Friday, 2016-02-26

*** haomaiwa_ has joined #openstack-cinder00:01
*** yangyape_ has quit IRC00:02
*** chlong_ has quit IRC00:03
*** angela-s has quit IRC00:05
*** haomaiwa_ has quit IRC00:09
*** merooney has quit IRC00:09
*** merooney has joined #openstack-cinder00:11
patrickeastjgriffith: hokay, so, now i can't repro that issue any more either00:12
patrickeastjgriffith: good news there00:12
jgriffithpatrickeast: butttt....?00:12
patrickeastjgriffith: no buts00:12
patrickeastjgriffith: :D00:12
jgriffithpatrickeast: pheww00:12
*** krtaylor has joined #openstack-cinder00:12
patrickeastjgriffith: the service query patch works great too00:12
* jgriffith steps back from off the ledge00:12
patrickeastjgriffith: my last question is.... now that i've failed over, i can do stuff and things are using the secondary target, did we decide on a sort of recommended path to get back?00:13
jgriffithpatrickeast: I plan to use failover again00:13
jgriffithpatrickeast: but need to add some things to make that work correctly00:14
jgriffithpatrickeast: something like "failover_host xxxx primary"00:14
jgriffithpatrickeast: that's what I was thinking anyway, don't know if folks have better suggestions00:14
patrickeastjgriffith: mm yea that should work for mine too, guess i coud just add the 'primary' as one of the secondary targets we report00:14
jgriffithpatrickeast: how come?00:15
jgriffithpatrickeast: so my thought was some "keyword" that meant, just get the default from config file00:15
patrickeastjgriffith: ohh i see00:15
patrickeastjgriffith: yea that would work00:15
jgriffithpatrickeast: and that would clear "active_backend_Id" as well00:15
jgriffithpatrickeast: I just need a clever "code" name00:16
jgriffithpatrickeast: heck... maybe I call it "graham-cracker"00:16
jgriffithpatrickeast: as in the delicious foundation of Cheesecake00:16
patrickeastjgriffith: perfect!00:16
jgriffithpatrickeast: I suppose others would rather I just use something like "default"00:17
patrickeastjgriffith: i guess if the naming gets weird it could just be a flag thats like 'undo-failover'00:17
jgriffithsmcginnis: no joy on eharneys patch yet I guess huh00:17
patrickeastjgriffith: default works as well00:17
jgriffithpatrickeast: yeah... but if we don't do it now then we have an rpc version bump and all that mess :(00:18
jgriffithpatrickeast: I'll work up a patch now that uses "default" as a key work00:18
jgriffithword even00:18
patrickeastjgriffith: sounds good to me00:18
jgriffithaorourke: would that work ok for you?  ^^00:18
aorourkejgriffith, let me read through the convo00:19
openstackgerritNate Potter proposed openstack/cinder: Don't allow operations on deleting volumes
aorourkejgriffith, patrickeast would we be expected to configure cinder.conf with a value for this? Or are we settling on a default?00:22
jgriffithaorourke: the default00:22
*** dramakri has quit IRC00:22
aorourkejgriffith, sound good to me00:23
jgriffithaorourke: patrickeast like this:
jgriffithaorourke: patrickeast so it set status back to ENABLED, clears the active_backend_id and disabled settings00:24
jgriffithaorourke: patrickeast then it's up to the driver what they need/want to do when they see "default" come back in00:24
aorourkejgriffith, would it be possible to allow a call to the driver first? That way we can ensure the admin properly reset everything back up on their end?00:25
jgriffithaorourke: the call to the driver is first actually00:25
aorourkejgriffith, oh it does... right?00:25
jgriffithaorourke: yep00:25
*** IlyaG has quit IRC00:25
aorourkejgriffith, great. Is there an exception or something we can raise to basically say not ready?00:25
*** laughterwym has joined #openstack-cinder00:25
jgriffithaorourke: same one we use for driver.failover_host00:26
aorourkejgriffith, cool00:26
jgriffithaorourke: that could use some refinement, but at least there's something there currently00:26
*** sdague has quit IRC00:26
jgriffithpatrickeast: I'll wait to get your buy in... and Swanson if you're still following this madness :)00:26
*** laughterwym has joined #openstack-cinder00:27
jgriffithso personally I like it because it's relatively "clean"00:27
*** daneyon has quit IRC00:28
patrickeastjgriffith: yea it looks good to me, i'll have to figure out for my driver how much we want to support... in theory we can try and set up replication back to the default (if its still alive) and make that call to default automate moving everything back over (sort of a fail-back behavior)00:31
patrickeastjgriffith: buuut that adds a lot of complexity : /00:31
patrickeastmaybe i'll just change the current array id and assume the admin move stuff back over on their own00:31
jgriffithpatrickeast: yeah, we talked in RTP about failover being something possibly best left to an admin00:32
jgriffithpatrickeast: the main thing as I've argued is that the use case is that your cluster ate itself00:33
jgriffithand doesn't exist any longer00:33
patrickeastjgriffith: yea def00:33
jgriffithbut perhaps you rolled something else in set it up, replicated to it and modified your config00:33
jgriffithto me this seemed like the only way to really recover... or by the way... this is how you switch the secondary over to now be advertised as the primary00:34
patrickeastits probably safest to assume the admin fixed things and is ready to go back00:34
patrickeastdoing it automatically is just asking to lose someones data00:34
jgriffithpatrickeast: yeah, that's what got us in to trouble with the old code00:34
patrickeastor more importantly, having it be my fault that the data is lost :D00:34
jgriffithpatrickeast: TAKE NO RESPONSIBILITY!!!!00:35
jgriffithpatrickeast: just have your driver raise an exception if backend_id is == default00:35
*** mylu has quit IRC00:35
jgriffiththat's what I'm going to do00:35
jgriffithat least for now00:35
patrickeastjgriffith: i'll probably swap my like current_id stuff we keep track of, but do nothing else00:35
jgriffithpatrickeast: yeah, that's the most I'll do probably.. and document it00:36
jgriffiththere's another one of those stupid query problems00:39
jgriffithoh.. never mind :)00:39
aorourkejgriffith, btw, that change seems to work fine for me as well. active_backend_id is available in kwargs upon init00:39
jgriffithaorourke: excellent00:39
*** The_Ball has joined #openstack-cinder00:41
*** ntpttr has quit IRC00:43
*** mylu has joined #openstack-cinder00:43
patrickeastjgriffith: lemmie know if my response on makes sense, i don't think we need provider info for anything on the volumes00:44
patrickeastjgriffith: i had considered updating volume replication status or setting ones that were left behind to an error state... but its proven very easy to get in a weird state where they aren't in sync with reality00:44
jgriffithpatrickeast: yeah, makes sense and that's what I kind of expected, but just thought I'd point it out00:45
*** salv-orlando has joined #openstack-cinder00:45
*** mylu has quit IRC00:46
patrickeastactually... as i think about it more maybe I will set left behind ones to error00:47
patrickeasteasy enough for an admin to do a reset-state if they are actually ok at some point in the future (like after doing a failover to default), kind of err on the side of assuming things are broken00:47
jgriffithpatrickeast: Yup!00:48
jgriffithpatrickeast: much better than just 'random' failures for user00:49
jgriffithdang... Erics patch to fix the log.warning mess is still 8 hours out just for the check queue :(00:51
patrickeastjgriffith: does anyone have magic powers to just merge it?00:53
jgriffithpatrickeast: not really.. the only way to really do that is pretty ugly.  Kill everything, clear the queues, promote that change then turn everything back on00:53
jgriffithpatrickeast: at least IIRC that's how they would do it00:54
patrickeastjgriffith: gotcha00:54
*** qeelee has joined #openstack-cinder00:56
*** mylu has joined #openstack-cinder00:58
*** gouthamr has joined #openstack-cinder00:58
jgriffithpatrickeast: aorourke
jgriffitha few changes to that gist I sent and a change to the query in the API01:00
patrickeastjgriffith: ship it!01:00
aorourkejgriffith, loogs good!01:02
aorourkejgriffith, I briefly scanned over your convo yesterday, what was the recommendation for finding replication targets/backend_id ?01:03
*** mylu has quit IRC01:03
jgriffithaorourke: get_capabilities01:03
aorourkejgriffith, ah ok. Is there a way to list them still?01:03
aorourkeor view them01:03
*** gouthamr_ has joined #openstack-cinder01:03
*** gouthamr has quit IRC01:03
jgriffithaorourke: so that would be the method to use... report the configured backends in your volume stats update01:04
jgriffithaorourke: then when an admin runs get_capabilities they'll now show up01:04
*** mylu has joined #openstack-cinder01:04
aorourkejgriffith, ah, i see that in the code now. Cool01:05
aorourke| replication_targets |                           []                          |01:05
aorourkei need to populate that01:05
jgriffithaorourke: exact-o-mundo01:05
jgriffithaorourke: here's my fake:
*** mylu has quit IRC01:06
jgriffithdon't forget to make it a member of the data dict01:06
jgriffithanyway.. you get the idea01:06
*** haomaiwa_ has joined #openstack-cinder01:06
*** salv-orlando has quit IRC01:07
jgriffithThere.. updated01:07
aorourkejgriffith, that was easy :)01:09
*** haomaiwa_ has quit IRC01:13
*** aorourke is now known as aorourke-afk01:13
*** yangyapeng has joined #openstack-cinder01:13
*** sheel has joined #openstack-cinder01:14
*** mylu has joined #openstack-cinder01:15
*** EinstCrazy has joined #openstack-cinder01:16
*** mylu has quit IRC01:17
*** edtubill has quit IRC01:18
*** mylu has joined #openstack-cinder01:19
*** IlyaG has joined #openstack-cinder01:19
openstackgerritJohn Griffith proposed openstack/cinder: Add ability to failback for replication V2.1
*** mtanino has quit IRC01:21
*** davechen_afk is now known as grassy01:22
*** cknight has joined #openstack-cinder01:22
*** davechen1 has joined #openstack-cinder01:22
*** mylu has quit IRC01:23
*** haomaiwang has joined #openstack-cinder01:25
openstackgerritPatrick East proposed openstack/cinder: Update Pure replication to cheesecake
smcginnisStill waiting for the gate fix to make it through.01:26
*** daneyon has joined #openstack-cinder01:28
*** garthb has quit IRC01:28
*** Lee1092 has joined #openstack-cinder01:28
*** garthb_ has quit IRC01:29
*** laughterwym has quit IRC01:29
*** laughterwym has joined #openstack-cinder01:30
*** davechen1 is now known as davechen01:30
openstackgerritPatrick East proposed openstack/cinder: Update Pure replication to cheesecake
smcginnisHmm, and it's not even in any of the queue right now.01:30
*** daneyon_ has joined #openstack-cinder01:30
*** haomaiwang has quit IRC01:33
*** daneyon has quit IRC01:33
*** uberjay has quit IRC01:35
*** uberjay has joined #openstack-cinder01:37
*** Suyash has quit IRC01:38
openstackgerritOpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements
*** yangyapeng has quit IRC01:39
*** EinstCrazy has quit IRC01:40
*** mylu has joined #openstack-cinder01:41
*** EinstCrazy has joined #openstack-cinder01:42
*** qeelee_ has joined #openstack-cinder01:42
*** qeelee has quit IRC01:42
*** qeelee_ is now known as qeelee01:42
*** leeantho has quit IRC01:45
*** cheneydc has joined #openstack-cinder01:48
*** dims has quit IRC01:49
*** chlong_ has joined #openstack-cinder01:50
*** laughterwym has quit IRC01:52
*** laughterwym has joined #openstack-cinder01:52
*** laughterwym has quit IRC01:53
*** laughterwym has joined #openstack-cinder01:53
*** annasort has joined #openstack-cinder01:55
*** mylu has quit IRC02:00
*** jungleboyj has joined #openstack-cinder02:00
*** mylu has joined #openstack-cinder02:08
*** laughterwym has quit IRC02:14
*** laughterwym has joined #openstack-cinder02:14
*** igornsa has joined #openstack-cinder02:20
*** laughterwym has quit IRC02:20
*** laughterwym has joined #openstack-cinder02:21
*** haomaiwang has joined #openstack-cinder02:21
*** knikolla has joined #openstack-cinder02:21
openstackgerritMatt Smith proposed openstack/cinder: Updating Datera DataFabric Driver to v2 of Datera DataFabric API
*** igornsa has quit IRC02:26
*** laughterwym has quit IRC02:29
*** laughterwym has joined #openstack-cinder02:29
*** qeelee has quit IRC02:29
*** qeelee has joined #openstack-cinder02:31
*** jwcroppe has joined #openstack-cinder02:31
*** mylu has quit IRC02:34
*** mylu has joined #openstack-cinder02:35
*** mylu has quit IRC02:39
openstackgerritLisaLi proposed openstack/cinder: Make nullable of fields in db model and object match
*** laughterwym has quit IRC02:47
*** laughterwym has joined #openstack-cinder02:47
*** laughterwym has quit IRC02:48
*** laughterwym has joined #openstack-cinder02:48
*** gcb has quit IRC02:48
*** bardia_ has quit IRC02:52
*** IlyaG has quit IRC02:52
*** laughterwym has quit IRC02:53
*** wanghao has quit IRC02:54
*** wanghao has joined #openstack-cinder02:55
*** delatte has quit IRC02:56
*** haomaiwang has quit IRC03:01
*** haomaiwang has joined #openstack-cinder03:01
*** cknight has quit IRC03:02
*** laughterwym has joined #openstack-cinder03:03
*** IlyaG has joined #openstack-cinder03:05
*** gcb has joined #openstack-cinder03:05
openstackgerritJay Bryant proposed openstack/cinder: Tests: Don't assert on LOG.warn
openstackgerritJay Bryant proposed openstack/cinder: Add CGSnapshotStatus enum field
*** dims has joined #openstack-cinder03:10
*** apoorvad has quit IRC03:11
openstackgerritLisaLi proposed openstack/cinder: unit test fix in netapp
*** knikolla has quit IRC03:15
lixiaoy1_hi, any cores here? I just fix a bug which blocks CI. please have a look:
*** boris-42 has quit IRC03:24
*** gouthamr_ has quit IRC03:26
*** links has joined #openstack-cinder03:26
*** sheel has quit IRC03:27
*** jwcroppe has quit IRC03:37
*** jwcroppe has joined #openstack-cinder03:37
openstackgerritLisaLi proposed openstack/cinder: Tests: Don't assert on LOG.warn
*** jwcroppe has quit IRC03:42
*** qeelee has quit IRC03:50
*** markvoelker has quit IRC03:51
*** daneyon_ has quit IRC03:54
*** erlon has quit IRC03:56
openstackgerritLisaLi proposed openstack/cinder: Tests: Don't assert on LOG.warn
*** haomaiwang has quit IRC04:01
*** haomaiwang has joined #openstack-cinder04:01
*** edtubill has joined #openstack-cinder04:02
*** cknight has joined #openstack-cinder04:05
*** edtubill has quit IRC04:09
*** yhayashi has quit IRC04:24
*** Poornima has joined #openstack-cinder04:26
*** gcb has quit IRC04:29
*** IlyaG_ has joined #openstack-cinder04:35
*** itzdilip has joined #openstack-cinder04:37
*** IlyaG has quit IRC04:38
*** jwcroppe has joined #openstack-cinder04:39
*** IlyaG_ has quit IRC04:39
*** mylu has joined #openstack-cinder04:41
*** qeelee has joined #openstack-cinder04:41
*** gcb has joined #openstack-cinder04:41
*** ntpttr has joined #openstack-cinder04:45
*** markvoelker has joined #openstack-cinder04:51
*** dustins has quit IRC04:56
*** markvoelker has quit IRC04:56
*** qeelee has quit IRC04:57
*** haomaiwang has quit IRC05:01
*** haomaiwang has joined #openstack-cinder05:01
*** shyama has joined #openstack-cinder05:03
*** shyama has quit IRC05:04
*** yhayashi has joined #openstack-cinder05:08
*** jwcroppe has quit IRC05:12
*** gcb has quit IRC05:13
*** edtubill has joined #openstack-cinder05:18
openstackgerritNate Potter proposed openstack/cinder: Don't allow operations on deleting volumes
*** mylu has quit IRC05:25
*** Roamer` has quit IRC05:26
*** Roamer` has joined #openstack-cinder05:26
*** ntpttr has quit IRC05:28
*** gcb has joined #openstack-cinder05:30
*** laughterwym has quit IRC05:30
*** cknight has quit IRC05:31
*** garthb_ has joined #openstack-cinder05:31
*** garthb has joined #openstack-cinder05:31
*** mylu has joined #openstack-cinder05:35
*** IlyaG has joined #openstack-cinder05:35
*** laughterwym has joined #openstack-cinder05:36
*** merooney has quit IRC05:37
openstackgerritLisaLi proposed openstack/cinder: Add backup update function
*** IlyaG has quit IRC05:40
*** laughterwym has quit IRC05:42
*** laughterwym has joined #openstack-cinder05:43
*** laughterwym has quit IRC05:43
*** laughterwym has joined #openstack-cinder05:43
*** laughterwym has joined #openstack-cinder05:43
*** qeelee has joined #openstack-cinder05:52
*** cheneydc has quit IRC05:54
*** laughterwym has quit IRC06:00
*** haomaiwang has quit IRC06:01
*** haomaiwang has joined #openstack-cinder06:01
*** wanghao has quit IRC06:05
*** wanghao has joined #openstack-cinder06:06
*** wanghao has quit IRC06:07
*** wanghao has joined #openstack-cinder06:08
*** sgotliv has joined #openstack-cinder06:14
openstackgerritDinesh Bhor proposed openstack/cinder: Provide user friendly messages for db sync
*** jwcroppe has joined #openstack-cinder06:19
*** laughterwym has joined #openstack-cinder06:22
openstackgerritMatt Smith proposed openstack/cinder: Updating Datera DataFabric Driver to v2 of Datera DataFabric API
*** abhishekk_ has joined #openstack-cinder06:24
*** Lallana has quit IRC06:26
*** links has quit IRC06:26
*** Lallana has joined #openstack-cinder06:26
*** sgotliv has quit IRC06:33
*** sgotliv has joined #openstack-cinder06:33
*** ChubYann has quit IRC06:34
*** IlyaG has joined #openstack-cinder06:36
*** nkrinner has joined #openstack-cinder06:37
*** dims has quit IRC06:37
*** IlyaG has quit IRC06:41
*** sheel has joined #openstack-cinder06:43
*** edtubill has quit IRC06:44
*** jeanfabrice has joined #openstack-cinder06:48
*** jwcroppe has quit IRC06:52
*** haomaiwang has quit IRC06:52
*** markvoelker has joined #openstack-cinder06:53
*** haomaiwang has joined #openstack-cinder06:53
*** markvoelker has quit IRC06:57
*** ankit_ag has joined #openstack-cinder06:58
*** haomaiwang has quit IRC07:01
*** haomaiwang has joined #openstack-cinder07:01
openstackgerritCao ShuFeng proposed openstack/python-cinderclient: Fix return type in consistencygroups docstring
*** yangyapeng has joined #openstack-cinder07:04
*** rcernin has joined #openstack-cinder07:05
openstackgerritPeter Wang proposed openstack/cinder: VNX: Update replication for v2.1
*** sgotliv has quit IRC07:10
*** garthb_ has quit IRC07:19
*** garthb has quit IRC07:19
*** chlong_ has quit IRC07:26
*** arecknag has joined #openstack-cinder07:26
*** lpetrut has joined #openstack-cinder07:27
*** anshul has joined #openstack-cinder07:27
*** salv-orlando has joined #openstack-cinder07:29
*** [1]Thelo has joined #openstack-cinder07:32
*** abhishekk has quit IRC07:34
*** Thelo has quit IRC07:34
*** [1]Thelo is now known as Thelo07:34
*** jeanfabrice has quit IRC07:36
*** frickler has joined #openstack-cinder07:44
*** ildikov has joined #openstack-cinder07:46
openstackgerritLisaLi proposed openstack/cinder: Add necessary fields to volume creation
*** laughterwym has quit IRC07:48
openstackgerritLisaLi proposed openstack/cinder: Add backup update function
*** jwcroppe has joined #openstack-cinder07:49
*** laughterwym has joined #openstack-cinder07:51
openstackgerritLisaLi proposed openstack/cinder: Add necessary fields to volume creation
*** laughterwym has quit IRC07:52
dulekgeguileo, DuncanT: Hi, you've read through the log from last night? Looks like gate fix is still not approved:
dulekgeguileo, DuncanT: I don't know why jungleboyj rebased it while it was in the queue, but we're with non-functional gate at the moment. :(07:54
*** cheneydc has joined #openstack-cinder07:56
*** mylu has quit IRC07:56
*** haomaiwang has quit IRC08:01
*** haomaiwang has joined #openstack-cinder08:01
*** zhangjn has quit IRC08:02
*** laughterwym has joined #openstack-cinder08:04
*** zhangjn has joined #openstack-cinder08:07
*** belmoreira has joined #openstack-cinder08:08
*** laughterwym has quit IRC08:08
duleklixiaoy1_: Why patchset 5 of was failing pep8?08:10
duleklixiaoy1_: This file even doesn't had that many linesā€¦08:10
lixiaoy1_dulek: I don't know. I didn't find any useful info from the logs08:18
lixiaoy1_dulek: pep8 succeeds in my machine08:18
duleklixiaoy1_: Sure, I was mistaken then. Eh, we have the py27, py34 blocked anyway. :( Thanks for updating the release note.08:20
lixiaoy1_dulek: np08:20
*** jwcroppe has quit IRC08:22
*** adrianofr_ has quit IRC08:25
*** adrianofr_ has joined #openstack-cinder08:27
openstackgerritxiaoqin proposed openstack/cinder: IBM Storwize with pool-aware-cinder-scheduler
openstackgerritVictor Stinner proposed openstack/cinder: Enable all unit tests on Python 3.4
haypo^^ i rebased my approved change since master was broken. maybe i wasn't needed. it looks like i lost the approval :-/08:38
dulekhaypo: Gate is dead until merges.08:39
haypo21:46 < smcginnis> haypo: Fix is in queue.08:39
haypodulek: ^^ the fix is not merged yet?08:39
dulekhaypo: Nope, it was stuck in the queue for a long time and I think this is why jungleboyj tried to rebase it.08:40
dulekhaypo: Which removed it from the queue.08:40
haypodulek: ah is the fix? ok08:40
haypobut it lost its "Patch Set 1: Code-Review+2 Workflow+1" too08:41
dulekhaypo: Yeah. I hoped geguileo, DuncanT or e0ne could fast-approve it, but looks like they aren't here yet.08:41
*** jeanfabrice has joined #openstack-cinder08:42
haypodulek: hum, i don't understand why it was rebased. the patch was ready to be merged, it was at the last step "Starting gate jobs."08:42
haypodulek: i heard that we can ask the infra to reorder jobs when a gate is broken08:42
haypoonce the change is +2a again, we can try to ask that to repair the gate quickly08:42
haypo(anyway, running new jobs on cinder is a loss of time, since all jobs will fail)08:43
dulekhaypo: From the IRC logs I've seen it was hanging in docs queue for an incredibly long time.08:43
dulekhaypo: True.08:43
dulekhaypo: At least the weather here marks a nice Friday. ;)08:43
hayporight now, zuul is closer to idle :)
dulekhaypo: Yeah, it would be great to have it approved now, it would get merged in about an hour.08:44
haypoit's fun to see the size of the queue depending on the hour, we can see that openstack is not really developed in all timezones08:44
haypoIMHO we depend a lot on america timezones :)08:44
dulekhaypo: Yes we do. :) But I like the fact my patches are checked super-fast in my working hours. ;)08:45
dulekhaypo: And get merged overnight. :)08:45
*** lpetrut has quit IRC08:47
openstackgerritMichal Dulko proposed openstack/cinder: Move deprecation release notes to correct section
*** jordanP has joined #openstack-cinder08:49
*** markvoelker has joined #openstack-cinder08:53
*** markvoelker has quit IRC08:57
*** haomaiwang has quit IRC09:01
*** haomaiwa_ has joined #openstack-cinder09:01
sheeldulek: hi09:07
sheeldulek: need some clarification on
sheeldulek: could we discuss that now?09:07
duleksheel: Sure.09:08
sheeldulek: ok, regarding your comment09:08
sheeldulek: I think tests will fail without my fix09:08
sheelas I have specifically written one more test09:08
sheel                       '_volume_glance_metadata_key_to_id',09:08
sheel                       return_value = '1')09:08
sheel    def test_volume_glance_metadata_key_to_id_called(self,09:08
sheel                                                     metadata_key_to_id_mock)09:08
sheelthis checks if mocked function is called once or not09:09
sheelif fix is not there test will fail09:09
sheeldulek: is this suffice your comment?09:09
sheeldulek: you can refer complete testcase here :
duleksheel: Oh, yeah, but that's like a white-box testing. If someone reworked volume_metadata_delete to still use _volume_glance_metadata_key_to_id but changed filter_by clause to use key=key again, it would still pass.09:11
sheeldulek: haha09:11
duleksheel: And it wouldn't find the regression.09:11
sheeldulek: yes09:12
sheeldulek: you are right09:12
duleksheel: Question is - is this testableā€¦?09:12
sheeldulek: I think hard to test09:12
sheeldulek: will still have to devote much more time for this09:12
sheeldulek: I think this is very rare case where someone will change key in code, though possible09:13
duleksheel: You can probably mock filter_by and check if it was called with correct params.09:14
duleksheel: But I've removed -1. :)09:14
sheeldulek: oh thanks09:14
*** rcernin has quit IRC09:14
sheeldulek: I think we can let this go as this suffice the case upto some extent09:14
sheeldulek: later on i will add tests09:15
sheeldulek: is it ok?09:15
dulekYeah, cool.09:15
sheeldulek: thanks ..09:15
duleksheel: BTW - you've seen post on the openstack-dev about case sensitivity issues?09:15
sheeldulek: no09:15
duleksheel: They have similar problem, we probably should watch how they solve it to be consistent.09:16
sheeldulek: ah, yes09:16
sheeldulek: i will follow this09:17
duleksheel: Thanks!09:18
sheeldulek: but we are able to create things with seperate cases..09:18
sheeldulek: will check further09:18
*** jwcroppe has joined #openstack-cinder09:19
*** salv-orlando has quit IRC09:19
*** salv-orlando has joined #openstack-cinder09:19
*** laughterwym has joined #openstack-cinder09:23
*** jistr has joined #openstack-cinder09:27
*** laughterwym has quit IRC09:28
*** rcernin has joined #openstack-cinder09:28
*** _alastor1 has joined #openstack-cinder09:32
*** aagrawal has joined #openstack-cinder09:32
*** jordanP has quit IRC09:32
*** laughterwym has joined #openstack-cinder09:33
*** jwcroppe_ has joined #openstack-cinder09:34
*** ishant has joined #openstack-cinder09:35
*** jwcroppe has quit IRC09:35
*** _alastor_ has quit IRC09:35
*** ankit_ag has quit IRC09:35
*** wanghao has quit IRC09:35
*** wanghao has joined #openstack-cinder09:35
*** yangyape_ has joined #openstack-cinder09:42
*** asselin__ has joined #openstack-cinder09:44
*** yangyapeng has quit IRC09:45
*** jordanP has joined #openstack-cinder09:47
openstackgerritMichal Dulko proposed openstack/cinder: Adding general notes about rolling upgrades
*** yangyape_ has quit IRC09:47
*** asselin_ has quit IRC09:48
*** ndipanov has joined #openstack-cinder09:50
*** tpsilva has joined #openstack-cinder09:50
*** jwcroppe_ has quit IRC09:53
*** EinstCrazy has quit IRC09:55
*** davechen has left #openstack-cinder09:55
*** grassy is now known as davechen_afk09:56
*** salv-orlando has quit IRC10:00
*** haomaiwa_ has quit IRC10:01
*** cheneydc has quit IRC10:01
*** haomaiwang has joined #openstack-cinder10:01
*** laughterwym has quit IRC10:03
*** laughterwym has joined #openstack-cinder10:04
*** salv-orlando has joined #openstack-cinder10:08
*** laughterwym has quit IRC10:08
*** qeelee has quit IRC10:09
*** salv-orlando has quit IRC10:11
*** salv-orlando has joined #openstack-cinder10:12
*** haomaiwang has quit IRC10:23
*** lpetrut has joined #openstack-cinder10:26
*** sdague has joined #openstack-cinder10:27
*** ishant has quit IRC10:27
*** haomaiwang has joined #openstack-cinder10:29
openstackgerritMichal Dulko proposed openstack/cinder: Add devref on rolling upgrades
openstackgerritVipin Balachandran proposed openstack/cinder: Mark oslo.vmware as optional dependency
*** erlon has joined #openstack-cinder10:48
*** sgotliv has joined #openstack-cinder10:49
*** jwcroppe has joined #openstack-cinder10:50
*** markvoelker has joined #openstack-cinder10:54
*** salv-orlando has quit IRC10:57
*** markvoelker has quit IRC10:58
*** yhayashi has quit IRC11:00
*** haomaiwang has quit IRC11:01
*** haomaiwang has joined #openstack-cinder11:01
yuriy_n17geguileo: Hi! Very sorry for distusbing. Please review whenever you have free time. Thanks a lot for advance.11:03
yuriy_n17DuncanT:  Hi! Very sorry for distusbing. Please review whenever you have free time. Thanks a lot for advance.11:04
*** sgotliv has quit IRC11:08
openstackgerritjaveme proposed openstack/cinder-specs: Implement the oslo_db enginefacade proposal
*** haomaiwang has quit IRC11:14
*** eantyshev has joined #openstack-cinder11:19
*** eantyshev has left #openstack-cinder11:19
*** eantyshev has joined #openstack-cinder11:20
*** qeelee has joined #openstack-cinder11:20
*** jwcroppe has quit IRC11:23
*** Poornima has quit IRC11:25
openstackgerritjaveme proposed openstack/cinder: Remove RequestBodySizeLimiter from middleware
eantyshevHello! Are Cinder gate tests broken? I don't see anything merged in the past 32 hours11:26
*** EinstCrazy has joined #openstack-cinder11:32
dulekeantyshev: Yup.11:36
dulekeantyshev: - this will fix it, but still not +A.11:36
eantyshevdulek: Thanks! Will give it my +1 :^)11:39
*** lprice has joined #openstack-cinder11:41
*** lprice1 has quit IRC11:41
*** smoriya_ has quit IRC11:56
*** qeelee has quit IRC11:58
openstackgerritScott DAngelo proposed openstack/python-cinderclient: api-microversion support for python-cinderclient
*** gcb has quit IRC12:10
*** gcb has joined #openstack-cinder12:11
*** Lee1092 has quit IRC12:12
*** ccesario has quit IRC12:12
*** krotscheck_dcm is now known as krotscheck12:13
*** haomaiwang has joined #openstack-cinder12:14
*** cdelatte has joined #openstack-cinder12:18
*** haomaiwang has quit IRC12:19
*** jwcroppe has joined #openstack-cinder12:20
*** salv-orlando has joined #openstack-cinder12:20
*** zhipeng has quit IRC12:21
*** ccesario has joined #openstack-cinder12:25
*** raildo-afk is now known as raildo12:32
*** laughterwym has joined #openstack-cinder12:33
*** jordanP has quit IRC12:35
*** jordanP has joined #openstack-cinder12:36
*** markvoelker has joined #openstack-cinder12:40
*** dalgaaf has joined #openstack-cinder12:40
*** takedakn has joined #openstack-cinder12:43
*** markvoelker has quit IRC12:44
*** julim has joined #openstack-cinder12:45
*** takedakn has quit IRC12:50
*** abhishekk_ has quit IRC12:52
*** jwcroppe has quit IRC12:53
*** gcb has quit IRC12:55
*** aagrawal has quit IRC12:57
geguileoyuriy_n17: Done13:18
*** markvoelker has joined #openstack-cinder13:20
openstackgerritScott DAngelo proposed openstack/cinder: Allow api_version_request.matches to accept a string
scottdasheel: ^^^13:20
scottdaI did not add the defaults for "min_verson=None, max_version=None"13:21
sheelscottda: great13:21
sheelscottda: so i need to pass 3.1, None13:21
sheelor may be only min version13:22
scottdaI want to think on that and run some tests first13:22
scottdasheel: I think for now you need "3.1", "3.1"13:22
openstackgerritAvishay Traeger proposed openstack/cinder: List manageable volumes and snapshots
scottdaBut test with passing "3.1", None13:22
*** ric has joined #openstack-cinder13:23
*** cknight has joined #openstack-cinder13:23
scottdaI'm just not sure about setting the defaults. But you can at least list that patch as a dependency and carry on with your testing, etc.13:23
sheelyes, sure13:23
sheeli ll keep that as dependent13:24
sheelscottda: will test and ping u in second half if all went well13:25
*** cknight has quit IRC13:25
yuriy_n17geguileo: Thank you ever so much.13:25
*** merooney has joined #openstack-cinder13:30
*** merooney has quit IRC13:34
*** cbader has quit IRC13:34
*** gouthamr has joined #openstack-cinder13:35
*** porrua has joined #openstack-cinder13:35
openstackgerritJordan Pittier proposed openstack/os-brick: Mock time.sleep() in 3 unit tests
*** baumann has joined #openstack-cinder13:35
*** akerr has joined #openstack-cinder13:37
*** haomaiwang has joined #openstack-cinder13:38
*** alkhodos has quit IRC13:42
*** merooney has joined #openstack-cinder13:43
*** ric has quit IRC13:47
*** cbader has joined #openstack-cinder13:47
*** jwcroppe has joined #openstack-cinder13:50
*** edmondsw has joined #openstack-cinder13:50
*** dims has joined #openstack-cinder13:54
*** jwcroppe has quit IRC13:58
openstackgerritScott DAngelo proposed openstack/cinder: Allow api_version_request.matches to accept a string
*** merooney has quit IRC13:59
scottdasheel: ^^ Fixed a bug, and added unit test for passing in api_version_request.matches("3.2", None) or (None, "3.2")13:59
scottdaThat should be good. I don't think we need default values = None for the args to matches()14:00
*** haomaiwang has quit IRC14:01
*** haomaiwang has joined #openstack-cinder14:01
sheelscottda: ok, nice14:02
openstackgerritJordan Pittier proposed openstack/os-brick: Actually run the RemoteFSClient unit tests
sheelscottda: sounds good14:03
jordanPI am on a "fix unit tests" spree :)14:03
jordanP(that's to be nice with the community, in the hope that by next patch, which is vendor specific, is going to be reviewed with kindess :p)14:05
scottdajordanP: You get a sticker for that. I cannot promise any kindness from anyone around here :)14:05
*** rlrossit has joined #openstack-cinder14:06
*** zhangjn has quit IRC14:08
*** zhangjn has joined #openstack-cinder14:09
*** jungleboyj has quit IRC14:11
openstackgerritMerged openstack/cinder: Tests: Don't assert on LOG.warn
openstackgerritMannu proposed openstack/cinder: Return BadRequest for invalid Unicode names
*** alyson_ has joined #openstack-cinder14:12
*** takedakn has joined #openstack-cinder14:13
*** takedakn has quit IRC14:14
*** EinstCrazy has quit IRC14:14
*** avishay has joined #openstack-cinder14:15
*** xyang1 has joined #openstack-cinder14:17
*** diablo_rojo has joined #openstack-cinder14:22
*** xyang has joined #openstack-cinder14:24
*** laughterwym has quit IRC14:25
*** anshul has quit IRC14:25
*** yangyapeng has joined #openstack-cinder14:26
*** nkrinner has quit IRC14:26
*** cknight has joined #openstack-cinder14:27
*** merooney has joined #openstack-cinder14:30
*** timcl has joined #openstack-cinder14:32
openstackgerritVictor Stinner proposed openstack/cinder: Enable all unit tests on Python 3.4
*** dansmith is now known as superdan14:36
*** laughterwym has joined #openstack-cinder14:37
*** jgregor has joined #openstack-cinder14:39
*** eharney has joined #openstack-cinder14:42
*** dustins has joined #openstack-cinder14:43
*** knikolla has joined #openstack-cinder14:43
openstackgerritEric Harney proposed openstack/cinder: Delete volumes with snapshots
smcginnisgeguileo: If there's another gate blocker, feel free to +A right away to get those through.14:45
*** larsks has joined #openstack-cinder14:45
*** chhavi has joined #openstack-cinder14:45
larskseharney: Does cinder have anything like nova's rpc version pinning, so that a version N-1 cinder-volume can talk to a version N cinder-api/scheduler/etc?14:46
*** gcb has joined #openstack-cinder14:46
openstackgerritEric Harney proposed openstack/cinder: Add missing requirements
*** crose has joined #openstack-cinder14:51
xyangsmcginnis: is anyone looking at this error in cinder.tests.unit.volume.drivers.netapp.dataontap.test_block_base.NetAppBlockStorageLibraryTestCase.test_get_reserved_percentage.14:52
akerrxyang: there's a patch up already14:52
akerrxyang: I think it just merged14:52
xyangakerr: great14:52
xyangakerr: thanks14:53
*** jwcroppe has joined #openstack-cinder14:54
xyangmerooney: thanks14:54
*** _alastor1 has quit IRC14:55
openstackgerritGage Hugo proposed openstack/cinder: Removed duplicated code in VNX cinder driver
openstackgerritScott DAngelo proposed openstack/cinder: Allow api_version_request.matches to accept a string
*** cknight has quit IRC14:58
*** kvidvans has quit IRC14:58
*** kvidvans has joined #openstack-cinder14:58
*** mriedem has joined #openstack-cinder14:58
scottdaWhen I post a patch with "Closes-Bug: <bug>" , isn't that supposed to show up in the bug report?14:59
eharneylarsks: yes, it does, there has been a fair amount of mitaka work to add that15:00
larskseharney: I saw the bp and some patches, but I wasn't sure if the feature was entirely implemented yet.  Sounds like yes?15:00
smcginnisscottda: Should. I noticed another one that didn't though.15:00
smcginnisscottda: I wonder if something broke there.15:00
*** edtubill has joined #openstack-cinder15:00
*** haomaiwang has quit IRC15:01
scottdasmcginnis: I'll keep an eye on it. Might just be a lag.15:01
eharneylarsks: that i'm not sure about -- i'm not sure what may be remaining still15:01
*** haomaiwang has joined #openstack-cinder15:01
larskseharney: okay, thanks.15:01
*** laughterwym has quit IRC15:01
*** arecknag has quit IRC15:02
sheelscottda: hey, I too noticed this thing15:02
sheelscottda: but after 2-3 houts it was showing15:03
sheelscottda: so seems some lag sometimes15:03
scottdasheel: Thanks. BTW, with your filters patch, are there any existing unit tests for this function? It would be nice to add/change UT for different microversion behaviour.15:04
*** delattec has joined #openstack-cinder15:04
sheelscottda: ah yes15:05
sheelscottda: let me check15:05
*** cdelatte has quit IRC15:06
*** laughterwym has joined #openstack-cinder15:06
sheelscottda: here you go15:08
sheelvery small tests though15:08
*** merooney has quit IRC15:08
scottdasheel: So, how does the behaviour change with your fix? Can this be unit tested?15:09
*** arch-nemesis has joined #openstack-cinder15:09
sheelscottda: actuall before this fix, filtering was considering everything as True15:10
sheelscottda: except if we explicitely pass False15:10
scottdasheel: If we can test the change, we should be able to test 2 ways: 1 with no_microversion | old_microversion and old behaviour, and one with new_microversion and new behaviour15:10
scottdasheel: OK, it's something we can think on.15:11
openstackgerritYuriy Nesenenko proposed openstack/cinder: Check for service existance in capabilities API
sheelscottda: but in this patch it will raise error for anything passed other than o/false/False/1/true/True15:11
*** jwcroppe has quit IRC15:11
sheelscottda: right i need to right test for this15:11
sheel I need to write15:12
sheelscottda: give me 15 min15:12
sheelscottda: will add to it15:12
sheelscottda: ok?15:12
scottdasheel: sounds goog15:13
sheeloh, I was checking my dictionary if it has goog15:13
sheeljust kidding15:13
*** laughterwym has quit IRC15:14
sheelscottda: here i go..15:17
*** JoseMello has joined #openstack-cinder15:17
sheelscottda: will add to patch15:18
sheelin a while15:18
mc_nairknow we're racing towards feature freeze, but anyone who gives a review will get some BBQ or beer in Austin15:18
mc_nairpure bribery time15:18
*** cknight has joined #openstack-cinder15:19
openstackgerritRonald Bradford proposed openstack/cinder: Define context.roles with base class
*** merooney has joined #openstack-cinder15:21
*** IlyaG has joined #openstack-cinder15:22
*** aorourke-afk is now known as aorourke15:24
openstackgerritScott DAngelo proposed openstack/cinder: Allow api_version_request.matches to accept a string
*** IlyaG has quit IRC15:27
*** delatte has joined #openstack-cinder15:28
*** lprice has quit IRC15:30
*** gcb has quit IRC15:30
*** delattec has quit IRC15:30
*** Yogi1 has joined #openstack-cinder15:33
*** lcurtis has joined #openstack-cinder15:39
*** belmoreira has quit IRC15:43
openstackgerritJacob Gregor proposed openstack/cinder: Storwize SVC multiple management IPs
sheelsmcginnis: hi there15:43
sheel turned green15:43
smcginnissheel: OK, thanks.15:44
smcginnissheel: Looks like the recheck is still running.15:45
sheelsmcginnis: all voting tests passed15:45
smcginnissheel: Oh, still working through but all tests are passing?15:46
sheelsmcginnis: yes all are passing, only one test is left, will hardly take atmost 30 seconds15:46
smcginnisThe review itself has the last jenkins result from about 10 hours ago.15:46
smcginnissheel: OK, thanks.15:46
*** mriedem has quit IRC15:49
sheelsmcginnis: thank you15:49
openstackgerritAvishay Traeger proposed openstack/python-cinderclient: List manageable volumes and snapshots
openstackgerritAvishay Traeger proposed openstack/python-cinderclient: List manageable volumes and snapshots
openstackgerritJordan Pittier proposed openstack/os-brick: Fix Scality SOFS support
*** avishay has quit IRC15:53
*** mriedem has joined #openstack-cinder15:53
*** mtanino has joined #openstack-cinder15:54
*** mriedem is now known as mriedem_meeting15:54
*** yangyapeng has quit IRC15:56
*** yangyapeng has joined #openstack-cinder15:56
*** yangyapeng has quit IRC15:56
*** Yogi1 has quit IRC15:58
*** salv-orl_ has joined #openstack-cinder15:59
*** jeanfabrice has quit IRC16:00
*** arch-nemesis has quit IRC16:00
*** lprice has joined #openstack-cinder16:00
*** martyturner has joined #openstack-cinder16:00
*** haomaiwang has quit IRC16:01
*** haomaiwang has joined #openstack-cinder16:01
*** salv-orlando has quit IRC16:01
*** hemnafk is now known as hemna16:04
*** haplo37 has joined #openstack-cinder16:06
*** _alastor1 has joined #openstack-cinder16:08
*** mriedem_meeting has quit IRC16:08
sheelxyang: yes its merged16:09
sheelxyang: sorry, other group16:09
smcginnishemna: Morning! Curious, did you ever see a speed improvement with all that caching work you did?16:11
hemnastill haven't plugged it into our CI yet16:11
hemnathe box that had my mirror died16:12
smcginnishemna: Bleh16:12
smcginnishemna: Any non-CI, local testing show an improvement?16:12
hemnamy local apt-get stuff worked well16:12
hemnaand was much faster16:12
hemnano 503 errors either16:12
smcginnishemna: Nice.16:13
smcginnisMaybe I should give it another try here.16:13
sheeleharney: hi16:13
eharneysheel: hi16:13
smcginnishemna: What packages did you end up setting up again? apt-cache-ng?16:13
hemnaapt-mirror is what I used16:13
hemnaI can send you my conf16:13
smcginnishemna: Cool, thanks for the pointer.16:13
*** logan- has quit IRC16:14
sheeleharney: regarding your review on
*** logan- has joined #openstack-cinder16:14
sheeleharney: i uploaded this patch set for verbose that you wanted to add16:14
*** mriedem has joined #openstack-cinder16:14
*** laughterwym has joined #openstack-cinder16:14
eharneysheel: this is not the right approach16:14
hemnathat pulls 64bit and 32bti16:14
hemnaand puts everything in /mnt/apt-mirror16:14
smcginnishemna: Is 32bit needed?16:15
sheeleharney: is this adding some complexity?16:15
sheeleharney: just out of curiosity..16:15
eharneysheel: yes? all of the code related to parsing Unicode etc...16:15
eharneysheel: which we don't need to achieve the right result16:15
smcginnishemna: Good thing I've got some storage laying around. ;)16:16
hemnasmcginnis, after you have that mirrored, then it's a simple matter of apache pointing to it.16:16
hemnasmcginnis, :)16:16
sheelsheel: oh16:16
sheeleharney: oh16:16
hemnasmcginnis, I used a 3PAR iSCSI volume for it...good thing since the box serving it up burned.16:16
smcginnishemna: Yeah, at least you don't need to pull it all down again.16:16
*** sgotliv has joined #openstack-cinder16:17
*** ntpttr has joined #openstack-cinder16:17
sheeleharney: so could you suggest some approach so that i could mould it16:17
sheeleharney: if possible only as good gesture..16:17
sheeleharney: thanks in advance16:17
yuriy_n17hemna: Hi! Very sorry for disturbing. Please review whenever you have free time.16:17
openstackgerritEric Harney proposed openstack/cinder: Return BadRequest for invalid Unicode names
*** mriedem has left #openstack-cinder16:18
*** mriedem has joined #openstack-cinder16:18
eharneysheel: i'm about to add it to this patch, all we need is a LOG.exception there and we're good ^16:19
hemnayuriy_n17, I have, and I kind think that DuncanT is right16:19
eharney(rebased to kick it out of the gate since i approved it before realizing that was missing)16:19
sheeleharney: ok16:19
*** laughterwym has quit IRC16:19
hemnaI think maybe we should just run sudo cinder local-attach16:19
*** yuriy_n17 has quit IRC16:21
*** dims has quit IRC16:22
*** _alastor1 has quit IRC16:23
*** jdurgin1 has joined #openstack-cinder16:24
openstackgerritEric Harney proposed openstack/cinder: Return BadRequest for invalid Unicode names
mriedemnew ceph job bug
openstackLaunchpad bug 1550409 in Cinder ""Detected volume stuck in downloading status, setting to ERROR." in ceph job" [Undecided,New]16:26
eharneysheel: try out that patch and see what you think ^16:26
sheeleharney: it seems ok16:27
sheeleharney: but no information about which field is craeting problem16:27
sheel"Error writing field to database"16:27
sheeleharney: its quite generalized16:27
sheeleharney: may be we should be a bit specific16:27
eharneysheel: the logged exception will give that info16:27
openstackgerritMichael Price proposed openstack/cinder: NetApp: Add Consistency Group support for E-Series
sheeleharney: you mean, it will output field name as well?16:28
sheeleharney: if so then quite good16:28
*** garthb has joined #openstack-cinder16:28
*** garthb_ has joined #openstack-cinder16:28
eharneysheel: yes, try it out and you will see a message in the log that says "Incorrect string value: '\\xF0\\x9F\\xA0\\x80' for column 'display_name'"16:28
sheeleharney: great16:29
eharneysheel: oslo.db etc already handles this stuff for us, we don't have to code for it too16:29
sheeleharney: i guess you suggest this approach16:29
sheeleharney: to committer for sure16:29
sheeleharney: nice fix16:29
*** Lee1092 has joined #openstack-cinder16:29
*** merooney has quit IRC16:30
*** rcernin has quit IRC16:31
*** darosale has joined #openstack-cinder16:34
*** mss has joined #openstack-cinder16:34
*** timcl has quit IRC16:35
*** garthb has quit IRC16:36
*** garthb_ has quit IRC16:36
openstackgerritNate Potter proposed openstack/cinder: Don't allow operations on deleting volumes
*** merooney has joined #openstack-cinder16:39
*** diablo_rojo has quit IRC16:40
openstackgerritMerged openstack/cinder: Correct iscsi lun type
*** garthb has joined #openstack-cinder16:44
jgriffithntpttr: ping16:47
openstackgerritDavid Rosales proposed openstack/cinder: Change Fail to Failed in error messages
jgriffithntpttr: sorry for the -2... but let me know if the comments don't make sense:
ntpttrjgriffith: hey16:47
ntpttrjgriffith: no problem, I was pretty sure there was a better way to go about it16:48
jgriffithntpttr: yeah...16:48
jgriffithntpttr: so sadly we have a lack of discipline in how we do things so there's various things scattered all over the place.  If I happen to see them I chime in16:48
jgriffithntpttr: it was your lucky day :)16:48
ntpttrjgriffith: what do you mean by various things?16:49
jgriffithntpttr: so status checks are a good example16:49
jgriffithntpttr: rather than just by convention making those checks in the API layer... we sometimes do them there, sometimes in the manager and sometimes in both :(16:50
ntpttrjgriffith: ah yeah your comment does make sense, and I was thinking that was needed - hence my comment after uploading the first patch set - but I wasn't sure what check would go there16:50
jgriffithntpttr: which is one of the reasons we have issues with races the way we do IMO16:50
ntpttrjgriffith: ah yes, like the quota checks from that bug I was looking at a while back16:50
jgriffithntpttr: YES!!!  Exactly!!!!16:50
jgriffithntpttr: we have the same sort of problem with other calls too16:51
jgriffithntpttr: ie we left things open to interpretation and now there's some really "weird" code in various places for everybody's "one-off" patch16:51
ntpttrjgriffith: so, before the delete call is issued from the API, we need to be checking the status of the volume to make sure nothing else is going on with it16:51
jgriffithntpttr: yes, that would be the correct thing here IMO16:51
ntpttrjgriffith: but things like initialize_connection don't change volume status do they?16:51
ntpttrjgriffith: that's why I wasn't sure what check to include16:52
jgriffithntpttr: although I'm sure somebody removed that at some point for their *special case* they wanted to solve16:52
jgriffithntpttr: they don't need to.. because prior to calling initialize-connection you need to call "reserve" which changes it16:52
jgriffithntpttr: so that's a confusing thing16:52
jgriffithand I think part of why understanding has diverged so much16:53
ntpttrjgriffith: okay that makes sense, so when the volume is reserved it's no longer available - so we need to check to see that it's not in some status other than available or error before deleting16:53
jgriffithntpttr: so when it comes to attach/detach... basically the only two calls that were every intended to "do" anything were intialize and terminate connection16:53
dulekIt's Friday evening, so I'm unable to look into it, but I've seen 2 failures of gate-tempest-dsvm-full-ceph on rechecked patches ( - one of them).16:53
dulekJust a heads up.16:53
jgriffithntpttr: the "reserve/attach" and "begin term / detach" were just DB status updates and call-backs16:54
jgriffithntpttr: but that's changed dramatically16:54
jgriffithntpttr: sorry... "reserve/attach" and "begin_detaching/detach" are the analogous calls I was referring to16:55
ntpttrjgriffith: ah okay that's starting to come together now. I didn't know about that!16:55
ntpttrjgriffith: thanks for the help/info :)16:55
jgriffithntpttr: sure16:55
jgriffithntpttr: BTW
jgriffithntpttr: I'm going to update it with the detach stuff right now16:56
* jgriffith should've done it years ago16:56
ntpttrjgriffith: awesome, I'll check it out16:56
*** Yogi1 has joined #openstack-cinder16:56
ntpttrit would be cool to have recommended reading stuff like this linked in the wiki somewhere maybe, otherwise it's not easy for new folks to find it all16:56
jgriffithntpttr: yeah... I wanted to start updating the devref docs in tree16:57
jgriffithntpttr: but I've found that that's not so good because they get neglected and nobody updates them16:57
jgriffithntpttr: incorrect info is worse than no info16:57
ntpttrjgriffith: true16:58
scottdajgriffith: ++ . We could stand to document these architectural/design standards and best practices, so we can have consistency for new code.16:58
Swansonjgriffith, make updating the wiki a requirement to merge?16:58
scottdaSwanson: or updating the devref16:58
jgriffithscottda: NOOO wikis!16:58
jgriffitherrr... Swanson16:58
jgriffithscottda: +116:59
Swansonjgriffith, scottda,  yeah, devref.16:59
jgriffithscottda: I was just looking there because I know I submitted that write up there as well... not sure it ever merged16:59
scottdaIMO, devref changes should be included in patches with major functional changes, so they can be reviewed at the same time and gate merging the feature or fix.16:59
Swansonjgriffith, v2.1 is never going to merge.  look forward to seeing it in newton.16:59
*** asselin_ has joined #openstack-cinder16:59
jgriffithscottda: agreed!16:59
jgriffithSwanson: huh?  Why do you say that?17:00
jgriffithSwanson: of course if that's how it goes I'm fine with it17:00
*** merooney has quit IRC17:00
*** haomaiwang has quit IRC17:01
jgriffithSwanson: but I thought after the unit test fix was in a recheck would allow it to just "breeze on by" :)17:01
*** apoorvad has joined #openstack-cinder17:01
jgriffithscottda: ntpttr I'll work up a devref docs patch for the attach/detach stuff right now17:01
*** haomaiwang has joined #openstack-cinder17:01
*** asselin__ has quit IRC17:01
Swansonjgriffith, I'm just a firm believer that gerrit is always darkest right before it fails completely.  So I have to assume it'll be another 29 hours before we see the next ceph failure.  Then god help us all.17:02
jgriffithWell that could be17:03
*** jgregor has quit IRC17:03
jgriffithand all the additions to eharney 's patch last night didn't help the timeline :)17:03
eharneythat thing did take a while to squeak through17:04
jgriffitheharney: yeah, then a rebase at the 11'th hour, followed by another add and then another17:05
jgriffitheharney: I was watching it in ancicipation and it drove me to drink :)17:05
*** jordanP has quit IRC17:06
*** mss is now known as _alastor_17:07
ntpttrjgriffith: would the best way to go about fixing the delete issue be to more the checks here up to the top of the function? ? Or were those put there for some reason? Seems odd that if the check is needed it would come after a possible deletion happens earlier in the function17:08
jgriffithntpttr: yes, IMO that's what we should have17:09
*** dims has joined #openstack-cinder17:09
*** leeantho has joined #openstack-cinder17:09
jgriffithntpttr: so we never get to a point where we try and delete and attached/attaching volume17:09
*** dims has quit IRC17:10
*** chris_morrell has joined #openstack-cinder17:12
*** jklare_ has joined #openstack-cinder17:12
*** melwitt_ has joined #openstack-cinder17:12
*** rhagarty__ has joined #openstack-cinder17:12
*** xek__ has joined #openstack-cinder17:13
*** IlyaG has joined #openstack-cinder17:13
hemnasmcginnis, ping17:14
*** j_king_ has joined #openstack-cinder17:15
*** Trident has joined #openstack-cinder17:15
*** diegows_ has joined #openstack-cinder17:16
*** guitarza1 has joined #openstack-cinder17:16
*** mikal_ has joined #openstack-cinder17:16
_alastor_hey folks.  I'm trying to get this patch ( to pass the gate, but I'm getting failures in drivers unrelated to mine (mostly it seems to be netapp)17:16
*** tonyb_ has joined #openstack-cinder17:16
*** spring_ has joined #openstack-cinder17:16
*** jeanfabrice has joined #openstack-cinder17:16
*** Nakato_ has joined #openstack-cinder17:17
_alastor_For example, the failure listed in this gate check ( is showing a failure in the netapp driver17:17
*** jungleboyj has joined #openstack-cinder17:18
eharney_alastor_: it's been fixed in master now, just issue a "recheck"17:18
_alastor_Ok, thanks!17:18
*** gouthamr_ has joined #openstack-cinder17:19
_alastor_Is the easiest way to recheck to `git commit --amend` then `git review` or is there a way to do it on the web UI?17:20
*** edmondsw has quit IRC17:20
*** gouthamr has quit IRC17:20
*** xek_ has quit IRC17:20
*** zul has quit IRC17:20
*** rhagarty_ has quit IRC17:20
*** ekarlso- has quit IRC17:20
*** guitarzan has quit IRC17:20
*** Nakato has quit IRC17:20
*** anteaya has quit IRC17:20
*** krotscheck has quit IRC17:20
*** davechen_afk has quit IRC17:20
*** _fortis has quit IRC17:20
*** jklare has quit IRC17:20
*** j_king has quit IRC17:20
*** tonyb has quit IRC17:20
*** mikal has quit IRC17:20
*** mkoderer__ has quit IRC17:20
*** thurloat has quit IRC17:20
*** strictlyb has quit IRC17:20
*** bswartz has quit IRC17:20
*** afazekas has quit IRC17:20
*** bapalm has quit IRC17:20
*** Tridde has quit IRC17:20
*** diegows has quit IRC17:20
*** melwitt has quit IRC17:20
*** krotscheck has joined #openstack-cinder17:20
*** mkoderer___ has joined #openstack-cinder17:20
hemnatype recheck in the review17:20
eharney_alastor_: you can just leave a comment in gerrit that says "recheck"17:21
*** afazekas has joined #openstack-cinder17:21
*** bapalm has joined #openstack-cinder17:21
*** strictlyb has joined #openstack-cinder17:21
_alastor_eharney: ok17:21
*** thurloat has joined #openstack-cinder17:21
*** dims has joined #openstack-cinder17:22
openstackgerritNate Potter proposed openstack/cinder: Stop deleting operation on busy volumes
ntpttr^^ something went wrong there, it didn't upload the correct changes. I'll try it again17:23
openstackgerritNate Potter proposed openstack/cinder: Stop deleting operation on busy volumes
*** jeanfabrice has quit IRC17:25
*** rlrossit has quit IRC17:26
*** ekarlso- has joined #openstack-cinder17:27
*** edmondsw has joined #openstack-cinder17:27
*** zul has joined #openstack-cinder17:27
*** timcl has joined #openstack-cinder17:28
*** anteaya has joined #openstack-cinder17:28
*** _fortis_ has joined #openstack-cinder17:28
*** anshul has joined #openstack-cinder17:31
*** lpetrut has quit IRC17:32
*** [1]Thelo has joined #openstack-cinder17:33
*** Thelo has quit IRC17:35
*** _fortis_ is now known as _fortis17:35
*** [1]Thelo is now known as Thelo17:35
*** lpetrut has joined #openstack-cinder17:37
*** porrua has quit IRC17:40
*** jdurgin1 has quit IRC17:41
*** laughterwym has joined #openstack-cinder17:44
*** IlyaG has quit IRC17:47
*** laughterwym has quit IRC17:48
*** jistr has quit IRC17:49
*** merooney has joined #openstack-cinder17:50
*** xyang has quit IRC17:52
*** xyang has joined #openstack-cinder17:52
openstackgerritSheel Rana proposed openstack/cinder: Added Bootable filter for cinder list command
_alastor_eharney: I posted the 'recheck' comment, but I can't see anything showing that a recheck is scheduled.  It doesn't show on tintri, or on the review itself17:54
sheelDuncanT: ^^17:54
sheel_alastor_: you wont see anything on review page for recheck17:55
DuncanTsheel: I'm waiting for scottda's +1 on that before I go through it, he's our current expert17:57
sheeloh I mistakenly mentioned you17:57
sheelDuncanT: I was also thinking of pinging scottda17:58
_alastor_sheel: ok, thanks.  I'll sit tight and hope for the best :)17:58
sheelDuncanT: sorry17:58
sheel_alastor_: hey no, you can see this on zool17:58
sheelwhats your review id17:58
_alastor_sheel: 28496717:58
DuncanTsheel: I need to apologise, I was just letting you know that I was waiting. Worth getting the first one as right as we can17:59
openstackgerritAdriano Freires Rosso proposed openstack/cinder: WIP: HNAS drivers refactoring
sheelDuncanT: no no, its ok...really i too wanted to get reviewed by scottda as we were discussing it since last day18:00
sheelDuncanT: sorry for making you feel like that18:00
sheelDuncanT: actually i am expecting -1 on this due to tests18:00
sheelDuncanT: :)18:00
openstackgerritJohn Griffith proposed openstack/cinder: Add basic workflow of attach/detach to devref
*** haomaiwang has quit IRC18:01
jgriffithntpttr: scottda smcginnis ^^18:01
jgriffiththere's a start at any rate18:01
*** haomaiwang has joined #openstack-cinder18:01
sheelDuncanT: once scottda passes it, i will ping you for review...18:01
openstackgerritAdriano Freires Rosso proposed openstack/cinder: WIP: HNAS drivers refactoring
*** ndipanov has quit IRC18:02
*** sgotliv has quit IRC18:02
sheel_alastor_: hey its in rechecking state18:03
_alastor_sheel: cool. Thanks for checking18:04
sheel_alastor_: a lot of requests are there, so may be take some long time to complete18:04
_alastor_sheel: Any way I can see that queue?18:04
sheel_alastor_: no problem, you can also check on
*** ChubYann has joined #openstack-cinder18:05
*** porrua has joined #openstack-cinder18:06
scottdasheel: Sorry, I'm heading out early today, so won't get to it until monday.18:06
sheelscottda: no issues18:06
sheelscottda: happy weekend18:06
ntpttrjgriffith: thanks, I'll check it out18:06
sheelscottda: take rest, its been long microversionnig ..18:07
scottdasheel: For unit tests, I was thinking you could use one of the patterns in cinder/tests/unit/api/test_versions.py18:07
scottdasheel: In that way, you could see that the behaviour differed base on what version of the api was requested.18:07
sheelscottda: ok, sure18:07
scottdasheel: Thanks!18:08
*** IlyaG has joined #openstack-cinder18:08
mc_nairDuncanT: know it's probably too late for you today, but think you'll get to review before the feature freeze?18:08
sheelscottda:  :)18:08
*** dims is now known as dimsum__18:09
DuncanTmc_nair: Yeah, I'll take a look. I'd really like to see some really thorough tests for this quota code, including some that really show all the ways it's supposed to work - we made a mistake of merging without that once, and it was rather painful18:09
sheelhey all18:10
sheelneed some eyes on ....18:10
sheelits waiting since long... :)18:10
sheel(quite small patch)18:10
*** lpetrut has quit IRC18:10
mc_nairDuncanT: yea, that's fair. I tried adding a decent amount of tests showing how it should work (it's over half the code in the patch), but there's nothing in Tempest currently.  I'm working on adding those as we speak. The saving grace of the nested quota stuff now is at least it's not on by default :)18:11
mc_nairDuncanT: but that also means I need to create a new Tempest job that uses the nested quota driver to get the coverage18:12
DuncanTmc_nair: Glad to hear the tempest stuff is being worked on. Sounds like a possible candidate for a nova-style big-ops test in the gate that just used the fake driver?18:12
mc_nairDuncanT: not aware of that.  Fake driver so it's not creating volumes and such? That would make sense.  I'll follow up with you once I get more of the coverage working18:14
sheelDuncanT: now i am looking towards you for some help in review ^^ ...18:14
openstackgerritMerged openstack/cinder: Huawei: Log the IP when login fails
mc_nairDuncanT: currently working though adding the hierarchical projects and being able to switch between which project we're using to create the volumes.  But that's starting to come along now18:15
openstackgerritMerged openstack/cinder: NetApp: Implement CGs for ONTAP Drivers
DuncanTmc_nair: great to hear, thanks.18:16
*** mriedem has quit IRC18:18
openstackgerritMerged openstack/cinder: Huawei: Don't fail when port group does not exist
openstackgerritMerged openstack/cinder: EMC VMAX - not cleaning up HW Resource WWPN initiators
openstackgerritJay Bryant proposed openstack/cinder: Add CGSnapshotStatus enum field
*** Yogi1 has quit IRC18:28
smcginnishemna: Still need something?18:28
*** timcl has quit IRC18:29
openstackgerritEric Harney proposed openstack/cinder: Delete volumes with snapshots
eharneyi'd appreciate some final review on -- i missed one thing on the last rebase, but it should be good now18:30
*** Suyash has joined #openstack-cinder18:31
*** bjornar__ has joined #openstack-cinder18:31
bjornar__How can I figure what options in cider.conf are spesific to api/volume/backup ?18:32
bjornar__Is this documented at all?18:32
sheelbjornar__: Hey18:35
sheelbjornar__: i think its clearly mentioned in cinder.conf18:35
sheelbjornar__: please refer "18:35
sheel# Options defined in cinder.backup.driver"18:35
sheel" Options defined in cinder.backup.drivers.ceph"18:35
sheel# Options defined in cinder.wsgi18:36
sheel# Options defined in cinder.volume.api18:36
sheelbjornar__: hope it will help in differentiating18:36
bjornar__sheel, but does this mean that section includes all the options for example cinder-volume will care about, or just the additional ones?18:37
*** timcl has joined #openstack-cinder18:37
sheelbjornar__: i think only more specific options18:38
bjornar__sheel, Thats what I mean..18:38
sheelbjornar__: for additional options I think i can point you to some other page18:38
sheelbjornar__: give me a sec, finding18:38
*** knikolla has quit IRC18:40
*** knikolla has joined #openstack-cinder18:40
openstackgerritApoorva Deshpande proposed openstack/cinder: Tintri image direct clone
sheelbjornar__: please refer
sheelbjornar__: for additional options of cinder.conf18:42
sheelbjornar__: you will find all options here in it, including different drivers etc18:44
bjornar__sheel, but still no documentation on what config options apply for the different services?18:44
sheelbjornar__: ah, only small  referential comments will be there in it18:45
*** fthiagogv has quit IRC18:46
sheellike "Table 2.66. Description of backups configuration options"18:46
*** haomaiwang has quit IRC18:46
bjornar__sheel, no basically "noone" knows what config options cinder-backup uses?18:46
*** AJaeger has joined #openstack-cinder18:47
bjornar__I mean, it probably uses database and mq.. so needs those options..18:47
openstackgerritAndreas Jaeger proposed openstack/cinder: always use pip constraints
bjornar__but probably only api needs [keystone_authtoken] .. and so on18:47
AJaegercinder cores, please review the change above together with the infra change
*** haomaiwang has joined #openstack-cinder18:48
AJaegersmcginnis: these two changes remove the constraints jobs and enable the new way of using them for cinder. A review of the infra change would be appreciated!18:48
smcginnisAJaeger: Will take a look now. Thanks!18:49
AJaegersmcginnis: thanks! If you have questions ask. No need to merge this *now* but let's not waste extra nodes with these jobs longer than needed...18:50
sheelbjornar__: some basic for backup,  for mq - backup_topic = cinder-backup, for backup host -> backup_driver = cinder.backup.drivers.swift18:50
smcginnisAJaeger: That would be nice to get rid of those.18:50
sheelbjornar__: but, yes these options are not much detailed anywhere18:50
bjornar__sheel, that is a big problem for understanding the services, I think18:50
*** arch-nemesis has joined #openstack-cinder18:50
AJaegersmcginnis: the same change I'm proposing has been done for nova and glance already18:50
*** bardia has joined #openstack-cinder18:51
sheelbjornar__: agree with you18:51
bjornar__sheel, I dont even know, without ss .. who is talking to who18:51
smcginnisAJaeger: OK, good. Always nice when someone else takes the leap before you. ;)18:51
sheelbjornar__: hmm18:51
smcginnisbjornar__: Are you looking for something like this?
bjornar__smcginnis, no, I want to know all the config options/sections that are of any interest to the different services18:52
smcginnisbjornar__: Well, the docs will tell you what relevant config settings you need.18:53
bjornar__smcginnis, so for example, [keystone_authtoken] .. is that needed by -backup or -volume ?18:53
bjornar__I do not think so, but I cant say for sure18:53
*** lpetrut has joined #openstack-cinder18:54
AJaegerthanks, smcginnis !18:55
*** bswartz has joined #openstack-cinder18:55
sheelbjornar__: may be you can follow up with architecture and guess upto some extent, like keystone authtok will be for api only which is part of dev docs18:57
bjornar__smcginnis, and, is it only cinder-api that needs database access, or also -volume? .. what about -backup ?18:57
sheelbjornar__: i am sharing architecture doc with you18:58
sheelyou will get some idea about that18:58
bjornar__this one?
sheelbjornar__: ah yes19:00
bjornar__yeah, that was great!19:00
*** haomaiwang has quit IRC19:01
bjornar__Is this really the most information that is out there on these topics?19:01
*** 92AAAEL3R has joined #openstack-cinder19:01
sheelbjornar__:  well you may go through openstack.org19:03
sheelbjornar__: a lot of documentation is present which will surely help you in getting answers to your query19:03
bjornar__I think I have seen most19:03
bjornar__So if you cant give me more help here, who can?19:04
sheelbjornar__: do refer
sheelbjornar__: I think i tried to help you most but still you can ask someone in group19:05
Swansonoooo...  275797 looks like it has finally kicked off!  There is hope!19:05
mtaninoHi :), Can I get an attention for this bug fix?
*** permalac has joined #openstack-cinder19:05
sheelbjornar__: simply you can float your query in group and see if someone have some more information ...there would be i think19:05
bjornar__sheel, lets start with: does -backup and -volume access database directly?19:06
*** delattec has joined #openstack-cinder19:06
bjornar__I mean if people in here does not know that, there are huge problems here!19:06
*** sdague has quit IRC19:06
sheelbjornar__: dont worry these are simple things19:07
sheelbjornar__: everyone knows and can share with you19:07
*** delatte has quit IRC19:08
*** delatte has joined #openstack-cinder19:09
sheelbjornar__: oh i missed your query19:10
sheelbjornar__: backup and volume should not access directly19:10
*** jwcroppe has joined #openstack-cinder19:11
bjornar__"should not"?19:11
*** delattec has quit IRC19:11
*** Lee1092 has quit IRC19:12
eharneybackup and volume services both talk to the database19:13
bjornar__sheel, I feel this is a huge issue with documentation. It is not possible (afaik) to figure this out from docs19:13
bjornar__eharney, so they connect directly to the database?19:13
eharneybjornar__: yes19:13
bjornar__eharney, an this you are sure about?19:13
bjornar__What would they do in the database?19:14
*** melwitt_ is now known as melwitt19:14
guitarza1bjornar__: they perform lots of db queries...19:14
guitarza1state updates for one easy example19:14
bjornar__guitarza1, ok.. what do they ask for?19:15
bjornar__I would believe they just did what they were told to do over mq.. but I dont know19:15
*** jgregor has joined #openstack-cinder19:15
*** alkhodos has joined #openstack-cinder19:15
alkhodosHi guys! Don't want to be annoying, but can 1 more core review this please? Hanging there for some time now :)19:16
guitarza1bjornar__: of course, but most of those actions involve them updating the state of a volume19:16
*** dimsum__ has quit IRC19:16
bjornar__guitarza1, so can you please point me to the architecture document where this is described in detail?19:16
guitarza1bjornar__: I don't understand19:17
guitarza1what architecture document?19:17
bjornar__guitarza1, but do they update the state themselves, or send rpc message to api or wherever.19:17
guitarza1why would they make an rpc call to the api to update the db? they just make the db update themselves19:17
*** chhavi has quit IRC19:17
bjornar__guitarza1, The document where I can see the workflow of for example cinder-volume19:17
bjornar__guitarza1, Because they have probably been given a job by rpc, then the answer to that job would probably be the new state.. and so on..19:18
guitarza1bjornar__: it's not that granular19:18
guitarza1the api doesn't know "Now I need to update the volume to state X"19:18
guitarza1the volume manager manages all the interactions with the backend drivers19:18
bjornar__But do you say: "there is no documentation describing the architecture _at all_"?19:19
smcginnisIf you want to see the architecture, you were given that. If you want to know the fine details of how things are done, look at the code.19:19
guitarza1I definitely didn't say that19:20
smcginnisIf you actually have a constructive question, please ask.19:20
bjornar__The constructive question was: "where can I find a complete reference of the config options in use by the different daemons"19:20
guitarza1bjornar__: maybe you could start here:
bjornar__..that  will tell me alot of what is going on.19:20
smcginnisYou've been given the config options for backup.19:21
bjornar__smcginnis, no, only those spesific to backup19:21
smcginnisYou now seem to want to complain that that doesn't branch out into every other possible option that could be touched by it.19:21
bjornar__smcginnis, I guess it needs rpc as well.. maby database.. what else?19:21
guitarza1bjornar__: the document you're looking for does not exist19:21
bjornar__guitarza1, dont you see that as a problem?19:21
bjornar__ok. good answer19:22
eharneylots of people want better documentation, few people want to write it :)19:22
guitarza1bjornar__: maybe you want this?
bjornar__I really hope you atleast agree that it should be possible to figure out what options apply where.19:22
*** ndipanov has joined #openstack-cinder19:24
openstackgerritMerged openstack/cinder: Huawei: Fix create volume with prefetch value error
*** anshul has quit IRC19:26
kvidvanssmcginnis: appreciate if you could take a look at had given a +2 but I had to rebase19:26
bjornar__guitarza1, how could I from the code extract the various config options for example -backup tries to access?19:27
*** cknight has quit IRC19:27
*** ntpttr has quit IRC19:28
guitarza1bjornar__: I'm not sure that's possible. The services load the entire config. Determining what options they use would be a pretty large task.19:29
guitarza1you can certainly trace the code and find out if you wish though19:30
*** cknight has joined #openstack-cinder19:30
kvidvansthanks smcginnis!19:31
kvidvansany core willing to review and hopefully help merge ..would appreciate the help19:34
*** porunov has joined #openstack-cinder19:38
*** martyturner has quit IRC19:41
hemnakvidvans, do you have CI for the NFS driver ?  I don't see it19:42
kvidvanshemna...yes..     you should also see it if you click toggle ci is the 7th item from the bottom..19:44
kvidvanswe have a problem with the way we report the result...we over write the existing label. but the logs are still available in the CI results table19:44
*** laughterwym has joined #openstack-cinder19:45
openstackgerritMerged openstack/cinder: EMC ScaleIO - fix bug in extend volume
smcginniskvidvans: Are you working on splitting that out into two separate accounts? Or combining the results reporting so that doesn't happen anymore?19:46
kvidvansyes...we will split it into separate accounts... the two CI system for NFS and iSCSI are not in sync so combined reporting is not possible19:47
*** xyang has quit IRC19:48
*** erhudy has joined #openstack-cinder19:48
hemnathis is confusing19:48
hemnayou should all of your CI reporting under the same acct19:49
*** laughterwym has quit IRC19:49
kvidvanshemna, ok..we will figure something out19:50
*** xyang has joined #openstack-cinder19:50
*** jwcroppe has quit IRC19:54
*** ntpttr has joined #openstack-cinder19:56
kvidvanshemna, ZFS NFS does report, its in the list of CI reports for the review19:58
hemnait's just not showing up in the list up top19:58
kvidvansright..but the CI is reporting...we will fix the reporting19:59
kvidvanswe need to combine the results19:59
*** sgotliv has joined #openstack-cinder19:59
*** 92AAAEL3R has quit IRC20:01
kvidvanshemna, thanks! I will work on the reporting20:01
*** haomaiwang has joined #openstack-cinder20:01
hemnaok np20:01
*** eharney has quit IRC20:02
openstackgerritJohn Griffith proposed openstack/cinder: Get disabled services on init
openstackgerritJohn Griffith proposed openstack/cinder: Replication v2.1 (Cheesecake)
smcginnisjgriffith: Oh noes. Rebase?20:04
jgriffithsmcginnis: yeah20:06
jgriffithno worries.. what could go wrong ;)20:06
smcginnis* queue total infra crash...20:07
jgriffithsmcginnis: uhoh... you "said it" out loud20:07
jgriffithwe're doomed I say... doomed20:07
bswartzWelcome to merge conflict hell! Your room will be right next to the river of failed rechecks...20:07
jgriffithbswartz: haha!  Nice20:08
smcginnisbswartz: Hah20:08
smcginnisbswartz: We can check out any time we want, but we can never leave?20:08
jgriffithhmm... I'm not sure we've fixed tbarron 's wonderul log.warn tests20:08
jgriffithrebase -i master.... run tests; failed20:08
smcginnisjgriffith: Same failures?20:08
jgriffithsmcginnis: yeah20:08
openstackgerritNate Potter proposed openstack/cinder: Stop deleting operation on busy volumes
smcginnisjgriffith: Pull latest from origin?20:09
smcginnisThat's the one.20:09
jgriffithsmcginnis: lemme try that... surely it's fixed20:09
smcginnisWe've had a few go through, so I would be surprised.20:09
*** mdenny has quit IRC20:10
jgriffithAwww.. and Who put an ssh call in the unit tests!!!!20:11
jgriffith"The Pretzels are making me THIRSTY!!!!"20:11
smcginnisssh call to what?20:12
jgriffithsmcginnis: hell if I know... but run unit tests and you'll notice a lovely "ssh_known_hosts" file created in your test dir20:13
guitarza1hah, nice20:13
*** sdague has joined #openstack-cinder20:13
smcginnistox -epy27 passes for me locally on master at least.20:14
jgriffithguitarza1: "nice" is NOT the word I used20:14
tbarronjgriffith: nice, is there an authorized_keys there for me as well?20:14
guitarza1jgriffith: fair enough :)20:14
jgriffithtbarron: nope20:14
jgriffithtbarron: and the file is empty BTW20:14
tbarronlemme pull fresh and see what I see20:14
smcginnisjgriffith: That must be recent.20:14
jgriffithtbarron: please don't tell me you added a "assert ssh_known_hosts file exists" in a unit test to boost coverage numbers :)20:15
tbarroni have no recollection of any such20:15
jgriffithSenator Barron20:15
tbarronbut I want to look at this thing20:15
*** eharney has joined #openstack-cinder20:17
*** martyturner has joined #openstack-cinder20:19
jgriffithsmcginnis: pheww... ok, fresh clone/pull/rebase and we're good20:19
smcginnisOh good20:19
tbarronno ssh_known_hosts added either?20:20
openstackgerritJohn Griffith proposed openstack/cinder: Add ability to failback for replication V2.1
openstackgerritJohn Griffith proposed openstack/cinder: Replication v2.1 (Cheesecake)
smcginnistbarron: No, I get that on master20:20
jgriffithsmcginnis: tbarron well sofab...20:20
*** alejandrito has joined #openstack-cinder20:20
jgriffithtake a look at test_ssh_utils.py20:20
*** mdenny has joined #openstack-cinder20:22
jgriffithactually, those are probably ok20:22
jgriffithI'll find it... and when I do... look out!!!20:23
smcginnisBlame only shows a recent online update from Eric there, so I would have been surprised.20:23
tbarronwell I've been putting stuff in there so eric can get his commit numbers up, so he owes me20:24
smcginnistbarron: Hah!20:24
smcginnistbarron: It was all part of your master plan.20:24
*** guitarza1 is now known as guitarzan20:26
tbarronjgriffith: smcginnis: removing the file and running 'tox -epy27 -- --regex cinder.tests.unit.test_ssh_utils' does not produce a new file20:27
tbarronso that's not it20:27
smcginnistbarron: Yeah, I think it's probably somewhere else that forgot to mock something.20:27
tbarronyeah, we're looking for what part of the code to mock now :-)20:27
smcginnisIt's mocking us.20:28
*** kfarr has joined #openstack-cinder20:28
smcginnisI tested against the last commit on Feb 23 and it does not show up.20:29
smcginnisSo something in the last couple of days.20:29
smcginnisLast one on the 24th and it does show up.20:32
*** eharney_ has joined #openstack-cinder20:32
smcginnisSo it was something on the 24th.20:32
*** eharney has quit IRC20:33
*** timcl has left #openstack-cinder20:33
*** eharney_ is now known as eharney20:33
*** AJaeger has left #openstack-cinder20:34
*** sgotliv has quit IRC20:37
openstackgerritNate Potter proposed openstack/cinder: Stop deleting operation on busy volumes
smcginnisIntroduced after "Fix HBAS iSCSI driver attachment" on the 24th.20:40
tbarronsmcginnis: the great bisector!20:41
smcginnistbarron: ;) It's one way to track it down.20:41
tbarronit's a great way20:41
openstackgerritAlex O'Rourke proposed openstack/cinder: WIP - 3PAR: Update replication to v2.1
openstackgerritNate Potter proposed openstack/cinder: Stop deleting operation on busy volumes
jgriffithtbarron: yeah, that's what I meant "it's ok" :(20:42
jgriffithstill on the hunt20:42
smcginnisOh, microversion code went in after there. I can blame scottda .20:42
openstackgerritNate Potter proposed openstack/cinder: Stop deleting operation on busy volumes
jgriffithsmcginnis: nice20:46
*** dims has joined #openstack-cinder20:48
smcginnisThis did it:
*** cknight has quit IRC20:49
tbarronsmcginnis: jgriffith: i see, griffith's outfit :-)20:50
tbarronsmcginnis: jgriffith: actually, I +1-ed that20:51
smcginnishemna: 192.9 Gib, 4 hours and 7 minutes.20:52
jgriffithsmcginnis: tbarron "go figure"20:52
tbarronsmcginnis: jgriffith: superficially, it *looks* like ssh stuff is getting mocked out in test_netapp.py20:53
*** cknight has joined #openstack-cinder20:53
smcginnistbarron: Hah, I just finally figured out what you meant by Griffith's outfit! :)20:53
* smcginnis is slow20:53
tbarronsmcginnis: maybe tired, but never slow20:54
smcginnistbarron: Yeah, nothing is jumping out there as the culprit.20:54
*** annasort has quit IRC20:54
tbarroni'll play with it a bit20:54
jgriffithsmcginnis: I suspect something in the external lib (zapi_client)20:54
jgriffithsmcginnis: I hate this external lib shit.. have no idea what's going on20:54
smcginnisjgriffith: True20:55
smcginnisIt's something in test_netapp.20:55
jgriffithsmcginnis: or the netapp_api.SSHUtil (cuz cinder's isn't good enough).... or paramiko.SSHclient (cuz Netapps is better, but still not the best), or .... blah blah blah20:55
smcginnisThe only change there is the zapi_client bit.20:55
smcginnisRepro'd with tox -e py27 -- -n cinder.tests.unit.test_netapp20:56
*** dave-mccowan has quit IRC20:56
akerrwe don't have an external lib anymore20:56
tbarronyeah, I did that repro too20:56
tbarronakerr: ahah!20:56
tbarronso the mock fails to prevent ssh from really running.20:56
jgriffithakerr: external to Cinder is what I mean there :)20:56
akerrjgriffith: ok, I thought we beat that horse already :)20:57
tbarronoh, you *don't' have the external lib20:57
jgriffithakerr: "It's not dead yet" :)20:57
openstackgerritMatan Sabag proposed openstack/cinder: Changes in ScaleIO configurations options.
akerri can't even think off-hand what we'd use ssh for...20:59
tbarronsmcginnis: yeah, i reproed it there too21:00
akerrall our management calls are http21:00
jgriffithakerr: it's all over in your code21:00
tbarronakerr: chuck had to do some ssh stuff to do some iscsi chap tricks21:00
akerrtbarron: oh ya...21:00
akerrtbarron: that pile...21:00
tbarronakerr: he regular zapis didn't have the secret stuff21:00
jgriffithakerr: hehe...21:00
*** haomaiwang has quit IRC21:01
akerrwe blame cfouts then :)21:01
tbarronakerr: I am biting my tongue about l***al requriements21:01
*** haomaiwang has joined #openstack-cinder21:01
tbarronpoor cfouts was forced to do that stuff21:01
tbarronbut the question is why this mock isn't working21:01
tbarronit looks like it would on the face of it21:02
*** akerr has quit IRC21:04
smcginnistbarron: Ah, it's too late.21:04
smcginnisThe constructor for the driver's client includes a call to _init_ssh_client.21:04
*** laughterwym has joined #openstack-cinder21:04
smcginnisSo by this point, the test already has the driver, then sets up the mock.21:04
tbarronsmcginnis: that makes sense21:04
smcginnisSo I think it's this:21:05
*** xyang has quit IRC21:05
*** merooney has quit IRC21:05
eharneyhere's how to find the source of it:
*** laughterwym has quit IRC21:09
*** kfarr has quit IRC21:10
smcginniseharney: Ooh, nice!21:10
eharneythis should have an actual fix w/ mock, but we could also set a conf fixture for ssh_hosts_key_file to be /dev/null to prevent this in the future21:13
eharney(maybe that would just encourage bad tests, though)21:13
smcginniseharney: True. But the fixture could be good to have.21:14
*** delattec has joined #openstack-cinder21:15
*** alejandrito has quit IRC21:16
*** jungleboyj has quit IRC21:17
*** delatte has quit IRC21:17
cfoutsokay, I had it mocked out properly but when I switched to using SSH Pool I forgot to double check that.21:17
cfoutssmcginnis, eharney: I can create the bug and get it fixed21:18
*** JoseMello has quit IRC21:18
smcginniscfouts: Great, thanks!21:20
eharneysounds great21:20
*** delattec has quit IRC21:20
openstackgerritJohn Griffith proposed openstack/cinder: Change frozen error messages from _LE to _
*** delattec has joined #openstack-cinder21:22
openstackgerritJohn Griffith proposed openstack/cinder: Add basic workflow of attach/detach to devref
*** gouthamr_ has quit IRC21:26
*** Yogi1 has joined #openstack-cinder21:26
*** alkhodos has quit IRC21:28
*** garthb_ has joined #openstack-cinder21:28
cfoutssmcginnis, eharney: LOG.warn is breaking tests in stable/liberty on our CI. I should have a patch up soonish for
openstackLaunchpad bug 1550102 in Cinder "Unit test test_get_reserved_percentage fails" [Critical,Fix released] - Assigned to Eric Harney (eharney)21:28
smcginniscfouts: Cool, thanks. I didn't think about backports.21:28
cfoutscan one of you add stable/liberty to that?21:28
openstackgerritJacob Gregor proposed openstack/cinder: Storwize SVC multiple management IPs
smcginniscfouts: I will if I can remember how. ;)21:29
smcginniseharney: Thanks21:30
eharneysmcginnis: yeah no kidding, it took me 2 minutes to find the button21:30
smcginniseharney: Ah, now I see it.21:30
*** lprice1 has joined #openstack-cinder21:30
smcginnisLink a related branch.21:30
eharneyi think that does something very different :)21:30
eharneyit's actually "target to series"21:30
*** lprice has quit IRC21:30
*** garthb has quit IRC21:30
smcginniseharney: Oh, which one is it?21:30
*** IlyaG has quit IRC21:31
eharneywe can just cherry-pick the master patch straight from gerrit right?21:31
smcginniseharney, cfouts: I would think so. It was pretty self contained.21:31
*** IlyaG has joined #openstack-cinder21:31
bswartzeharney: yes, iff it's merged in master21:32
jgriffithjust curious... does anybody know if there's some reason people use test__ instead of test_ in their test names?  IE two underscores?21:32
tbarronstuck keyboard21:33
jgriffithdell, hp and netapp seem to all do it a fair bit21:33
bswartzbecause 2 is better than 121:33
eharneytoo much coffee, i assume21:33
jgriffithok  fair enough... I deserved all of those21:33
smcginnisOr it's test_ in front of something like _do_something.21:33
bswartzwell it you're testing a private method, then it would make sense21:33
tbarroni always just copied dell21:33
eharneythere's even test___eq__, neato21:33
jgriffithsmcginnis: yeah... that's what it is... but why?21:34
smcginniseharney: Really!?21:34
openstackgerritMerged openstack/cinder: LeftHand: Updating minimum client version
eharneywell, the method is called __eq__ :)21:34
bswartzbecuause test_x() tests x() and test__x() tests _x()21:34
cfoutseharney: I cherry picked your patch and fixed some LOG.warn in test_nfs. those tests were deprecated and removed in M21:34
jgriffithbswartz: ok... I'm officially sorry I asked :)21:34
smcginnisjgriffith: I think it's probably good to keep test_(actual name).21:34
*** apoorvad has quit IRC21:35
smcginnisThen if there is a do_something() that internally calls _do_something() it's clear which is the target of the test.21:35
jgriffithsmcginnis: yeah, the reason I was asking is I was grepping on some things and got "unexpected" results21:35
jgriffithsmcginnis: sure.... I guess... I dunno21:35
eharneycfouts: what landed didn't really look much like what i submitted, so i probably don't know all the details21:35
smcginnisYeah, meh21:35
jgriffithsmcginnis: I always thought that's what private methods (_do_something) were for21:36
jgriffithbut what do I know21:36
jgriffithI'm clearly not "hip"21:36
tbarronoh quit bragging :-)21:36
tbarronit's nice out, I'm going for a walk with my dog.  Enough double underscores for now.21:37
jgriffithtbarron: nice!!!21:37
jgriffithtbarron: you've got the right idea!  Enjoy!21:37
smcginnisNIce here too. Think I should probably get out before my last obligation of the week.21:38
smcginnistbarron: Have a good one.21:38
scottdasmcginnis: I broke stuff?21:39
smcginnisscottda: Just finding excuses to blame microversions for problems. :P21:39
openstackgerritOpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements
*** adrianofr_ has quit IRC21:40
*** jgregor has quit IRC21:42
*** jamielennox is now known as jamielennox|away21:43
*** Yogi1 has quit IRC21:43
*** porunov has quit IRC21:45
scottdaWhew, ok. I'm in the mountains. No internet and poor cell phone coverage. So, I won't be fixing anything until Monday :)21:51
*** apoorvad has joined #openstack-cinder21:52
smcginnisscottda: Nice! Sounds fun.21:54
hemnascottda, +A21:54
_alastor_I think my CI is hitting a race condition, but I'm not sure how to prevent it.21:55
smcginnis_alastor_: What's happening?21:55
_alastor_If you take a look at this log:
_alastor_I'm getting a MismatchError as the exception21:56
*** scottda is now known as scottda_is_outta21:56
_alastor_ testtools.matchers._impl.MismatchError: [u'cinder-scheduler', u'cinder-volume'] != [u'cinder-scheduler']21:56
*** mdenny has quit IRC21:57
smcginnisThat's a new one to me.21:57
_alastor_It only happens sometimes (in the past 24 hours I've this pass a few times)21:57
*** dustins has quit IRC21:59
_alastor_But if the cinder-volume service wasn't running, I'd get a lot more test failures than just this one.21:59
*** salv-orlando has joined #openstack-cinder21:59
smcginnis_alastor_: I noticed I could access the c-vol log when I poked around.22:00
smcginnis_alastor_: The name was there, but when I click on it I get an error.22:00
_alastor_Yeah, the index is getting miscreated, there's a double slash in the url that should be a single slash22:00
smcginnis_alastor_: But yeah, it should have more than two failures if that's the case. So I think you're right.22:00
*** haomaiwang has quit IRC22:01
_alastor_Here's the c-vol log:
*** eharney has quit IRC22:01
*** haomaiwang has joined #openstack-cinder22:01
*** salv-orl_ has quit IRC22:02
*** crose has quit IRC22:02
smcginnis_alastor_: No, it does look like a c-vol failure:22:04
smcginnis2016-02-26 12:58:33.129 ERROR cinder.volume.manager [req-a19dadf7-4753-4846-9656-83d5d9ab59cf None] [driver-VolumeManager] Failed to initialize driver22:04
smcginnisDateraAPIException: Request to Datera cluster returned bad status: 400 | Bad Request22:04
_alastor_I think that might be the exception that happens before I re-initialize the driver with the code that's currently in review.22:05
*** martyturner has quit IRC22:05
_alastor_It also has to fail a request to the backend before it logs in with the provided creds.22:05
_alastor_So there's always at least one of those :)22:05
smcginnisOK, yeah, I see goodness after that.22:06
*** baumann has left #openstack-cinder22:06
*** merooney has joined #openstack-cinder22:06
*** tpsilva has quit IRC22:07
*** sheel has quit IRC22:07
_alastor_The exception happens at 13:03, but in the c-vol logs I don't see anything out of the ordinary22:10
*** lprice1 has quit IRC22:10
*** jungleboyj has joined #openstack-cinder22:11
smcginnis_alastor_: Yeah, I didn't either.22:11
smcginnis_alastor_: Not sure what that could be. I never ran into that before.22:11
smcginnis_alastor_: You said it's only periodic?22:11
*** knikolla has quit IRC22:11
_alastor_Yeah, the most recent run passed:
_alastor_But I've seen that error on a few of the runs in the past 2422:13
smcginnis_alastor_: Maybe ask in the channel next week and see if anyone else has seen it.22:13
smcginnis_alastor_: I have a feeling most are gone by now.22:13
*** lprice has joined #openstack-cinder22:15
_alastor_smcginnis: btw, is the Datera CI reporting enough to allow review 284967 to make it through provided it passes the gate?22:17
smcginnis_alastor_: Yeah, as long as it runs against those changes to validate it passes tempest, that should be good.22:18
openstackgerritHelen Walsh proposed openstack/cinder: EMC VMAX - Changing PercentSynced to CopyState in isSynched
smcginnis_alastor_: Looks like it hit the gate issues we had over the last day.22:19
_alastor_smcginnis: It's running them right now, I patch the CI with those changes for every patch request it gets22:19
_alastor_smcginnis: Yeah, but I think it's catching up now22:19
smcginnis_alastor_: Just triggered a recheck to see if we get a +1 from Jenkins.22:19
smcginnis_alastor_: Assuming Jenkins and your CI are happy, and the code looks good, shouldn't be a problem to get it in at this point.22:19
_alastor_smcginnis: cool22:19
*** darosale has quit IRC22:21
*** merooney has quit IRC22:23
*** porrua has quit IRC22:30
*** merooney has joined #openstack-cinder22:30
*** edtubill has quit IRC22:33
*** lprice has quit IRC22:38
jungleboyjsmcginnis: Anything critical that needs more eyes right now?22:39
smcginnisjungleboyj: Maybe a gander through the open client patches?22:40
smcginnisjungleboyj: I think there were a few waiting a second set of eyes.22:40
smcginnisjungleboyj: Deadlines early next week. Yikes!22:40
jungleboyjsmcginnis: Time flies when you are having fun?22:41
smcginnisjungleboyj: Heh, something like that. ;)22:41
jungleboyjI wish someone would get working again.22:42
smcginnisjungleboyj: I have one working22:42
smcginnisjungleboyj: All kinds of goodies:
*** IlyaG has quit IRC22:43
jungleboyjsmcginnis: Oh, you run the reports too!22:43
* jungleboyj wants to hug smcginnis 22:43
smcginnisjungleboyj: Yeah, I was playing around with some other things using some of the code from that.22:43
smcginnisjungleboyj: Figured I might as well throw it out somewhere.22:44
*** ndipanov has quit IRC22:44
* tbarron averts his eyes22:45
smcginnistbarron: Back from dog walking already?22:45
tbarronthe dogs were better behaved22:45
jungleboyjsmcginnis:I need those numbers to motivate me.22:45
smcginnisjungleboyj: I haven't run it in awhile, but this was interesting:
*** merooney has quit IRC22:46
*** marcusvrn_ has quit IRC22:47
jungleboyjsmcginnis: He he he.  Still staying ahead of you.22:47
jungleboyjOne day I will pass up jgriffith22:47
*** zhonghua has joined #openstack-cinder22:48
smcginnisjungleboyj: Could be getting close. I should run that one again. I'm sure I've beat you by now. :P22:50
jgriffithjungleboyj: huh?22:50
jgriffithOh.. that!22:50
openstackgerritMerged openstack/cinder: Fix NoneType error in service_get_all
smcginnisjgriffith: He's comparing size.22:51
jgriffithjungleboyj: I've been on vacation for 6 months... now's your chance!!!22:51
* jgriffith provides a subdued but deep chuckle22:51
* jungleboyj is working on it.22:51
jgriffithas coffee spurts out his nose22:51
jungleboyjjgriffith: Ouch!22:51
*** zhonghua-lee has quit IRC22:51
smcginnisMy work here is done.22:51
jgriffithsmcginnis: well played!22:52
*** alyson_ has quit IRC22:52
jungleboyjsmcginnis: According to stackalytics for all time reviews I am behind John by about 500 reviews and you are behind me by about 600.22:54
jungleboyjI am the meanest of the three.22:54
smcginnisWell, we have at least that out there. I think you're goal for the weekend is to review all outstanding patches. :)22:54
jungleboyjsmcginnis: Ha, I don't think my boys would appreciate that.22:55
jungleboyjIs this one we want in?
jungleboyjThe BP link doesn't work.22:56
smcginnisjungleboyj: Server side of that is still pending.22:56
*** merooney has joined #openstack-cinder22:57
smcginnisjungleboyj: But I think we can approve that and it will just hang until the support goes in to cinder.22:57
smcginnisjungleboyj: It was previously approved I see.22:58
jungleboyjsmcginnis: Yeah, I just saw that.22:58
jungleboyjsmcginnis: Weird, not sure why the link to the BP isn't working.  There is one out there.22:59
smcginnisjungleboyj: Format right?22:59
jungleboyjLooks like it.23:00
jungleboyjsmcginnis: Not worth holding it up for.23:00
smcginnisI've seen that some times where it just doesn't link right. I'm not sure why.23:00
jungleboyjWant me to +A?23:00
*** haomaiwang has quit IRC23:01
jungleboyjOr wait for the server side to go in?23:01
*** haomaiwa_ has joined #openstack-cinder23:01
openstackgerritGerald McBrearty proposed openstack/cinder: Revert "Remove Cisco FC Zone Manager Driver"
*** IlyaG has joined #openstack-cinder23:04
jungleboyjWhat is Gerald doing?23:04
*** NightKhaos has quit IRC23:06
*** dims has quit IRC23:07
*** wanghao has quit IRC23:08
*** wanghao has joined #openstack-cinder23:09
*** xyang1 has quit IRC23:11
*** chris_morrell has quit IRC23:13
*** merooney has quit IRC23:13
jungleboyjOops, way to push to the wrong tree.23:13
*** haplo37 has quit IRC23:24
aorourkejgriffith, is there a strict deadline for the python-cinderclient?23:26
*** IlyaG has quit IRC23:27
mtaninojungleboyj: ping,23:30
*** lpetrut has quit IRC23:31
mtaninojungleboyj: If you have a chance now, could you revisit this quota fix?
smcginnisaorourke: client deadline is next Tuesday, IIRC.23:31
aorourkesmcginnis, cool. Just making sure it was not today or anything :)23:32
*** dims has joined #openstack-cinder23:34
jungleboyjmtanino: Sure, I need to drop off for now but can take a look later.23:34
mtaninojungleboyj: thank you, Review on Monday is fine for me :)23:34
*** yangyapeng has joined #openstack-cinder23:35
*** erlon has quit IRC23:36
*** jungleboyj has quit IRC23:38
*** edmondsw has quit IRC23:39
*** haplo37 has joined #openstack-cinder23:40
*** erhudy has quit IRC23:40
*** ntpttr has quit IRC23:53
*** aorourke is now known as aorourke-afk23:55

Generated by 2.14.0 by Marius Gedminas - find it at!