org/21389101:25
*** andreykurilin__ has quit IRC01:25
*** salv-orlando has joined #openstack-cinder01:29
*** esker has joined #openstack-cinder01:30
*** rick__ has joined #openstack-cinder01:33
openstackgerritOpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements  https://review.openstack.org/21389101:33
*** salv-orlando has quit IRC01:35
*** esker has quit IRC01:36
*** bill_az has quit IRC01:37
*** _cjones_ has quit IRC01:37
openstackgerritAdriano Freires Rosso proposed openstack/cinder: Adds manage/unmanage methods for HNAS drivers.  https://review.openstack.org/19799901:37
*** rick__ has quit IRC01:37
*** rick-chen has joined #openstack-cinder01:38
openstackgerritxing-yang proposed openstack/cinder: EMC VMAX Create CG from CG Snapshot  https://review.openstack.org/19528001:38
*** lxg has joined #openstack-cinder01:38
openstackgerritLi Yingjun proposed openstack/cinder: Add the ability to update type public status  https://review.openstack.org/20678801:39
*** takedakn has joined #openstack-cinder01:42
*** takedakn has quit IRC01:42
*** wilson has quit IRC01:42
*** vincent_hou has joined #openstack-cinder01:45
openstackgerritDavanum Srinivas (dims) proposed openstack/cinder: Fix test failure under python 2.7.9  https://review.openstack.org/21396201:58
openstackgerritXinXiaohui proposed openstack/cinder: Calculate virtual free capacity and notify  https://review.openstack.org/20692302:01
*** sakaYK has quit IRC02:02
openstackgerritThang Pham proposed openstack/cinder: Register RPC and object versions  https://review.openstack.org/20970102:05
*** haomaiwang has joined #openstack-cinder02:06
openstackgerritThang Pham proposed openstack/cinder: Register RPC and object versions  https://review.openstack.org/20970102:09
*** mtanino has quit IRC02:11
openstackgerritThang Pham proposed openstack/cinder: Register RPC and object versions  https://review.openstack.org/20970102:13
*** p0rtal_ has quit IRC02:14
*** davechen has joined #openstack-cinder02:14
*** vervet has joined #openstack-cinder02:20
*** vervet has quit IRC02:21
*** vincent_hou has quit IRC02:22
openstackgerritThang Pham proposed openstack/cinder: Conversion to volume object  https://review.openstack.org/17705402:23
*** zhonghua-lee has quit IRC02:27
*** smoriya has joined #openstack-cinder02:29
*** dims has quit IRC02:30
*** tiny-hands has quit IRC02:30
openstackgerritVincent Hou proposed openstack/cinder: Use cinder internal tenant to create the destination volume  https://review.openstack.org/21023702:31
*** baojg has joined #openstack-cinder02:31
*** bkopilov has quit IRC02:32
*** vervet has joined #openstack-cinder02:39
*** bswartz has joined #openstack-cinder02:42
*** logan2 has quit IRC02:44
*** akerr has joined #openstack-cinder02:50
*** esker has joined #openstack-cinder02:51
*** akerr_ has joined #openstack-cinder02:52
*** hakimo has joined #openstack-cinder02:52
*** hakimo_ has quit IRC02:54
*** akerr has quit IRC02:55
*** haomaiwang has quit IRC03:01
*** baojg has quit IRC03:02
*** haomaiwa_ has joined #openstack-cinder03:02
*** apoorvad has quit IRC03:02
*** esker has quit IRC03:02
*** esker has joined #openstack-cinder03:04
openstackgerritOpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements  https://review.openstack.org/21389103:05
*** diogogmt has joined #openstack-cinder03:13
*** simondodsley has joined #openstack-cinder03:15
*** baojg has joined #openstack-cinder03:15
*** HoloIRCUser has joined #openstack-cinder03:18
*** salv-orlando has joined #openstack-cinder03:19
*** simondodsley has quit IRC03:19
*** HoloIRCUser has quit IRC03:23
*** HoloIRCUser has joined #openstack-cinder03:24
*** diogogmt has quit IRC03:26
*** esker has quit IRC03:26
*** salv-orlando has quit IRC03:26
*** esker has joined #openstack-cinder03:26
openstackgerritThang Pham proposed openstack/cinder: Continue conversion to volume object  https://review.openstack.org/20140403:28
*** skylerberg has joined #openstack-cinder03:30
*** changbl has quit IRC03:30
openstackgerritDave Chen proposed openstack/python-cinderclient: Clean the unused methods in `volumes.py`  https://review.openstack.org/21028703:32
*** garthb has joined #openstack-cinder03:32
*** logan2 has joined #openstack-cinder03:34
*** akerr_ has quit IRC03:34
*** esker has quit IRC03:37
*** esker has joined #openstack-cinder03:38
openstackgerritWilson Liu proposed openstack/cinder: Add hypermetro support for Huawei driver  https://review.openstack.org/20202303:39
*** sakaYK has joined #openstack-cinder03:41
*** markvoelker has quit IRC03:42
*** sakaYK has quit IRC03:42
openstackgerritWilson Liu proposed openstack/cinder: Add volume retype support for Huawei driver  https://review.openstack.org/20148503:43
*** baojg has quit IRC03:49
*** jgregor has joined #openstack-cinder03:50
*** jgregor has quit IRC03:54
*** bkopilov has joined #openstack-cinder03:58
*** esker has quit IRC03:58
*** esker has joined #openstack-cinder04:00
*** haomaiwa_ has quit IRC04:01
*** haomaiwa_ has joined #openstack-cinder04:02
*** zhenguo has quit IRC04:06
*** mylu has joined #openstack-cinder04:09
*** coolsvap|away is now known as coolsvap04:10
*** esker has quit IRC04:13
*** esker has joined #openstack-cinder04:13
*** esker has quit IRC04:14
*** esker has joined #openstack-cinder04:14
*** baojg has joined #openstack-cinder04:16
*** bill_az has joined #openstack-cinder04:18
openstackgerritwanghao proposed openstack/cinder: Implement function to manage/unmanage snapshots  https://review.openstack.org/14459004:19
*** esker has quit IRC04:20
*** mylu has quit IRC04:21
*** lixiaoy1 has quit IRC04:22
openstackgerritwanghao proposed openstack/cinder: Implement function to manage/unmanage snapshots  https://review.openstack.org/14459004:32
*** HoloIRCUser has quit IRC04:35
*** deepakcs has joined #openstack-cinder04:35
*** shyama has joined #openstack-cinder04:38
*** markvoelker has joined #openstack-cinder04:42
*** mylu has joined #openstack-cinder04:42
*** vilobhmm has joined #openstack-cinder04:43
openstackgerritwanghao proposed openstack/cinder: Implement function to manage/unmanage snapshots  https://review.openstack.org/14459004:44
*** apoorvad has joined #openstack-cinder04:46
*** markvoelker has quit IRC04:47
*** apoorvad has quit IRC04:54
*** Apoorva has joined #openstack-cinder04:55
*** lixiaoy1 has joined #openstack-cinder04:58
*** haomaiwa_ has quit IRC05:01
*** haomaiwa_ has joined #openstack-cinder05:02
*** lixiaoy11 has joined #openstack-cinder05:04
*** nkrinner has joined #openstack-cinder05:06
*** lixiaoy1 has quit IRC05:07
openstackgerrityuji hagiwara proposed openstack/cinder: Remove duplicate keys from dictionary  https://review.openstack.org/21399905:12
*** dims has joined #openstack-cinder05:20
*** smoriya has quit IRC05:21
*** garthb has quit IRC05:22
*** dims has quit IRC05:26
*** Apoorva has quit IRC05:28
*** p0rtal has joined #openstack-cinder05:30
*** p0rtal has joined #openstack-cinder05:31
*** salv-orlando has joined #openstack-cinder05:37
*** p0rtal has quit IRC05:45
*** p0rtal has joined #openstack-cinder05:48
*** salv-orlando has quit IRC05:50
*** ankit_ag has joined #openstack-cinder05:52
*** haomaiwa_ has quit IRC06:01
*** haomaiwang has joined #openstack-cinder06:02
openstackgerritTina Tang proposed openstack/cinder: Return multiple iSCSI portals in VNX Cinder driver  https://review.openstack.org/21358306:08
*** skylerberg has quit IRC06:14
*** BharatK has quit IRC06:14
winston-d_zZZDuncanT: around?06:17
*** winston-d_zZZ is now known as winston-d06:17
*** salv-orlando has joined #openstack-cinder06:21
*** coolsvap is now known as coolsvap|away06:22
*** vlaza has joined #openstack-cinder06:24
*** zhenguo has joined #openstack-cinder06:24
openstackgerritwanghao proposed openstack/cinder: Add pagination to backups  https://review.openstack.org/20449306:30
*** zhenguo has quit IRC06:32
*** p0rtal has quit IRC06:32
*** mylu has quit IRC06:35
*** mylu has joined #openstack-cinder06:36
*** anshul has joined #openstack-cinder06:37
*** vilobhmm has quit IRC06:39
*** mylu has quit IRC06:40
*** zhonghua-lee has joined #openstack-cinder06:40
*** openstackgerrit_ has joined #openstack-cinder06:43
*** markvoelker has joined #openstack-cinder06:43
*** bill_az has quit IRC06:44
*** markvoelker has quit IRC06:48
*** BharatK has joined #openstack-cinder06:50
*** yrabl has joined #openstack-cinder06:52
*** mylu has joined #openstack-cinder06:57
*** mylu has quit IRC06:58
*** BharatK has quit IRC06:58
*** haomaiwang has quit IRC07:01
*** haomaiwang has joined #openstack-cinder07:02
*** rushiagr_away is now known as rushiagr07:03
*** geguileo has joined #openstack-cinder07:08
*** lihit has joined #openstack-cinder07:11
*** lihit has quit IRC07:11
*** BharatK has joined #openstack-cinder07:12
*** markus_z has joined #openstack-cinder07:14
*** alexschm has joined #openstack-cinder07:15
openstackgerritTina Tang proposed openstack/cinder: Fix the virtual port support in VNX driver  https://review.openstack.org/21404407:20
*** zhongjun has joined #openstack-cinder07:23
*** vervet has quit IRC07:39
*** vervet has joined #openstack-cinder07:39
*** pschaef has joined #openstack-cinder07:40
*** pschaef has quit IRC07:41
*** IanGovett has joined #openstack-cinder07:42
*** gardenshed has joined #openstack-cinder07:42
*** ronis has joined #openstack-cinder07:44
*** ronis_ has joined #openstack-cinder07:46
*** dims has joined #openstack-cinder07:48
*** sgotliv has joined #openstack-cinder07:48
*** ronis has quit IRC07:50
*** dims has quit IRC07:53
*** agarciam has joined #openstack-cinder07:53
lixiaoy11:wanghao here?07:55
*** jistr has joined #openstack-cinder07:55
openstackgerrityuji hagiwara proposed openstack/cinder: Remove duplicate keys from dictionary  https://review.openstack.org/21399907:58
wanghaolixiaoy11: yes07:59
*** haomaiwang has quit IRC08:01
*** haomaiwang has joined #openstack-cinder08:02
lixiaoy11:wanghao I am just looking at your patch : https://review.openstack.org/#/c/207696/908:05
*** jordanP has joined #openstack-cinder08:05
lixiaoy11:wanghao one question:         image_size = utils.as_int(image_meta['size'], quiet=False)08:06
lixiaoy11        image_size_in_gb = (image_size + GB - 1) / GB08:06
lixiaoy11You set image size as 1 in unit test, image_size_in_gb = 1. So it is not zero.08:06
lixiaoy11wanghao: based on your update, it shuold not raise invalidinput exception here08:07
wanghaolixiaoy11: I remembered it was set zero in fake_image_service.... Well I will check the code later.08:15
lixiaoy11wanghao: ok08:16
winston-dlixiaoy11: Lisa?08:17
lixiaoy11winston-d: yes08:17
winston-dlixiaoy11: https://review.openstack.org/#/c/213616/, shouldn't we fix Cinder to save encryption metadata to Glance, rather than stopping uploading encrypted volume to image?08:18
lixiaoy11winston-d: as the liberty is going to freeze. three things need to do: 1. save the encryption metadata into Glance, 2. Create volumes with image, 3. Work with Nova08:20
*** zhongjun has quit IRC08:20
lixiaoy11winston-d: i am not sure it is worthy to save encryption image in Glance. How do you think08:21
*** andreykurilin__ has joined #openstack-cinder08:22
winston-dlixiaoy11: well, I think it depends on how the volume is encypted.08:22
lixiaoy11winston-d: what do you mean?08:23
winston-dif the volume is encrypted by the storage backend, which means when the data leaves backend, it should be decrypted first.08:24
lixiaoy11winston-d: I am afraid that we don't care about this scenario. As both from cinder /nova, it encrypts/unencrypts in backend storage, and they read same data08:25
*** gardensh_ has joined #openstack-cinder08:26
*** gardenshed has quit IRC08:26
winston-dwhy not? you change will stop user from uploading such volume to glance.08:26
*** bluex-pl has joined #openstack-cinder08:27
*** bluex-pl has quit IRC08:28
*** bluex-pl has joined #openstack-cinder08:28
lixiaoy11winston-d: no. As currently this just cares about the encryption by dm-crypt.  that is a volume_type is encrypted.08:28
winston-dvolume_type.is_encrypted != volume is encrypted by dm-crypt08:30
lixiaoy11winston-d: I mean this encryption uses the third software to do encryption, Cinder needs to record the encryption algorithm, key, etc in db.08:32
lixiaoy11winston-d: the encryption in backend storage, which is transparent to cinder node/compute node08:33
lixiaoy11winston-d: if encryption is done by backend storage, no metadata is recorded in cinder, is_encrypted returns false08:36
winston-dlixiaoy11: if 'encryption is transparent to cinder/nova', then i agree, but i am not sure that is true for all backend level encryption.08:37
lixiaoy11winston-d: let me have a check08:39
*** shausy has joined #openstack-cinder08:42
*** gardensh_ has quit IRC08:43
*** markvoelker has joined #openstack-cinder08:45
*** zhongjun has joined #openstack-cinder08:45
*** chenying has joined #openstack-cinder08:47
openstackgerritYuriy Nesenenko proposed openstack/python-cinderclient: Add support '--all-tenants' for cinder backup-list  https://review.openstack.org/20746908:49
lixiaoy11winston-d: from cinder code, no encryption info is trasferred to backend storage. if encryption is done by backend, it should know such things. Any way, wait for others confirmed08:49
*** d0ugal has joined #openstack-cinder08:50
*** markvoelker has quit IRC08:50
winston-dlixiaoy11: sure, thx for confirming.08:52
lixiaoy11winston-d: my concern is that using different encryption algorithm, so may be create a encrypted image with encryption algorithm 1, and another same image with encryption algorithm 2. is it useful for customers08:53
*** Lee1092 has joined #openstack-cinder08:55
*** chenying has quit IRC08:58
lixiaoy11winson-d: how about I set work in progress -1 for above patch, and I am trying to save encryption metadata to Glance, if it can be completed in libirty, it is ok, Or else use this patch to prevent08:58
*** haomaiwang has quit IRC09:01
*** shikel has quit IRC09:01
*** haomaiwang has joined #openstack-cinder09:02
*** zhongjun has quit IRC09:05
*** zhongjun has joined #openstack-cinder09:06
*** sh1kel has joined #openstack-cinder09:06
*** lixiaoy11 has quit IRC09:08
openstackgerritYuriy Nesenenko proposed openstack/python-cinderclient: Add support '--all-tenants' for cinder backup-list  https://review.openstack.org/20746909:12
openstackgerritGorka Eguileor proposed openstack/cinder: Fix backup list all_tenants=0 issue  https://review.openstack.org/21408609:32
*** gardenshed has joined #openstack-cinder09:33
*** rushiagr is now known as rushiagr_away09:37
*** afazkas has joined #openstack-cinder09:42
*** zhongjun has quit IRC09:42
*** davechen has left #openstack-cinder09:54
*** gardenshed has quit IRC09:54
*** tiny-hands has joined #openstack-cinder09:55
openstackgerritGorka Eguileor proposed openstack/cinder: Make migration's volume source deletion async  https://review.openstack.org/21125710:00
*** tiny-hands has quit IRC10:00
*** haomaiwang has quit IRC10:01
*** haomaiwang has joined #openstack-cinder10:02
*** baojg has quit IRC10:06
*** baojg has joined #openstack-cinder10:06
*** baojg has quit IRC10:07
*** baojg has joined #openstack-cinder10:07
*** baojg has quit IRC10:07
*** baojg has joined #openstack-cinder10:07
*** rick-chen has quit IRC10:08
*** baojg has quit IRC10:17
*** baojg has joined #openstack-cinder10:18
*** baojg has quit IRC10:23
*** sayali has quit IRC10:24
*** sayali has joined #openstack-cinder10:25
*** gardenshed has joined #openstack-cinder10:36
* DuncanT returns from vacation and starts climbing to email mountain10:38
*** gardenshed has quit IRC10:44
*** gardenshed has joined #openstack-cinder10:45
*** markvoelker has joined #openstack-cinder10:46
*** zhonghua-lee has quit IRC10:46
*** gardenshed has quit IRC10:47
*** tiny-hands has joined #openstack-cinder10:47
*** gardenshed has joined #openstack-cinder10:47
*** zhonghua-lee has joined #openstack-cinder10:48
*** sh1kel has quit IRC10:48
*** markvoelker has quit IRC10:50
*** dims has joined #openstack-cinder10:51
*** tiny-hands has quit IRC10:52
BharatKDuncanT, Please review the patch https://review.openstack.org/#/c/182254/ when you have some time. Thanks.10:53
*** haomaiwang has quit IRC10:54
*** salv-orlando has quit IRC10:57
openstackgerritGorka Eguileor proposed openstack/cinder: Fix backup list all_tenants=0 issue  https://review.openstack.org/21408610:58
DuncanTBharatK: An impressively small patch, I'll review it now10:58
BharatKDuncanT, Thanks :D10:59
DuncanTBharatK: Code looks good, and looking at a couple of the CI failures they don't seem related to the patch11:01
BharatKDuncanT, Yes11:02
BharatKDuncanT, They are failing for all Cinder patches.11:03
DuncanTBharatK: Thanks. I'm approving now11:03
BharatKDuncanT, Thank you very much.11:03
DuncanTBharatK: Thanks for an easy to review patch11:04
BharatKDuncanT, np. :-)11:05
*** shausy has quit IRC11:09
*** shausy has joined #openstack-cinder11:09
*** delattec has quit IRC11:10
*** cdelatte has quit IRC11:10
*** gardenshed has quit IRC11:14
*** salv-orlando has joined #openstack-cinder11:19
*** esker has joined #openstack-cinder11:20
*** marcusvrn has joined #openstack-cinder11:27
*** dims_ has joined #openstack-cinder11:28
*** dims has quit IRC11:30
*** zhongjun has joined #openstack-cinder11:31
*** cdelatte has joined #openstack-cinder11:32
*** lpetrut has joined #openstack-cinder11:36
*** DuncanT has quit IRC11:38
*** marcusvrn has quit IRC11:38
*** markvoelker has joined #openstack-cinder11:46
*** esker has quit IRC11:50
*** esker has joined #openstack-cinder11:51
*** markvoelker has quit IRC11:51
*** bkopilov has quit IRC11:58
*** dims has joined #openstack-cinder12:00
*** dims has quit IRC12:00
*** dims_ has quit IRC12:00
*** dims has joined #openstack-cinder12:01
*** esker has quit IRC12:01
*** dims has quit IRC12:05
*** andreykurilin__ has quit IRC12:05
*** markvoelker has joined #openstack-cinder12:12
*** Lee1092 has quit IRC12:15
*** dims has joined #openstack-cinder12:15
*** Lee1092 has joined #openstack-cinder12:16
*** salv-orlando has quit IRC12:18
*** raildo-afk is now known as raildo12:18
*** bswartz has quit IRC12:20
*** annegentle has joined #openstack-cinder12:24
*** zhenguo has joined #openstack-cinder12:29
*** zhongjun has quit IRC12:30
*** nkrinner has quit IRC12:32
*** Adriano__ has quit IRC12:32
*** Adriano__ has joined #openstack-cinder12:34
*** sgotliv_ has joined #openstack-cinder12:38
*** marcusvrn has joined #openstack-cinder12:38
*** diogogmt has joined #openstack-cinder12:39
*** agarciam1 has joined #openstack-cinder12:39
*** agarciam has quit IRC12:40
*** sgotliv has quit IRC12:40
*** dims_ has joined #openstack-cinder12:41
*** dims has quit IRC12:41
*** zhonghua-lee has quit IRC12:41
*** agarciam has joined #openstack-cinder12:43
*** porrua has joined #openstack-cinder12:44
*** agarciam1 has quit IRC12:44
*** lpetrut has quit IRC12:45
*** zhongjun has joined #openstack-cinder12:45
*** akerr has joined #openstack-cinder12:46
*** edmondsw has joined #openstack-cinder12:47
*** marcusvrn has quit IRC12:49
*** chlong has joined #openstack-cinder12:52
*** Adriano__ is now known as adrianofr12:55
*** diogogmt has quit IRC12:56
*** Yogi1 has joined #openstack-cinder13:00
*** lpetrut has joined #openstack-cinder13:00
*** dims has joined #openstack-cinder13:01
*** dims_ has quit IRC13:02
*** bswartz has joined #openstack-cinder13:02
*** haomaiwang has joined #openstack-cinder13:03
*** vervet has quit IRC13:03
*** Yogi1 has quit IRC13:04
openstackgerritSean McGinnis proposed openstack/cinder: Dell SC: Fix error causing missed log message  https://review.openstack.org/21417813:11
*** dims_ has joined #openstack-cinder13:15
*** dims has quit IRC13:17
openstackgerritJon Bernard proposed openstack/cinder: Add support for file I/O volume migration  https://review.openstack.org/18727013:19
*** changbl has joined #openstack-cinder13:19
*** dustins has joined #openstack-cinder13:20
*** bill_az has joined #openstack-cinder13:21
*** xyang1 has joined #openstack-cinder13:21
*** jgregor has joined #openstack-cinder13:22
*** mriedem has joined #openstack-cinder13:23
*** cburgess has quit IRC13:23
*** julim has joined #openstack-cinder13:24
*** dustins has quit IRC13:26
*** cburgess has joined #openstack-cinder13:26
*** dustins has joined #openstack-cinder13:26
*** simondodsley has joined #openstack-cinder13:27
*** Yogi11 has joined #openstack-cinder13:29
*** diogogmt has joined #openstack-cinder13:30
*** dims_ has quit IRC13:31
*** dims has joined #openstack-cinder13:31
*** ankit_ag has quit IRC13:38
*** JoseMello has joined #openstack-cinder13:40
*** dims has quit IRC13:40
*** eharney has quit IRC13:41
*** changbl has quit IRC13:41
*** jungleboyj has quit IRC13:42
*** dims has joined #openstack-cinder13:43
openstackgerritScott DAngelo proposed openstack/cinder: force_detach terminate_connection needs connector  https://review.openstack.org/21386713:47
*** deepakcs has quit IRC13:48
*** zhongjun has quit IRC13:48
*** bkopilov has joined #openstack-cinder13:52
*** marcusvrn_ has joined #openstack-cinder13:53
*** zhongjun has joined #openstack-cinder13:53
*** thangp has joined #openstack-cinder13:56
*** bluex-pl has quit IRC13:57
openstackgerritSzymon Wróblewski proposed openstack/cinder: Tooz locks  https://review.openstack.org/18353713:59
*** breitz has quit IRC13:59
*** breitz has joined #openstack-cinder14:00
*** tiny-hands has joined #openstack-cinder14:00
*** haomaiwang has quit IRC14:01
*** haomaiwang has joined #openstack-cinder14:02
*** DuncanT has joined #openstack-cinder14:03
*** dims has quit IRC14:04
*** dims has joined #openstack-cinder14:05
*** jecarey has joined #openstack-cinder14:11
*** esker has joined #openstack-cinder14:12
*** martyturner has joined #openstack-cinder14:13
*** IlyaG has joined #openstack-cinder14:14
*** eharney has joined #openstack-cinder14:14
*** diogogmt has quit IRC14:14
*** jungleboyj has joined #openstack-cinder14:17
*** annegentle has quit IRC14:20
openstackgerritWilson Liu proposed openstack/cinder: Enhance FC zone support for Huawei driver  https://review.openstack.org/20157814:20
*** lcurtis has joined #openstack-cinder14:22
*** afazkas has quit IRC14:27
*** zul has joined #openstack-cinder14:29
*** salv-orlando has joined #openstack-cinder14:30
*** vlaza has left #openstack-cinder14:31
*** salv-orlando has quit IRC14:34
*** theanalyst has quit IRC14:35
*** dims_ has joined #openstack-cinder14:35
*** jseiler__ has joined #openstack-cinder14:36
*** dims has quit IRC14:37
*** jseiler_ has quit IRC14:39
*** porrua has quit IRC14:44
*** annegentle has joined #openstack-cinder14:44
*** julim_ has joined #openstack-cinder14:51
*** aorourke_ has joined #openstack-cinder14:51
*** hodos has joined #openstack-cinder14:52
hodoshi guys, can anyone tell if iSCSI cinder driver/brick connector expect it to be just 1 LUN per target?14:52
hodosor it's OK to have many volumes on the same target, 1 LUN per volume?14:54
*** erlon has joined #openstack-cinder14:54
*** zul has quit IRC14:54
*** mtanino has joined #openstack-cinder14:55
*** bswartz has quit IRC14:57
smcginnishodos: That's the way our storage works. We have multiple LUNs exposed via the same target.14:58
*** alexschm has quit IRC14:58
*** martyturner has quit IRC14:59
hodossmcginnis: what's the driver class?14:59
*** lcurtis has quit IRC14:59
*** julim has quit IRC14:59
*** tobasco_ has quit IRC14:59
*** breitz has quit IRC14:59
*** skraynev has quit IRC14:59
*** aorourke has quit IRC14:59
smcginnishodos: The DellStorageCenterxxxDrivers.14:59
*** tobasco has joined #openstack-cinder14:59
hodossmcginnis: Thanks. We've tried the same approach, but hit the issue: sometimes /dev/disk/by-path does not have all the available nodes15:00
*** haomaiwang has quit IRC15:01
smcginnishodos: Hm, don't think I've seen that.15:01
hodossmcginnis: I've tried to walk through what brick does manually, with no luck15:01
*** dims has joined #openstack-cinder15:01
smcginnishodos: Maybe once Swanson is on he might have some input. He's done much more hands-on testing with our driver.15:01
*** haomaiwang has joined #openstack-cinder15:02
hodossmcginnis: thanks, ok. I'll try to figure this out while15:02
*** dims_ has quit IRC15:03
*** julim has joined #openstack-cinder15:04
*** ronis_ has quit IRC15:04
*** martyturner has joined #openstack-cinder15:05
*** breitz has joined #openstack-cinder15:05
*** julim_ has quit IRC15:06
*** lcurtis has joined #openstack-cinder15:07
jgriffithhodos: as smcginnis pointed out you just modify the lun field in the model info15:07
jgriffithsmcginnis: I had a question for you....15:07
jgriffithsmcginnis: but I can't remember what it was :(15:07
jgriffithsmcginnis: it was VERY important, the future of Cinder was quite possible determined by it.....15:07
jordanPjgriffith, could have a look at https://review.openstack.org/#/c/206935/ It's the backport of "Add deactivate step to extend_lv". I know you co-authored it but as you are not the main author, maybe you can still +A it :)15:08
jordanP(bend the rules... :p)15:09
jgriffithjordanP: looking15:09
*** zul has joined #openstack-cinder15:09
*** crose has joined #openstack-cinder15:10
jgriffithjordanP: in the *good old days* I used to submit to master, approve, back port and approve all in the same morning :)15:11
jgriffithjordanP: all my own patches :)15:12
*** skraynev has joined #openstack-cinder15:12
jordanPjgriffith, haha ! I am sure you miss this time.. I would :)15:12
jgriffithjordanP: might explain all the typos :)15:12
*** theanalyst has joined #openstack-cinder15:12
smcginnisjgriffith: Hah!15:12
smcginnisjgriffith: Was you're question around config file vs extra specs?15:13
jgriffithhemnafk: FYI, I switched my CI to pull os-brick from HEAD and that patch fixes the image-copy issue I was seeing as well it appears15:13
jgriffithhemnafk: last 50 runs were all succesful15:13
jgriffithhemnafk: In other words "we need to release that os-brick version" :)15:14
jordanPjgriffith, you used LIBS_FROM_GIT="os-brick" ?15:14
*** diogogmt has joined #openstack-cinder15:14
jgriffithjordanP: correct15:15
jgriffithjordanP: for some reason every since we split out brick I've been seeing an intermittent disconnect issue on image-copy-to-volume15:15
jgriffithjordanP: The only thing that still bothers me is *why* I never hit it before moving to the lib15:16
jgriffithjordanP: since then hemnafk proposed a patch that removes the rescan on detach15:16
jgriffithjordanP: I'm still slightly nervous.... and TBF I do have to switch back to my other system and make sure the behavior is the same there15:17
jgriffithjordanP: but it's promising thus far15:17
*** bswartz has joined #openstack-cinder15:17
*** hodos|2 has joined #openstack-cinder15:17
*** dannywilson has joined #openstack-cinder15:18
*** tsekiyama has joined #openstack-cinder15:19
*** hodos has quit IRC15:20
*** zhongjun has quit IRC15:21
*** ociuhandu has joined #openstack-cinder15:23
*** dannywilson has quit IRC15:23
*** dannywil_ has joined #openstack-cinder15:23
hodos|2jgriffith: that was exactly what i've been talking about15:24
*** dannywil_ has quit IRC15:24
jgriffithhodos|2: Oh?  That's the same problem you were seeing?15:24
jgriffithhodos|2: sounds like you, me and erlon may have all been in the same boat15:24
hodos|2jgriffith: i get device not connected on volume from image15:25
hodos|2jgriffith: or device too small15:25
*** dannywilson has joined #openstack-cinder15:25
jgriffithhodos|2: yup, sounds like the same deal15:25
hodos|2jgriffith: let me test with lib from git15:26
jgriffithdevice to small is completely different, don't know what to make of that15:26
smcginnishodos|2: That one I have seen intermittently in our CI runs.15:26
*** garthb has joined #openstack-cinder15:26
*** garthb_ has joined #openstack-cinder15:26
jgriffithhodos|2: just incase you need a reference15:26
smcginnisThe best I could diagnose it was it was some kind of network communication issue or something.15:26
smcginnisOnly happens on our iSCSI driver, not our FC one.15:27
*** garthb has quit IRC15:27
jgriffithsmcginnis: that's interesting...15:27
openstackgerritWalter A. Boring IV (hemna) proposed openstack/os-brick: Add support for --interface option in iscsiadm  https://review.openstack.org/19345115:27
hodos|2smcginnis: smcginnis: thanks; i'll report on the result15:27
smcginnisjgriffith: Yeah, never could nail it down to an exact thing.15:27
smcginnisSo it's just my gut feel.15:27
smcginnisOnce we fixed up some network settings it seemed to go away, though not completely.15:27
jgriffithsmcginnis: do you have the message?  I'm trying to find it in the code15:28
jgriffithsmcginnis: I had assumed it's the emu-convert but won't to verify15:28
smcginnisjgriffith: No, it's been a while.15:28
jgriffithsmcginnis: hodos|2 in other words that one probably doesn't have anything to do with your device15:28
smcginnisjgriffith: Assumed it was some kind of network error that caused a SCSI command to fail, then causing it to seem like an out of space condition or something.15:28
jgriffithsmcginnis: hodos|2 you may be running out of space on your Instance to do the conversion15:28
smcginnisjgriffith: Agree15:28
*** dims has quit IRC15:29
smcginnisjgriffith: That's what I initially thought, but I did verify I had plenty of space.15:29
smcginnisAdded a lot of debug code to print out info and it had room.15:29
jgriffithsmcginnis: hodos|2 I believe that's where I wrote that message... and the dd running out of space gives a clear stack-trace IIRC15:29
jgriffithweird, well if anybody sees that again feel free to shoot me a log15:29
*** dims has joined #openstack-cinder15:30
smcginnisjgriffith: I'll watch for it. (i.e. Swanson will watch for it and tell me)15:30
Swansonum, what?15:30
smcginnisSwanson: Just signing you up for things. :)15:30
jgriffithSwanson: smcginnis just signed you up to rewrite all of the Cinder drivers in golang15:30
smcginnisHAH! :)15:31
smcginnisjgriffith: You've joked about that enough now that I'm starting to think you're not entirely joking. ;)15:31
*** p0rtal has joined #openstack-cinder15:31
*** gouthamr has joined #openstack-cinder15:31
jgriffithsmcginnis: haha!  Nah... I just have it on the brain because the other project I'm working on is all golang15:31
jgriffithsmcginnis: and the swift stuff in go just makes it funny to me15:32
jgriffithI swear I'm not suggesting it :)15:32
smcginnisCool. Always fun to get a little more experience in another language.15:32
jgriffithsmcginnis: indeed... and I have to admit that I really kinda like it as a lang15:33
jgriffithsmcginnis: it's good for crusty old C folks like myself :)15:33
*** salv-orlando has joined #openstack-cinder15:33
jgriffithsmcginnis: but it does make you appreciate python!!!15:33
*** DuncanT has quit IRC15:36
*** marcusvrn_ has quit IRC15:36
*** erlon has quit IRC15:36
*** shyama has quit IRC15:39
*** jecarey has quit IRC15:43
*** sayali has quit IRC15:45
*** dims has quit IRC15:45
openstackgerritThang Pham proposed openstack/cinder: Continue conversion to volume object  https://review.openstack.org/20140415:45
openstackgerritThang Pham proposed openstack/cinder: Conversion to volume object  https://review.openstack.org/17705415:45
*** zul has quit IRC15:46
*** dims has joined #openstack-cinder15:47
*** lpetrut has quit IRC15:47
gouthamrHi, could I have one more core-reviewer review this patch: https://review.openstack.org/#/c/211659/ - thanks for the CR, thingee15:49
*** erlon has joined #openstack-cinder15:49
*** shyama has joined #openstack-cinder15:50
*** p0rtal has quit IRC15:50
*** p0rtal has joined #openstack-cinder15:50
*** afazkas has joined #openstack-cinder15:51
*** dflorea has joined #openstack-cinder15:53
*** _cjones_ has joined #openstack-cinder15:53
*** p0rtal has quit IRC15:55
*** hemna has joined #openstack-cinder15:56
*** ccesario_ has quit IRC15:57
*** eharney has quit IRC15:58
*** Apoorva has joined #openstack-cinder15:58
*** marcusvrn_ has joined #openstack-cinder15:59
*** eharney has joined #openstack-cinder15:59
*** Apoorva has quit IRC15:59
*** ccesario_ has joined #openstack-cinder16:00
*** martyturner has quit IRC16:00
*** haomaiwang has quit IRC16:01
*** Apoorva has joined #openstack-cinder16:01
*** haomaiwang has joined #openstack-cinder16:01
*** porrua has joined #openstack-cinder16:01
*** Yogi11 has quit IRC16:02
openstackgerritGoutham Pacha Ravi proposed openstack/cinder: Configure space reservation on NetApp Data ONTAP  https://review.openstack.org/21165916:02
*** sayali has joined #openstack-cinder16:03
*** bswartz has quit IRC16:04
*** haomaiwang has quit IRC16:04
*** bswartz has joined #openstack-cinder16:06
*** jistr has quit IRC16:08
*** BharatK has quit IRC16:10
jordanPthingee, hi. I've fixed my CI issue in https://review.openstack.org/#/c/200610/ (Rework Scality SOFS driver to use RemoteFS class), if you have a chance to have a look again, that would be great. thanks16:12
*** DuncanT has joined #openstack-cinder16:12
*** vilobhmm has joined #openstack-cinder16:16
*** skylerberg has joined #openstack-cinder16:22
eharneyjungleboyj: so about https://review.openstack.org/#/c/214191/ ...  i think in the past we've considered it better to backport a harmless patch as a dependency to get a clean backport in two patches rather than manually resolving conflicts in the patch fixing the bug?16:22
*** leeantho has joined #openstack-cinder16:25
dannywilsonjgriffith: ping, got a replication question16:26
openstackgerritScott DAngelo proposed openstack/cinder: force_detach terminate_connection needs connector  https://review.openstack.org/21386716:27
*** Yogi1 has joined #openstack-cinder16:28
*** agarciam has quit IRC16:30
openstackgerritPatrick East proposed openstack/cinder: Implement manage/unmanage snapshot in Pure drivers  https://review.openstack.org/21385516:31
*** akerr is now known as akerr_away16:32
*** esker has quit IRC16:34
jgriffithdannywilson: sure, what's up?16:35
openstackgerritWilson Liu proposed openstack/cinder: Enhance FC zone support for Huawei driver  https://review.openstack.org/20157816:35
*** dims has quit IRC16:36
*** mylu has joined #openstack-cinder16:37
*** IanGovett1 has joined #openstack-cinder16:38
*** dims has joined #openstack-cinder16:39
*** shausy has quit IRC16:39
dannywilsonjgriffith: post replication_failover should we be marking any snapshots that are on original host as not available?16:40
*** mylu has quit IRC16:40
dannywilsonjgriffith: snapshots from the failed over volume of course16:40
*** IanGovett has quit IRC16:40
*** mylu has joined #openstack-cinder16:41
*** annegentle has quit IRC16:41
*** diogogmt_ has joined #openstack-cinder16:41
*** yrabl has quit IRC16:42
*** afazkas has quit IRC16:42
*** krtaylor has quit IRC16:43
jgriffithdannywilson: so that's a good question... it depends on whether they're actually available or not16:43
*** akerr_away is now known as akerr16:43
*** diogogmt has quit IRC16:43
*** diogogmt_ is now known as diogogmt16:43
jgriffithdannywilson: in my case I've treated it as "if the backend is actually available and I can get to them... great"16:43
*** ronis_ has joined #openstack-cinder16:44
jgriffithdannywilson: if it's a "true" failure, then they're gone16:44
jgriffithdannywilson: make sense?16:44
dannywilsonjgriffith: almost16:44
dannywilsonjgriffith: do they get marked as not available at any point?  or just fail because you can't reach the array?16:44
*** mylu has quit IRC16:45
*** mylu has joined #openstack-cinder16:45
jgriffithdannywilson: fail because you can't reach the array16:45
jgriffithdannywilson: cinder does it for you16:46
jgriffithdannywilson: because the host is marked as 'down'16:46
jgriffithdannywilson: it's kinda redundant to add something else at this point16:46
jgriffithdannywilson: it's certainly something somebody could enhance in the future though16:46
patrickeastjgriffith: i think one thing we are trying to figure out is what to do with the original volume once failed over, for us we need to keep it around to keep its snapshots alive, but if we keep it around its essentially orphaned16:47
*** martyturner has joined #openstack-cinder16:47
jgriffithpatrickeast: I'd certainly keep it around16:47
patrickeastjgriffith: in the case where it isn't a true failure, but like a test failover16:47
jgriffithpatrickeast: replication fail over doesn't mean you delete anything IMHO16:47
jgriffithpatrickeast: it just provides access, that's all16:48
openstackgerritVipin Balachandran proposed openstack/cinder: Rename free_virtual in capacity filter  https://review.openstack.org/21427616:48
patrickeastjgriffith: true, i guess its up to the admin then to go clean that up later manually16:48
jgriffithpatrickeast: in other words... "I can't get to it from path A... no big deal, go to path B"16:48
jgriffithpatrickeast: now doing all the directional stuff is your problem :)16:48
dannywilsonjgriffith: does this implementation consider failover, enable_replication, failover again to put volume back on original backend?  or is that for later?16:49
jgriffithdannywilson: it does, that was a requirement by DuncanT16:49
jgriffithdannywilson: failover can just be called multiple times to toggle the backend16:49
patrickeastdannywilson: hmm so that case we would leave behind a orphan volume on each failover XD16:50
jgriffithpatrickeast: so you can't just switch primary/secondary back and forth?16:50
patrickeastjgriffith: pure does snapshot based replication, so the 'replicated' volume is actually just snapshots up until we failover and create a volume from the latest one16:50
patrickeastjgriffith: when we go back we could delete the original and re-make it from the replicated snapshot16:51
patrickeastso maybe not too big of a deal16:51
patrickeastjust one potential orphan on each array16:51
jgriffithpatrickeast: yeah, this whole replication thing is just kind of a pain :(16:51
patrickeasthaha truth16:51
patrickeasti think we can make this work though16:52
jgriffithpatrickeast: agreed16:52
dannywilsonjgriffith: more than "kind of" a pain, but necessary :)16:52
jgriffithdannywilson: :)16:52
jgriffithdannywilson: patrickeast so honestly I'm not set on the idea of toggling back and forth without some admin intervention16:53
jgriffithdannywilson: patrickeast that's pretty ambitious IMO16:53
*** dims has quit IRC16:54
*** krtaylor has joined #openstack-cinder16:56
*** dims has joined #openstack-cinder16:57
*** sgotliv_ has quit IRC16:59
*** martyturner has quit IRC17:00
*** martyturner has joined #openstack-cinder17:01
*** Yogi11 has joined #openstack-cinder17:02
*** jordanP has quit IRC17:02
*** dims_ has joined #openstack-cinder17:02
*** dflorea has quit IRC17:02
openstackgerritYuriy Nesenenko proposed openstack/python-cinderclient: Add support '--all-tenants' for cinder backup-list  https://review.openstack.org/20746917:03
*** dims has quit IRC17:04
*** Yogi1 has quit IRC17:04
*** zhenguo has quit IRC17:05
*** markus_z has quit IRC17:07
*** p0rtal has joined #openstack-cinder17:08
*** dims has joined #openstack-cinder17:08
*** dims_ has quit IRC17:09
openstackgerritVictor Stinner proposed openstack/cinder: Port more unit tests to Python 3  https://review.openstack.org/21381717:09
*** IanGovett1 has quit IRC17:10
*** dims_ has joined #openstack-cinder17:11
mriedemcinder multi-attach is the same volume attached to multiple instances right?17:13
*** dims has quit IRC17:14
*** bswartz has quit IRC17:15
scottdamriedem: Yes, and also the same volume attached to the same instance >1 time17:15
yuriy_n17geguileo: Hi! Please review https://review.openstack.org/#/c/207469/17:15
mriedemscottda: so how does the volume get response look if you have multiple attachments? is instance_uuid a list rather than a single string?17:16
mriedemscottda: i was looking at that here https://review.openstack.org/#/c/184537/9/nova/cmd/manage.py17:16
mriedemif vol['instance_uuid'] != instance.uuid17:16
hemnamriedem: you get multiple attachments in a volume17:16
mriedemsure, but what does the response body look like from the volume GET call?17:17
* mriedem checks the v2 view builder17:17
mriedemah there is an 'attachments' key17:17
*** bswartz has joined #openstack-cinder17:18
*** ndipanov has joined #openstack-cinder17:18
mriedemisn't that kind of bogus if you have more than one?17:18
mriedemi guess you're expected to just call the volume API each time until attachments is 017:19
mriedemif you're trying to detach all instanes17:19
*** dims_ has quit IRC17:20
*** nikeshm has joined #openstack-cinder17:20
*** boris-42 has quit IRC17:20
scottdaYou mean detach all volumes?17:20
*** pv has joined #openstack-cinder17:20
*** dims has joined #openstack-cinder17:20
*** yuriy_n17 has quit IRC17:20
hemnaiI don’t think that’s bogus17:20
*** dims has quit IRC17:20
scottdaYes, I'm not sure if there is a function for that in the not-yet-landed nova patch for multi-attach17:21
*** dims has joined #openstack-cinder17:21
hemnaso the current model in the cinder volume manager is that you aren’t required to pass in an attachment_id17:22
hemnaif you don’t, it checks how many attachments there are17:22
hemnaif there is 1, it works17:22
hemnaif there is >1, it raises17:22
hemnait was unclear what was the ‘safest’ thing to do17:22
mriedemthat seems right17:23
mriedemfor backwards compat17:23
hemnaso I pushed to make this safe and raise instead of detaching all17:23
mriedemif yo'ure doing multiattach cases then you need to opt into detaching a specific thing17:23
mriedemif the caller wants to do all, they have to iterate that on their side i guess17:23
hemnaso if you want to detach all, currently you need to loop over attachments and call detach on each.17:23
mriedemunless you had like a special case attachment_id == 'all' case or something17:24
hemnaprobably not a bad idea to make a change to detach all17:24
hemnajust not sure what the ‘right’ API change is to support that17:24
scottdaWe talked about an admin action to detach_all for help with cleanup when we don't have a Nova instance or attachement17:24
hemnadetach_all ?17:24
hemnaor add an additional argument to detach, all=False17:24
*** Yogi11 has quit IRC17:25
pvim getting a volume-attach error in my nova compute logs at  File "/usr/local/lib/python2.7/dist-packages/cinderclient/v2/volumes.py", line 448, in initialize_connection {'connector': connector})[1]['connection_info']17:25
pvthat point17:25
pvshould i ask on nova irc or here17:25
hemnascottda: that was slightly different though, as it was for force detach, but w/o connectors :(17:25
*** ebalduf has joined #openstack-cinder17:25
*** Yogi1 has joined #openstack-cinder17:25
mriedempv: you're going to need a paste of the full stacktrace17:25
scottdaso, if the Nova instance is gone, but Cinder still thinks it's attached, we'd have something to terminate_connection for the volume in general, without a specific attachement_id17:25
scottdahemna: right17:25
mriedempv: and level of code you're on, environment, backends used, etc etc17:26
mriedemno dead cats17:26
hemnascottda: we should still think about updating the attachments table to store the connector dict17:26
scottdaYes, indeed17:26
hemnathen it would be possible to do a force detach where the nova instance is gone17:26
scottdaI'll add that to the list of Mitaka stuff17:27
pvim running devstack on an ubuntu server testing my driver's initialize/terminate connection code17:27
hemnascottda: ok.  I can definitely help with that17:28
pvits an FC driver and FC enabled machine17:28
hemnapv, that’s the first error?17:29
hemnapv did nova complain about not finding the volume?17:29
*** IanGovett has joined #openstack-cinder17:29
hemnapv: I have a feeling that nova complained before that trace17:29
pvyeah from what im seeing on the logs17:30
pvlet me check the overall logfile17:30
pvone sec17:30
pvthe first error in the log is 2015-08-18 10:10:12.088 ERROR nova.compute.manager [req-17165518-6dcd-446c-9aab-53bd8fdc7aaa None None] No compute node record for host cli17:31
hemnamriedem: scottda, so this addition to detach seems like a good first change after microversions land17:31
pvthen its the failed to attach volume error17:31
hemnapv: check the cinder log around the same time17:32
hemnapv looks like something puked in Cinder’s call to initialize_connection17:33
mriedemhemna: scottda: sure, i'm not actually even asking for this, just asking questions about how it works today17:33
hemnamriedem: ok np17:33
hemnathe nova multiattach code still hasn’t landed17:33
scottdahemna: Yes Microversions Will Fix All®17:33
hemnaso no one is really using multiattach in Cinder17:33
hemnaso I think we can make changes as needed in M17:33
hemnascottda: who is going to do it :)17:34
*** martyturner has quit IRC17:34
hemnascottda: thingee knows the API better than anyone afaik17:34
pvERROR cinder.volume.manager [req-0ff37083-9611-483c-b5a6-9bf7aeb33b0a admin] [volume-dc88faa0-b710-4c5a-baaf-d8f2a450efe9] Driver initialize connection failed (error: expected string or buffer).17:34
pvcool theres the issue17:34
scottdaI'm going to write the spec, just as soon as I'm done with Nova-Cinder API spec17:34
pvthanks a lot17:34
hemnapv np17:34
*** martyturner has joined #openstack-cinder17:35
scottdaThen I'll see if I can s̶t̶e̶a̶l̶ leverage code from Manila or Ironic17:35
*** dflorea has joined #openstack-cinder17:38
hemnascottda: +117:38
*** p0rtal has quit IRC17:40
*** angela-s has joined #openstack-cinder17:40
erlonhemna: ping17:40
hemnaerlon: hey17:41
*** p0rtal has joined #openstack-cinder17:42
*** breitz has quit IRC17:42
*** dflorea has quit IRC17:42
erlonhemna: hey, so, I'm testing that patch(that removes the scan on disconnect) in a multinode setup, and the devices still are note removed from the hosts17:42
*** dflorea has joined #openstack-cinder17:42
skylerbergI saw on the etherpad that third party CI systems will be expected to run against changes to os-brick. Is there more information on this somewhere?17:43
hemnaskylerberg: for folks that have custom connectors in os-brick yes17:43
erlonhemna: do you have any hint why?17:43
hemnaerlon: so what kind of use case causes this?17:44
skylerberghemna: Thanks. I misread the etherpad and thought "all drivers" refered to Cinder drivers.17:45
hemnaskylerberg: since os-brick does the volume detection and removal, if you want to make sure that any changes to os-brick doesn’t break your backend, then it’s a good idea to add your CI for your array against os-brick patches17:45
erlonhemna: create a VM, boot from image(creates a new volume), with instance count >= 417:45
hemnaskylerberg: that’s the upside to having your CI added to os-brick patches.17:45
*** changbl has joined #openstack-cinder17:46
hemnaerlon: the volume is a n-cpu lvm volume though no?17:46
erlonhemna: ??17:47
hemnaboot from image isn’t a cinder volume afaik17:48
*** JoseMello has quit IRC17:49
erlonhemna: yes if you select 'boot from image(creates a new volume)'17:49
hemnaerlon: afaik that volume is an lvm volume from the nova compute host itself, not cinder17:50
*** breitz has joined #openstack-cinder17:52
*** ociuhandu has quit IRC17:52
erlonhemna: I don't think so, I've been using this to test some features on HNAS, and if the BE is configured as default_volume_type, nova will boot from the volumes on the backend17:53
erlonhemna: I actually not even have LVM configured in my setup17:53
hemnajungleboyj: yes17:53
*** simondodsley has quit IRC17:53
hemnaerlon: LVM on the compute host.17:53
jungleboyjhemna: The boot from volume creates a volume in whatever storage you have configured, mounts it on the control node, copies the inmage and then attaches it to the compute node for boot.17:54
hemnaboot from volume yes17:54
hemnaboot from image ?17:54
jungleboyjOh, Sorry.17:54
hemnaboot from volume, boots from a pre-existing volume on Cinder17:54
jungleboyjhemna: Ok, so in that case, yes the image is put in storage on the Compute node.  Where ever images are being held for the hypervisor.17:55
erlonhemna: yes, there's the 3 options in horizon, boot from volume, boot from image and boot from image creating a new volume17:55
*** garthb_ has quit IRC17:55
hemnayou can do a boot from snapshot17:57
hemnawhicih creates a new cinder volume from the snapshot17:57
hemnabut boot from image is a local nova compute lvm volume17:58
hemnaunless I’m drunk and have lost my mind17:58
*** dflorea has quit IRC17:59
*** dflorea has joined #openstack-cinder17:59
jungleboyjeharney: Updated my vote on those backports.  You are right, doing the two backports makes more sense.18:00
erlonhemna: https://www.dropbox.com/s/ri2bjlme9fsw2eu/horizon.png?dl=018:01
eharneyjungleboyj: thanks :)18:01
hemnaerlon: k that’s a horizon thing.  not sure what it does actually18:02
hemnaI’m looking at the nova api18:02
hemnacould be a wrapper to call cinder first18:02
jungleboyjeharney: Welcome.  I was spinning back and forth between those this morning.  In the end, you are right.18:03
*** mylu has quit IRC18:03
erlonhemna: I don't know what happens under the hoods in nova, but in cinder it creates a volume (with blank name), attach on volume node, copy the image from glance, and then disconnect the volume18:05
erlonhemna: only after that nova proceeds booting18:05
hemnaok, that must be a wrapper in horizon then to call cinder first to create volume from image18:06
*** chlong has quit IRC18:06
hemnado you get a choice for volume type?18:06
hemnajust get a random volume backend ?18:06
erlonhemna: no, it always use the default_volume_type18:06
hemnaick ok18:06
*** IanGovett1 has joined #openstack-cinder18:07
hemnaso you boot a vm, then destroy it and you have orphaned volumes on the compute host ?18:07
jgriffithhemna: erlon yes18:07
jgriffithhemna: erlon horizon provides a boot from volume wrapper18:08
jgriffithhemna: erlon it does NOT provide a method to specify type (which is rtes-lame)18:08
hemnayah that’s lame stuff right there18:08
jgriffithhemna: erlon it always uses the configured default type or "no" type18:08
jgriffithhemna: it's been there for 3 or 4 releases :(18:08
jgriffithhemna: erlon honestly I suggest to my customers to NOT use it18:09
hemnaerlon: ok I’ll test this out on my backend tomorrow.  I’m WAH today and can’t really test it at the moment.18:09
*** chlong has joined #openstack-cinder18:09
*** IanGovett has quit IRC18:09
david-lylejgriffith: hemna, so would the suggestion be to drop in the create volume workflow there, or another selector for type enough?18:10
jgriffithdavid-lyle: type selector would solve it18:10
erlonhemna: great, Ill keep trying to understand what happens18:10
jgriffithdavid-lyle: i tried to hack it in there myself but screwed everything up :)18:10
jgriffithdavid-lyle: I think you and I talked about this in FTC last summer... not sure18:11
*** openstackgerrit_ has quit IRC18:11
david-lylemost likely, and I probably spaced it, :(18:11
jgriffithdavid-lyle: just adding the type selector would make the functionality identical18:11
jgriffithdavid-lyle: me too LOL18:11
david-lyleok, I can take a look at that18:11
jgriffithdavid-lyle: there is one other problem though.....18:11
jgriffithdavid-lyle: the timing in all the wrapped rpc calls18:11
jgriffithdavid-lyle: I can't remember exactly who's call ends up owning it, but it times out frequently for larger images/volumes18:12
*** salv-orlando has quit IRC18:12
hemnaick yah large volumes can take a while to dd from the image :(18:12
jgriffithdavid-lyle: because there's a parent call that own the timeout, and it's not long enough for the horribly long download/convert/copy-to-volume/boot process18:12
jgriffithdavid-lyle: I couldn't quite figure out how to mess with that18:13
david-lylejgriffith: so even with the type selector it would only work for small images?18:13
jgriffithdavid-lyle: but I'm happy to help out later this week or next if you like18:13
david-lyleunless we tackle the second part?18:13
jgriffithdavid-lyle: yeah... there's a time issue for larger issues.  At least last I checked18:13
jgriffithdavid-lyle: in other words "yes, you're statement is correct"18:13
david-lyleI doubt we did anything to fix it18:14
jgriffithNever know :)18:14
david-lyleok, I can start on the selector and then work on timing issues18:14
jgriffiththat should be "there's a time issue for larger images"18:14
jgriffithdavid-lyle: thanks!!!18:14
david-lyleincremental progress at least18:14
jgriffithdavid-lyle: let me know if I can help18:14
jgriffithdavid-lyle: yeah, that alone would be awesome18:15
*** IanGovett has joined #openstack-cinder18:15
david-lylejgriffith: will do. I'll finish up the current bug fix first :)18:15
david-lylewhy are there so many?18:15
*** dflorea has quit IRC18:16
jgriffithdavid-lyle: hemna erlon just to verify we're all on the same page:   https://www.dropbox.com/s/shugdv640lj553l/Screenshot%202015-08-18%2012.16.00.png?dl=018:16
jgriffith"Boot form image (creates a new volume)" is what I'm referring to18:17
*** IanGovett1 has quit IRC18:17
openstackgerritAngela Smith proposed openstack/cinder: Adds VF support with HTTPS southbound connector  https://review.openstack.org/18479918:17
david-lyleis there a bug already, or should I open one18:17
erlonjgriffith: yes18:17
jgriffithdavid-lyle: I think I logged one at one point... lemme see if I can find it18:17
erlonjgriffith: that's the way I found to simulate the 'very busy' systems situation18:18
jgriffithdavid-lyle: for me and LP it's easier to just log a new one LOL18:18
david-lylejgriffith: that's an LP feature18:18
jgriffithdavid-lyle: haha... they implemented it VERY well18:18
jgriffithdavid-lyle: doesn't look like I ever logged one18:19
jgriffithdavid-lyle: you want me to log it?18:19
david-lylejgriffith: that would help18:20
*** dannywilson has quit IRC18:20
david-lylejgriffith: and I'll pick it up18:20
*** tim_o_ has joined #openstack-cinder18:22
erlondavid-lyle: is it possible to set a timeout based on the size of the image? I mean, not a fixed timeout, but something like, image_mb_timeout, that will leead to an actual 'image.size() * image_mb_timeout' timeout18:23
*** IanGovett has quit IRC18:24
david-lyleerlon: I have no idea, if the APIs support passing it in, the calculation could be done in Horizon18:25
*** IanGovett has joined #openstack-cinder18:25
*** dannywilson has joined #openstack-cinder18:25
jgriffitherlon: that would be super tricky.  We just need to make the create volume the parent.. but we can tackle that one later :)18:25
erlonjgriffith: mhm18:26
erlonjgriffith: I don't  know may be that would just be a couple lines18:27
*** dflorea has joined #openstack-cinder18:27
*** IanGovett1 has joined #openstack-cinder18:28
jgriffithdavid-lyle: https://bugs.launchpad.net/horizon/+bug/148617818:29
openstackLaunchpad bug 1486178 in OpenStack Dashboard (Horizon) "Boot from image (creates a new volume) Doesn't allow specification of volume-type" [Undecided,New]18:29
jgriffitherlon: ok, go for it :)18:29
jgriffitherlon: you might have a great idea that I haven't thought of18:29
*** martyturner has quit IRC18:29
david-lylejgriffith: on it. Thanks!18:29
jgriffithdavid-lyle: thank you sir!18:29
jgriffithdavid-lyle: I'll dig into the timing thing later and file something there if I get more info/details18:30
jgriffithdavid-lyle: or maybe erlon has a cool proposal to deal with it18:30
*** dflorea_ has joined #openstack-cinder18:30
erlonjgriffith: :)18:30
*** IanGovett has quit IRC18:30
david-lylejgriffith: let me know what I can do to help on that one.18:31
tim_o_asselin, i  have a couple of follow up questions can we chat?18:31
erlonjgriffith: have you seen what I told to Walter?18:31
jgriffithdavid-lyle: thanks man!18:31
jgriffitherlon: I'll scroll back....18:31
jgriffitherlon: you told him lots of things it looks like :)18:31
erlonjgriffith: the fix merged into brick yesterday is not 'fixing' in my multinode setup18:32
jgriffitherlon: oh... which problem are you seeing?  Logs?18:32
erlonjgriffith: there are still lost /dev/sdx lost18:32
*** martyturner has joined #openstack-cinder18:33
jgriffithCan somebody please explain to me HOW I can have an Iron Maiden radio channel on Spotify and it plays the firkin Beatles every once in a while!!  WTF?18:33
tim_o_asselin, got pulled into a meeting -- will ping you later18:33
jgriffithI mean... I HATE the Beatles, and they're not even related!!!18:33
erlonno, I create a VM, booting from image + creating the image, when I remove then, at the same time, the block devices are not removed18:33
*** dflorea has quit IRC18:33
*** dflorea has joined #openstack-cinder18:34
jgriffitherlon: not sure I follow exactly; but I think that's very different than what I was worried about18:34
jgriffitherlon: which block devices?  nova's?18:34
jgriffitherlon: there's the attach and then the mapped block device on the Nova side AIR18:35
erlonhmm, yes, and the one used on c-volume to copy from the image18:35
jgriffitherlon: I had to filter things on LVM conf for this very problem18:35
*** skiafa has joined #openstack-cinder18:35
erlonjgriffith:  filter?18:36
hemnajgriffith: I get worse with Pandora.   I fire up a Queens of the Stone Age channel and get Dave Mathews Band.   *sigh*18:36
pvspotify radio is pretty on point18:36
hemnayah it’s pretty good.18:36
hemnaPandora is junk IMHO18:36
*** dflorea_ has quit IRC18:36
erlonjgriffith: try to get back to last.fm :)18:37
jgriffitherlon: jeje18:37
*** tim_o_ has quit IRC18:37
skiafahi folks, I was wondering if someone is working on the manage/unmanage volume tests... does anybody know?18:38
jgriffithskiafa: not that I know of18:38
erlonskiafa: we have just talked about that today18:39
skiafatks j. yep... I saw that... rs so this is a tempest project specific or is a cinder specific?18:40
erlonskiafa: I don't know as well, would be nice to have that18:40
openstackgerritErickson Filipe Guedes dos Santos proposed openstack/cinder: Nested Quota Driver: Get Project Hierarchy  https://review.openstack.org/20617118:40
openstackgerritErickson Filipe Guedes dos Santos proposed openstack/cinder: Cinder Nested Quota Driver  https://review.openstack.org/20536918:40
erlonskiafa: tempest18:41
*** shyama has quit IRC18:43
skiafaerlon: cool... I´ll look into it then... tks18:43
*** mylu has joined #openstack-cinder18:46
*** dflorea has quit IRC18:50
*** dflorea has joined #openstack-cinder18:53
*** mylu has quit IRC18:53
*** mylu has joined #openstack-cinder18:54
adrianofrhemna: Hi... we addressed your comment in https://review.openstack.org/#/c/197999/18:54
adrianofrhemna: it would be nice if you could take a look on this18:54
*** dflorea has quit IRC18:55
*** mylu_ has joined #openstack-cinder18:55
*** mylu has quit IRC18:56
*** ociuhandu has joined #openstack-cinder18:56
hemnaadrianofr: ok18:57
adrianofrhemna: tks :)18:57
*** mylu_ has quit IRC18:58
*** mylu has joined #openstack-cinder18:59
*** sgordon has quit IRC18:59
*** sgordon has joined #openstack-cinder18:59
openstackgerritJohn Griffith proposed openstack/cinder: Cinder replication V2  https://review.openstack.org/20026918:59
*** mylu has quit IRC19:01
*** mylu has joined #openstack-cinder19:01
*** Lee1092 has quit IRC19:03
*** dflorea has joined #openstack-cinder19:03
*** annegentle has joined #openstack-cinder19:06
*** bill_az has quit IRC19:06
*** e0ne has joined #openstack-cinder19:07
*** ndipanov has quit IRC19:09
*** mylu has quit IRC19:14
*** mylu has joined #openstack-cinder19:14
*** esker has joined #openstack-cinder19:15
*** salv-orl_ has joined #openstack-cinder19:16
*** martyturner has quit IRC19:16
*** annegentle has quit IRC19:21
*** skylerberg has quit IRC19:22
*** mylu has quit IRC19:23
*** mylu has joined #openstack-cinder19:23
*** simondodsley has joined #openstack-cinder19:24
*** annegentle has joined #openstack-cinder19:26
*** mylu has quit IRC19:27
*** crose has quit IRC19:30
*** skylerberg has joined #openstack-cinder19:32
*** simondodsley has quit IRC19:33
*** mylu has joined #openstack-cinder19:36
*** belmoreira has joined #openstack-cinder19:36
*** e0ne has quit IRC19:44
openstackgerritTom Swanson proposed openstack/cinder: Dell SC: Better exception handling in init_conn  https://review.openstack.org/21431619:45
openstackgerritAdriano Freires Rosso proposed openstack/cinder: Adds the randomic option to cinder retry function  https://review.openstack.org/21431719:46
hodos|2jgriffith: VolumeDeviceNotFound: Volume device not found at [u'/dev/disk/by-path/ip-']19:51
*** belmoreira has quit IRC19:51
hodos|2jgriffith: is this the error that os-brick (new one) fixes ?19:51
hodos|2jgriffith: it's happening during copy image to volume19:52
hodos|2jgriffith: in the os_brick/initiator/connector.py 41319:52
*** gouthamr has quit IRC19:56
*** annegentle has quit IRC19:57
*** Yogi1 has quit IRC19:57
*** gouthamr has joined #openstack-cinder19:58
*** Yogi1 has joined #openstack-cinder19:58
*** Yogi1 has quit IRC19:58
*** annegentle has joined #openstack-cinder20:00
jgriffithhodos|2: yes, that looks similar to what I enountered20:00
hodos|2jgriffith: hmm didn't help me with os-brick from git20:01
jgriffithhodos|2: lemme switch to my other device, maybe it doesn't really fix it for me either20:01
hodos|2jgriffith: when i do ls -la /dev/disk/by-path/ I see a stale device which is linked to /dev/mapper/....20:02
hodos|2jgriffith: when i do ls -la /dev/disk/by-path/ I see a stale device which is linked to /dev/mapper/....20:02
hodos|2jgriffith: I've read erlon's troubles looks similar...20:03
*** dflorea has quit IRC20:04
jgriffithhodos|2: there certainly seems to be something that we broke somewhere20:05
jgriffithhodos|2: race condition of some sort or another20:05
hemnahodos|2: /dev/mapper entries are usually created by multipath20:05
hemnaI’m on a call at the moment…can’t help much20:06
hemnahodos|2: have you tried the latest os-brick from github?   We recently landed a fix to remove iscsi rescan on detach, which was known to cause orphaned entries in /dev/disk/by-path20:08
hodos|2hemna: yes20:10
hemnahodos|2: did you clean up /dev/disk/by-path prior to doing a test?20:10
hemnajust to make sure those weren’t stale before the test20:11
hodos|2hemna: yes. i can spin up an instance and tell you the version. yes it was new VM20:11
hemnaif you know those aren’t valid volume paths20:12
hemnayou can nuke them20:12
*** ronis_ has quit IRC20:12
hemnaand manually issue an iscsi rescan20:12
openstackgerritTom Swanson proposed openstack/cinder: Dell SC: Better exception handling in init_conn  https://review.openstack.org/21431620:12
hemnasee if they come back20:12
hemnaif they do, then it’s because those volumes are still being exported by the backend20:12
openstackgerritJay Bryant proposed openstack/cinder: Register the volume_opts config options in remotefs.py  https://review.openstack.org/21432520:15
hodos|2hemna: ok, so what happens is that when I run multiple 'create volume from image' (which is done also in tempest suite) sometimes i get this exception that "device is not found" while  i have these stale mappings in the /dev/disk/by-path20:16
hemnahodos|2: hrmm sounds familair20:16
hemnasounds like something that erlon is seeing related to create volume from image20:16
openstackgerritThang Pham proposed openstack/cinder: Continue conversion to volume object  https://review.openstack.org/20140420:16
openstackgerritThang Pham proposed openstack/cinder: Conversion to volume object  https://review.openstack.org/17705420:16
hodos|2hemna: and because it happens only after several create/delete operations I don't really know which volume is supposed to be there20:17
hodos|2hemna: or not20:17
openstackgerritWalter A. Boring IV (hemna) proposed openstack/os-brick: WIP FC Stop calling multipath command line  https://review.openstack.org/21338920:17
*** mylu has quit IRC20:17
hemnahodos|2: so is this multiple create volume from image commands directly to cinder20:18
hemnaor is it also nova boot from image (created from volume) ?20:18
hodos|2hemna: yes20:18
*** boris-42 has joined #openstack-cinder20:19
hodos|2hemna: i think it's cinder20:19
hemnaok only cinder20:19
hodos|2hemna: I was able to reproduce without booting20:19
hemnaI’ll fire up a new vagrant and see if I can reproduce this20:20
*** gouthamr has quit IRC20:20
*** bswartz has quit IRC20:24
*** zul has joined #openstack-cinder20:28
*** dflorea has joined #openstack-cinder20:40
*** changbl has quit IRC20:40
*** mylu has joined #openstack-cinder20:41
openstackgerritEric Harney proposed openstack/cinder: Revert "mark oslo.vmware as optional dependency"  https://review.openstack.org/21434020:46
openstackgerritEric Harney proposed openstack/cinder: Revert "Enhance unit tests for zfssa drivers"  https://review.openstack.org/21434120:46
*** esker has quit IRC20:47
*** mylu has quit IRC20:50
*** IanGovett1 has quit IRC20:50
hemnahodos|2: just ran a simple test with a cirros image and it worked w/o an orphaned volume20:51
hemnahodos|2: what image are you using?20:52
hodos|2hemna: cirros20:52
hodos|2hemna: did you do just a single volume? maybe do a mix of simple volumes and volumes from images20:53
hemnayah a single volume from image20:53
hodos|2hemna: it works with single all the time20:53
erlonhemna: yesterday I tested the new patch from os-brick and it solved the problem in my onenode devstack setup20:54
hemnaerlon: ok great20:54
smcginnisAnyone know how to pull stats on gate tests? Curious as to how many of the sheepdog test runs have actually passed.20:54
smcginnisthingee: ^^20:54
erlonhemna: today I tried multinode and the problem appeared again20:55
erlonhemna: and 1 volume at time always works ok20:55
*** simondodsley has joined #openstack-cinder20:56
hemnasmcginnis: https://github.com/jogo/lastcomment20:56
*** cdelatte has quit IRC20:57
smcginnishemna: Cool, thanks!20:57
*** dustins has quit IRC20:57
hemna:) np20:57
*** dflorea has quit IRC20:57
smcginnishemna: Hmm, that's by user though. Sheepdog is done by jenkins. So need to pull out all jenkins comments for openstack/cinder with the results for sheepdog.20:58
*** xyang1 has quit IRC20:58
smcginnisWill have to play around with this and see what it shows.20:58
hemnasmcginnis: I think it supports by CI20:58
hemnaoh yah I think this is for 3rd party CI20:58
boris-42jgriffith: hemna hi there20:58
hemnaboris-42: hey20:59
*** julim has quit IRC21:00
boris-42I would like to continue work on osprofiler21:00
boris-42hemna: ^21:00
boris-42I would like to change that way how it is configured via cinder.conf21:00
*** zul has quit IRC21:00
boris-42hemna: ^ do we need deprecation and other stuff for old conf options21:01
openstackgerritScott DAngelo proposed openstack/cinder-specs: WIP draft of Cinder-Nova-API-changes  https://review.openstack.org/21178121:01
*** dflorea has joined #openstack-cinder21:01
*** krtaylor has quit IRC21:01
hemnaif you mean to change the options used, then yah, you’ll have to deprecate the existing options and continue to support them at least a release AFAIK21:02
smcginnishemna, boris-42: Would that policy apply to osprofiler?21:03
boris-42hemna: ok21:03
smcginnisSeems like we have a little more leeway with that one.21:03
boris-42smcginnis: ya it's not core functionallity21:03
smcginnisNot like it's functionality most end users are expecting to stay consistent.21:03
boris-42it can be treat as experimental functionallity21:03
hemnaprobably still not a good idea to just stop supporting them w/o deprecating in a release no?21:03
*** dflorea has quit IRC21:03
smcginnisI doubt many know about it.21:03
smcginnishemna: True.21:03
boris-42hemna: ok I will try to make via deprecation21:03
boris-42in any case it is broken now =(21:04
hemnaif it’s impossible, then we should discuss21:04
smcginnisHah, then nothing to deprecate! :P21:04
hemnamight be a good meeting topic for tomorrow ?21:04
*** simondodsley has quit IRC21:07
boris-42hemna: I am going to move this in mailing list21:07
boris-42hemna: it's not only about Cinder21:07
hemnaboris-42: ok coolio21:07
hodos|2hemna: so you just tried 1 volume?21:09
hemnahodos|2: yah21:09
*** raildo is now known as raildo-afk21:09
*** krtaylor has joined #openstack-cinder21:14
openstackgerritAngela Smith proposed openstack/cinder: Adds VF support with HTTPS southbound connector  https://review.openstack.org/18479921:15
eharneyjgriffith: i'd rather keep it, but keeping it requires tox 1.7, thingee says we can't move to tox 1.7 -- i'd also like to be able to actually run tox :)21:19
eharneyjgriffith: nobody said previously that we couldn't move to tox 1.721:20
jgriffitheharney: haha... what a mess21:20
jgriffitheharney: ok, so I'll approve your revert then21:20
eharneyjgriffith: i'm not really sure why we can't, but, regardless, we need something that at least runs :)21:20
jgriffitheharney: I didn't know you were around or would've just pinged you21:20
smcginnisIs there a reason we can't move to 1.7?21:20
eharneyi think at least one other project is already on 1.7, so, i don't really think so21:20
jgriffithsmcginnis: that's a great question :)21:21
jgriffitheharney: yes, neutron I believe is already there21:21
jgriffithso it's being used anyway21:21
smcginnisSo I wonder if this is just a misconception that it needs to be a cross-project lock step move.21:21
*** zul has joined #openstack-cinder21:22
openstackgerritEric Harney proposed openstack/cinder: Test whether sample config generation works  https://review.openstack.org/21435621:22
eharneysmcginnis: maybe, but either way, currently tests are broken on a significant number of setups, so, we need to land something to fix that if we can't decide about upgrading21:22
smcginniseharney: I suppose we can always revisit it later.21:23
eharneysmcginnis: the optional dependency thing is something we can live without for the moment, as nice as it would be... we've lived without it for a while already21:23
*** thangp has quit IRC21:23
smcginnisVery true21:23
eharneyyet other stuff related to tests in that patch ^21:23
smcginniseharney: check_uptodate.sh: can not find config file21:24
eharneysmcginnis: i figured something would blow up there... so many different envs to check against :)21:25
smcginnisfrom ~/.../OpenStack/cinder$ ./run_tests.sh -p21:25
jgriffithI honestly don't even know what to think any more... I should probably just give up21:25
eharneysmcginnis: err, well, the point is that the current tree is broken21:25
eharneysmcginnis: pull Jay's patch too if you want to see it succeed :)21:25
eharneysmcginnis: wait... no, i have no idea about that one21:26
smcginniseharney: Yours is dependent on his, so it should have his changes.21:26
eharneysmcginnis: i see what i did wrong there... -p doesn't generate the config file, but it asks to check it21:27
eharneysmcginnis: will update in a sec21:27
smcginniseharney: Awesome21:29
smcginniseharney: Bonus points from me for even remembering to update run_tests.sh :)21:29
eharneysmcginnis: i just need to move more of the old code into the "else" section21:29
openstackgerritEric Harney proposed openstack/cinder: Test whether sample config generation works  https://review.openstack.org/21435621:31
*** garthb has joined #openstack-cinder21:34
*** cbader has quit IRC21:35
*** tiny-hands has quit IRC21:36
smcginniseharney: Much better.21:36
*** porrua has quit IRC21:39
*** akerr has quit IRC21:40
*** sileht has quit IRC21:43
*** edmondsw has quit IRC21:44
SwansonIf anyone with a +2 hammer could take a look at this https://review.openstack.org/#/c/214316/ it would be great.  Minor exception handling change to give me a chance to find a failure.21:45
openstackgerritYuriy Nesenenko proposed openstack/python-cinderclient: Add support '--all-tenants' for cinder backup-list  https://review.openstack.org/20746921:45
jungleboyjeharney: Can we add that in for those who still use run_tests.sh as well?21:47
eharneyjungleboyj: it is21:47
eharney(i use run_tests.sh actually)21:47
smcginnisjungleboyj: That's what I tested with. Works good now.21:47
*** mriedem has quit IRC21:48
jungleboyjHmmm, I tried it and it failed for tox but not run_tests.21:48
jungleboyjLet me try again.21:48
*** zul has quit IRC21:49
jungleboyjAh, I see it now.  It was because I had another pep8 fail with run_tests.sh that it didn't continue to run your code.  All good.21:50
*** jgregor has quit IRC21:51
*** jgregor has joined #openstack-cinder21:51
*** garthb has quit IRC21:55
*** jgregor has quit IRC21:55
*** dannywil_ has joined #openstack-cinder21:57
*** dannywil_ has quit IRC21:58
*** dannywil_ has joined #openstack-cinder21:58
*** martyturner has joined #openstack-cinder22:00
*** dannywilson has quit IRC22:01
*** dannywil_ has quit IRC22:02
*** garthb has joined #openstack-cinder22:03
*** garthb has quit IRC22:03
*** garthb has joined #openstack-cinder22:03
*** jgregor has joined #openstack-cinder22:04
*** jungleboyj has quit IRC22:05
openstackgerritTom Swanson proposed openstack/cinder: Dell SC: Better exception handling in init_conn  https://review.openstack.org/21431622:07
*** andreykurilin__ has joined #openstack-cinder22:08
*** gouthamr has joined #openstack-cinder22:14
openstackgerritEric Harney proposed openstack/cinder: Tests: Fix zfssa TestRestClientURL.test_request  https://review.openstack.org/21436822:16
*** annegentle has quit IRC22:17
hodos|2jgriffith: hemna: you were right guys, os-brick fix helps. just for some reason LIBS_FROM_GIT=os-brick didn't install latest22:20
hemnahodos|2: that’s good to know. :)22:21
hemnahodos|2: did you get /opt/stack/os-brick ?22:21
*** jgregor has quit IRC22:22
hodos|2hemna: yes it was there22:22
*** bswartz has joined #openstack-cinder22:22
hodos|2hemna: but when i checked the code the fix wasn't rhe22:22
hodos|2hemna: so i installed it from https://github.com/openstack/os-brick22:23
smcginnisDoes os-brick need to be in ENABLED_SERVICES/22:27
SwansonIt isn't a service, is it?22:29
hemnasmcginnis: no, it’s in nova and cinder’s requirements.txt22:29
*** dflorea has joined #openstack-cinder22:29
*** niknakpaddywak has joined #openstack-cinder22:35
openstackgerritVilobh Meshram proposed openstack/cinder: Cinder Nested Quota Driver  https://review.openstack.org/20536922:38
openstackgerritJohn Griffith proposed openstack/cinder: Cinder replication V2  https://review.openstack.org/20026922:45
*** hodos|2 has quit IRC22:45
*** cbader has joined #openstack-cinder22:47
*** eharney has quit IRC22:47
*** cbader has quit IRC22:53
openstackgerritVictor Stinner proposed openstack/cinder: Port image_utils to Python 3  https://review.openstack.org/21381722:55
*** jungleboyj has joined #openstack-cinder22:56
*** angela-s has quit IRC22:57
*** dflorea has quit IRC23:00
*** dflorea has joined #openstack-cinder23:01
*** cbader has joined #openstack-cinder23:02
hemnapatrickeast: ping23:06
*** annegentle has joined #openstack-cinder23:08
*** annegentle has quit IRC23:09
openstackgerritJohn Griffith proposed openstack/cinder: Update devref unit tests doc  https://review.openstack.org/21437623:09
*** annegentle has joined #openstack-cinder23:10
*** dannywilson has joined #openstack-cinder23:10
*** dflorea has quit IRC23:13
*** ebalduf has quit IRC23:18
*** dims_ has joined #openstack-cinder23:20
*** dims has quit IRC23:22
patrickeasthemna: pong23:23
hemnapatrickeast: any luck messing aroudn with SR-IOV yet?23:23
hemnaI’m hacking on my blade to see if I can enable it23:23
hemnalooks like the card supports it, just isn’t enabled23:23
patrickeasthemna: nah i haven't even got an order for an HBA that can support it23:24
patrickeasthemna: ooo sick23:24
patrickeasthemna: ive still got it on my backlog... somewhere...23:24
hemnait’s a bunch of trickery23:24
hemnait’ll be a miracle if it even works23:24
hemnaI think my card supports 64 VF’s per port23:25
patrickeastnice, that should be a pretty big boost in testing throughput23:25
patrickeasti remember looking through some of the instructions to do it and the commands to configure them were not very straight forward23:26
*** lcurtis has quit IRC23:26
*** dannywilson has quit IRC23:28
*** cbader has quit IRC23:28
*** chlong has quit IRC23:29
openstackgerritVictor Stinner proposed openstack/cinder: Port test_nfs to Python 3  https://review.openstack.org/21438423:30
openstackgerritAlex O'Rourke proposed openstack/cinder: LeftHand: Adding Consistency Group Support  https://review.openstack.org/21438623:31
*** changbl has joined #openstack-cinder23:31
*** vilobhmm has quit IRC23:32
hemnapatrickeast: yah it’s not obvious.   basically lspci -vvv23:33
hemnaif you don’t do the 3 v’s you don’t get to see it23:34
hemnapatrickeast: http://pastebin.com/wDbBsyiv23:34
hemnaVFs of 023:35
hemnabasically means the card is capable, but it’s disabled in BIOS23:35
patrickeastahh yea i remember seeing something about that23:36
patrickeasti figured out how to do it for the supermicro initiators i've got23:36
patrickeastjust a toggle on kind of deal23:36
patrickeasthemna: i had this page bookmarked, https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/Virtualization_Host_Configuration_and_Guest_Installation_Guide/sect-Virtualization_Host_Configuration_and_Guest_Installation_Guide-SR_IOV-How_SR_IOV_Libvirt_Works.html23:38
patrickeasti assume its a similar process with a FC device23:38
hemnayah it is23:38
hemnathe emulex  driver pdf talks about it as well23:38
hemnaPage 42, setting up SR-IOV23:39
hemnapatrickeast: also FWIW, I have a WIP patch in os-brick that you might want to take a look at when you have time.23:39
hemnabasically changes the mechanism for discovering multipath devices for FC (and eventually iSCSI)23:40
hemnayour CI passed against it :P23:40
patrickeastoo neat23:40
patrickeasti mean, if the ci passes that means its ready to ship, right?23:40
hemnaship it!23:41
hemnawe have a customer that has hundreds of mpath volumes23:41
hemnaon the same box23:41
hemnaand running multipath -ll <path> gets REALLY slow23:41
patrickeastahh yea i bet23:41
hemnalike exponentially slower to run on every attach23:41
hemnaso my patch avoids even calling multipath -ll at all23:42
hemnaI’d like to get some other eyes on it to see if what I’m doing is sane23:42
*** gouthamr_ has joined #openstack-cinder23:42
hemnaI’ve done a bunch of digging on how multipath daemon creates it’s devices and it’s multipath id’s23:42
patrickeastat first glance it seems like a pretty solid win, i'll take a closer look and play around with it later today or tomorrow23:43
hemnathe multipath id is nothing more than the volume WWN which is at page 0x83 of every volume.23:43
patrickeastawesome, i had wondered about that23:43
hemnaso I read that page, to get the WWN, and then look for /dev/mapper/<WWN> to show up.23:43
hemnathat path shows up after the multipath daemon creates the mpath device23:43
* hemna still waits for HP CI on that patch......23:44
*** gouthamr has quit IRC23:44
patrickeastneed to get that sr-iov stuff working!23:44
*** gouthamr has joined #openstack-cinder23:45
hemnaas a side effect of this patch, I’ll be able to get rid of all the multipath_id pooping all over nova23:45
hemnaso it will clean up nova a bit, which will also help live migration as well.23:45
hemnaso I’m hoping this is the ticket.23:45
*** gouthamr_ has quit IRC23:47
*** diogogmt has quit IRC23:48
*** ndipanov has joined #openstack-cinder23:50
patrickeasthemna: so with that change what happens if they have user_friendly_names set? i thought that changed the /dev/mapper/ entry to something like /dev/mapper/mpath123:51
patrickeastwould we just say that for openstack you can't enable that?23:52
*** ndipanov has quit IRC23:52
*** ndipanov has joined #openstack-cinder23:52
patrickeasti got burned by this once copying a multipath.conf file that had it enabled and broke some scripts : /23:52
openstackgerritKurt Martin proposed openstack/cinder: LeftHand driver is ignoring reserved_percentage  https://review.openstack.org/21439023:56
hemnadoes that change the dev/mapper path it creates?23:58
patrickeasthemna: check out the first section here https://help.ubuntu.com/lts/serverguide/multipath-devices.html23:59
hemnaI’ll have to enable that and see how much stuff I break23:59
patrickeastexplains better than i can23:59
*** tiny-hands has joined #openstack-cinder23:59
*** ndipanov has quit IRC23:59

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!