Monday, 2015-03-09

openstackgerritThang Pham proposed openstack/cinder: Raise exception for invalid mock assert calls
*** _cjones_ has quit IRC00:16
openstackgerritThang Pham proposed openstack/cinder: Switch to oslo_versionedobjects
*** ho has joined #openstack-cinder00:22
*** e0ne has quit IRC00:49
*** bill_az has joined #openstack-cinder00:50
*** tshefi has quit IRC00:51
*** Lee1092 has joined #openstack-cinder00:59
*** topshare has joined #openstack-cinder01:00
*** _cjones_ has joined #openstack-cinder01:12
*** takedakn has joined #openstack-cinder01:14
*** takedakn has quit IRC01:17
*** mriedem1 has quit IRC01:26
*** _cjones_ has quit IRC01:31
*** bill_az has quit IRC01:37
*** liusheng has quit IRC01:46
*** _cjones_ has joined #openstack-cinder01:51
*** qwebirc49866 has joined #openstack-cinder01:58
*** qwebirc49866 has quit IRC01:58
*** DaveChen has joined #openstack-cinder02:01
openstackgerritwanghao proposed openstack/cinder: Implement function of import/export snapshots
*** krtaylor has quit IRC02:04
*** Ilja has quit IRC02:06
*** kaufer has joined #openstack-cinder02:09
*** diegows has quit IRC02:11
openstackgerritVincent Hou proposed openstack/cinder: Delete the temporary volume if migration fails
openstackgerritVincent Hou proposed openstack/cinder: Test.
openstackgerritVincent Hou proposed openstack/cinder-specs: Enhance list operations with the additional keys and next link
*** kaisers has joined #openstack-cinder02:19
*** kaisers1 has quit IRC02:20
*** [1]Thelo has joined #openstack-cinder02:24
*** hflai has joined #openstack-cinder02:26
*** haomaiwang has joined #openstack-cinder02:26
*** Thelo has quit IRC02:26
*** [1]Thelo is now known as Thelo02:26
*** Krast has joined #openstack-cinder02:32
*** Thelo has left #openstack-cinder02:33
*** Thelo has joined #openstack-cinder02:33
*** coolsvap has joined #openstack-cinder02:34
*** _cjones_ has quit IRC02:37
*** krtaylor has joined #openstack-cinder02:55
*** Longgeek has joined #openstack-cinder02:55
*** kaufer has quit IRC02:57
*** jcru has joined #openstack-cinder02:58
*** vilobhmm has joined #openstack-cinder02:59
*** vilobhmm1 has joined #openstack-cinder03:04
*** vilobhmm has quit IRC03:05
*** buster_604 has quit IRC03:13
*** jcru has quit IRC03:15
*** buster_604 has joined #openstack-cinder03:18
*** buster_604 has quit IRC03:19
*** Mandell has joined #openstack-cinder03:19
*** buster_604 has joined #openstack-cinder03:42
*** _cjones_ has joined #openstack-cinder03:44
*** sandywalsh has quit IRC03:45
*** _cjones_ has quit IRC03:55
*** vilobhmm1 has quit IRC04:15
*** vilobhmm has joined #openstack-cinder04:16
*** buster_604 has quit IRC04:26
*** esker has quit IRC04:28
*** esker has joined #openstack-cinder04:28
*** Ilja has joined #openstack-cinder04:43
*** BharatK has joined #openstack-cinder04:48
*** _cjones_ has joined #openstack-cinder04:52
*** _cjones_ has quit IRC04:56
*** buster_604 has joined #openstack-cinder04:56
*** buster_604 has quit IRC04:58
*** Ilja has quit IRC05:00
*** sgotliv has quit IRC05:11
*** Ilja has joined #openstack-cinder05:26
*** lpetrut has joined #openstack-cinder05:28
*** BharatK has quit IRC05:52
*** BharatK has joined #openstack-cinder05:56
*** pradipta has joined #openstack-cinder05:58
openstackgerritPeter Wang proposed openstack/cinder: Fixes VNX NotImplementedError of unmanage
openstackgerritOpenStack Proposal Bot proposed openstack/cinder: Imported Translations from Transifex
*** anshul has joined #openstack-cinder06:09
*** deepakcs has joined #openstack-cinder06:09
*** BharatK has quit IRC06:32
*** BharatK has joined #openstack-cinder06:34
*** nlevinki has joined #openstack-cinder06:35
*** sgotliv has joined #openstack-cinder06:39
*** beecee has quit IRC06:43
*** sgotliv has quit IRC06:44
*** bkopilov has quit IRC06:50
*** _cjones_ has joined #openstack-cinder06:53
openstackgerritPeter Wang proposed openstack/cinder: Fixes VNX NotImplementedError of unmanage
openstackgerritPeter Wang proposed openstack/cinder: Fixes VNX NotImplementedError of unmanage
*** rwsu has joined #openstack-cinder06:57
*** sgotliv has joined #openstack-cinder06:58
*** _cjones_ has quit IRC06:58
*** rwsu is now known as rwsu-afk06:58
*** Ilja has quit IRC06:59
openstackgerritAbhijeet Malawade proposed openstack/cinder: Get volume from db again before updating it's status
*** ankit_ag has joined #openstack-cinder07:08
*** afazekas has joined #openstack-cinder07:09
openstackgerritHuang Zhiteng proposed openstack/cinder: DO NOT MERGE: Scheduler Raises NoValidHostNotComplete when it doesn't have complete view
*** nkrinner has joined #openstack-cinder07:17
*** vilobhmm has quit IRC07:18
*** ankit8188 has joined #openstack-cinder07:31
*** ankit_ag has quit IRC07:33
openstackgerritLiu Xinguo proposed openstack/cinder: Huawei driver check before associating LUN to a LUN group
*** Miouge has joined #openstack-cinder07:44
*** bkopilov has joined #openstack-cinder07:45
*** TobiasE1 has joined #openstack-cinder07:47
openstackgerritLiu Xinguo proposed openstack/cinder: Huawei driver fix problems under multipath
*** bkopilov has quit IRC07:50
*** bkopilov has joined #openstack-cinder07:50
*** bkopilov has quit IRC07:53
*** bkopilov has joined #openstack-cinder07:54
*** _cjones_ has joined #openstack-cinder07:55
*** Ilja has joined #openstack-cinder07:58
*** _cjones_ has quit IRC08:00
*** nellysmitt has joined #openstack-cinder08:00
*** Ilja has quit IRC08:07
*** chlong has quit IRC08:07
*** nellysmitt has quit IRC08:12
*** nellysmitt has joined #openstack-cinder08:12
*** lpetrut has quit IRC08:20
openstackgerritTina Tang proposed openstack/cinder: Attach/detach batch processing in VNX driver
*** ronis has joined #openstack-cinder08:22
*** lpetrut has joined #openstack-cinder08:30
*** enterprisedc has quit IRC08:30
*** bkopilov has quit IRC08:31
*** Ilja has joined #openstack-cinder08:31
*** Ilja has quit IRC08:32
*** Ilja has joined #openstack-cinder08:32
*** lpetrut has quit IRC08:33
*** Ilja1 has joined #openstack-cinder08:35
*** Ilja has quit IRC08:36
*** RanZ has joined #openstack-cinder08:38
*** Mandell has quit IRC08:39
openstackgerritwanghao proposed openstack/cinder: Implement function of import/export snapshots
openstackgerritLiu Xinguo proposed openstack/cinder: Huawei driver remove LUN controller change
wanghaohi guys: there is a patch:'', has already got +2 for a week, could you help me to merge it? THX08:47
*** jordanP has joined #openstack-cinder08:48
*** Kennan_ has joined #openstack-cinder08:48
Kennan_hi all, could I ask one question ?08:49
Kennan_in devstack08:49
Kennan_2015-03-09 16:49:47.391 TRACE cinder.volume.manager ProcessExecutionError: Unexpected error while running command. 2015-03-09 16:49:47.391 TRACE cinder.volume.manager Command: sudo cinder-rootwrap /etc/cinder/rootwrap.conf env LC_ALL=C LVM_SYSTEM_DIR=/etc/cinder vgs --noheadings --unit=g -o name,size,free,lv_count,uuid --separator : --nosuffix stack-volumes-lvmdriver-1 2015-03-09 16:49:47.391 TRACE cinder.volume.manager Exi08:49
Kennan_but when I run08:50
Kennan_sudo vgs --noheadings --unit=g -o name,size,free,lv_count,uuid --separator : --nosuffix stack-volumes-lvmdriver-108:50
Kennan_it have output08:50
Kennan_I don't know why cinder-volum run with sudo cinder-rootwrap /etc/cinder/rootwrap.conf08:51
Kennan_do you know ?08:51
*** _cjones_ has joined #openstack-cinder08:56
openstackgerritwanghao proposed openstack/cinder: Fix the unicode encode error when create volume
*** briancline has quit IRC09:00
*** nlevinki has quit IRC09:01
*** _cjones_ has quit IRC09:02
*** sdague has quit IRC09:04
*** dobson has quit IRC09:04
*** karimb has joined #openstack-cinder09:04
*** nlevinki has joined #openstack-cinder09:04
*** briancline has joined #openstack-cinder09:06
*** sdague has joined #openstack-cinder09:06
*** dobson has joined #openstack-cinder09:06
wanghao@kennan, Is there more detial error message about cinder-volume running command?09:10
*** jistr has joined #openstack-cinder09:13
*** lpetrut has joined #openstack-cinder09:14
RanZHi. Quick question: is there a reason why "id" is not a valid search_opts when searching for Volumes? I'm referring to it missing here:
*** karimb has quit IRC09:25
*** karimb has joined #openstack-cinder09:27
wanghao@RanZ, For common user, display_name is more useful for querying, I think.09:27
*** Kennan_ has left #openstack-cinder09:31
openstackgerritSHIGEMATSU Mitsuhiro proposed openstack/cinder: Fix wrong log output in cinder/cinder/volume/drivers/
RanZ@wanghao Is there any specific reason why "id" is not available though? in my use case I often use display_name, but then in other times I'd like to use ID09:43
wanghao@RanZ, It's unkown for me, but I agree with you. It's should be configurable for common user to query volumes by more choice.09:49
*** Ilja1 has quit IRC09:50
wanghaothere is a bp about this:'', I hope it will be more flexibility.09:50
RanZ@wanghao I see that you created this ticket. Did you create one for Nova as well? I've noticed it's the same with querying servers by ID09:55
*** Ilja has joined #openstack-cinder09:56
*** _cjones_ has joined #openstack-cinder09:59
*** _cjones_ has quit IRC10:09
*** haomaiwang has quit IRC10:09
wanghao@RanZ.Not yet.I'am thinking a question,because we already have ability to show the resource by ID. Is it necessary for index or detail query? I mean that filter by az or host is ok, but ID is unique for resource.10:11
*** coolsvap has quit IRC10:17
pradip_vedamshi all, can anyone tell me for VPN as service we need to openstack environment or we can create two diffrent tenant in same environment....????????10:17
pradip_vedamssorry guys by mistake i post neutron question here....10:19
*** RanZ has quit IRC10:20
*** topshare has quit IRC10:24
*** TobiasE has joined #openstack-cinder10:26
*** TobiasE1 has quit IRC10:27
openstackgerritrajiv proposed openstack/python-cinderclient: cinderclient accepts arguments after metadata without -- separator
*** RanZ has joined #openstack-cinder10:35
*** bkopilov has joined #openstack-cinder10:38
openstackgerritwanghao proposed openstack/cinder: Fix response when query host detail by host name
*** chlong has joined #openstack-cinder11:03
*** nlevinki has quit IRC11:04
*** _cjones_ has joined #openstack-cinder11:05
openstackgerritwanghao proposed openstack/cinder: Implement function of import/export snapshots
*** markvoelker has joined #openstack-cinder11:13
*** nlevinki has joined #openstack-cinder11:20
*** kallebe has joined #openstack-cinder11:21
*** ho has quit IRC11:24
*** IanGovett has joined #openstack-cinder11:36
*** _cjones_ has quit IRC11:38
*** cdelatte has joined #openstack-cinder11:44
*** diegows has joined #openstack-cinder11:46
*** markvoelker has quit IRC11:47
*** markvoelker has joined #openstack-cinder11:48
openstackgerrityogeshprasad proposed openstack/cinder: Added provider_id to volume and snapshot object
*** anik80 has quit IRC11:51
*** markvoelker has quit IRC11:52
openstackgerrityogeshprasad proposed openstack/cinder: Added provider_id to volume and snapshot object
*** Anticimex has quit IRC11:57
*** Anticimex has joined #openstack-cinder11:57
*** Miouge_ has joined #openstack-cinder12:01
*** Miouge has quit IRC12:01
*** Miouge_ is now known as Miouge12:01
*** ganso_ has joined #openstack-cinder12:03
*** IanGovett has quit IRC12:04
*** Ilja has quit IRC12:06
*** IanGovett has joined #openstack-cinder12:10
*** Ilja has joined #openstack-cinder12:14
*** Ilja1 has joined #openstack-cinder12:14
*** Ilja has quit IRC12:14
*** marcusvrn has joined #openstack-cinder12:16
*** sgotliv has quit IRC12:17
*** bswartz has quit IRC12:22
*** dulek has joined #openstack-cinder12:23
*** TobiasE1 has joined #openstack-cinder12:24
*** [1]Thelo has joined #openstack-cinder12:24
*** TobiasE has quit IRC12:25
*** Thelo has quit IRC12:26
*** [1]Thelo is now known as Thelo12:26
*** topshare has joined #openstack-cinder12:30
*** sgotliv has joined #openstack-cinder12:32
*** marcusvrn has quit IRC12:33
*** marcusvrn has joined #openstack-cinder12:35
*** _cjones_ has joined #openstack-cinder12:36
*** annegentle has joined #openstack-cinder12:36
*** Mandell has joined #openstack-cinder12:40
*** topshare has quit IRC12:43
*** aix has joined #openstack-cinder12:43
*** Mandell has quit IRC12:44
*** Miouge has quit IRC12:50
*** Miouge has joined #openstack-cinder12:50
*** ndipanov has joined #openstack-cinder12:53
*** ndipanov has quit IRC12:53
*** bswartz has joined #openstack-cinder12:56
*** deepakcs has quit IRC12:57
*** Yogi1 has joined #openstack-cinder13:00
*** TobiasE has joined #openstack-cinder13:01
*** akerr has joined #openstack-cinder13:01
*** Yogi1 has left #openstack-cinder13:02
*** TobiasE1 has quit IRC13:02
*** nlevinki has quit IRC13:03
*** chlong has quit IRC13:04
*** Miouge has quit IRC13:04
*** Yogi1 has joined #openstack-cinder13:06
*** topshare has joined #openstack-cinder13:08
*** _cjones_ has quit IRC13:09
*** nlevinki has joined #openstack-cinder13:17
*** chlong has joined #openstack-cinder13:17
*** timcl has joined #openstack-cinder13:21
*** kaufer has joined #openstack-cinder13:25
*** dustins has joined #openstack-cinder13:28
*** superdan is now known as dansmith13:34
*** eharney has quit IRC13:42
*** tbarron has joined #openstack-cinder13:43
*** anki80 has joined #openstack-cinder13:48
*** yamada-h has quit IRC13:49
*** sgotliv_ has joined #openstack-cinder13:53
*** sgotliv has quit IRC13:53
*** Ilja1 has quit IRC13:57
*** topshare has quit IRC14:00
*** gmurphy has left #openstack-cinder14:03
*** topshare has joined #openstack-cinder14:03
*** _cjones_ has joined #openstack-cinder14:06
*** Ilja has joined #openstack-cinder14:06
*** Anticimex has quit IRC14:07
*** marcusvrn1 has joined #openstack-cinder14:07
*** Anticimex has joined #openstack-cinder14:08
*** marcusvrn has quit IRC14:09
*** annegentle has quit IRC14:10
*** markvoelker has joined #openstack-cinder14:11
*** thangp has joined #openstack-cinder14:12
*** annegentle has joined #openstack-cinder14:15
*** jcru has joined #openstack-cinder14:17
*** jcru has quit IRC14:17
*** ebalduf has joined #openstack-cinder14:17
*** jcru has joined #openstack-cinder14:17
*** markvoelker has quit IRC14:18
*** IanGovett has quit IRC14:19
*** markvoelker has joined #openstack-cinder14:19
*** mwichmann has joined #openstack-cinder14:22
*** markvoelker has quit IRC14:23
*** Guest82873 has quit IRC14:23
*** nellysmitt has quit IRC14:24
*** markvoelker has joined #openstack-cinder14:24
*** r-daneel has joined #openstack-cinder14:28
*** xyang has joined #openstack-cinder14:30
*** akerr is now known as akerr_away14:31
*** eharney has joined #openstack-cinder14:31
*** akerr has joined #openstack-cinder14:32
*** akerr is now known as akerr_away14:32
*** akerr_away has quit IRC14:33
*** emagana has joined #openstack-cinder14:33
*** akerr has joined #openstack-cinder14:34
*** alonmarx has quit IRC14:36
*** alonmarx has joined #openstack-cinder14:37
*** ankit8188 has quit IRC14:37
*** _cjones_ has quit IRC14:39
*** mriedem has joined #openstack-cinder14:39
*** mtanino has joined #openstack-cinder14:42
openstackgerritThang Pham proposed openstack/cinder: Raise exception for invalid mock assert calls
*** beekneemech is now known as bnemec14:49
*** annegentle has quit IRC14:54
openstackgerritStefan Amann proposed openstack/cinder: Adjust Cinder to support FCP on System z systems
*** ndipanov has joined #openstack-cinder14:56
*** patrickeast has joined #openstack-cinder14:59
*** alonmarx_ has joined #openstack-cinder14:59
*** alonmarx_ has quit IRC15:00
*** alonmarx_ has joined #openstack-cinder15:00
*** alonmarx has quit IRC15:01
openstackgerritSHIGEMATSU Mitsuhiro proposed openstack/cinder: Fix wrong log output in cinder/cinder/volume/drivers/
*** ndipanov has quit IRC15:02
openstackgerritOpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements
openstackgerritKurt Martin proposed openstack/cinder: HP lefthand driver filter and evalautor function
*** topshare has quit IRC15:06
*** asselin_ has joined #openstack-cinder15:06
*** nellysmitt has joined #openstack-cinder15:08
*** asselin_ has quit IRC15:08
*** asselin_ has joined #openstack-cinder15:08
*** Mandell has joined #openstack-cinder15:10
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder: Add volume multi attach support
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder: Add multiattach capabilities injection
*** ndipanov has joined #openstack-cinder15:17
*** dannywilson has joined #openstack-cinder15:20
*** dannywil_ has joined #openstack-cinder15:21
*** dannywilson has quit IRC15:22
*** rmesta has joined #openstack-cinder15:22
*** hemnafk is now known as hemna15:22
*** annegentle has joined #openstack-cinder15:23
*** afazekas has quit IRC15:26
openstackgerritThang Pham proposed openstack/cinder: Switch to oslo_versionedobjects
*** tsekiyama has joined #openstack-cinder15:26
*** emagana has quit IRC15:28
openstackgerritxing-yang proposed openstack/cinder: Don't override extra specs with config in VMAX
*** kallebe has quit IRC15:30
*** markvoelker has quit IRC15:30
*** IanGovett has joined #openstack-cinder15:31
*** markvoelker has joined #openstack-cinder15:31
jungleboyjhemna: Ping.15:31
*** emagana has joined #openstack-cinder15:31
*** jdurgin has joined #openstack-cinder15:32
*** harlowja_at_home has joined #openstack-cinder15:32
*** markvoelker has quit IRC15:35
*** _cjones_ has joined #openstack-cinder15:35
*** esker has quit IRC15:36
*** anshul has quit IRC15:37
*** david-lyle_afk is now known as david-lyle15:38
*** markvoelker has joined #openstack-cinder15:38
*** Ilja has quit IRC15:41
*** rwsu-afk is now known as rwsu15:41
*** diegows has quit IRC15:42
*** markvoelker has quit IRC15:43
*** markvoelker has joined #openstack-cinder15:44
*** emagana has quit IRC15:44
*** patrickeast has quit IRC15:44
openstackgerritThang Pham proposed openstack/cinder: Switch get_all_snapshots to use objects
*** Mandell has quit IRC15:46
hemnajungleboyj, ?15:48
*** nellysmitt has quit IRC15:49
*** markvoelker has quit IRC15:49
openstackgerritxing-yang proposed openstack/cinder: Fixed a concurrency issue in VMAX driver
openstackgerritxing-yang proposed openstack/cinder: Fixed a problem in recreating masking view in VMAX
jungleboyjhemna: Are any of the zone managers able to handle dual fabric environments?15:51
hemnameaning 2 brocade switches ?15:51
hemnaor brocade + cisco at the same time ?15:51
hemnaafaik brocade and cisco don't play well together15:51
hemnayou can define 2 brocade zones and use em in cinder.conf15:52
*** asselin_ has quit IRC15:52
*** nellysmitt has joined #openstack-cinder15:52
jungleboyjhemna: The question was ' can they handle a dual Fabric environment (A and B fabric).15:53
hemnaafaik yes15:53
jungleboyjOk, that was what I had assumed but wanted to check with someone smarter than me.  ;-)15:53
*** emagana has joined #openstack-cinder15:53
jungleboyjhemna: Thanks!15:55
*** diegows has joined #openstack-cinder15:56
*** sgotliv has joined #openstack-cinder15:58
*** sgotliv_ has quit IRC15:58
*** kallebe has joined #openstack-cinder15:58
*** _cjones_ has quit IRC16:01
openstackgerritMarkus Zoeller (markus_z) proposed openstack/cinder: Adjust Cinder to support FCP on System z systems
*** nlevinki has quit IRC16:05
*** annegentle has quit IRC16:06
*** esker has joined #openstack-cinder16:07
*** TobiasE has quit IRC16:11
*** nellysmitt has quit IRC16:11
*** _cjones_ has joined #openstack-cinder16:14
*** _cjones_ has quit IRC16:14
*** jaypipes has joined #openstack-cinder16:14
*** _cjones_ has joined #openstack-cinder16:14
openstackgerritMarkus Zoeller (markus_z) proposed openstack/cinder: Adjust Cinder to support FCP on System z systems
*** markus_z has joined #openstack-cinder16:17
*** nlevinki has joined #openstack-cinder16:19
*** emagana has quit IRC16:20
*** nellysmitt has joined #openstack-cinder16:21
hemnaline 72516:22
hemnaI didn't touch that16:22
*** Guest82873 has joined #openstack-cinder16:22
*** ronis has quit IRC16:23
*** thrawn01 has joined #openstack-cinder16:25
hemna        cinderclient(context).volumes.attach(volume_id, instance_uuid,16:26
hemna                                             mountpoint, mode=mode)16:26
hemnaso I tried to fix that by adding the hostname to populate it16:27
openstackgerritMichal Dulko proposed openstack/cinder: Backup object
*** Longgeek has quit IRC16:29
*** leeantho has joined #openstack-cinder16:30
*** jdurgin has quit IRC16:33
*** RanZ has quit IRC16:34
*** krtaylor has quit IRC16:36
*** harlowja_at_home has quit IRC16:39
openstackgerritJean-Daniel Bonnetot proposed openstack/cinder-specs: Add the possibility for admin to list volume-backend-names declared in cinder.conf.
*** dulek has quit IRC16:41
*** Yogi1 has quit IRC16:46
openstackgerritMarkus Zoeller (markus_z) proposed openstack/os-brick: Adjust os-brick to support FCP on System z systems
*** leeantho has quit IRC16:47
*** leeantho has joined #openstack-cinder16:47
*** nlevinki has quit IRC16:47
*** krtaylor has joined #openstack-cinder16:48
*** lpetrut has quit IRC16:48
*** leeantho has quit IRC16:49
*** leeantho has joined #openstack-cinder16:49
*** Apoorva has joined #openstack-cinder16:51
*** Krast has quit IRC16:55
*** Krast has joined #openstack-cinder16:56
hemnaso....I'm not sure what I should do16:57
*** emagana has joined #openstack-cinder17:03
*** _cjones_ has quit IRC17:05
*** markus_z has quit IRC17:06
*** rmesta has left #openstack-cinder17:06
*** annegentle has joined #openstack-cinder17:07
*** tsekiyama has quit IRC17:07
*** nellysmitt has quit IRC17:11
*** nellysmitt has joined #openstack-cinder17:13
*** nellysmitt has quit IRC17:17
*** _cjones_ has joined #openstack-cinder17:24
*** ebalduf has quit IRC17:27
*** sgotliv has quit IRC17:28
*** ronis has joined #openstack-cinder17:28
*** annashen has joined #openstack-cinder17:30
*** lpetrut has joined #openstack-cinder17:30
*** jdurgin has joined #openstack-cinder17:35
*** harlowja has joined #openstack-cinder17:40
*** vilobhmm has joined #openstack-cinder17:41
*** harlowja_ has joined #openstack-cinder17:43
*** ndipanov has quit IRC17:45
*** harlowja has quit IRC17:45
*** Mandell has joined #openstack-cinder17:47
*** vilobhmm1 has joined #openstack-cinder17:48
*** patrickeast has joined #openstack-cinder17:51
*** enterprisedc has joined #openstack-cinder17:51
*** vilobhmm has quit IRC17:52
openstackgerritPatrick East proposed openstack/cinder: PureISCSIDriver consistency group updates.
*** annegentle has quit IRC17:55
*** annegentle has joined #openstack-cinder17:55
patrickeastjungleboyj: haha thanks for catching that log message typo… i’ve probably looked at it 100 times by now and didn’t notice it was the wrong method name17:57
patrickeasttoo much copy pasta17:57
*** jistr has quit IRC17:59
jungleboyjpatrickeast: Welcome.  Was going to ping you when I saw you come on.18:01
jungleboyjFelt bad -1'ing at this point.18:01
jungleboyjhemna: xyang Do you know if things can still be merged tomorrow or is today the last chance?18:02
hemnaI'm unclear on that as well18:02
hemnaI'm desperately trying to get multi-attach in today18:03
hemnabut my manager wants me to work on slides instead.18:03
jungleboyjYeah, I was going to look at that in a few minutes.18:03
*** pradipta has quit IRC18:03
hemnayet another rebase....damnit18:05
*** chlong has quit IRC18:05
hemnaI'm in rebase hell18:06
jungleboyjMorphine has made my cat hungry and paranoid.18:06
SwansonKilo-3 in 10 days, right?  RC's start 3 weeks after that?18:07
hemnaSwanson, we have a self imposed dealine  though18:08
*** tsekiyama has joined #openstack-cinder18:08
*** annegentle has quit IRC18:08
Swansonhemna: I'm just worried about testing timelines.18:12
hemnaonce our self imposed deadline passes, we are all supposed to be testing.18:13
*** tsekiyama has quit IRC18:13
Swansonwe've been testing for a while.  I just want to make sure people know the timelines18:14
jungleboyjhemna: They got the updates for System Z pushed up:  When ever you have a chance to look.18:14
*** lcurtis has joined #openstack-cinder18:15
*** bswartz has quit IRC18:16
*** bswartz has joined #openstack-cinder18:22
hemnahe's been absent a lot due to traveling to conferences doing python presentations and such18:26
hemnaI'm not sure where everyone is.18:26
xyangjungleboyj: I'm not sure if things can still be merged tomorrow.18:26
* jungleboyj thinks if we are doing reviews today and asking for changes it should be allowed.18:27
jungleboyjBut that is me.18:27
*** karimb has quit IRC18:27
kallebeHey, I was trying to find some way to run hacking tests locally but could not find it. Does anyone know how to do it?18:27
hemnaI'm sure that'll be meaningless tomorrow.18:27
jungleboyjhemna: Let me know when the multi-attach change is good to look at again.18:28
jungleboyjkallebe: ./ -V -p18:28
hemnajungleboyj, ok thanks.  just waiting on local tox18:28
jungleboyjIt is part of pep8.18:28
kallebejungleboyj ok. thanks for the help18:28
jungleboyjhemna: Then you can go look at the SystemZ patch while I do that.  ;-)18:28
hemnajungleboyj, ok!  url ?18:29
*** jordanP has quit IRC18:31
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder: Add volume multi attach support
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder: Add multiattach capabilities injection
hemnaxyang, jungleboyj ok I removed the setting of the attached_host.18:32
hemnait now acts as it did previously, which will always have a value of None or Null in the DB column.18:32
xyanghemna: ok, I'll take a look18:33
*** tsekiyama has joined #openstack-cinder18:33
mtaninoanish: ping,18:34
*** sgotliv has joined #openstack-cinder18:38
hemnajungleboyj, do we have any cert results for a run on the s390 system?18:40
jungleboyjLet me find those.18:41
*** mgagne is now known as mgagne_PHL18:42
hemnajungleboyj, thanks18:42
*** _cjones_ has quit IRC18:44
*** rushil has joined #openstack-cinder18:44
jungleboyjhemna: I think this is the one:
openstackLaunchpad bug 1422423 in Cinder "Cinder Certification results with zKVM platform support." [Low,Confirmed]18:49
hemnajungleboyj, ok, the issue with that log is that it doesn't show cinder.conf18:52
hemnajungleboyj, how can I tell that this ran on a s390 system ?18:52
jungleboyjGood question.  It also probably hasn't been run since their latest update.  In fact, I know that.18:54
openstackgerritBharat Kumar Kobagana proposed openstack/cinder: GlusterFS: Support over subscription in thin provisioning
xyanghemna: looks good to me no. just waiting for Jenkins18:56
jungleboyjI am looking right now xyang hemna18:57
hemnathanks guys18:58
jungleboyjhemna: None of the System Z guys are on right now.  So I guess we will have to make a comment and hope that we can still merge tomorrow.18:59
jungleboyjCan you explain to them exactly what is missing.18:59
hemnaok I can try.18:59
jungleboyjI guess you are looking for something I haven't worried about before.18:59
hemnamost of the cert runs dump the contents of cinder.conf at the top18:59
hemnathe upside is that the change isn't breaking our FC 3rd party CI runs19:00
hemnawhich the patch directly changes in brick19:00
hemnaso, I Guess as long as it doesn't break that it's cool.19:00
jungleboyj:-)  That is a good sign.19:00
*** emagana has quit IRC19:00
hemnaif it's broken on S390,  meh19:00
*** emagana has joined #openstack-cinder19:01
*** tbarron has quit IRC19:01
*** dustins_ has joined #openstack-cinder19:01
hemnalooking through the tempest logs to see if it even tests copy volume <---> image19:01
openstackgerritGorka Eguileor proposed openstack/cinder: Preserve usage and reservations on quota deletion
*** tbarron has joined #openstack-cinder19:03
*** tbarron has quit IRC19:05
*** dustins has quit IRC19:05
openstackgerritKallebe Monteiro proposed openstack/cinder: Replace assertEqual(True, *) -> assertTrue(*)
*** emagana has quit IRC19:06
*** Longgeek has joined #openstack-cinder19:08
anishmtanino: pong19:09
hemnajungleboyj, crap, it looks like our FC driver CI tests are skipping the test_volume_upload tempest tests19:09
mtaninoanish: hi19:09
hemnajungleboyj, so I take back what I said about it not breaking FC volume <--> image operations.19:09
hemnajungleboyj, it's unknown now.19:10
hemnajungleboyj, I'm trying to track down why we are skipping those.19:10
hemnajungleboyj, I'll check out the fc storwize output and see if you guys have it19:10
mtaninoanish: Could you let me know what the problem happens at
openstackLaunchpad bug 1422095 in Cinder "tgt driver does not handle file open/read errors correctly" [Undecided,In progress] - Assigned to Anish Bhatt (anish7)19:11
hemnajungleboyj, crap, storewize CI is also skipping it.19:11
hemnajungleboyj, so I have no way to verify this hasn't broken FC copy volume<-->image looking at 3rd party CI results19:11
mtaninoanish: I'm thinking about your comment on my IET patch, and I couldn't get the reason why we need to handle IOError.19:12
*** Longgeek has quit IRC19:12
*** _cjones_ has joined #openstack-cinder19:12
anishmtanino: if the file exists but open fails, we proceed on assuming file does not exist19:12
mtaninoanish: more comments are helpful for my understanding :)19:12
*** sgotliv has quit IRC19:13
*** Ilja has joined #openstack-cinder19:13
anishmtanino: so I hit this issue when I was writing the cxt helper19:14
jungleboyjhemna: :-(  Well, lets see if they can run again and get us results.19:14
anishbut I do not have a good way to reproduce it (because I am not sure what caused it in the first place)19:14
mtaninoanish: permission denied or someting like that?19:14
*** BharatK has quit IRC19:15
anishmtanino: sadly cannot recall :(19:15
hemnajungleboyj, ok19:15
mtaninoanish: What happens if we proceed to create new file and new chap auth if open fails?19:16
*** sgotliv has joined #openstack-cinder19:16
hemnajungleboyj, xyang lunch....bbiab.19:16
anishmtanino: well that is the issue19:16
anishif you look at create_iscsi_target, it opens in w+ mode19:16
anishso it will overrite the file if it exists19:17
anishbut if open failed because of something like persmissions19:17
anishthis will fail as well19:17
anishmaybe if the file was just corrupted or something19:17
*** bkopilov has quit IRC19:17
anishit won't19:17
anishkinda makes the behaviour unpredictable19:17
anishso I thought catching it in get_target_chap_auth() makes more sense19:17
mtaninoanish: as a result?19:17
jungleboyjhemna: Ok.  +2 from me on multi-attach.19:18
anishmtanino: so basically trying to avoid unpredictable behaviour19:18
mtaninoanish: hmm19:19
mtaninoanish: I understand you intended to change the get_target_chap_auth() behaviour using raise when fails to handle the file.19:22
eharneyxyang: i thought we were avoiding creating blueprints for each "driver X implements feature Y"19:22
mtaninoanish: let me consider.19:22
mtaninoanish: thanks19:22
anishanytime !19:23
anishmtanino: let me know if you come up with a better approach, as is still under review as well19:24
xyangeharney: I think that was changed in Kilo.  we have lots of blueprints19:26
xyangeharney: I'm ok with it either way.  it is a very small change. but should check with thingee19:27
smcginnisI thought we just didn't need specs for these kinds of things.19:27
eharneyxyang: so the expectation is each driver gets its own thin provisioning changes blueprints..?19:27
jungleboyjeharney: Yeah, that was John's approach but not Mike's.19:27
eharneyyeah and it was a mess and i thought we decided to do something different...19:27
xyangeharney: so both pure and vnx have blueprints for that19:27
xyangeharney: personally I'm ok with or without blueprint19:27
jungleboyjxyang: +119:28
eharneywe need to write this stuff down somewhere :)19:28
*** BharatK has joined #openstack-cinder19:28
xyangeharney: +119:29
kallebeHello. So is tomorrow really the last day for patches to be merged? My commit fixes a relatively serious bug in SAN driver. I would appreciate if some other core reviewer gave the other +2 Thanks19:29
xyangkallebe: that is for new features, not for bug fix19:29
kallebexyang oh, sorry. My bad. Thanks for the info.19:30
jungleboyj:-)  Fear not, only part of the sky is falling.19:30
hemnajungleboyj, xyang back19:32
jungleboyjhemna: Front19:32
*** tbarron has joined #openstack-cinder19:33
anishmtanino: interestingly, ietadm is already present in :)19:33
mtaninoanish: so we have to implement it :)19:34
*** tbarron has quit IRC19:35
*** e0ne has joined #openstack-cinder19:38
*** IanGovett has quit IRC19:39
*** tbarron has joined #openstack-cinder19:41
xyanghemna, jungleboyj: still waiting for Jenkins19:42
xyangjungleboyj, hema: about this  There's no cinderclient patch yet. I don't how he tested it19:43
*** tbarron has quit IRC19:43
*** thingee has joined #openstack-cinder19:44
*** emagana has joined #openstack-cinder19:45
jungleboyjYeah, Not sure about that one.19:45
openstackgerritGorka Eguileor proposed openstack/cinder: Preserve usage and reservations on quota deletion
hemnagate-cinder-python27 failed19:46
jungleboyjJust the fact that it doesn't pass unit test at this point is bad.19:46
xyangjungleboyj: I asked about it but he said he doesn't have any.  hemna: yes, unit test failed too19:46
thingeeeharney, xyang: I asked for bps for each driver feature, because I felt like we were oversubscribing ourselves and not realizing it19:46
thingeethere's no tracking otherwise19:46
eharneyah, ok19:46
*** annegentle has joined #openstack-cinder19:46
hemnayah that has to get fixed19:46
xyangjungleboyj, hemna: unit test failure  was from the new tests he added19:47
*** annashen has quit IRC19:47
thingeeeharney: but I don't need a spec. just a simple thing for me to target and keep track of is good enough for me19:47
xyangthingee: ok19:47
jungleboyjThere is thingee !19:47
eharneythingee: ok, thanks19:47
hemnathingee, !!19:47
jungleboyjthingee: Can things still merge tomorrow or today is the last day?19:47
thingeejungleboyj: tomorrow is fine :)19:48
hemnashhh chill guys, dad is back home...19:48
xyangthingee: eharney: I'm fine with the oversubscription one as it is so small19:48
thingeehemna: lol19:48
* thingee is at the ops midcycle meetup19:48
jungleboyjthingee: Excellent.  Don't feel bad about the comments I am making today.19:48
thingeeI have some good feedback that I would like to share with you all once I have a chance.19:48
jungleboyjhemna: I picture us all running to thingee ... thingee 's home!19:48
xyangthingee: :)19:49
thingeek3 is looking pretty good. ttx was really happy with how much progress we made.19:49
thingeemight be a record :P19:49
xyangthingee: jungleboyj, hemna and I were just talking about the import/export snapshot patch.  we don't know how he tested this without cinderclient change19:50
hemnaxyang, manually constructed curl calls?19:50
*** kallebe has left #openstack-cinder19:50
xyanghemna: that's possible19:50
thingeexyang: is there cinderclient changes proposed?19:51
xyangthingee: there is none.19:52
xyangthingee: I asked for it. He said he will propose it after the server side is merged19:52
xyangthingee: normally people will propose both19:52
thingeeok, yeah curl is all I can think of. Though you shouldn't use curl. that'll hurt your head. I recommend httpie :)
thingeexyang: yeah that's weird19:53
xyangthingee: so I'm worried that he didn't really test it19:53
jungleboyjAlso kind-of a big change to be dropping right now.19:54
thingeeyeah doesn't make much sense. People will unlikely use. I would just ask for the cinderclient stuff now.19:54
thingeeit shouldn't be too hard since there's already code for import/export volumes19:54
xyangthingee: sure.19:55
openstackgerritBrianna Poulos proposed openstack/cinder: Add project_id to barbican keymgr wrapper
xyangthingee: multiattach looks good now. you want to take a look?
*** leeantho_ has joined #openstack-cinder19:58
*** tbarron has joined #openstack-cinder19:58
thingeexyang: thanks, yeah I haven't looked at it yet. Do we feel really comfortable with this late?19:58
thingeejungleboyj, hemna ^19:59
thingeeas oppose to first thing in L for nova support?19:59
jungleboyjthingee: Which one?  Multi-attach?19:59
thingeejungleboyj: yea19:59
*** leeantho has quit IRC19:59
jungleboyjthingee: I am ok with it.  Hell of a headache for hemna to keep carrying it.20:00
thingeejungleboyj: understood20:00
thingeeis it tested?20:00
thingeeI understand there is unmerged nova changes, but it's still possible to test20:00
jungleboyjthingee: The PowerVC team has been working with it. They are going to cherry-pick the pieces and start using it.20:00
xyangthingee: I tested it. It works with the current code20:00
hemnahey jenkins finally came through20:01
thingeexyang: oh tested with an emc driver?20:01
xyangthingee: no, I tested it with lvm driver20:01
hemnait passes the storewize 3rd party CI fwiw20:01
*** BharatK has quit IRC20:01
hemnaand the 3par CI20:01
thingeehemna: 3par and storwize have support for it already?20:02
thingeehow does it work if the nova changes haven't merged?20:02
hemnaI've personally tested against 3par with all of the nova patches and cinderclient patches20:02
hemnato test actual multiple volume attaches20:02
hemnaand it works20:02
hemnabut we wont' have: nova side, or cinderclient side for Kilo20:03
*** Apoorva has quit IRC20:03
jungleboyjPowerVC has been using it.  Pulled the Nova and Cinder patches in and it has been working for their multi-attach needs.20:03
hemnaso what we really need to ensure, is that it doesn't break existing paths20:03
hemnamy follow up patch adds scheduler changes needed to prevent volume creates for drivers that don't report up multiattach: True20:03
hemnawhich there will be none for Kilo20:04
hemnait defaults to False20:04
hemnaso if someone injects the multiattach flag in the cinder create API, it will report no host found.20:04
thingeehemna: yes, that's what I'm worried about with this being so late. There's not really a gain right now except it being there without a doubt to the nova team. But I think we can commit together to get this in first thing of L to back you up on the nova side.20:04
hemnathat's the part that Duncan wanted.20:04
hemnathe gain is that I can focus on getting the nova and cinderclient part to land early L20:04
hemnaotherwise we are back to ground zero yet again20:05
hemnasame issue I've faced the last 3 OpenStack releases.20:05
*** tbarron has quit IRC20:06
hemnathe shadow tables patch landing in K is going to cause even more pain with this patch.20:06
thingeeto be fair, two releases ago was april. that was also pretty late. Don't remember the history with juno though.20:06
*** e0ne has quit IRC20:06
hemnawhich is yet another reason I'm trying to get this to land before the shadow tables patch lands20:06
*** e0ne has joined #openstack-cinder20:06
hemnathat basically doubles any effort anyone has to do with DB changes.20:06
hemnawhich completely sucks.20:06
thingeerather, it was submitted in april for icehouse, which was late.20:07
thingeejuno I'm not sure what happened.20:07
hemnait's the same thing that happens every release20:07
hemnanova's time schedule is impossible to meet.20:07
jungleboyjthingee: It is always Nova that has held it out.20:07
hemnaand cinder patches are held hostage until nova lands20:07
hemnabut nova won't land unless cinder lands20:07
hemnarinse, lather, repeat.20:07
hemna3 releases later...20:07
thingeeyeah but we won't. We'll can merge it first thing in L. I'm just really worried about existing stuff still working :(20:07
jungleboyjthingee: My $0.02 is that showing good faith in it now will help the Nova cause.20:08
hemnaexisting stuff does work20:08
hemnaor CI would be failing20:08
*** ndipanov has joined #openstack-cinder20:08
*** Apoorva has joined #openstack-cinder20:09
xyangthingee: so I tested attach volume and launch VM from boot volume works with existing stuff20:09
thingeehemna: you just said earlier "13:03:17           hemna | so what we really need to ensure, is that it doesn't break existing paths"20:09
xyangCI tests them as well20:09
hemnaand for the record, it's only "late" because no one has helped me do reviews on it.20:09
thingeehemna: that scares me :)20:09
hemnait's not like the patch hasn't been up for...oh...a year for people to look at.20:09
thingeeespecially a day and a half before cut20:10
hemnathingee, that's what CI is for no?20:10
hemnaand CI is passing20:10
*** Lee1092 has quit IRC20:10
hemnaif we don't trust out CI, then we have other problems.20:10
thingeewell then why did you say that? It just worried me is all20:10
thingeeI'm taking your words :P20:10
*** tbarron has joined #openstack-cinder20:11
hemnaIt was just a logical explanation of where we are with it, and what the importance of the patch is today.20:11
hemnasince we aren't getting the nova side to land and not the cinderclient side to land20:11
hemnahence, the importance on not breaking existing code.20:11
hemnawhich is why we look at the CI results20:11
hemnaif we find issues, I fix them or we revert it.20:12
hemnasimple as that.20:12
thingeeI get there is a CI hemna. I'm just taking your words of you still wanting to "ensure" that it doesn't break existing paths.20:12
hemnabut if your primary concern is the timing of landing patches right now, then all new feature patches should get -2 today then.20:12
thingeehemna: this is a big patch + touches a crucial part of Cinder.20:13
thingeekind of different than import/export snapshots20:13
thingeewhich is one endpoint and standalone20:13
thingeenot attaching20:13
hemnaso what do you want me to do then?20:13
hemnaI've been very attentive to this patchset trying to get people to review it and land.20:14
*** tbarron has quit IRC20:14
smcginnisJust my 2c, I'd like to see multiattach land now so we have time to work with it right away in L.20:14
* smcginnis goes back to his corner.20:14
thingeehemna: I think you've done everything you can. I think the community needs to make a decision though and it seems like people are ok with this going in regardless and feel comfortable.20:14
hemnaif you are worried, pull the patch down and test it yourself against lvm20:15
hemnaxying did that20:15
hemnaI've done it20:15
hemnaI just don't know what else to do, to get a patch to land in Cinder.20:15
* hemna is frustrated20:15
thingeeI believe you all tested it. I just haven't looked at the code myself. I know what it touches though, so it was my concern of something big like this might be scary. and again I'll repeat for 1000th time I was taking your own words of how you felt about it20:17
thingeeI'm not trying to twist your words. You wanted to ensure exisitng paths still work.20:18
hemnayes of course, I am the first person to tell you I don't want to break anything20:18
hemnabut I also don't want to work on this for another 3 releases20:18
openstackgerritErlon R. Cruz proposed openstack/cinder: Adds pool aware scheduling for HNAS drivers
hemnadoing the cinder <--> Nova, chicken and egg dance20:18
hemnait's just not worth it at a certain point.20:19
*** leeantho_ has quit IRC20:19
*** leeantho has joined #openstack-cinder20:19
thingeeI'll still review it, but it's my recommendation to wait to early L. I don't think the rebase and push immediate in L for testing a how dev cycle is really a bad idea.20:19
thingeetesting for a whole*20:19
hemnathen, I'm going to -2 the shadow tables patch then20:20
hemnabecause that causes a lot of problems for devs doing DB work20:20
xyangthingee: I think the problem is people always tend to wait until last minute.  so even if you target L-1, it could be pushed to L-3 again20:20
thingeexyang: can we not all commit to making this happen early for hemna ?20:21
hemnathe rebase with the shadow tables will be hell.20:21
xyangxyang: I support it, but that is not enought:)20:21
hemnakatco, I have to run to a meeting.  you guys decide what you think is right.20:22
thingeexyang: I really avoid making promises, but I would promise to review it early and get it merged. that would two +2's with the both us.20:22
*** hemna is now known as hemnafk20:22
thingeeif people are unsure about shadow tables this late, I agree as core you should say not right now.20:23
katcodid i get pinged on mistake?20:23
katco*by mistake?20:23
thingeekatco: yea20:23
katcoah ok20:23
katcohappy hacking :)20:23
e0nethingee: the main consern about shadow tables - it makes db migrations more complicated:(20:24
thingeejungleboyj, hemnafk: xyang: that's just my recommendation though. The community doesn't haven't to act on my thoughts and I'll still review this patch today, but I really would rather have us all commit to getting it merged as soon as the cinder version change commit goes through for L.20:24
*** rajinir has joined #openstack-cinder20:26
*** tbarron has joined #openstack-cinder20:26
openstackgerritBrianna Poulos proposed openstack/cinder: Add project_id to barbican keymgr wrapper
e0nei'm ok if we delay my patch for L, but this won't make db migration patches more easy after my patch lands20:26
rajinirIs there a plan to merge pending reviews before code freeze?20:26
e0nerajinir: what reviews do you mean?20:27
rajinirI want to get mine merged
thingeerajinir: just bug fixes, so that should be fine20:27
e0nerajinir: it's a bugfix. it will be acceptable after feature code freeze20:27
rajinirok, thanks20:27
thingeerajinir: see the full announcement here
jungleboyjI have to drop for a bit.  I am still in support of hemnafk 's patch going in now, but of course will help it go in first thing in L if that is what is decided.20:28
jungleboyjBack in about 30 min.20:29
*** tbarron has quit IRC20:30
*** IanGovett has joined #openstack-cinder20:30
*** aix has quit IRC20:31
thingeejungleboyj, hemnafk, xyang:
thingeeI've posted my thoughts, not a review yet.20:33
*** jungleboyj has quit IRC20:33
*** annashen has joined #openstack-cinder20:35
*** tbarron has joined #openstack-cinder20:35
thingeetbarron: you're running out of time with this :)20:38
xyangthingee: ok, thanks.20:41
*** Yogi1 has joined #openstack-cinder20:42
*** rlucio has joined #openstack-cinder20:45
*** Mandell has quit IRC20:46
*** akerr has quit IRC20:47
*** bswartz has quit IRC20:52
*** emagana has quit IRC21:02
*** Yogi1 has quit IRC21:02
*** emagana has joined #openstack-cinder21:02
*** jungleboyj has joined #openstack-cinder21:03
thingeeshow and tell happening at the ops midcycle meetup. lets see if Cinder gets mentioned =S21:04
jungleboyjthingee: :-)21:05
jungleboyjWhat else could they talk about?21:06
*** emagana has quit IRC21:07
*** lpetrut has quit IRC21:10
jungleboyjtbarron: How is your refactoring going?21:11
*** thangp has quit IRC21:14
*** leeantho_ has joined #openstack-cinder21:15
jungleboyjhemnafk: When you have a chance, can you look at this again:
*** leeantho has quit IRC21:17
*** dustins_ has quit IRC21:27
*** hodos has joined #openstack-cinder21:29
*** hemnafk is now known as hemna21:33
hemnaso jungleboyj looks like you need to update then  and it's ok ?21:33
jungleboyjYeah.  I am testing the patch right now.21:33
jungleboyjFor some reason a couple of those oslo.utils were inlcuded in the list of changes but didn't make it to the patch.21:34
* e0ne doesn't sleep yet21:37
e0nehemna: hi. do you hace question about multi attash or shadow tables?21:38
hemnaso, yah the shadow tables thing keeps bugging me21:38
e0newhat do you mean?21:39
hemnathe entire purpose of it is to allow an admin to offload the 'deleted' rows in the db to other tables21:39
hemnathe help performance, indexing etc yah?21:39
hemnaso, is there a way we can do that w/o putting 2x the burden on the cinder devs ?21:39
e0neand this is needed if admin/operator doesn't want to delete it at all21:39
*** hodos|2 has joined #openstack-cinder21:39
hemnawhy not a cinder-manage tool that does this in one shot21:39
hemnameaning it basically does this as a one time shot, right now, when an admin asks for it.21:40
hemnahave the tool ask for another db instance and then it goes through all the work of creating shadow tables for the existing schema and migrate the data21:41
hemnainstead of forcing every cinder patch that touches the DB to always keep 2 copies of everything in sync.21:41
hemnait's going to be hell for patches like multiattach21:42
hemnaso, the tool does all the work, when asked to do it21:42
hemnaand cinder doesn't actually create any shadow tables, and migration itself.21:42
hemnaonly the tool does it at the time the admin asks for it.21:42
e0nepersonaly, i like the idea to make such changes once per release21:42
*** hodos has quit IRC21:43
hemnathen there are no changes to cinder's development cycle, unit tests, db migration, db migration tests, etc.21:43
hemnawe could even add a unit test for the shadow table tool21:43
jungleboyjhemna: If there is a way to do that +221:43
e0nei'm not sure that modification production db schema is acceptable for many of operators21:43
hemnaI really don't like the idea of forcing cinder dev's to continually maintain cinder tables and shadow tables, schema, data21:44
hemnae0ne, if that's the case, then doing upgrades is also not acceptable21:44
hemnaas upgrades do exactly that.21:44
e0nebut you don't do upgrates often21:45
* jungleboyj is watching my son trying to learn to dive and can't figure out why kicking his ligs out is so baffling to him.21:45
e0neit's a planned maintenance21:45
hemnaevery time you upgrade a version of cinder.21:45
hemnaand that's also to my point.21:45
hemnaevery upgrade is going to automatically touch their data, and move their data21:45
hemnashouldn't we leave that up to the operator to decide when they want their data moved?21:45
*** chlong has joined #openstack-cinder21:46
hemnahence the tool21:46
hemnaput all the burden on the tool21:46
hemnanot on cinder21:46
hemnait seems very strange to force devs to be updating shadow tables21:47
hemnawhich are really archive tables a the time you archive the data.21:47
hemnabecause in the case of multiattach21:47
hemnawe are creating a new table, and moving data to it.21:47
hemnait doesn't make sense to have to migrate old shadow table data21:47
hemnawhen the new table never existed at the time those rows in the volume table were moved to the volume shadow table.21:48
openstackgerritMike Perez proposed openstack/cinder: Update Datera's export to avoid deprecated keys
hemnathat seems wrong to me.21:48
*** asselin_ has joined #openstack-cinder21:48
hemnaso a side affect of the tool would be that it only archives the scheme and data that exist in the main cinder db now.21:48
hemnameaning that if you run it today, you get shadow tables for today's schema.21:49
*** thingee has quit IRC21:49
hemnaand then when you run it 6 months from now, you get the schema of 6 months from now.21:49
openstackgerritJay Bryant proposed openstack/cinder: Namespace updates for _i18n and imageutils & fileutils
hemnaand there is no migration from today's shadow tables to 6+ months now shadow tables.21:49
hemnathe tool is crazy simple21:49
hemnait's basically what you have in the patch today.21:50
hemnaanyway, I just thought about this today in my frustration about the multiattach patch dying.21:50
*** pradip_vedams has quit IRC21:51
smcginnishemna: That does seem safer long term.21:51
*** kaufer has quit IRC21:51
*** pradip_vedams has joined #openstack-cinder21:51
hemnayah I think so as well21:51
jungleboyjYeah, good though and then it is only an impact for those who want to use it.21:51
e0nei feel like something went wrong with shadow tables21:51
hemnawell I think what shadow tables is trying to solve is good21:52
*** jdurgin has quit IRC21:52
hemnahelp DB performance from deteriorating over time.21:52
hemnano reason to index rows that will never come up21:52
jungleboyjYeah, it is a good idea.21:52
hemnabut adding 2x the burden on devs to maintain it forever forward, is a huge cost IMHO.21:52
*** asselin_ has quit IRC21:53
e0nehemna: db performance it's consequence rather than a cause21:53
e0newe've got db purge in master21:53
e0neas i understand, we need shadow tables for some cinder users who makes any analytics based on db data. e.g for billing21:54
hemnaand those same queries can run against the tool created db and tables21:55
e0nein such case, imo, we don't need uptodate shadow tables21:55
jungleboyjhemna: So you are saying they would go into a separate db?>21:56
jungleboyjI.E. cinder_shadow ?21:56
hemnayah could be21:56
hemnadepends on what options you pass into the tool21:56
hemnaadd db connectivity params and possible table prefix as another option21:56
jungleboyjAh, ok, Yeah, give the user flexibility.21:57
*** chlong has quit IRC21:57
hemnaso if they really wanted to, they could just use the cinder db, and change their prefix for every run21:57
hemna6 months from now21:58
*** chlong has joined #openstack-cinder21:58
hemnaand the tool ignores any table starting with _21:58
jungleboyjSo, new tables with that prefix would be created as the shadow tables?21:59
jungleboyjIn the existing DB.21:59
hemnaall existing tables in the current db (at the time of running the tool) would get an equivalent _09_2015_<table name> table22:00
jungleboyjIf they don't want to use a different prefix each time, can the data just be appended to existing tables?22:00
hemnaand it's archived data22:00
hemnajungleboyj, so no22:00
hemnabecause the schema changes22:00
hemnaand there in lies the rub22:00
hemnaand why it sucks to put this on devs now.22:00
hemnain the current patch set for shadow tables22:00
hemnaevery cinder developer that needs to do db table changes22:00
hemnaALSO has to do shadow table changes22:01
hemnain order for the current patchset's scheme to work.22:01
openstackgerritJoshua Harlow proposed openstack/cinder: Attempt to see what happens with close-fds patch
hemnawhich is very strange22:01
jungleboyjAh ... gotcha ... So basically turn it into a DB snapshot.22:01
hemnabecause you are changing a schema for a table who's data has already been archived, and hence it was never in that schema format.22:01
hemnavery very strange22:01
hemnajungleboyj, exactly!22:01
jungleboyjhemna: Seems reasonable.  One would hope what ever data analysis tooling they have can check more than one set of tables.22:02
anishany chance I get this reviewed pre-cutoff ?
anish(it had 1 +2, but lost it on rebase)22:03
jungleboyjanish: No hurry on that.  We aren't cutting off for bugs.22:04
*** chlong has quit IRC22:04
jungleboyjI will take a look later thought.  I have to drop of for a bit.22:05
*** eharney has quit IRC22:06
e0nehemna: i'm almost agree with you22:06
anishaah, cool. makes sense bugs aren't affected by cutoff (I just wasn't aware)22:06
hemnae0ne, :)22:06
*** jungleboyj has quit IRC22:06
hemnaanish, no unit tests22:06
hemnayet there are logic changes.22:07
hemnaand changes in expected exception behavior22:07
anisherm, this path is already covered by unit tests22:07
e0nehemna: one more question22:07
hemnathe new exceptions being raised where there was none previously?22:07
hemnaline 143 now raises an exception22:08
hemnawhere there was none previously22:08
anishI can add a test if you want22:08
hemnayes please22:08
anishbut I'm not sure how to fake errors22:08
hemnajust need to mock the conditions that lead to those errors22:08
e0nehemna: did you see
hemnathat you expect22:08
hemnae0ne, sure22:09
e0ne"Shadow tables could have blob field to store some "deleted" data and to not impose restrictions on database schema changes."22:09
hemnathat will get pulled when the tool runs as well.22:09
hemnablobs are ugly for archiving22:09
hemnabecause they can be very difficult to search/index on22:10
anishhemna: can you point me to an example of such a negative test ?22:10
*** gugl has joined #openstack-cinder22:10
hemnathere are all kinds of negative tests in test_volume.py22:11
e0nehemna: agree about index/search in almost all DBs22:11
anishhemna: to make sure I am getting this correctly. The test case would be raise foo error by mocking open, then asserting self.assertRaises(foo)22:14
anishwhich just seems like a generic way to check that the raise wasn't swallowed somewhere in the call path22:14
anishis that correct ?22:14
hemnathere are 2 conditions there22:15
hemnaone happens at IOError22:15
*** xyang has quit IRC22:15
hemnaif e_fnf.errno != 222:15
anishyes. I'll be checking both22:15
hemnaand a more general exception22:15
hemnashould test for both of those happening.22:15
anishko !22:15
hemnaa la raise a non IOError22:15
hemnamake sure you see that22:15
hemnaand then test the != 2 raising of the IOError22:16
hemnagood to go22:16
*** ganso_ has quit IRC22:17
*** hodos|2 has quit IRC22:17
*** rushiagr_away has quit IRC22:18
*** annegentle has quit IRC22:19
*** annegentle has joined #openstack-cinder22:21
*** theanalyst has quit IRC22:21
e0nehemna: i'm almost agree with you about developers impact. it's not a best solution22:21
e0nefrom other point of view, 'db snapshot tool' is out of scope of cinder. it should/could be based on oslo.db or be a stand-alone tool22:22
e0nei need to re-think current implementaion22:23
rhagartyeOne: if you get a chance, could you revisit - you note that you will +2 once the blueprint is approved, which it has22:23
*** openstack has joined #openstack-cinder22:26
*** annegentle has quit IRC22:26
*** annegentle has joined #openstack-cinder22:27
*** Thelo has quit IRC22:27
*** [1]Thelo is now known as Thelo22:27
e0nerhagarty: hi. i'll take a look on it tomorrow morning22:29
rhagartyThanks - much appreicated22:29
e0nehemna: i'll update my patch with comments tomorrow. i can't get better idea now than do something we have for migration tests in the past22:33
e0nei mean something like
*** rushil has quit IRC22:35
*** Mandell has joined #openstack-cinder22:36
hemnae0ne, also, I had just thought of another thing for the tool22:40
hemnaat the time you run the tool you know the schema version22:40
hemnayou can compare that to the schema version of any existing shadow tables22:40
hemnafrom the last run22:40
hemnaif they are the same, then bingo, you can just reuse the previous run's tables.22:41
*** IanGovett has quit IRC22:41
openstackgerritOpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements
*** jcru has quit IRC22:42
e0nehemna: i'm so sorry, need to sleep for a while :(22:51
hemnae0ne, no worries.  night man!22:51
*** lcurtis has quit IRC22:51
e0nei'll try to rebase your patch on my one and implement migrations as nova does22:51
e0neand wee need to think how it could be implemented in oslo.db22:52
hemnaok sounds good.  lets think about the tool idea as well22:52
e0neto try remove developers impact22:52
hemnathat's a really big burden22:52
e0nethe tool should be implemented out of scope of cinder, imho22:53
hemnacould be22:53
e0nebecause we've got the same troubles in other projects with soft deletes22:53
hemnaa general purpose tool/solution would be awesome in that case then.22:53
e0neabsolutely agree22:54
hemnaI know nova is polluted with shadow_tables22:54
hemnakinda sucks for doing raw db debugging22:54
hemnashow tables....100 tables22:54
*** esker has quit IRC22:54
anishif I'm mock_foo.side_effect = exception.IOError, anyone know how to set IOError.errno ?22:54
anish*if I'm setting22:54
*** bswartz has joined #openstack-cinder22:56
e0nehemna: so, is current "nova way" an option for you?22:57
hemnaI'm not sure I know what the nova way is22:57
hemnaI'd prefer to never see shadow tables in cinder22:57
hemnaand make it an admin thing22:57
hemnabut give admins the tools to do it.22:57
e0nei mean how migrations in nova is implemented22:58
hemnahonestly there is nothing from preventing smart admins that care from doing it today manually.22:58
hemnae0ne, I'm not overly familiar with how nova is doing it22:58
hemnais it the same as your current proposal?22:58
hemnaputting the changes in each db migration version?22:59
e0neyes, i've posted a sample:
hemnaso that's also modifying existing shadow tables22:59
hemnait looks like22:59
hemnawhich I think is really strange to do22:59
hemnaI think of them as snapshot tables23:00
hemnaso that mechanism works if all you are doing is adding a new column23:00
hemnabut when you are dropping a column from one table and moving it to another23:01
hemnakinda gets nasty23:01
hemna(which is what I have to do for multiattach)23:01
e0neunderstood, i'll test it tomorrow23:02
*** ronis has quit IRC23:02
*** annegentle has quit IRC23:03
* e0ne bb in ~10-12 hours23:03
*** Mandell has quit IRC23:05
*** rajinir has quit IRC23:07
*** tbarron has quit IRC23:07
*** hemna is now known as hemnafk23:08
*** e0ne has quit IRC23:13
*** annashen has quit IRC23:14
*** annashen has joined #openstack-cinder23:14
*** Mandell has joined #openstack-cinder23:15
*** openstack has joined #openstack-cinder23:25
vilobhmm1thangp : u there ?23:34
*** annashen has quit IRC23:39
*** annashen has joined #openstack-cinder23:39
*** Guest82873 has quit IRC23:41
*** chlong has joined #openstack-cinder23:43
*** annashen has quit IRC23:44
*** topshare has joined #openstack-cinder23:45
*** _cjones_ has quit IRC23:45
*** annashen has joined #openstack-cinder23:45
*** _cjones_ has joined #openstack-cinder23:48
*** topshare has quit IRC23:52
*** topshare has joined #openstack-cinder23:52
*** topshare has quit IRC23:53
*** tbarron has joined #openstack-cinder23:57
openstackgerritBob Callaway proposed openstack/cinder: NetApp implement lun and initiator os config options

Generated by 2.14.0 by Marius Gedminas - find it at!