Friday, 2015-08-28

*** sghanekar_ has quit IRC00:00
*** setmason has left #openstack-cinder00:00
*** breitz has quit IRC00:01
*** dims has joined #openstack-cinder00:03
*** bnemec has quit IRC00:03
*** salv-orlando has quit IRC00:04
*** dims has quit IRC00:05
*** dims has joined #openstack-cinder00:05
*** takedakn has quit IRC00:07
*** takedakn has joined #openstack-cinder00:08
*** vilobhmm1 has joined #openstack-cinder00:09
*** sghanekar has joined #openstack-cinder00:09
*** tsekiyam_ has joined #openstack-cinder00:09
*** dims has quit IRC00:10
*** mtanino has quit IRC00:12
*** gouthamr has quit IRC00:12
*** tsekiyama has quit IRC00:13
*** gouthamr has joined #openstack-cinder00:13
*** tsekiyam_ has quit IRC00:13
*** leeantho has quit IRC00:16
*** dims has joined #openstack-cinder00:17
*** takedakn has quit IRC00:22
*** gouthamr has quit IRC00:26
*** mudassirlatif has quit IRC00:32
*** akerr has joined #openstack-cinder00:33
*** garthb has quit IRC00:34
*** akerr_ has joined #openstack-cinder00:38
*** markvoelker has joined #openstack-cinder00:39
*** akerr has quit IRC00:42
*** esker has joined #openstack-cinder00:43
*** _cjones_ has quit IRC00:44
*** vilobhmm1 has quit IRC00:56
*** EinstCrazy has joined #openstack-cinder00:56
*** zhangjn has joined #openstack-cinder00:57
*** zhenguo has joined #openstack-cinder00:57
*** dannywilson has quit IRC01:03
*** lixiaoy1 has joined #openstack-cinder01:04
*** smoriya has joined #openstack-cinder01:10
winston-d_zZZhemnafk: ping01:11
*** winston-d_zZZ is now known as winston-d01:11
openstackgerritGorka Eguileor proposed openstack/cinder: Remove API races from delete methods
openstackgerritGorka Eguileor proposed openstack/cinder: Add atomic conditional updates to objects
openstackgerritGorka Eguileor proposed openstack/cinder: WIP: Remove more API races
openstackgerritGorka Eguileor proposed openstack/cinder: Improve metadata update operations
openstackgerritGorka Eguileor proposed openstack/cinder: Remove API races from attach and detach methods
openstackgerritGorka Eguileor proposed openstack/cinder: Move get_by_id to CinderObject
*** lixiaoy11 has joined #openstack-cinder01:15
openstackgerritWilson Liu proposed openstack/cinder: Enhance FC zone support for Huawei driver
*** lixiaoy1 has quit IRC01:17
*** akerr_ has quit IRC01:23
guitarzanjlk: nice!01:24
*** davechen has joined #openstack-cinder01:28
*** setmason has joined #openstack-cinder01:29
*** Guest62749 has joined #openstack-cinder01:34
lixiaoy11geguileo: Hi Gorka, are you still here?01:34
*** Guest62749 has quit IRC01:36
*** Lee1092 has joined #openstack-cinder01:39
*** akerr has joined #openstack-cinder01:45
*** sghanekar has quit IRC01:47
*** wilson1 has joined #openstack-cinder01:50
openstackgerritPatrick East proposed openstack/cinder: Replication V2 for Pure Storage
*** hw__chenying has joined #openstack-cinder01:53
openstackgerritVincent Hou proposed openstack/cinder: Volume status management during migration
*** wilson has quit IRC01:53
*** setmason has quit IRC01:58
*** p0rtal_ has quit IRC02:00
*** dims has quit IRC02:06
*** jwcroppe has joined #openstack-cinder02:07
*** boris-42 has quit IRC02:10
*** jwcroppe has quit IRC02:11
*** Apoorva has quit IRC02:19
*** haomaiwang has joined #openstack-cinder02:23
*** haomaiwang has quit IRC02:24
*** haomaiwang has joined #openstack-cinder02:25
*** angerson has quit IRC02:26
*** angerson has joined #openstack-cinder02:29
*** dims has joined #openstack-cinder02:33
*** sparr has quit IRC02:34
*** wilson-1 has joined #openstack-cinder02:34
*** wilson1 has quit IRC02:38
*** sparr has joined #openstack-cinder02:42
openstackgerritchenying proposed openstack/python-cinderclient: Add a period for the description string of a argument
*** haomaiwang has quit IRC02:44
*** hakimo has joined #openstack-cinder02:52
*** hakimo_ has quit IRC02:54
*** haomaiwang has joined #openstack-cinder02:58
*** haomaiwang has quit IRC03:01
*** haomaiwang has joined #openstack-cinder03:02
*** raildo-afk has quit IRC03:05
*** mriedem has quit IRC03:06
*** bill_az has quit IRC03:07
*** raildo-afk has joined #openstack-cinder03:08
*** haomaiwang has quit IRC03:18
*** dims has quit IRC03:23
*** vaidy has quit IRC03:27
*** daneyon has joined #openstack-cinder03:27
*** vaidy has joined #openstack-cinder03:27
*** swamireddy has quit IRC03:27
*** haomaiwang has joined #openstack-cinder03:27
*** tbarron has quit IRC03:28
*** daneyon_ has quit IRC03:28
*** mgagne has quit IRC03:28
*** swamireddy has joined #openstack-cinder03:29
*** salv-orlando has joined #openstack-cinder03:30
*** tbarron has joined #openstack-cinder03:32
*** mgagne has joined #openstack-cinder03:34
*** haomaiwang has quit IRC03:36
*** salv-orlando has quit IRC03:36
*** harlowja_at_home has joined #openstack-cinder03:40
*** akerr is now known as akerr_away03:42
*** subhadeep_de has joined #openstack-cinder03:47
*** akerr_away is now known as akerr03:48
*** akerr has quit IRC03:49
subhadeep_deHi Folks, I am quite new to Cinder project. Let me know what would be the best way to contribute03:49
scottdasubhadeep_de: Have you seen this:
subhadeep_deyes, i did few bug fixes in cinder client and nova for the cinder for volume api in Juno03:53
openstackgerritTina Tang proposed openstack/cinder: Clone cg support in VNX driver
subhadeep_deshould I start looking at the reviews for the existing bug fixes03:55
subhadeep_deto get more understanding on the ccinder driver?03:55
anishsubhadeep_de: you might find more things to fix here
*** jwcroppe has joined #openstack-cinder03:59
*** harlowja_at_home has quit IRC04:01
subhadeep_dethanks anish I will look into this, is this more of discovery module on the compute node04:02
*** hemna has joined #openstack-cinder04:05
anishsubhadeep_de: no, it combines the iscsi code from nova/cinder into a separate library04:07
anishit's relatively new so probably has more things to do than regular cinder04:07
*** links has joined #openstack-cinder04:09
subhadeep_deokay, thanks I will go through the library details. does it come deployed with devstack or we need to install it seperately?04:10
*** subhadeep_de has quit IRC04:14
openstackgerritPatrick East proposed openstack/cinder: Generic image-volume cache
*** martyturner has joined #openstack-cinder04:18
*** baojg has joined #openstack-cinder04:19
*** martyturner has quit IRC04:22
*** hw__chenying has quit IRC04:23
*** haomaiwang has joined #openstack-cinder04:26
*** vilobhmm1 has joined #openstack-cinder04:27
*** frickler has quit IRC04:30
openstackgerritVincent Hou proposed openstack/python-cinderclient: Volume status managment for voluem migration
openstackgerritPatrick East proposed openstack/cinder: Replication V2 for Pure Storage
*** subhadeep_de has joined #openstack-cinder04:35
subhadeep_deHi anish, does it come deployed with devstack?04:35
*** subhadeep_de has quit IRC04:36
*** subhadeep_de has joined #openstack-cinder04:37
*** BharatK has joined #openstack-cinder04:42
*** p0rtal has joined #openstack-cinder04:47
*** p0rtal has quit IRC04:48
*** p0rtal has joined #openstack-cinder04:48
*** haomaiwang has quit IRC04:50
*** deepakcs has joined #openstack-cinder04:56
*** subhadee_ has joined #openstack-cinder04:57
*** subhadee_ has quit IRC05:00
*** p0rtal has quit IRC05:00
*** subhadee_ has joined #openstack-cinder05:00
*** subhadeep_de has quit IRC05:01
*** subhadee_ has quit IRC05:02
*** theanalyst has quit IRC05:02
*** subhadeep_de has joined #openstack-cinder05:02
*** theanalyst has joined #openstack-cinder05:05
openstackgerritMitsuhiro Tanino proposed openstack/cinder: Allow specified backend capabilities to be retrieved
*** hemna has quit IRC05:15
*** setmason has joined #openstack-cinder05:15
*** setmason_ has joined #openstack-cinder05:18
*** setmason has quit IRC05:20
*** setmason_ is now known as setmason05:20
*** shyama has joined #openstack-cinder05:25
openstackgerritLisaLi proposed openstack/cinder: Data transfer between encrypted volumes and images
*** salv-orlando has joined #openstack-cinder05:31
*** haomaiwang has joined #openstack-cinder05:33
*** haomaiwang has quit IRC05:35
*** salv-orlando has quit IRC05:36
*** jwcroppe has quit IRC05:41
*** jwcroppe has joined #openstack-cinder05:42
*** skylerberg has quit IRC05:47
*** jwcroppe has quit IRC05:48
*** BharatK has quit IRC05:51
*** setmason has quit IRC05:51
*** setmason has joined #openstack-cinder05:52
*** agarciam has joined #openstack-cinder05:53
*** haomaiwang has joined #openstack-cinder06:02
*** BharatK has joined #openstack-cinder06:06
*** setmason has quit IRC06:17
*** salv-orlando has joined #openstack-cinder06:18
*** haomaiwang has quit IRC06:23
*** anshul has joined #openstack-cinder06:28
*** lixiaoy11 has quit IRC06:32
*** lixiaoy1 has joined #openstack-cinder06:32
*** baojg has quit IRC06:33
*** p0rtal has joined #openstack-cinder06:35
*** p0rtal has quit IRC06:35
*** p0rtal has joined #openstack-cinder06:35
*** ankit_ag has joined #openstack-cinder06:42
*** JoeStack has joined #openstack-cinder06:42
*** jwcroppe has joined #openstack-cinder06:44
*** lixiaoy11 has joined #openstack-cinder06:45
*** lixiaoy1 has quit IRC06:47
*** nkrinner has joined #openstack-cinder06:48
openstackgerritLisaLi proposed openstack/cinder: Raise HTTPBadRequest when invalid filters are given
openstackgerritLisaLi proposed openstack/cinder: Raise HTTPBadRequest for invalid filters
*** vilobhmm1 has quit IRC07:02
*** vincent_hou has joined #openstack-cinder07:03
*** ronis has joined #openstack-cinder07:05
*** p0rtal has quit IRC07:09
*** vincent_hou has quit IRC07:10
*** vincent_hou has joined #openstack-cinder07:11
*** stefan_amann has joined #openstack-cinder07:12
*** EinstCrazy has quit IRC07:13
*** zhangjn has quit IRC07:15
*** jwcroppe has quit IRC07:18
*** subhadee_ has joined #openstack-cinder07:21
*** subhadeep_de has quit IRC07:21
*** haomaiwang has joined #openstack-cinder07:38
*** bluex-pl has joined #openstack-cinder07:42
*** bluex-pl has quit IRC07:42
*** bluex-pl has joined #openstack-cinder07:43
openstackgerritXinXiaohui proposed openstack/cinder: Calculate virtual free capacity and notify
*** sgotliv has joined #openstack-cinder07:44
*** haomaiwang has quit IRC07:44
*** haomaiwa_ has joined #openstack-cinder07:45
*** haomaiw__ has joined #openstack-cinder07:49
*** haomaiwa_ has quit IRC07:50
*** ndipanov has joined #openstack-cinder07:51
*** ndipanov has quit IRC07:52
*** ndipanov has joined #openstack-cinder07:52
*** haomaiw__ has quit IRC07:58
*** ronis has quit IRC08:02
*** leopoldj has joined #openstack-cinder08:04
*** shyama has quit IRC08:05
*** yuriy_n17 has joined #openstack-cinder08:08
*** sgotliv has quit IRC08:11
*** jistr has joined #openstack-cinder08:12
*** jordanP has joined #openstack-cinder08:13
*** hw__chenying has joined #openstack-cinder08:17
openstackgerritVincent Hou proposed openstack/python-cinderclient: Volume migration progress
openstackgerritVincent Hou proposed openstack/python-cinderclient: Volume migration progress
*** ronis has joined #openstack-cinder08:24
*** IanGovett has joined #openstack-cinder08:25
openstackgerritVincent Hou proposed openstack/python-cinderclient: Volume status managment for volume migration
openstackgerritVincent Hou proposed openstack/python-cinderclient: Volume migration progress
*** lpetrut has joined #openstack-cinder08:34
*** lpetrut has left #openstack-cinder08:39
openstackgerritVincent Hou proposed openstack/cinder: Adds migration abortion
openstackgerritVincent Hou proposed openstack/cinder: Volume status management during migration
openstackgerritTakeaki Matsumoto proposed openstack/cinder: Fix url in API response to get original
*** shyama has joined #openstack-cinder08:47
openstackgerritVincent Hou proposed openstack/python-cinderclient: Volume status management for volume migration
openstackgerritVincent Hou proposed openstack/python-cinderclient: Volume migration progress
*** amit213 has quit IRC08:50
*** amit213 has joined #openstack-cinder08:51
*** vincent_hou has quit IRC08:52
*** baojg has joined #openstack-cinder08:53
*** e0ne has joined #openstack-cinder08:54
*** skraynev has quit IRC09:00
*** e0ne has quit IRC09:03
*** jordanP has quit IRC09:07
*** jordanP has joined #openstack-cinder09:07
*** takedakn has joined #openstack-cinder09:23
*** e0ne has joined #openstack-cinder09:25
*** subhadeep_de has joined #openstack-cinder09:25
*** subhadeep_de has quit IRC09:26
*** subhadee_ has quit IRC09:26
*** subhadeep_de has joined #openstack-cinder09:26
*** baojg has quit IRC09:26
*** skraynev has joined #openstack-cinder09:33
*** takedakn has quit IRC09:33
*** haomaiwang has joined #openstack-cinder09:34
*** skraynev has quit IRC09:38
*** salv-orlando has quit IRC09:42
*** davechen has left #openstack-cinder09:43
*** skraynev has joined #openstack-cinder09:45
openstackgerritLisaLi proposed openstack/cinder: Minor optimization
*** rushiagr_away is now known as rushiagr09:53
*** lixiaoy11 has quit IRC09:56
*** haomaiwang has quit IRC10:01
*** haomaiwa_ has joined #openstack-cinder10:02
*** bluex-pl has quit IRC10:06
*** bluex-pl has joined #openstack-cinder10:06
*** bluex-pl has quit IRC10:07
*** bluex-pl has joined #openstack-cinder10:08
*** bluex-pl has quit IRC10:08
*** bluex-pl has joined #openstack-cinder10:08
*** jwcroppe has joined #openstack-cinder10:22
*** lpetrut1 has joined #openstack-cinder10:25
*** haomaiwa_ has quit IRC10:27
*** yuriy_n17 has quit IRC10:27
*** mgagne has quit IRC10:27
*** swamireddy has quit IRC10:27
*** daneyon has quit IRC10:27
*** vaidy has quit IRC10:27
*** hakimo has quit IRC10:27
*** aix has quit IRC10:27
*** jgriffith has quit IRC10:27
*** jamielennox has quit IRC10:27
*** sayali has quit IRC10:27
*** d0ugal has quit IRC10:27
*** links has quit IRC10:27
*** tbarron has quit IRC10:27
*** markvoelker has quit IRC10:27
*** jwang has quit IRC10:27
*** Roamer` has quit IRC10:27
*** lprice1 has quit IRC10:27
*** afazekas has quit IRC10:27
*** leseb has quit IRC10:27
*** anteaya has quit IRC10:27
*** lpetrut1 is now known as lpetrut10:27
*** links has joined #openstack-cinder10:29
*** tbarron has joined #openstack-cinder10:29
*** markvoelker has joined #openstack-cinder10:29
*** lprice1 has joined #openstack-cinder10:29
*** jwang has joined #openstack-cinder10:29
*** Roamer` has joined #openstack-cinder10:29
*** afazekas has joined #openstack-cinder10:29
*** leseb has joined #openstack-cinder10:29
*** anteaya has joined #openstack-cinder10:29
*** lpetrut is now known as 18VAAKDGR10:29
*** haomaiwa_ has joined #openstack-cinder10:30
*** yuriy_n17 has joined #openstack-cinder10:30
*** mgagne has joined #openstack-cinder10:30
*** swamireddy has joined #openstack-cinder10:30
*** vaidy has joined #openstack-cinder10:30
*** daneyon has joined #openstack-cinder10:30
*** hakimo has joined #openstack-cinder10:30
*** aix has joined #openstack-cinder10:30
*** jgriffith has joined #openstack-cinder10:30
*** jamielennox has joined #openstack-cinder10:30
*** sayali has joined #openstack-cinder10:30
*** d0ugal has joined #openstack-cinder10:30
*** haomaiwa_ has quit IRC10:30
*** yuriy_n17 has quit IRC10:30
*** mgagne has quit IRC10:30
*** swamireddy has quit IRC10:30
*** daneyon has quit IRC10:30
*** vaidy has quit IRC10:30
*** hakimo has quit IRC10:30
*** aix has quit IRC10:30
*** jgriffith has quit IRC10:30
*** jamielennox has quit IRC10:30
*** sayali has quit IRC10:30
*** d0ugal has quit IRC10:30
*** yuriy_n17 has joined #openstack-cinder10:32
*** mgagne has joined #openstack-cinder10:32
*** swamireddy has joined #openstack-cinder10:32
*** vaidy has joined #openstack-cinder10:32
*** daneyon has joined #openstack-cinder10:32
*** hakimo has joined #openstack-cinder10:32
*** aix has joined #openstack-cinder10:32
*** jgriffith has joined #openstack-cinder10:32
*** jamielennox has joined #openstack-cinder10:32
*** sayali has joined #openstack-cinder10:32
*** d0ugal has joined #openstack-cinder10:32
*** hw__chenying has quit IRC10:34
*** zhenguo has quit IRC10:40
*** bluex-pl has quit IRC10:51
*** dimsum__ has joined #openstack-cinder10:53
*** takedakn has joined #openstack-cinder10:56
*** bluex-pl has joined #openstack-cinder11:01
*** bluex-pl has quit IRC11:02
*** tbarron has left #openstack-cinder11:03
*** tbarron has joined #openstack-cinder11:03
*** bluex-pl has joined #openstack-cinder11:03
*** bluex-pl has quit IRC11:04
*** bluex-pl has joined #openstack-cinder11:04
openstackgerritWilson Liu proposed openstack/cinder: Enhance FC zone support for Huawei driver
openstackgerritGorka Eguileor proposed openstack/cinder: Remove API races from delete methods
openstackgerritGorka Eguileor proposed openstack/cinder: Add atomic conditional updates to objects
openstackgerritGorka Eguileor proposed openstack/cinder: WIP: Remove more API races
openstackgerritGorka Eguileor proposed openstack/cinder: Move get_by_id to CinderObject
openstackgerritGorka Eguileor proposed openstack/cinder: Improve metadata update operations
openstackgerritGorka Eguileor proposed openstack/cinder: Remove API races from attach and detach methods
*** aix has quit IRC11:14
*** takedakn has quit IRC11:24
*** takedakn has joined #openstack-cinder11:26
*** takedakn has quit IRC11:35
*** porrua has joined #openstack-cinder11:39
*** shyama has quit IRC11:40
*** esker has quit IRC11:40
*** porrua has quit IRC11:46
*** wilson1 has joined #openstack-cinder11:47
*** wilson-1 has quit IRC11:50
*** julim has joined #openstack-cinder11:51
openstackgerritIvan Kolodyazhny proposed openstack/cinder-specs: Remove hardcoded releases list from unit tests
*** tpsilva has joined #openstack-cinder11:59
openstackgerritTina Tang proposed openstack/cinder: Clone cg support in VNX driver
*** rushiagr is now known as rushiagr_away12:03
*** jwcroppe has quit IRC12:06
*** deepakcs has quit IRC12:14
*** raildo-afk is now known as raildo12:16
*** martyturner has joined #openstack-cinder12:20
*** aix has joined #openstack-cinder12:24
*** akerr has joined #openstack-cinder12:28
*** salv-orlando has joined #openstack-cinder12:30
jordanPgeguileo, hi. Is someone working on the UT failure in VMware ?12:31
jordanPI am also affected by this12:31
*** salv-orlando has quit IRC12:32
geguileojordanP: UT?12:33
jordanPgeguileo, unit test12:33
geguileojordanP: You mean the _SentinelObject issue?12:33
jordanPgeguileo, yes12:34
*** chlong has joined #openstack-cinder12:34
geguileojordanP: I have been looking at it at it's not an easy problem  :-(12:34
jordanPyeah but it's blocking cinder...12:35
geguileojordanP: It only happens when tests are run concurrently12:35
geguileojordanP: And not always12:35
geguileojordanP: It appeared on the 24th12:35
erlongeguileo: does it use something in CONF?12:35
jordanPgeguileo, do you have an elasticsearch query already ?12:35
erlongeguileo: I have run into problems using conf before12:35
geguileoerlon: The problem is that a Mock suddenly is not into place12:36
erlongeguileo: ??12:36
geguileojordanP: I did a quick on with just the message12:36
geguileoerlon: You run the tests with ./test_run -> Everyting OK12:36
geguileoerlon: You run it with tox (uses parallel execution) -> Fails 1 every N runs12:37
erlongeguileo: they happen with --concurrency>1?12:37
geguileoerlon: Yes12:37
erlongeguileo: hmm12:37
geguileoerlon: And what happens is that a method that should have been Mocked, is not mocked12:37
eikkegeguileo: I believe test_create_backing_from_stream_optimized_file mock's builtin 'open', which is a global effect, and can indeed affect other tests under parallel execution (in a single python runtime)12:38
geguileoerlon: I think it is not a inter-test dependency12:38
erlongeguileo: hmm, ill try to reproduce here12:38
erlongeguileo: what is the tox command?12:38
jordanPthis line is suspicious also:
jordanPfile_open is the mock of builtin open12:39
eikkewhich is used as a context manager all the time in cinder.volume.drivers.vmware.vmdk.open12:39
geguileojordanP: Why suspicious?12:40
eikkeso within the context, tmp_file will be 'the file object returned by the open call'12:40
eikkeso if you have concurrently running tests, and another test performs 'open' (like, linecache does when looking at the backtrace), things go wrong12:40
jordanPgeguileo, that's not how I would mock 'open'12:40
geguileojordanP: Neither would I12:41
jordanPI would use mock.mock_open12:41
geguileojordanP: But that hasn't change lately (I believe)12:41
eikkejordanP: that'd have the same problem here, fwiw12:41
geguileojordanP: By the way, thanks for the Tempest review :-)12:42
*** bluex-pl has quit IRC12:42
*** BharatK has quit IRC12:42
*** bluex-pl has joined #openstack-cinder12:43
jordanPgeguileo, np :)12:43
jordanPthat volume backup is indeed broken12:43
*** bluex-pl has quit IRC12:43
geguileoerlon: jordanP I have to start testing migration patches that have been submitted, so I can't work more on that right now, I'll try to have a look later tonight if it isn't resolved by then12:43
*** bluex-pl has joined #openstack-cinder12:44
*** gouthamr has joined #openstack-cinder12:44
erlongeguileo: ok, I won't have that much time too, but ill try to give a look12:44
geguileoerlon: OK, thanks12:45
*** haomaiwa_ has joined #openstack-cinder12:45
*** zul has quit IRC12:49
*** Yogi1 has joined #openstack-cinder12:49
*** rushiagr_away is now known as rushiagr12:51
*** zul has joined #openstack-cinder12:52
*** takedakn has joined #openstack-cinder12:52
*** timcl has joined #openstack-cinder12:53
jordanPeikke, you said "<eikke> jordanP: that'd have the same problem here, fwiw"12:54
jordanPbut mock_open has a readlines() method12:55
jordanPguys a logstash query:
*** takedakn has quit IRC12:56
jordanP100% failure12:57
geguileojordanP: What?12:57
*** merooney has joined #openstack-cinder12:57
geguileojordanP: Is it 100% now?12:57
jordanPif the message "SentinelObject' object has no attribute 'readline" is in the console log then the job fails12:58
*** nkrinner has quit IRC12:58
jordanPI am not saying cinder is 100% blocked :)12:58
tbarronjordanP: geguileo: eikke:
openstackLaunchpad bug 1488690 in Cinder "test_create_backing_from_stream_optimized_file fails with "AttributeError: '_SentinelObject' object has no attribute 'readlines'"" [Undecided,In progress] - Assigned to Jon Bernard (jbernard)12:59
jordanPjust that my query is a good marker12:59
*** takedakn has joined #openstack-cinder12:59
tbarronjordanP: geguileo: eikke:
geguileotbarron: Thanks12:59
*** dustins has joined #openstack-cinder13:00
geguileotbarron: Although I never like skipping tests13:00
jordanP"nuke it from orbit"13:00
jbernardi dont' either, but ive really failed at getting it to fail locally13:00
tbarronjordanP: geguileo: eikke: also
*** haomaiwa_ has quit IRC13:01
geguileotbarron: Yeah, that's more like it :-)13:01
jordanPI am not sure13:01
tbarrongeguileo: but it doesn't work.  It was my guess too though.13:01
jordanPopen was mocked13:01
geguileoAlthough I'll have to talk with Eric13:01
geguileoBecause it should be the same thing13:01
geguileoAnd as far as I could tell13:02
geguileoThe problem is not there13:02
*** haomaiwang has joined #openstack-cinder13:02
jordanPSo the "skip" patch is stuck in the gate...13:02
*** nkrinner has joined #openstack-cinder13:03
tbarrongeguileo: jordanP: eikke: jbernard: so what eikke was saying above sounds plausible.13:03
erlongeguileo: what BEs will use to test the migration?13:04
geguileoerlon: I was just asking myself that same question   XD13:04
geguileoerlon: Probably LVM, NFS and Ceph13:04
erlongeguileo: lol13:04
tbarronI really don't like the way that test combines a positive and negative test case in the same test either.13:05
eikkeI think here's the problem13:05
erlongeguileo: Im not familiar with Ceph, does it need a hardware?13:05
jbernarderlon: nope13:05
eikke removed the use of fileutils13:05
geguileoerlon: It needs a cluster13:06
eikkebefore, was mocked, which doesnt affect 'open'13:06
eikkeafter, global 'open' is mocked13:06
geguileoerlon: Or I think devstack can deploy it as well (that one I haven't tried)13:06
tbarronwell, leaving while I drive off to work.  I'll catch up on the backlog when I get there.  eikke: sounds like you are on it!13:06
eikkebut now somewhere in the vmware driver (in _create_backing_from_stream_optimized_file), an exception is thrown13:06
eikkewhich is caught and re-thrown through oslo_utils' rethrow context manager13:06
eikkewhich then uses traceback & linecache, which uses open, which fails etc13:07
erlongeguileo: hmmm, we will probably need to test migration in our drivers as well, and I want to setup a environment with as many BE as possible13:07
eikkehiding the original exception13:07
erlongeguileo: will you use single or multi node?13:07
geguileoerlon: Single13:08
eikkejordanP: sorry, coming back to the fact mock_open() has a 'readlines' attribute: sure, but it won't return what linecache expects it to return. Really, mocking something global like 'open' isi rather bad practice, because someone in any call chain could require open to behave non-mocked, and then you're screwed13:09
geguileoeikke: Yes, something like pdb.set_trace()  ;-)13:09
eikkejordanP: remember how I advised you to mock 'as high up in the chain as possible'? ;-)13:10
*** aix has quit IRC13:10
openstackgerritOpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements
*** kbyrne has quit IRC13:13
jordanPpepperidge farm remembers13:15
*** jwcroppe has joined #openstack-cinder13:16
*** annegentle has joined #openstack-cinder13:17
openstackgerritEdmund Rhudy proposed openstack/cinder: Adds allow_availability_zone_fallback option to Cinder
*** kbyrne has joined #openstack-cinder13:19
*** kbyrne has quit IRC13:20
smcginnise0ne, geguileo, jgriffith: This looks good to me, but would like at least one more set of eyes on it:
eikkebon, added some comments to the bug report13:21
*** kbyrne has joined #openstack-cinder13:21
geguileosmcginnis: I'll have a look13:21
geguileosmcginnis: I reviewed previous path, so it should be quick13:21
smcginnisgeguileo: Yeah, just would like more sanity check. Thanks!13:22
*** asselin has joined #openstack-cinder13:22
e0nesmcginnis: thanks for reminder:)13:23
*** leopoldj has quit IRC13:24
*** breitz has joined #openstack-cinder13:28
*** links has quit IRC13:29
*** guest1 has joined #openstack-cinder13:29
*** aix has joined #openstack-cinder13:29
*** IanGovett has quit IRC13:30
geguileosmcginnis: Do you want me to +1 the workflow or do you want to do it yourself?13:31
*** IanGovett has joined #openstack-cinder13:31
*** merooney has quit IRC13:33
smcginnisgeguileo: If you're good with it then go ahead.13:34
geguileosmcginnis: Ok, done13:34
smcginnisgeguileo: Awesome13:34
*** bill_az has joined #openstack-cinder13:34
smcginnisgeguileo: Noticed your comment on there. The logger handles conversion if it's unicode.13:35
*** lcurtis has joined #openstack-cinder13:35
*** JoeStack has quit IRC13:36
*** mriedem has joined #openstack-cinder13:37
*** salv-orlando has joined #openstack-cinder13:37
geguileosmcginnis: Yep, I was thinking of %13:37
*** setmason has joined #openstack-cinder13:37
geguileosmcginnis: I'm a little overwhelmed with all the stuff I have to do  %-D13:38
*** setmason has quit IRC13:38
*** thangp has joined #openstack-cinder13:38
smcginnisgeguileo: Hah, yeah. Suddenly after L-3 it will slow down. :)13:39
geguileosmcginnis: I hope so!!13:39
*** merooney has joined #openstack-cinder13:39
smcginnisgeguileo: And then the summit and it all starts back up again.13:40
geguileosmcginnis: XD XD13:40
e0neguys, what are your thoughts about
e0nepersonaly, i'm going to -1 on it13:40
openstackgerritTina Tang proposed openstack/cinder: Clone cg support in VNX driver
*** setmason has joined #openstack-cinder13:41
smcginnise0ne: Looking...13:41
smcginnise0ne: Not sure if that is the right fix for what the bug is describing.13:42
*** dansmith is now known as superdan13:42
* smcginnis 's 20 second analysis13:42
e0neagree. posting my comments there13:42
smcginnisAnd updating pot file.13:43
e0nesmcginnis: great skilll "20s analysis for everything"13:43
smcginnisDon't touch the pot. Unless you're in one of those states.13:43
smcginnisSorry, that was bad. :D13:43
e0ne"I'm not a fun of the such solution at all. IMO, if we need to disable AZ, we need to do it in a scheduler or even disable AZ filter at all." - copy&paste from gerrit13:44
smcginnisTo you, not the patch. :)13:44
*** garthb has joined #openstack-cinder13:45
*** bnemec has joined #openstack-cinder13:46
*** edmondsw has joined #openstack-cinder13:48
*** dimsum__ has quit IRC13:49
*** dimsum__ has joined #openstack-cinder13:50
*** rushiagr is now known as rushiagr_away13:51
duleke0ne: Disabling AvailabilityZoneFilter will not work.13:52
duleke0ne: AZ is validated on an c-api level.13:52
*** eharney has joined #openstack-cinder13:53
*** guest1 has quit IRC13:53
e0nedulek: it means we've got other issue13:55
duleke0ne: Why do you think that?13:55
e0nedulek: if we have AvailabilityZoneFilter in scheduler, we must validate AZ in scheduler13:56
duleke0ne: AZ is validated in the API in the same way source_volid is validated.13:56
smcginnisWOw, the gate queue is going to take a long time to drain I think.13:57
duleke0ne: Besides you can skip scheduler if for example you're creating volume from source.13:57
*** merooney has quit IRC13:58
*** smoriya has quit IRC13:59
duleke0ne: Yup, then you have CONF.cloned_volume_same_az option - this is checked in API, because such call will bypass the scheduler.13:59
*** ankit_ag has quit IRC13:59
duleke0ne: That's why I think the fallback fix is valid *for this release*.13:59
e0nedulek: oh.. it's crazy14:00
duleke0ne: Otherwise we may create backward incompatibilities, which are probably unacceptable at this time in the cycle.14:00
*** setmason has quit IRC14:00
duleke0ne: I don't like it either. ;)14:00
openstackgerritwanghao proposed openstack/cinder: Add pagination to backups
*** haomaiwang has quit IRC14:01
e0nedulek: i mean our current AZ validation14:01
duleke0ne: I'm working on a spec for Mitaka to sort this stuff out.14:01
*** jwcroppe has quit IRC14:01
duleksmcginnis: Any chance of 20 second analysis of my previous messages? ;)14:01
*** haomaiwang has joined #openstack-cinder14:02
smcginnisdulek: Sorry, gotta wrap up something else first. Then I'll take a look. 20s with a few minutes latency. :)14:02
*** kaisers has quit IRC14:02
duleksmcginnis: Got it, superpowers have its cooldown time. ;)14:02
e0nedulek: adding new hack, that introduce new conf parameter which breaks AZ concept only to make deployment easier it not an option for me14:03
*** merooney has joined #openstack-cinder14:03
e0nedulek: it we need to get instance and volume in different AZs, we need to move validation from API to scheduler14:03
e0nedulek: but i don't understand such use case14:04
*** andreykurilin__ has quit IRC14:05
duleke0ne: Actually we don't want to do instance and volume in different AZs.14:08
e0nedulek: as i understand, it's more a deployment issue: to put c-vol in some AZ14:08
duleke0ne: We want Cinder to *ignore* AZs.14:08
duleke0ne: Problem was created when Nova started passing AZ to Cinder when creating a volume.14:08
duleke0ne: This was commit done back in Juno. Our ops run into problems with it when running internal setup here just a few weeks ago.14:09
duleke0ne: And since then 2 ML threads about the issue started + bugreport and fix by Edmund.14:10
e0nedulek: if we need igrore it, imo we should do it in a scheduler and remove validation from the api14:11
duleke0ne: Ah, we're running Fuel here internally, so fuel 6.1 is affected by the issue.14:11
smcginnisdulek: So not having the full background, if we want Cinder to ignore AZs, should we just stop doing it? What was the reason it was added?14:11
e0nedulek: i know about this bug in our distro:)14:11
duleke0ne: Sorry. ;)14:12
duleksmcginnis: We all know purpose of AZs, current concept works well for LVM, where c-vol acts like nova-compute.14:12
e0nedulek: but don't want to have a hack for it nor in our disto, neither in cinder master14:12
smcginnisdulek: So it's really only relevant in Cinder when using LVM?14:13
duleksmcginnis: Probably no, when we're running multiple c-vols with multiple backends we may have multiple AZs - one per c-vol.14:13
duleksmcginnis: Now there were no problems until this commit:
duleksmcginnis: Nova started to pass AZ to Cinder (hey, before that nobody cared!).14:14
smcginnisdulek: Just confused by this comment: < dulek> e0ne: We want Cinder to *ignore* AZs.14:14
openstackgerritdane-fichter proposed openstack/cinder: Data transfer between encrypted volumes and images
e0nedulek, smcginnis: why we can't move AZ filtering from API to scheduler like nova does?14:16
duleksmcginnis: Oh, sorry. So what I've meant is that you can have a deployment where you create AZs for VMs. But sometimes you don't want to care about AZs for volumes, because it is done on a different level.14:16
smcginnisdulek: Just seems overall broken in my opinion then.14:16
smcginnisWe care except when we don't, but we don't _really_ know when we're supposed to care or not.14:17
e0nesmcginnis: +114:17
duleksmcginnis: Well, it's option that administrator sets.14:17
duleksmcginnis: A way of saying - my Ceph have crash map properly configured.14:17
smcginnisdulek: This sets the option to fallback to a different zone. But what if it's one of the cases where we really should enforce it?14:18
duleksmcginnis: It depends on deployment - that's why fallback it's a configuration option.14:18
smcginnisStill doesn't seem right to me.14:19
dulekGetting back to e0ne's question - we cannot skip validation in API and do it in scheduler.14:19
dulekWe skip scheduler in some cases.14:19
dulekAnd we have CONF.cloned_volume_same_az which checks if requested AZ is same as volume being cloned, or snapshot we're creating a volume from.14:20
duleks/which checks/which enforces a check14:21
*** crose has joined #openstack-cinder14:21
duleksmcginnis: Well as a long term solution it probably isn't. We may stick to how it's done currently and accept that ops will hack Cinder manually.14:22
*** zhenguo has joined #openstack-cinder14:22
duleksmcginnis: That what ops at my company are doing and that's what bloomberg guys are doing also.14:22
smcginnisdulek: I'm a little concerned putting work arounds into the official code. It makes it easier to forget it's just a workaround.14:24
smcginnisdulek: But I see what you're saying.14:24
duleksmcginnis: I'm working on a spec to sort out AZs stuff in Cinder, but that's more work and impossible to be done in L>14:24
*** aix has quit IRC14:25
smcginnisdulek: Do you think it would be ready for M?14:25
duleksmcginnis: That's why I'm looking for a way to fix it now somehow.14:25
e0nedulek: did you try to move AZ filtering from api to scheduler? how much such changes affect cinder code?14:25
*** kvidvans has joined #openstack-cinder14:25
tbarroneikke: I added a note as followup to yours in
openstackLaunchpad bug 1488690 in Cinder "test_create_backing_from_stream_optimized_file fails with "AttributeError: '_SentinelObject' object has no attribute 'readlines'"" [Undecided,In progress] - Assigned to Jon Bernard (jbernard)14:25
duleksmcginnis, e0ne: Maybe there's better solution? I'm really open to ideas.14:25
duleke0ne: Okay, let me estimate that...14:25
smcginnisdulek: Sorry, I'm just a critic. :)14:26
duleksmcginnis: Constructive one is always welcomed. :)14:26
eikketbarron: any cclue why it succeeds when running tox on my local machine?14:26
tbarroneikke: no.  How much concurrency do you have, or do you not think that's relevant?14:27
eikkeat first I thought it'd be relevant, but I think I was wrong back then14:27
smcginnisdulek: I guess at a minimum I would like that changed from Closes-Bug to Partial-Bug so the actual issue isn't lost sight of.14:27
eikkebasically: if 'concurrent test execution' would be in a single process/address space, any mocking would be impossible14:27
*** crose has quit IRC14:27
guitarzandulek: it's an interesting problem I think14:28
eikketbarron: hence (and thats also how at least nose does it), tests run in multiple processes, and as such a mock in one testcase doesnt affect others14:28
guitarzandulek: setting up a cloud that cares about compute AZs but not volume AZs14:28
dulekguitarzan: I think Ceph users does it all the time. They have different mechanisms to achieve resiliency.14:28
eikketbarron: honestly, I think the question is more 'Why did it work in the past', rather than 'Why is it broken now'14:29
guitarzandulek: yes, but as DuncanT noted in the mailing list, you still have single points of failure, just not at the actual block device level14:29
guitarzanit is kind of breaking the basic idea of AZs, but I'm not saying I completely disagree with you :)14:30
tbarroneikke: and, as you asked, why does it work fine when run locally?14:30
dulekguitarzan: Have you seen my answer there? Can you comment if I'm reasonable.14:30
dulekguitarzan: I've meant - can you comment if answer is reasonable? ;)14:30
guitarzandulek: it doesn't really speak to DuncanT's point14:30
* smcginnis catches up on ML posts14:31
dulekguitarzan: So I'm probably mistaken somewhere. Can you elaborate?14:31
guitarzandulek: oh, I see what you're saying14:31
*** stefan_amann has quit IRC14:32
guitarzandulek: I agree14:32
guitarzanI read the wrong response :)14:33
*** timcl has quit IRC14:33
dulekGlad to hear that. :)14:33
dulekI'm thinking also if removal of these lines would hurt us much.14:34
dulekIn case of new volume creation - validation will be done by scheduler14:34
dulekIn case of validation if requested volume AZ matches the AZ of the snapshot the volume is created from - this is checked by lines 286-304.14:35
tbarroneikke: could it be a version-of-mock issue?  if my local version fills in readlines automagically but in gerrit we sometimes get a mock that refuses to do that for the SentinelObject?14:36
eikketbarron: that, I cant eexplain14:36
DuncanTdulek: Fair enough on the control plane if you don't aim for az resiliance at the control plane level, I guess14:36
dulekAh, there's the difference - user won't get feedback, that he provided wrong AZ.14:36
eikketbarron: ah, hmh, yeah14:36
DuncanTdulek: Thanks for taking a moment to explain14:36
smcginnisDuncanT: Aren't you supposed to not be working? :)14:36
dulekDuncanT: I do care, but don't think AZs are for that. :)14:36
tbarroneikke: I believe for a while we had a stricter version of mock that broke stuff for similar reasons14:37
DuncanTsmcginnis: Well, I'm balancing out some of the time that I should have been working ;-)14:37
smcginnisDuncanT: Hah!14:37
guitarzanDuncanT: haha!14:37
dulekDuncanT: :D14:37
eikketbarron: I wiped my .tox, so supposedly I should have the same mock version as gate, but still14:37
smcginnisOK, I guess if it's marked as a partial fix with a note in the bug that we need to evaluate the right approach, and you're working on an improved spec, it may be OK.14:38
DuncanTdulek: With a scheduler, db replica, api node and c-vol in every AZ, you can lose any one AZ (of 3, for example) and still be otherwise fully functional.... but not if all your c-vols are on one node14:38
*** aix has joined #openstack-cinder14:38
smcginnisStill don't like then that we would have config options we would need to deprecate and support for a few releases.14:38
guitarzanDuncanT: that's only true if HA cinder works14:38
DuncanTdulek: The cinder-has-no-AZs case still looks broken anyway14:38
openstackgerritMitsuhiro Tanino proposed openstack/cinder: Efficient volume copy for generic volume migration
DuncanTguitarzan: Nah, have 3 c-vols all treating ceph as if it is 3 separate backends14:39
guitarzanthere be dragons14:39
guitarzanthat's actually interesting though...14:39
dulekDuncanT: You would need 3 different API endpoints for Cinder to achieve that, but yes, it's possible.14:40
*** takedakn has quit IRC14:40
*** ronis has quit IRC14:40
dulekguitarzan: There won't be dragons I think. c-vol's should work in parallel connected to different backends.14:40
guitarzandulek: yeah, I caught up14:41
guitarzanexcept the idea of "different backends" is loaded here14:41
dulekDuncanT, guitarzan, smcginnis, e0ne: So what do you think of simple removal of these lines:
guitarzanthe problem is still that api calls aren't directed to an AZ14:42
guitarzandulek: would that actually help you?14:42
*** julim has quit IRC14:43
DuncanTguitarzan: An LB will just fail over to whatever is left up in terms of API servers...14:43
dulekguitarzan: Definitely. That's the place where it fails. Then if we're going through scheduler - it will validate the AZ. If not - validation will be done in API few lines lower (if cloned_volume_same_az is True).14:43
dulekThe only difference will be that user won't get feedback that the AZ he provided is incorrect.14:44
DuncanTdulek: I'll need to have a think about the consequences of that14:44
*** mtanino has joined #openstack-cinder14:44
dulekDuncanT: One I've just mentioned. Second would be when you have AZFilter disabled - nothing will validate if AZ is correct. But if you disabled it, then you probably don't care.14:45
*** kbyrne has quit IRC14:45
*** timcl has joined #openstack-cinder14:46
guitarzanDuncanT: so you imagine the api sending messages to any rabbit in any AZ?14:46
tbarroneikke: yeah, my theory doesn't hold.  I have mock 1.3.0 with requirements mock >= 1.2.  And that's exactly what I see with the failure with eharney's fix attempt:
mtaninosmcginnis: Hi, Could you visit this again? I've fixed some nits.14:47
tbarroneikke: much as I don't like skipping tests, I guess this one should be pulled and re-written.14:47
dulekguitarzan: Ah, good point. RabbitMQ need to be common. ;)14:47
*** setmason has joined #openstack-cinder14:47
eikketbarron: a patch to skip the test is waiting in gate14:47
eikkefor good reason14:47
eikke(not the waiting part :P)14:48
tbarroneikke: yeah, but it would be better to fix the problem if we could.14:48
*** aix has quit IRC14:48
tbarroneikke: that said, it has my +114:48
eharneyunfortunately i can't seem to even hit it locally14:48
tbarroneharney: me either.14:49
eikkemaybe a fix could be to mock save_and_reraise_exception :-P14:49
eikkeif the exception path is actually to be taken14:49
smcginnismtanino: Thanks! Just waiting for Jenkins.14:49
mtaninosmcginnis: I see. Thank you :)14:49
eharneytbarron: i think i just spotted another problem looking at it, though14:53
tbarroneharney: ?14:54
eharneytbarron: let me try a couple things, will hopefully submit another patch14:55
*** annasort has joined #openstack-cinder14:55
*** jungleboyj has joined #openstack-cinder14:56
*** rushiagr_away is now known as rushiagr14:57
*** ebalduf has joined #openstack-cinder14:57
*** kbyrne has joined #openstack-cinder14:57
openstackgerritXinXiaohui proposed openstack/cinder: Calculate virtual free capacity and notify
*** aix has joined #openstack-cinder15:00
*** haomaiwang has quit IRC15:01
openstackgerritTakeaki Matsumoto proposed openstack/cinder: Fix url in API response to get original
*** haomaiwang has joined #openstack-cinder15:02
DuncanTguitarzan: Rabbit would be a cluster, and normally things would work pan-az, just means you can loose a whole AZ and not loose access to the other volumes15:02
*** julim has joined #openstack-cinder15:02
*** skraynev has quit IRC15:03
*** mtanino has quit IRC15:03
guitarzanDuncanT: yeah, I run clustered rabbit. I'm just trying to find the mental block in my brain :)15:04
DuncanTguitarzan: I tried running a clustered brain but the rest of it voted me out of the cluster15:05
*** skraynev has joined #openstack-cinder15:05
guitarzanI think my brain isn't liking the idea of a volume that isn't actually offline15:05
*** hemnafk is now known as hemna15:06
guitarzanbut this idea seems pretty good I think15:06
*** garthb has quit IRC15:06
openstackgerritRyan McNair proposed openstack/cinder: Fix MITM vulnerability for Brocade FC SAN lookup
*** boris-42 has joined #openstack-cinder15:13
*** zhenguo has quit IRC15:14
SwansonGot a really minor change that could impact some random CI failures.  I'd like to have it running over the weekend.  1 +2 from hemna so it just needs another push.
*** edmondsw has quit IRC15:19
*** edmondsw has joined #openstack-cinder15:20
openstackgerritJohn Griffith proposed openstack/cinder: get_replication_updates call to driver is wrong
*** jdurgin1 has joined #openstack-cinder15:21
openstackgerritEric Harney proposed openstack/cinder: Tests: Fix mock/sentinel use in vmdk tests
*** jaypipes has joined #openstack-cinder15:23
*** jaypipes is now known as leakypipes15:24
*** merooney has quit IRC15:24
*** edtubill has joined #openstack-cinder15:25
openstackgerritWalter A. Boring IV (hemna) proposed openstack/os-brick: Add new Connector APIs for path validation
*** JoeStack has joined #openstack-cinder15:31
*** skraynev has quit IRC15:32
*** skylerberg has joined #openstack-cinder15:33
*** skraynev has joined #openstack-cinder15:33
openstackgerritdane-fichter proposed openstack/cinder: Data transfer between encrypted volumes and images
*** timcl has quit IRC15:36
dulekDuncanT: Just read joke on clustered-brain. Good one! ;)15:36
sdagueis anyone addressing the cinder vmware unit tests being racey - that's causing a bunch of gate resets?15:37
smcginnissdague: Yep, patch to skip going through soon.15:39
sdaguesmcginnis: is it posted already?15:39
sdaguewe can put that into the promote list15:39
smcginnissdague: There was some discussion earlier this morning. Let me track it down. tbarron, do you have that handy?15:39
*** chlong has quit IRC15:39
*** skraynev has quit IRC15:40
*** bluex-pl has quit IRC15:40
openstackgerritThang Pham proposed openstack/cinder: Conversion to volume object
openstackgerritThang Pham proposed openstack/cinder: Sync volume versionedobject to ORM
Swansoneharney: Thanks!15:41
*** p0rtal has joined #openstack-cinder15:41
*** p0rtal has quit IRC15:41
smcginnissdague: I believe this is it:
*** p0rtal has joined #openstack-cinder15:41
smcginnisWay down in the gate queue.15:42
sdaguesmcginnis: thanks15:42
smcginnissdague: np15:43
*** mtanino has joined #openstack-cinder15:46
openstackgerritTakeaki Matsumoto proposed openstack/cinder: Fix url in API response to get original
*** ronis has joined #openstack-cinder15:47
*** markstur_ has quit IRC15:47
*** yuriy_n17 has quit IRC15:49
*** skraynev has joined #openstack-cinder15:50
*** rushiagr is now known as rushiagr_away15:51
*** p0rtal has quit IRC15:51
*** ircuser-1 has quit IRC15:52
*** markstur has joined #openstack-cinder15:54
*** kvidvans has quit IRC15:55
*** haomaiwang has quit IRC15:58
*** haomaiwang has joined #openstack-cinder15:58
*** p0rtal has joined #openstack-cinder15:59
*** p0rtal has quit IRC16:00
*** p0rtal has joined #openstack-cinder16:00
*** haomaiwang has quit IRC16:01
*** jordanP has quit IRC16:01
*** haomaiwang has joined #openstack-cinder16:02
*** p0rtal has quit IRC16:03
*** sghanekar has joined #openstack-cinder16:03
*** p0rtal has joined #openstack-cinder16:03
*** jdurgin1 has quit IRC16:07
*** skraynev has quit IRC16:10
*** leakypipes has quit IRC16:11
*** vilobhmm1 has joined #openstack-cinder16:14
*** rhedlind has joined #openstack-cinder16:15
*** dannywilson has joined #openstack-cinder16:17
openstackgerritRyan McNair proposed openstack/cinder: Fix MITM vulnerability for Brocade FC SAN lookup
*** _cjones_ has joined #openstack-cinder16:19
*** 18VAAKDGR is now known as lpetrut116:20
*** leeantho has joined #openstack-cinder16:20
hemnalooks like the Mellanox CI is totally broken16:20
*** agarciam has quit IRC16:20
*** merooney has joined #openstack-cinder16:21
*** merooney has quit IRC16:21
*** chlong has joined #openstack-cinder16:22
*** merooney has joined #openstack-cinder16:22
*** skraynev has joined #openstack-cinder16:23
*** garthb has joined #openstack-cinder16:23
*** agarciam has joined #openstack-cinder16:24
guitarzanhemna: quick question, how do I know what interface nova and/or brick are going to use for iscsi?16:25
guitarzanhemna: hmm, actually this might be a matter of what my connection hash looks like... I'll check :)16:26
*** timcl has joined #openstack-cinder16:26
hemnathe outbound interface ?16:26
hemnaI presume that's a routing table issue no ?16:26
*** e0ne has quit IRC16:28
*** panbalag has joined #openstack-cinder16:28
guitarzanhemna: ya, i think so, I just need to figure out which of devstack's interfaces to use on the iscsi target side16:29
*** annegentle has quit IRC16:31
SwansonIf your job fails in gate you should have to mail donuts to everyone behind you in gate at the time.16:32
angela-shemna: i retriggered the CI job yesterday, still waiting in the queue, quite a backlog.  the last CI run for this was before we fixed the format of the comment message, so that's why it doesn't show up in the right hand side16:33
*** vilobhmm1 has quit IRC16:33
*** dflorea has joined #openstack-cinder16:34
panbalagHi.. can someone let me know the command to do a cinder volume retype?16:35
panbalagI don't see a 'cinder retype' in the command list.16:35
*** jistr has quit IRC16:35
*** salv-orlando has quit IRC16:36
*** esker has joined #openstack-cinder16:37
jgriffithpanbalag: maybe check "cinder --version" ?  perhaps your on Havanna ?16:37
panbalagjgriffith: I'm on 1.2.1 ..16:38
*** mriedem is now known as mriedem_lunch16:38
jgriffithpanbalag: then it should be there, unless it's disabled in policy file16:38
*** esker has quit IRC16:38
SwansonDid they recently disable it?  Cause it is showing up as not a valid subcommand when I do a cinder help retype16:39
jgriffithpanbalag: or if we borked something... I'm on 1.1.116:39
panbalagjgriffith: when I try "cinder help retype'" I get the error " not a valid subcommand" ...16:39
*** esker has joined #openstack-cinder16:39
panbalagjgriffith: how do I enable it ?16:39
*** annegentle has joined #openstack-cinder16:40
jgriffithpanbalag: not sure, it is in fact broken, looking now16:40
openstackgerritJay Bryant proposed openstack/python-cinderclient: Volume status management for volume migration
*** earlephilhower has joined #openstack-cinder16:40
earlephilhowerIs there a way to get the list of Jenkins merges w/timestamps for the entire OpenStack github?  Something changed 28hrs ago and it's causing 100% failure rate on my CI now. :(16:43
*** setmason_ has joined #openstack-cinder16:43
jgriffithpanbalag: Swanson cinder --os-volume-api-version 2 retype16:44
*** jwcroppe has joined #openstack-cinder16:44
* jgriffith honestly thought that the whole v1-v2 train wreck had been fixed :(16:44
panbalagjgriffith: That worked..Thanks16:44
*** chlong has quit IRC16:45
*** setmason has quit IRC16:45
*** setmason_ is now known as setmason16:45
jgriffithWhy the F are we setting V1 as default again ?16:45
Swansonjgriffith: Hmmmm...  I thought v2 was the default16:45
jgriffithSwanson: me too :)16:45
jgriffithSwanson: it appears we're messed up in devstack deploy again though16:46
*** agarciam has quit IRC16:46
jgriffithSwanson: honestly this whole thing has been a nightmare, between the keystone changes and the fact that we haven't been pushing cinderclient and testing it well16:46
*** agarciam has joined #openstack-cinder16:46
jgriffithSwanson: I have something to look at during bug fix mode :)16:46
Swansonjgriffith: :)  And here I was just worrying about seeing an openstack-dev/grenade change in gate ahead of me.  This should give me something else to be concerned about.16:47
tbarronsmcginnis: sdague: sorry, was at lunch.  As smcginnis said:
smcginnistbarron: Hah, no worries. Just knew you were part of that discussion so I thought I would ping you while I looked in case you still had it handy.16:48
sdaguetbarron: yeh, smcginnis hooked me up. We promoted that in the gate queue to hopefully reduce overall failure rate16:50
*** p0rtal has quit IRC16:51
tbarronsmcginnis: sdague: excellent, and thanks.16:51
*** Apoorva has joined #openstack-cinder16:51
*** gouthamr has quit IRC16:52
*** Apoorva has quit IRC16:52
eikkejay, gate succss for the py27 run of the vmdk skip patch16:52
*** agarciam has quit IRC16:53
*** Apoorva has joined #openstack-cinder16:53
*** agarciam has joined #openstack-cinder16:53
jgriffitheikke: ?16:53
eikkejgriffith: just that gate-cinder-python27 passed for 2178824, hopefully other jobs soon as well16:54
jgriffitheikke: ahh... got ya :)16:54
jgriffithIsn't this like the second or third time the vmware unit tests have hosed us?16:54
jgriffitheharney: ^^16:54
*** subhadeep_de has quit IRC16:57
*** mudassirlatif has joined #openstack-cinder16:57
*** timcl has quit IRC16:59
*** JoeStack has quit IRC17:00
*** mtanino has quit IRC17:00
*** haomaiwang has quit IRC17:01
smcginnisjgriffith: I'll beat them up next week when I'm out there. ;)17:01
raildohey guys, quickly question... I saw this documentation about cinder client explaning how to use the quota operations
jgriffithsmcginnis: haha17:02
raildoanyone know how can I do this in the openstack client?17:02
*** haomaiwang has joined #openstack-cinder17:02
raildois this already implemented in openstack client? I only find the quota commands for nova, and not for cinder =/17:03
hemnaangela-s, ok I'll wait then and watch.17:05
*** julim has quit IRC17:06
smcginnisraildo: You're talking python-openstackclient, not python-cinderclient, right?17:06
raildosmcginnis: python-openstackclient17:06
smcginnisraildo: I don't have much experience with that one.17:06
raildosmcginnis: btw I found this code
smcginnisraildo: It's possible it hasn;t been added there yet.17:07
jgriffithraildo: it looks like it might be in master... looking now17:07
*** setmason has quit IRC17:09
*** julim has joined #openstack-cinder17:09
*** setmason has joined #openstack-cinder17:10
*** erhudy has joined #openstack-cinder17:10
angela-shemna: thanks, or i'll ping you once it runs.17:11
hemnaangela-s, ok thanks for letting me know.17:11
raildosmcginnis: jgriffith hum.. I think that the 'opoenstack quota <command>' works for cinder too:
ericksonsantosraildo, I think that is here that openstackclient gets the quota for volume:
*** salv-orlando has joined #openstack-cinder17:12
raildoericksonsantos: yes! let's do test this :)17:12
ericksonsantosraildo, sure17:12
*** vilobhmm1 has joined #openstack-cinder17:13
vilobhmm1jungleboyj, smcginnis : thanks ! :)17:14
smcginnisvilobhmm1: np17:15
smcginnisraildo, ericksonsantos: Looks right!17:15
*** timcl has joined #openstack-cinder17:15
*** jlk has left #openstack-cinder17:16
*** e0ne has joined #openstack-cinder17:18
openstackgerritThang Pham proposed openstack/cinder: Conversion to volume object
ericksonsantossmcginnis, nice!17:18
*** lpetrut1 has quit IRC17:19
*** subhadeep_de has joined #openstack-cinder17:19
openstackgerritAnna Sortland proposed openstack/cinder: Earlier authority check for create volume API
*** dflorea has quit IRC17:20
*** aix has quit IRC17:20
*** dflorea has joined #openstack-cinder17:25
*** dflorea has quit IRC17:30
*** dflorea has joined #openstack-cinder17:31
*** dflorea has quit IRC17:32
*** Yogi1 has quit IRC17:33
*** changbl has quit IRC17:33
*** e0ne has quit IRC17:33
*** dflorea has joined #openstack-cinder17:35
jungleboyjvilobhmm1: Welcome!17:35
smcginnisCool, later.17:36
smcginnisOops. :-D17:36
*** ronis has quit IRC17:36
smcginnisSplit screen is going to kill me.17:36
raildosmcginnis: I have this same problem, some times I cam here and put "vim"17:37
smcginnisraildo: Yeah, that too. :)17:37
*** mriedem_lunch is now known as mriedem17:40
*** jungleboyj has quit IRC17:41
*** anshul has quit IRC17:42
*** vgridnev has joined #openstack-cinder17:43
*** merooney has quit IRC17:44
smcginnisAnyone else getting this on fresh devstack deployments: functions-common:466 Cloning not allowed in this configuration17:45
*** jaypipes has joined #openstack-cinder17:46
*** merooney has joined #openstack-cinder17:47
*** p0rtal has joined #openstack-cinder17:48
*** ebalduf has quit IRC17:49
*** ebalduf has joined #openstack-cinder17:50
*** Yogi1 has joined #openstack-cinder17:50
*** jwcroppe has quit IRC17:52
sdaguesmcginnis: so the cinder unit test fix died on a different cinder unit test race -
sdagueI'm about to call it a week, but that should get investigated17:52
*** mudassirlatif_ has joined #openstack-cinder17:54
*** e0ne has joined #openstack-cinder17:55
smcginnissdague: Thanks for the heads up. We'll look in to it.17:56
smcginnissdague: Have a good one.17:56
*** mudassirlatif has quit IRC17:57
*** mudassirlatif_ is now known as mudassirlatif17:57
*** ronis has joined #openstack-cinder17:57
smcginnisHmm, haven't seen that one before"17:59
smcginnisoslo_db.exception.DBConnectionError: (psycopg2.OperationalError) unknown error17:59
*** haomaiwang has quit IRC18:01
*** esker has quit IRC18:01
*** 20WABZC7L has joined #openstack-cinder18:02
tbarronsmcginnis: sdague: so can we do a recheck on 217882 right now and get it pushed to the head of queue again?18:02
smcginnistbarron: That's what I was just looking in to. :)18:03
tbarronsmcginnis: excellent18:03
smcginnistbarron: Doesn't look like it. :[18:04
tbarronsmcginnis: how did you find that answer?  ask on infra?18:04
smcginnistbarron: I'm a guy, I don't ask directions.18:04
smcginnistbarron: Just reading this:
smcginnistbarron: And trying things.18:05
tbarronsmcginnis: but what did sdague do last time?  my concern is that this one is sitting behind a bunch or rechecked jobs that are just going to fail again.18:06
smcginnistbarron: OK, it's still in the gate queue, but I was able to also kick it off in the check queue. We'll see if that will do it.18:06
smcginnistbarron: Yeah, once this makes it back to the gate queue we should see if someone in infra can do whatever he did to bump it up.18:07
*** shyama has joined #openstack-cinder18:10
SwansonGate appears to be running about 24 hours today.18:10
* tbarron watche #openstack-infra with admiration18:10
*** timcl has quit IRC18:13
*** kvidvans has joined #openstack-cinder18:13
tbarronsmcginnis: ty!18:13
smcginnistbarron: Thank you for the suggestion!18:14
smcginnistbarron: And chiming in over there.18:14
*** aix has joined #openstack-cinder18:14
smcginnisNow we know for the future if something like this happens again.18:15
smcginnisInfra to the rescue.18:15
*** lprice1 has quit IRC18:16
hemnajobs launched just took a nosedive18:16
smcginnisOh joy.18:16
*** dflorea has quit IRC18:16
tbarronyeah, we should only bug like that infrequently, but this one got workflow yesterday at 6:07 pm ET, tests are failing because of it quite frequently, and people are issuing rechecks on those failures which just fail all over again.18:17
*** dflorea has joined #openstack-cinder18:17
smcginnisYeah, definitely only when really needed.18:18
eikkegate 2.7 failed on 217882 :(18:19
*** shyama has quit IRC18:20
*** mtanino has joined #openstack-cinder18:23
smcginniseikke: Getting requed18:25
hemnacan we spike the football with another +A ?18:25
smcginnishemna: Already talked to infra and they are going to get it going again once the current run fails.18:26
smcginnisUnforatunately that looks like it will be at least 30 minutes out.18:26
tbarronlprice: all we can do right now is check if you are failing for some reason other than
openstackLaunchpad bug 1488690 in Cinder "test_create_backing_from_stream_optimized_file fails with "AttributeError: '_SentinelObject' object has no attribute 'readlines'"" [Undecided,In progress] - Assigned to Jon Bernard (jbernard)18:31
tbarronsorry, wrong window18:32
eikkeI'm still puzzled what causes that to be triggered so consistently now18:32
*** p0rtal_ has joined #openstack-cinder18:35
*** dflorea has quit IRC18:37
openstackgerritEric Harney proposed openstack/cinder: Remove executable bits on files
openstackgerritEric Harney proposed openstack/cinder: Detect addition of executable files
*** p0rtal has quit IRC18:38
*** jaypipes has quit IRC18:40
openstackgerritMitsuhiro Tanino proposed openstack/cinder-specs: Fix Get Volume Driver Capabilities Spec
*** skraynev has quit IRC18:41
*** e0ne has quit IRC18:42
*** skraynev has joined #openstack-cinder18:42
smcginnistbarron, hemna, eikke: Been requeued in the gate queue at the top of the heap!18:43
tbarronsmcginnis: :-)18:44
SwansonThat poor nova change.  It is going to go over 24 hours in the queue.18:44
hemnatop of the chain baby18:45
*** mudassirlatif has quit IRC18:47
*** xyang1 has joined #openstack-cinder18:52
*** mudassirlatif has joined #openstack-cinder18:55
*** hodos has joined #openstack-cinder18:55
smcginnispy27 run has passed this time. Don't want to jinx it, but things are looking good.18:56
*** timcl has joined #openstack-cinder18:56
hemna27 and 34 passed18:57
*** rady has joined #openstack-cinder19:00
* tbarron is looking the other way, fingers crossed19:00
*** 20WABZC7L has quit IRC19:01
*** haomaiwang has joined #openstack-cinder19:02
openstackgerritxing-yang proposed openstack/cinder: Fix a merge problem in VMAX driver
*** dannywilson has quit IRC19:09
smcginnisxyang1: ping19:11
xyang1smcginnis: Hi19:11
smcginnisxyang1: Hey! I saw you signed up (or were signed up) for this patch:
smcginnisxyang1: Jay and I have +2'd it, but I didn't want to +A until I talked to you.19:12
*** tpsilva has quit IRC19:12
smcginnisxyang1: Just wanted to see if you had any input on it first or if you wanted to review it yet.19:12
xyang1smcginnis:  I have not looked at the latest.  I can take a look19:13
smcginnisxyang1: Cool, thanks!19:14
xyang1smcginnis: Np19:14
*** dannywilson has joined #openstack-cinder19:17
vgridnevhey cinder folks, could you please review the change: . It's required for the future BDD CI19:18
*** Lee1092 has quit IRC19:23
*** Apoorva_ has joined #openstack-cinder19:25
*** Apoorva has quit IRC19:25
*** hodos|2 has joined #openstack-cinder19:28
*** hodos has quit IRC19:31
*** changbl has joined #openstack-cinder19:37
*** devlaps has joined #openstack-cinder19:37
*** IanGovett has quit IRC19:39
*** Apoorva_ has quit IRC19:41
smcginnistbarron, hemna, eikke: It made it!19:41
*** rushil has joined #openstack-cinder19:42
*** eharney has quit IRC19:43
*** thangp has quit IRC19:46
*** Apoorva has joined #openstack-cinder19:47
*** bill_az has quit IRC19:51
*** Apoorva has quit IRC19:51
panbalagjgriffith: do you know anyone who works on cinder retype? I'm receiving an exception when trying to retype.. the retype is successful, but the source volume is not deleted..I see an exception in wondering if anyone here who could help19:51
jgriffithpanbalag: so you're doing a retype with migration?  Migration doesn't work19:52
SwansonIt doesn't?19:52
jgriffithpanbalag: or depending on who you ask "it's just misunderstood"19:52
panbalagjgriffith: I used "cinder --os-volume-api-version 2 retype --migration-policy on-demand vol2 nfs"19:53
jgriffithpanbalag: Swanson here's  a few patches trying to fix things:,n,z19:54
*** rushil has quit IRC19:54
*** rushil has joined #openstack-cinder19:54
jgriffithpanbalag: You might want to check with vincent; I've sort of washed my hands of that at this point and tell most people "just don't do it"19:55
jgriffithpanbalag: we'll likely get it fixed up for this release, but currently it's sort of wonkie19:55
Swansonjgriffith:  poor vincent.  I swear to god it worked at one point.19:55
jgriffithSwanson: well... it "does" sometimes19:55
jgriffithSwanson: panbalag but it has all sorts of issues with cleanup and error handling19:56
*** lpetrut has joined #openstack-cinder19:56
jgriffithSwanson: panbalag and the bigger problem is, it "says" it worked... but look on your backend and make sure19:56
panbalagjgriffith: the exception I see happens when the source volume (attached to an instance) is being deleted..
panbalagjgriffith: I only verified the 'cinder list' output when I said it was successful19:57
jgriffithpanbalag: interesting...19:57
Swansonjgriffith: Thanks.  I know we worked our way through some failures at one point.  Looks like it kept being fixed after that.19:57
jgriffithEndpoint Not Found... that's a new one19:57
* jgriffith hasn't seen that one before 19:57
panbalagjgriffith: is there any configuration that needs to be enabled specifically for retype to work? wondering if I'm missing any configs19:58
panbalagjgriffith: ok. I checked the backend - it didn't get through..20:00
*** haomaiwang has quit IRC20:01
*** haomaiwang has joined #openstack-cinder20:02
jgriffithpanbalag: :(  Yeah, that's been my biggest complaint since day one20:03
jgriffithpanbalag: it'll silently fail... swap all the info in the db and say it's "good"20:03
jgriffithpanbalag: and you don't know until you actually try and use the volume that things didn't work20:03
tbarronsmcginnis: hurray!20:04
*** dflorea has joined #openstack-cinder20:07
*** merooney has quit IRC20:08
*** dflorea has quit IRC20:09
*** dflorea has joined #openstack-cinder20:09
patrickeastjgriffith: hey, did you ever get the image cache working with your backend? i just realized i never really followed up on your review (aside from fixing the permissions stuff)20:10
jgriffithpatrickeast: I did not :(20:10
jgriffithpatrickeast: continued to run into the glance permissions problem.20:11
jgriffithSomebody else tried it and saw the same thing IIRC20:11
patrickeastjgriffith: was that after i removed all the checks for permissions?20:11
jgriffithpatrickeast: errr... probably not :)20:12
jgriffithpatrickeast: but, it was the glance call that was failing20:12
jgriffithpatrickeast: not an op in the cinder code20:12
openstackgerritWalter A. Boring IV (hemna) proposed openstack/os-brick: Adding a tool that allows querying of LUNs
jgriffithpatrickeast: the actual glance client call was rejected20:12
patrickeastjgriffith: oh really?20:13
*** rushil has quit IRC20:13
jgriffithpatrickeast: yeah20:13
patrickeastjgriffith: the one i see on the review comment is "INFO cinder.image.cache [req-ee548891-ecd4-4c3c-b6c2-0bcfe2f2ef90 None] Unable to use ImageVolumeCache because image 5784e30f-64ee-4481-ab82-02eea526716a is not accessible by internal tenant."20:13
patrickeastwhich is from cinder20:13
patrickeastthats a check i added into the cache code20:13
jgriffithpatrickeast: I'll load it up and check it out again20:13
*** Yogi1 has quit IRC20:13
patrickeastjgriffith: that would be awesome... definitely want to fix that if there is still an issue20:13
jgriffithpatrickeast: but pretty sure I hacked some stuff out and got an error from the glance call itself20:14
*** Apoorva has joined #openstack-cinder20:14
jgriffithpatrickeast: sure, I'll have a look at it20:14
openstackgerritJon Bernard proposed openstack/cinder: Add support for file I/O volume migration
jgriffithpatrickeast: BTW... getting configs into the driver on init is proving to be a PITA20:14
patrickeastjgriffith: for replication?20:14
jgriffithpatrickeast: because it turns out we parse the config file for the driver on driver-init20:14
jgriffithpatrickeast: yeah20:14
patrickeastjgriffith: haha yea20:15
patrickeastjgriffith: we've pretty much decided we will only support unmanaged devices20:15
jgriffithpatrickeast: which means the only way I can really do this is wait til *all* volume managers are up and initialized, and then issue a call20:15
jgriffithpatrickeast: hehe... yeah, I'm kind of on the opposite side of that one20:15
jgriffithpatrickeast: too much change for all the associated calls fro this stage of the release cycle for my liking20:16
patrickeastjgriffith: oh no, sorry, didn't mean to change the apis20:16
jgriffithpatrickeast: adding an endpoint is easy enough but making sure I catch every single call that might have it switched sucks20:16
patrickeastjgriffith: just that our driver would ignore replication settings if they werent specified as unmanaged devices20:16
jgriffithpatrickeast: ?20:16
patrickeastjgriffith: because parsing the managed ones is proving troublesome20:16
jgriffithpatrickeast: oh.. yeah, I knew what you meant20:16
jgriffithyeah... totally cool20:16
openstackgerritWalter A. Boring IV (hemna) proposed openstack/os-brick: WIP: Set executor to Putils
jgriffithpatrickeast: I think I might add an admin API like "update_replication_settings" and pass it in that way... assuming nobody disagrees?20:17
* jgriffith is looking at patrickeast and dannywilson 20:17
jgriffithpatrickeast: or init_managed_replication20:18
patrickeastjgriffith: so what i don't really understand is why we can't just get to the config file contents directly in the driver code... i guess i don't really understand enough about how oslo_config works, but it seems like a wrapper around the python config file parsing20:18
jgriffithpatrickeast: yeah... so in theory you could do "CONF.backend-name.option"20:18
patrickeastjgriffith: ehh my only hesitation is that its kind of late in the game for that isn't it?20:18
jgriffithwe don't actually parse the config file and fill that in until the startup20:19
jgriffithpatrickeast: LOL20:19
jgriffithpatrickeast: Well it was my argument two days ago that it was too late in the game to implement it and you disagreed20:19
*** merooney has joined #openstack-cinder20:19
*** akerr has quit IRC20:19
jgriffithpatrickeast: besides... it's not like it will impact anybody :)20:20
*** bill_az has joined #openstack-cinder20:20
*** merooney has quit IRC20:20
patrickeastjgriffith: not exactly true... anyway, my point is that even if you put the review up today for it do you really think it would get merged by the freeze date?20:20
patrickeastjgriffith: like driver implementations aside20:20
jgriffithpatrickeast: what's not true?20:20
*** panbalag has quit IRC20:21
* jgriffith isn't following20:21
patrickeastjgriffith: a few things i guess, it will impact folks... like you and me writing the implementations, right? the core reviewers who are presumably swamped with the like 30 things still on the list for L3, etc20:22
patrickeastjgriffith: what i meant by too late in the game though is that the feature freeze is like monday, right?20:22
patrickeastjgriffith: adding a new api would have to merge in record time20:22
patrickeastjgriffith: thats all20:22
jgriffithpatrickeast: oh... yeah20:22
jgriffithpatrickeast: meh... I suppose, it's a single call but yeah... maybe not even worth considering20:23
patrickeastjgriffith: don't get me wrong, i like the idea of having that api20:23
patrickeastjgriffith: definitely not advocating that we shouldn't have that kind of ability20:23
jgriffithpatrickeast: my point about impacting anybody was you apparantly have a driver that doesn't need it, so you wouldn't use it anyway20:23
patrickeastjgriffith: well... we would use it if we could20:23
jgriffithpatrickeast: yeah... who knows.  We'll see how things shake out in M20:23
jgriffithpatrickeast: haha... Ok, fair enough20:24
*** p0rtal_ is now known as p0rtal20:24
smcginnisjgriffith, patrickeast, winston-d, thingee, xyang1: Looks like our summit session is the last block on Tuesday. 5:30pm20:24
patrickeastjgriffith: so back to the config thing... why can't we parse it earlier? or can we and we just dont?20:24
jgriffithpatrickeast: we can't without a lot of change20:25
xyang1smcginnis: Right20:25
patrickeastjgriffith: what if we parse it twice? :o20:25
jgriffithpatrickeast: because it goes through and sets the global CONF entries as it starts the services20:25
patrickeastjgriffith: ooh20:25
patrickeastjgriffith: i see20:25
jgriffithpatrickeast: yeah... that's kinda what I was wondering20:25
kmartinsmcginnis, almost prime time!20:25
jgriffithpatrickeast: startup20:25
*** julim has quit IRC20:25
jgriffiththat's where the magic happens20:25
smcginniskmartin: That will probably be the night where there's a good early vendor party. :)20:26
jgriffithpatrickeast: anyway.... no big deal, we'll look at it in M20:26
kmartinsmcginnis, less questions that way :)20:26
jgriffithsmcginnis: LOL... addind is a word!!!20:26
smcginnisjgriffith: Hah! :)20:27
jgriffithI don't mind changing it if you promise to put the +2/A back on it after I do so :)20:27
smcginnisjgriffith: Don't bother.20:27
openstackgerritJohn Griffith proposed openstack/cinder: get_replication_updates call to driver is wrong
smcginnisjgriffith: But I would.20:27
jgriffithtoo late20:27
smcginnisjgriffith: I mean I would put the +2/+A back, not that I would update the typo. :)20:28
jgriffithI can't remember if gerrit will auto add the votes back in for commit mesg updates or not20:28
smcginnisjgriffith: Doesn't look like it. I got it.20:28
kmartinsmcginnis, my 12 year old has an English paper due, can I put in gerrit for you to review? :)20:29
smcginnisI just pointed it out to be funny. I really don't care!20:30
smcginnisI don't -1 for those.20:30
*** dflorea has quit IRC20:30
smcginniskmartin: That would be pretty funny to use gerritt to do paper reviews. I guess we already do with docs, but would be awesome to see a school spin up an instance for that.20:31
smcginnisOh well, it's beer:30 here. Have a good weekend all. Might be scarce most of next week.20:32
kmartinsmcginnis, yeah jgriffith and I would not make core in that system20:32
kmartinenjoy your weekend20:32
jgriffithsmcginnis: enjoy!20:32
hemnasmcginnis, why not use a hacking check with libpspell ?20:32
smcginnishemna: Awesome. I'm going to propose it. :P20:33
hemnait'll auto -1 as smcginnis20:33
*** smcginnis is now known as smcginnis_afk20:34
asselin_All, sorry about all the 'unstable' checks on our ci. Our log server filled up. It's fixed now (but some more in the queue will report unstable)20:35
patrickeastanyone else seeing issues with their CI system and tempest.api.volume.admin.test_volumes_backup.VolumesBackupsV2Test.test_volume_backup_create_get_detailed_list_restore_delete?20:35
patrickeastonly my FC driver seems to be affected by it20:35
*** lpetrut has quit IRC20:36
*** dflorea has joined #openstack-cinder20:36
Swansonpatrickeast: I've not seen that one.20:37
patrickeastasselin_: do you guys have a monitoring system like nagios or something? problems like that are almost a thing of the past for me once i set one up20:38
patrickeastasselin_: although it kills my inbox :(20:38
asselin_patrickeast, not yet, we have someone who'll start working on that20:39
tbarronpatrickeast: the restore fails doing volume_file.write() with IOError: Operation not permitted20:40
patrickeasttbarron: yea, seems like a valid error if it can't write to the disk20:40
patrickeasttbarron: the interesting thing is that its intermittent20:40
tbarronpatrickeast: is it trying to write to the same volume as the adjacent 'Fibre volume not yet found' warning?20:40
patrickeasttbarron: good question... not sure20:41
tbarronpatrickeast: that's all I see there ...20:42
patrickeasttbarron: looks like it found the volume
*** dustins_ has joined #openstack-cinder20:44
*** dflorea has quit IRC20:44
*** dustins has quit IRC20:45
*** dustins_ is now known as dustins20:45
*** krtaylor has quit IRC20:51
*** edtubill has left #openstack-cinder20:51
tbarronpatrickeast: LOL the NetApp proxy-screener held up that earlier link for the longest time when I retrieved it20:54
tbarronpatrickeast: mebbe it saw PureFCDriver :-)20:54
patrickeasttbarron: clearly something sketchy20:55
*** jaypipes has joined #openstack-cinder20:55
tbarronpatrickeast: anyways, looks like volume  405e58ed-0bba-4614-9141-47a3bd1feb35 is having its export removed right before an attempt to write a chunk to it.20:55
patrickeasttbarron: hmm yea seems like that shouldn't have happened20:56
tbarronmaybe the restore is taking too long and there's a timeout and test cleanup has started ....20:57
tbarronon that note, I'm outta here.  It's east coast weekend now :-)20:57
patrickeasttbarron: ohh yea that would make sense20:58
*** esker has joined #openstack-cinder20:58
patrickeasttbarron: thanks for taking a look! have a good weekend20:58
*** esker has quit IRC20:58
tbarroncy'all later.  I'll be doing some reviews and checking mine over coffee in the morning.20:58
*** tbarron has left #openstack-cinder20:58
*** timcl has quit IRC20:58
*** esker has joined #openstack-cinder20:59
*** haomaiwang has quit IRC21:01
*** haomaiwang has joined #openstack-cinder21:01
*** raildo is now known as raildo-afk21:02
*** krtaylor has joined #openstack-cinder21:04
*** IlyaG has joined #openstack-cinder21:04
*** _cjones_ has quit IRC21:04
*** dflorea has joined #openstack-cinder21:05
*** IlyaG has quit IRC21:07
*** dustins has quit IRC21:10
*** dustins has joined #openstack-cinder21:10
*** dustins has quit IRC21:11
*** hemna has quit IRC21:11
*** IlyaG has joined #openstack-cinder21:18
*** dflorea has quit IRC21:25
*** jungleboyj has joined #openstack-cinder21:26
jungleboyjkmartin: Ping21:27
kmartinjungleboyj, yo21:27
jungleboyjkmartin: Need more help seeing up a Storwize.  ;-)21:27
*** eharney has joined #openstack-cinder21:28
jungleboyjkmartin: Just kidding.  Is there someone from HP that is working on code for the community to improve the ability to bring existing resources into Cinder's management?21:28
*** bkopilov has quit IRC21:28
kmartinjungleboyj, what type of resources?21:29
jungleboyjI am assuming existing volumes and/or backends with existing volumes.21:29
kmartinjungleboyj, not that I'm aware of21:30
*** dflorea has joined #openstack-cinder21:30
*** dflorea has quit IRC21:31
openstackgerritxing-yang proposed openstack/cinder: Fix a merge problem in VMAX driver
jungleboyjkmartin: Ok, interesting.  One of our developers thought that maybe such work was being done.  We have code to help.  :-)21:32
kmartinjungleboyj, I can ask around21:33
jungleboyjPlease do.  Let me know if you hear anything.21:34
*** bkopilov has joined #openstack-cinder21:34
*** IlyaG has quit IRC21:35
*** martyturner has quit IRC21:36
*** dflorea has joined #openstack-cinder21:36
*** mriedem has quit IRC21:37
*** dimsum__ has quit IRC21:38
*** cbader has quit IRC21:40
SwansonDid pure have a replication v2 patch in the pipeline?21:42
patrickeastSwanson: yea21:42
*** bkopilov has quit IRC21:42
Swansonpatrickeast: Thanks!  Wanted to see what an implementation looked like.21:43
*** martyturner has joined #openstack-cinder21:46
*** annegentle has quit IRC21:47
*** jaypipes has quit IRC21:47
*** ebalduf has quit IRC21:47
*** annegentle has joined #openstack-cinder21:47
*** martyturner has left #openstack-cinder21:49
*** jaypipes has joined #openstack-cinder21:49
*** hemna has joined #openstack-cinder21:51
*** jaypipes has quit IRC21:53
*** jaypipes has joined #openstack-cinder21:53
*** setmason has left #openstack-cinder21:54
eikkewhat's the reason for gate jobs to be 'aborted by anonymous'?21:55
Swansoneikke: Dunno.  Try #openstack-infra21:56
*** dimsum__ has joined #openstack-cinder21:56
*** Apoorva has quit IRC21:58
*** setmason has joined #openstack-cinder21:58
*** Apoorva has joined #openstack-cinder21:59
*** dims_ has joined #openstack-cinder22:00
*** xyang1 has quit IRC22:00
*** haomaiwang has quit IRC22:01
*** bkopilov has joined #openstack-cinder22:01
*** haomaiwang has joined #openstack-cinder22:02
*** dimsum__ has quit IRC22:02
*** setmason has quit IRC22:03
jgriffithSwanson: patrickeast just FYI, I'm about to give that a -2 :(22:03
jgriffithSwanson: patrickeast we kinda need to learn from past mistakes22:03
patrickeastjgriffith: so... no one gets replication then?22:04
jgriffithSwanson: patrickeast there's a number of things int the replication code that still need worked out22:04
jgriffithpatrickeast: that's what I'm thinking... if we can't make any additions/changes to what's there22:04
jgriffithpatrickeast: and since we said nobody can implement it after last Sunday (and the base code didn't even get submitted as final until then)22:05
Swansonjgriffith: Is there a way in v2 replication to move a volume from being on a secondary array for one backend to being on the primary array of another?22:05
SwansonOr is that just normal migration?22:05
jgriffithpatrickeast: You're statement earlier about "can't add an api call or any new methods" kind of got me thinking about it22:05
jgriffithSwanson: not sure I follow22:06
jgriffithpatrickeast: you made a good point about things impacting other peoples implementation22:06
jgriffithpatrickeast: and I suspect that as more people try and implement this they're likely going to have tweaks that they propose.22:07
patrickeastjgriffith: i kind of get where you are coming from... but i would like to make a few things clear (for the record as it were...) we met the freeze requirements thingee set out, and i didn't mean we couldn't add new stuff, just that i was skeptical there would be time to do it22:07
Swansonjgriffith: 2 arrays.  Each on a backend.  (dell1 and dell2).  dell1 uses dell2 as its replication array and dell2 uses dell1 as it's array.  Dell1 burns.  So the backend has failed over its volumes to the replication volumes on dell2.  Dell2 should probably just handle those volumes.22:07
patrickeastmaybe its just a process thing though... in my mind having it in the tree just makes things easier to iterate on and improve22:07
patrickeasthaving a bunch of out of tree secret impls22:08
jgriffithpatrickeast: haha!22:08
patrickeastand proposing changes/fixes to the core feature seems less good22:08
patrickeastthat being said22:08
patrickeasthaving 20+22:08
patrickeastin the tree22:08
patrickeastthats hard to iterate on22:08
jgriffithnothing secret on my side if that's what you meant22:08
*** jaypipes has quit IRC22:08
jgriffithnote the nova repo with the essex-replication branch, and the replication v2 branch in my cinder repo as well22:09
patrickeastjgriffith: haha, maybe wrong words to use22:09
jgriffithpatrickeast: I was hoping that was all that was22:09
patrickeastjgriffith: not secret, just less easy to see in one place22:09
SwansonGate is averaging 25 hours to get a job through.22:09
patrickeastjgriffith: anyway, if you don't agree with that i can understand22:09
jgriffithSwanson: :(22:09
jgriffithpatrickeast: nah... it's fine22:10
jgriffithpatrickeast: I don't think I"m getting my point across at all so that's ok22:10
jgriffithpatrickeast: black and white interpretation of the various deadlines and process rules, you're absolutely rigth22:11
jgriffithpatrickeast: although... I will say it's impossible that your driver "worked" because the method was broken and if it were called it would blow up until today22:12
jgriffithpatrickeast: thus the patch I pushed to fix it this afternoon22:12
patrickeastjgriffith: yea, and i mean, i'm not trying to be a rule nazi about that stuff, i understand they are more like guidelines we set for ourselves22:12
patrickeastjgriffith: no argument there, i'm taking care of that22:12
patrickeastjgriffith:  and i think i understand what you are saying, which is that its easier to modify the core feature without any implementations22:12
patrickeastjgriffith: and it prevents half-baked features from releasing22:13
jgriffithpatrickeast: yeah, that's all i was getting at22:13
jgriffithpatrickeast: and avoids any API changes ... or the whole situation we had in the first place that led to V2 to begin with (one impl at the end of the release)22:13
jgriffithbut.. that's cool22:13
jgriffithI'm surely not sweating it22:14
patrickeastjgriffith: i'm not sure if comparing it to the v1 replication is the same... but i see your point22:14
jgriffithpatrickeast: well... I'll go ahead and throw my version up on gerrit later22:15
jgriffithpatrickeast: won't merge for certain but at least something else as a reference will be available22:15
patrickeastjgriffith: i guess the only thing i want to argue is that having implementations in tree is useful, even ones that will have to change a bit as they settle down and the core feature is tweaked22:15
SwansonIt would be nice if one could add features to drivers of a stable branch.  Up until milestone 1 of the next release or something.22:16
patrickeastjgriffith: i wouldn't push on it if we had a lvm replication impl ;)22:16
jgriffithpatrickeast: haha!  It's useful for Pure to say "hey we have replication in Liberty" LOL22:16
jgriffithSwanson: yeah... I think we really need to look at how we do releases all the way around22:16
patrickeastjgriffith: honestly i was going to ensure it was labelled as experimental and shouldn't be used22:16
jgriffithSwanson: I was talking to bswartz abou tthis the other day22:16
jgriffithpatrickeast: I know.. that was a smart ass comment22:17
jgriffithpatrickeast: anyway... we'll carry on22:17
jgriffithno problem22:17
*** edmondsw has quit IRC22:17
Swansonjgriffith: The issue I see is that companies trail on releases.  So it would really be nice to be able to support everything in a release eventually.22:17
jgriffithSwanson: no doubt!22:18
jgriffithSwanson: the thing that's annoying is most distros/customers are at least n-122:18
*** ronis has quit IRC22:18
jgriffithSwanson: often n-222:18
jgriffithSwanson: so you work on this pretty new feature, it takes two cycles and never sees the light of day for 18-24 months22:18
jgriffithSwanson: until you do the backport for them :)22:19
Swansonjgriffith: Yep.  We showed up with Kilo.  Most people are juno or icehouse.22:19
jgriffithSwanson: cuz they "gotta have it"22:19
*** vgridnev has quit IRC22:19
jgriffithbut forget about upgrading their cloud, because we refuse as a dev community to make deployment and upgrades better22:19
Swansonupgrades are even possible?22:20
jgriffithSwanson: LOL... YESS!!!!!22:20
*** Apoorva has quit IRC22:24
*** Apoorva has joined #openstack-cinder22:24
*** ganso_ has quit IRC22:25
*** bnemec has quit IRC22:27
openstackgerritAnthony Lee proposed openstack/cinder: Fix QoS keys not being available to scheduler
openstackgerritMichael Price proposed openstack/cinder: Add support for volume groups and netapp_raid_type
openstackgerritMichael Price proposed openstack/cinder: Add additional SSC extra specs to E-Series driver
openstackgerritMichael Price proposed openstack/cinder: Implement AutoSupport for NetApp E-Series driver
openstackgerritMichael Price proposed openstack/cinder: Implement thin provisioning support for E-Series
openstackgerritAnish Bhatt proposed openstack/os-brick: Add support for --interface option in iscsiadm
hemnaok, off to do my community service at Knee Deep Brewery.   Beer isn't going to drink itself.22:39
*** hemna is now known as hemnafk22:39
*** dflorea has quit IRC22:39
*** dflorea has joined #openstack-cinder22:42
*** ndipanov has quit IRC22:42
*** dflorea has quit IRC22:44
*** eharney has quit IRC22:47
*** dflorea has joined #openstack-cinder22:53
*** lcurtis has quit IRC22:54
*** dflorea has quit IRC22:55
*** p0rtal has quit IRC22:58
*** p0rtal has joined #openstack-cinder22:59
*** jamielennox is now known as jamielennox|away23:00
*** haomaiwang has quit IRC23:01
*** vincent_hou has joined #openstack-cinder23:01
*** haomaiwa_ has joined #openstack-cinder23:02
*** p0rtal has quit IRC23:03
*** annegentle has quit IRC23:04
*** dannywilson has quit IRC23:05
*** dflorea has joined #openstack-cinder23:05
*** vincent_hou has quit IRC23:07
*** p0rtal has joined #openstack-cinder23:07
*** hodos|2 has quit IRC23:08
*** Apoorva has quit IRC23:13
*** dflorea has quit IRC23:15
*** Apoorva has joined #openstack-cinder23:17
*** dflorea has joined #openstack-cinder23:20
*** mtanino has quit IRC23:21
*** david-lyle has quit IRC23:24
*** Apoorva has quit IRC23:30
*** garthb has quit IRC23:32
*** Apoorva has joined #openstack-cinder23:32
*** ganso_ has joined #openstack-cinder23:35
*** dflorea has quit IRC23:39
*** Apoorva has quit IRC23:41
openstackgerritOpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements
*** salv-orlando has quit IRC23:45
openstackgerritOpenStack Proposal Bot proposed openstack/python-cinderclient: Updated from global requirements
*** p0rtal has quit IRC23:47
*** dimsum__ has joined #openstack-cinder23:52
*** dims_ has quit IRC23:55

Generated by 2.14.0 by Marius Gedminas - find it at!