Thursday, 2015-02-26

thingeehemna: I haven't looked lately. what's up with brick, os-brick, cinder-lib whatever00:00
anishmtanino: sure. just a headsup, I'm afk from tomorrow-monday00:00
hemnathanks for asking00:00
mtaninoanish: have a nice vacation?00:00
hemnaso, there are 2 outstanding patches,n,z00:00
hemnaI'd like other eyes looking at those00:01
hemnaI'm in the process of testing it against my WIP for Cinder here:
*** annashen has quit IRC00:01
hemnaso basically what needs to happen is this00:02
hemnaos-brick patches land.00:02
hemnaos-brick release to pypi00:02
hemnacinder Remove Brick from cinder patch lands.00:02
hemnaand well prior to step 3, global requirements updates and cinder requirements updates as well.00:02
thingeeOk, I mentioned earlier I'm going to really nervous about this if we can get things landed by this week.00:02
thingeeI really want churn in the gate on this00:03
hemnayah understandable00:03
hemnaeither way, we are in decent shape00:03
hemnaos-brick exists as a project00:03
hemnaif we get cinder using for K great, if not, L00:04
*** ebalduf has joined #openstack-cinder00:04
hemnaI just need others to help reviews and test.00:04
thingeespeaking of still that needs to land soon...00:04
thingeexyang1: hey00:04
hemnaI don't want to be the only one.00:04
xyang1thingee: Hi00:04
thingeexyang1: have you had a chance to test out and ?00:04
thingeexyang1: DuncanT signed off on it.00:05
thingeehemna: happy to make those first priorities in L, so you can have a better chance in Nova.00:05
thingeethose being multi-attach and brick if ends up that way00:06
hemnaok coolio.  thanks00:06
hemnaI'll keep working towards K for now.00:06
thingeeor be loud to whoever the PTL is :)00:06
xyang1thingee: I did test the patchset two days ago.  Worked for me.00:07
hemnathingee, so another question related to brick00:07
thingeexyang1: that's so great to hear00:07
thingeexyang1: can you tell me what exactly you did to test it?00:07
*** timcl has joined #openstack-cinder00:08
xyang1thingee: I tested migrate down and also add a column to snapshot table.  Make sure i can still create snapshot etc and can see the new field00:08
*** ebalduf has quit IRC00:08
xyang1thingee: Anything specific you are looking for00:09
thingeexyang1: excellent! I don't even think DuncanT covered snapshots, so that's great00:09
thingeexyang1: I think he just covered volumes00:09
thingeexyang1: no that sounds great00:09
thingeehemna: go ahead :)00:09
xyang1thingee: Great. The patch is about snapshot though00:09
hemnaso regarding patches that go against os-brick and cinder, we need to ensure that cinder tests are run for os-brick patches and vice versa00:10
hemnaso a change to os-brick doesn't break cinder in check/gate.00:10
thingeexyang1: snapshot and volumes00:10
thingeeit says in the title anyways00:10
xyang1thingee: I think he added volume later.  Ok00:12
*** nikesh_vedams has quit IRC00:12
*** timcl has quit IRC00:12
thingeexyang1: looks like backup object is added later00:12
anishmtanino: yep00:13
*** nikesh_vedams has joined #openstack-cinder00:13
*** cpallares has quit IRC00:13
*** david-lyle is now known as david-lyle_afk00:14
xyang1thingee: Let me go through the code again.  Somehow I thought volume is not completely covered.  Backup was added by another person00:15
*** Manshoon has joined #openstack-cinder00:17
*** Manshoon has quit IRC00:18
*** gary-smith_ has quit IRC00:18
*** gary-smith_ has joined #openstack-cinder00:18
thingeexyang1: sounds good00:20
thingeehemna: makes sense00:21
openstackgerritMitsuhiro Tanino proposed openstack/cinder: Implement IET target driver
thingeehemna: so a job that installs cinder, deps with the proposed os-brick patch, runs
thingeeincluding tempest00:23
thingeeon the cinder side, we should already be fine, since we're installing it from pypi each time.00:24
thingeeof the version in requirements00:24
hemnaright now it's just tox tests pep8 and py27) that are running.00:24
thingeeyeah it makes sense to me.00:25
hemnathingee, so what's the process of doing a lib release to pypi?00:26
hemnawho kicks that off ?00:26
hemnadoes thierry have to sign off on doing a release to pypi ?00:26
jungleboyjthingee: Sorry, had to step away.00:27
jungleboyjthingee: As far as what I think looks best, that was why I was leaning towards the file.00:27
thingeehemna: well we now have the PTL kick off things for cinderclient. should probably do the same here too.00:28
jungleboyjWe have 70+ files that have config entries.  Then we can make a hacking check that cleanly indicates how to do things properly.00:28
thingeeif it's under the cinder project00:28
hemnathingee, ok cool00:29
thingeejungleboyj: ok, I would still appreciate the convo in the friday meeting if you can bring that up.00:29
thingeeif there really is no consensus or no one really cares, we can just do our own thing00:29
*** mdbooth has quit IRC00:29
hemnathingee, ok so I'll do my best to wrap up my os-brick testing tonight or tomorrow00:30
jungleboyjthingee: Ok, which Friday meeting.  I thought oslo meeting was on Mondays.00:30
thingeejungleboyj: shows what I know. yeah the monday one :)00:30
jungleboyjOk.  I can do that.00:31
hemna is small and needs some love.00:31
thingeehemna: +200:32
hemnathank you00:32
jungleboyjthingee: I may try to get a patch up to use as an example for the discussion.  The bulk of the work will be the same.00:32
* jungleboyj is getting my loving rea... Oh00:32
jungleboyjI was too slow.00:32
thingeejungleboyj, hemna, eharney, DuncanT, winston-d_zZZ: need some help in reviews for k3 targets still
thingeeactually jungleboyj has quite a few merges already :)00:35
thingeeaccording to the list00:35
jungleboyjthingee: :-)  I was busy this weekend.00:35
*** mdbooth has joined #openstack-cinder00:35
thingeebut xyang1 and I have doing our best with the initial list.00:35
thingeehave been doing*00:36
thingeexyang1: thank you00:36
xyang1thingee: You are welcome:)00:36
*** tbarron has quit IRC00:36
thingeeeven if a third of them are emc related heh00:36
thingeein the merge list00:36
* jungleboyj pats xyang1 on the back00:36
*** tbarron has joined #openstack-cinder00:37
xyang1thingee: Cg patch was updated.  The 2nd patch was added to the list00:37
jungleboyjthingee: Have a quiet day tomorrow should be able to take another pass at reviews.00:37
thingeexyang1: yeah apologies. These last three weeks I've been doing my best with reviews, and I hope you've all been noticing. This week and the last two weeks have been really difficult for me, personal reasons.00:38
hemnaxyang1, I hadn't had time to look at it again00:38
hemnabeen trying to get my stuff in :(00:38
jungleboyjthingee: Dude, you are apologizing about 128 reviews in the last 30 days?00:39
xyang1thingee: hemna no problem.00:39
* jungleboyj has been trying to catch up. Evening doing reviews while losing blood. :-)00:39
xyang1hemna: There is another cg patch added to the list: create cg from cgsnapshot00:40
thingeexyang1: I'm hoping to potential race conditions are taken care of with create cg from cgsnapshot.00:40
*** jaypipes has quit IRC00:40
xyang1thingee: I think so.00:40
xyang1thingee: I am only creating db entry first00:41
openstackgerritPatrick East proposed openstack/cinder: PureISCSIDriver consistency group updates.
vilobhmm1thingee : ping00:41
xyang1thingee: Pure's patch is not on the list00:42
thingeevilobhmm1: hi00:42
vilobhmm1thingee : i droped and email regarding the exception request00:42
vilobhmm1did you get a chance to have a look at it00:42
thingeevilobhmm1: no because I haven't responded to it. ;)00:43
vilobhmm1can you please have a look :)00:43
vilobhmm1whenever you get time00:43
thingeepatrickeast, xyang1: done00:44
thingeexyang1: if I miss anything, feel free to add it.00:44
patrickeastthingee: thanks00:44
thingeeif the bp is already approved and targeted00:44
xyang1thingee: Ok, thanks00:44
vilobhmm1myself, dulek and e0ne met last week and thought of coming up a combined approach for using micro-states for validation and taskflow persistence; I am working on the writeup/blueprint for the same can that be done as part of the vancover design summit ? so that we have a conclusion on it so that everyone is there at the same time00:44
*** Mandell has quit IRC00:44
vilobhmm1thingee : ^^00:44
thingeevilobhmm1: is there a strong reason this needs to be in K?00:44
thingeevilobhmm1: take a look at
thingee49 targeted00:45
thingee18 implemented00:45
vilobhmm1thingee : it will help for improving state management00:45
vilobhmm1this is also implemented00:46
vilobhmm1and ready for review00:46
vilobhmm1thanks for the link00:47
thingeevilobhmm1: this was sent to the dev ML?00:48
thingeevilobhmm1: implemented in this context means merged00:49
vilobhmm1thingee : yes on 02/2000:49
*** rmesta has quit IRC00:50
vilobhmm1thingee : ok, the prototype for micro-states is ready if i can get 2 reviewers it can be merged in k3 ;00:51
vilobhmm1you did mention in cinder weekly about addtional test pipeline that the fix needs to go through; if we have steps documented for running addtional functional tests somewhere i can even run them and get the fix up and ready00:51
vilobhmm1thingee : the email is titled "Exception request : Making "Introducing micro_states for create workflow" part of K-3"00:52
thingeevilobhmm1: is this still use taskflow?00:52
thingeevilobhmm1: I think last we spoke, in the meeting, and midcycle meetup, people are unsure with the usage.00:52
thingeeI would like to start with that.00:52
thingeebefore we dive futher with this00:52
thingeeand not ignore what the community has already brought up00:53
thingeeI think dulek wrote a "not pretty solution" to an outstanding issue with allocations going wrong. jgriffith was unable to resolve this issue easily because of confusion with taskflow.00:53
*** Lee1092 has joined #openstack-cinder00:54
vilobhmm1thingee : what do you mean by "I would like to start with that" yes the micro-states approach depends on taskflow and so does the persistence thing00:54
thingeevilobhmm1: I mean what I said. We should discuss if we even want to continue using taskflow.00:54
thingeeI'm getting so annoyed with people ignoring what the community of Cinder has brought up and forcing these patches down and not wanting to discuss those issues.00:55
vilobhmm1i am ready for discussion00:55
vilobhmm1and whatever the community agress on we can follow with that00:56
vilobhmm1so as mentioned above i am working on a writeup/blueprint to combine these 2 approaches do you think a design summit talk will be good for it ; to come to some conclusion ?00:56
vilobhmm1we can talk about taskflow incluson or not in next cinder weekely if you think its ok00:56
thingeeyes. Here's an interesting thing that jgriffith brought up at the midcycle meetup. It seems like ever since we introduced taskflow, everytime we get together, there are people complaining and asking if we should consider using taskflow.00:57
thingeeI had to fight for taskflow in the Kilo summit, but the annoyance that people have been raising to me, and the lack of additional flows being implemented in taskflow *much* earlier in the kilo dev cycle has been concerning00:58
thingeeI understand stuff has been proposed now, but with the earlier allocation bug, the limitations of taskflow doing persistence with sqlite is not that great.00:59
thingeeThat buys us a host coming backup and resuming, which is not what we heard initially with taskflow00:59
vilobhmm1hmm…taskflow has its own benefits though..we can add more flows…once this micr-states infrastructure is in…or i can work on adding more flows00:59
thingeevilobhmm1: additional flows have been proposed. They're up for review now. Lazy to dig them up right now.01:00
hemnathingee, I think fwiw, I -1'd the persistence patch due to it's limitation on using local sqllite01:00
hemnathingee, and i believe they fixed that.01:00
vilobhmm1so when can we expect a decision on taskflow ?01:00
hemnaI haven't had time to reup on that review though.01:00
thingeehemna, vilobhmm1: it was some issue with sqlalchemy being mentioned?01:01
hemnathingee, I believe it was related to oslo.db01:01
*** mtanino has quit IRC01:01
vilobhmm1i think that was resolved in duleks blueprint01:01
hemnaand not tying taskflow to oslo.db01:01
thingeeI need to look at dulek's solution for the allocation bug still01:01
vilobhmm1will you be interested in this writeup/blueprint on state management with/without taskflow…or otherwise i can spend my cycles on helping for k301:02
vilobhmm1thingee : ^^01:02
thingeehere's a dumb question, how exactly does data get passed from one task's execute to another?01:03
*** ebalduf has joined #openstack-cinder01:04
hemnaI believe the return of execute is the input to the next task01:05
vilobhmm1every task returns something and that gets passed as input to next task01:05
hemnahence the requires in __init__ and provides01:05
vilobhmm1thingee : ^^01:05
vilobhmm1a (returns)===>B(input)-B(return)===>C(input)01:06
hemnathat defines what that task provides as output01:06
hemnathat specifies what that task needs01:07
thingeeso example, this returns nothing
thingeebut it expects volume_ref and volume_spec
hemnayah that task doesn't provide anything01:08
thingeein the next tasks01:08
hemnathat's where the db comes from01:08
*** ebalduf has quit IRC01:09
hemnadb, event_suffix01:09
vilobhmm1 is the way the tasks will be executed /scheduled01:09
hemnathe flow engine ensures that what the task requires gets injected into the tasks's execute()01:09
hemnadepending on what the flow lists as requires in __init__01:09
thingeehemna: but again it expects a volume_ref and spec
thingeewhat you linked says db, driver01:10
thingeedriver != spec01:10
vilobhmm1thingee, hemna01:10
hemnaexecute != __init__01:10
vilobhmm1governs the policy used for the flow01:10
vilobhmm1thingee : hemna is write execute != __init__01:10
* thingee can work on a filesystem, but not taskflow01:11
thingeeok let me try again01:12
hemnacrap, I have to run and get my kids01:12
hemnathingee, sorry01:12
hemnaharlowja_away, ping01:12
*** hemna is now known as hemnafk01:12
*** Mandell has joined #openstack-cinder01:13
vilobhmm1harlowja_away : is out today01:13
vilobhmm1hemna : ^^01:13
vilobhmm1thingee : sure01:14
thingeehemnafk, vilobhmm1: I realize __init__ != execute ... hemnafk brought it up :)01:14
*** Anticimex has quit IRC01:14
*** Anticimex has joined #openstack-cinder01:14
thingeemy earlier confusion again01:14
thingeereturns nothing01:14
thingeeso how does
thingeeget volume_ref and volume_specs01:15
vilobhmm1because these values are in mem01:16
vilobhmm1as of now01:16
thingeethat does not help me understand how positional arguments get passed in01:16
vilobhmm1duleks proposal was also to move these values into persistence thing01:17
*** cpallares has joined #openstack-cinder01:17
thingeevilobhmm1: please lets talk about persistence right now. I'm trying to understand just this part :)01:17
thingeelets not talk about*01:17
vilobhmm1the return values are in mem and passed to the next task depeding on the herirarchy proposed01:18
vilobhmm1so in this case01:18
*** patrickeast has quit IRC01:22
*** rwsu is now known as rwsu-afk01:27
*** Mandell has quit IRC01:31
*** _cjones_ has quit IRC01:31
*** Mandell has joined #openstack-cinder01:34
*** sigmavirus24 is now known as sigmavirus24_awa01:44
*** Thelo has joined #openstack-cinder01:46
*** IanGovett1 has quit IRC01:51
*** davechen has joined #openstack-cinder01:51
*** jdurgin has quit IRC01:53
openstackgerritwanghao proposed openstack/cinder: Implement function of import/export snapshots
*** nikesh_vedams has quit IRC01:57
*** nikesh_vedams has joined #openstack-cinder01:57
*** dannywilson has joined #openstack-cinder02:03
*** ebalduf has joined #openstack-cinder02:05
*** leeantho has quit IRC02:05
*** vilobhmm1 has quit IRC02:07
*** jdurgin has joined #openstack-cinder02:07
*** vilobhmm has joined #openstack-cinder02:08
*** ebalduf has quit IRC02:09
*** Mandell has quit IRC02:10
*** dannywilson has quit IRC02:13
*** dannywilson has joined #openstack-cinder02:13
*** annashen has joined #openstack-cinder02:13
*** bill_az has quit IRC02:14
*** diegows has quit IRC02:14
*** cpallares has quit IRC02:16
*** dannywilson has quit IRC02:18
*** Thelo has quit IRC02:18
*** kvidvans has quit IRC02:24
*** vilobhmm has quit IRC02:24
*** casusbel_ has joined #openstack-cinder02:35
*** dalgaaf has quit IRC02:37
*** Apoorva has quit IRC02:38
*** casusbelli has quit IRC02:38
*** chlong_ has joined #openstack-cinder02:43
*** kaisers has joined #openstack-cinder02:43
*** kaisers1 has joined #openstack-cinder02:43
*** kaisers2 has quit IRC02:44
*** annashen has quit IRC02:45
*** kaisers3 has quit IRC02:46
*** markvoelker has quit IRC02:47
*** markvoelker has joined #openstack-cinder02:47
*** markvoelker has quit IRC02:51
*** takedakn has joined #openstack-cinder02:52
openstackgerritwuyuting proposed openstack/python-cinderclient: Make cinderclient support snapshot-force-delete
*** annashen has joined #openstack-cinder03:01
*** Thelo has joined #openstack-cinder03:02
*** takedakn has quit IRC03:02
*** takedakn has joined #openstack-cinder03:03
*** ebalduf has joined #openstack-cinder03:06
*** ebalduf has quit IRC03:11
*** annashen has quit IRC03:17
openstackgerritwanghao proposed openstack/cinder: query volume detail support volume_glance_metadata
*** Anticimex has quit IRC03:24
*** Anticimex has joined #openstack-cinder03:25
*** tbarron has quit IRC03:29
*** coolsvap_ is now known as coolsvap03:33
*** bkopilov has quit IRC03:36
*** annashen has joined #openstack-cinder03:37
*** Mandell has joined #openstack-cinder03:40
*** markvoelker has joined #openstack-cinder03:43
*** annashen_ has joined #openstack-cinder03:44
*** annashen has quit IRC03:48
*** takedakn has quit IRC03:49
*** annashen has joined #openstack-cinder03:49
*** cpallares has joined #openstack-cinder03:50
*** annashen has quit IRC04:00
*** scottda has joined #openstack-cinder04:00
*** scottda has quit IRC04:01
*** pdeore_ has joined #openstack-cinder04:02
openstackgerritxing-yang proposed openstack/cinder: Change max_over_subscription_ratio default value
*** xyang has quit IRC04:06
*** ebalduf has joined #openstack-cinder04:07
*** annashen_ has quit IRC04:07
openstackgerritwanghao proposed openstack/cinder: Implement function of import/export snapshots
*** ebalduf has quit IRC04:11
*** scottda has joined #openstack-cinder04:14
*** scottda has quit IRC04:15
*** kaufer has joined #openstack-cinder04:16
*** kaufer has quit IRC04:19
*** Mandell has quit IRC04:31
*** patrickeast has joined #openstack-cinder04:33
*** Mandell has joined #openstack-cinder04:39
*** annashen has joined #openstack-cinder04:41
*** lpetrut has joined #openstack-cinder04:43
*** markvoelker has quit IRC04:44
*** markvoelker has joined #openstack-cinder04:45
*** deepakcs has joined #openstack-cinder04:47
*** ho has joined #openstack-cinder04:48
*** markvoelker has quit IRC04:49
*** bkopilov has joined #openstack-cinder04:56
*** coolsvap is now known as coolsvap_04:57
openstackgerritPatrick East proposed openstack/cinder: Add CHAP support to PureISCSIDriver
openstackgerritPatrick East proposed openstack/cinder: Add DB table for driver specific data
*** patrickeast has quit IRC04:58
*** xyang1 has quit IRC05:03
*** coolsvap_ is now known as coolsvap05:04
*** ebalduf has joined #openstack-cinder05:08
*** asselin_ has quit IRC05:11
*** anuragpalsule has joined #openstack-cinder05:11
*** ebalduf has quit IRC05:12
*** BharatK has joined #openstack-cinder05:20
*** cpallares has quit IRC05:24
*** anshul has joined #openstack-cinder05:29
*** jcru has joined #openstack-cinder05:34
*** lpetrut has quit IRC05:41
openstackgerritxing-yang proposed openstack/cinder: Add support to incremental backups in cinder
*** number80 has quit IRC05:55
*** avishay has joined #openstack-cinder05:56
openstackgerritOpenStack Proposal Bot proposed openstack/cinder: Imported Translations from Transifex
*** ebalduf has joined #openstack-cinder06:08
*** ebalduf has quit IRC06:12
*** annashen has quit IRC06:15
openstackgerritLiu Xinguo proposed openstack/cinder: Huawei driver check before associating LUN to a LUN group
*** fifieldt has joined #openstack-cinder06:26
*** takedakn has joined #openstack-cinder06:27
*** BharatK has quit IRC06:31
*** Ilja has joined #openstack-cinder06:32
*** thingee has quit IRC06:35
*** pck has quit IRC06:37
*** annashen has joined #openstack-cinder06:42
*** pck has joined #openstack-cinder06:42
openstackgerritLiu Xinguo proposed openstack/cinder: Huawei driver fix problems under multipath
*** BharatK has joined #openstack-cinder06:44
*** casusbel_ has quit IRC06:47
*** casusbelli has joined #openstack-cinder06:48
*** cnesa has joined #openstack-cinder06:49
*** avishay has quit IRC06:50
*** annashen has quit IRC06:50
openstackgerritLiu Xinguo proposed openstack/cinder: Huawei driver remove LUN controller change
*** nikesh_vedams has quit IRC06:57
*** cnesa has quit IRC07:01
*** nkrinner has joined #openstack-cinder07:01
*** avishay has joined #openstack-cinder07:03
*** takedakn has quit IRC07:04
*** cnesa has joined #openstack-cinder07:07
*** ebalduf has joined #openstack-cinder07:09
*** ebalduf has quit IRC07:13
*** TobiasE has joined #openstack-cinder07:17
openstackgerritAbhijeet Malawade proposed openstack/cinder: Get volume from db again before updating it's status
*** jdurgin1 has joined #openstack-cinder07:29
*** cnesa has quit IRC07:30
*** nikesh_vedams has joined #openstack-cinder07:33
*** sgotliv has quit IRC07:37
*** chlong_ has quit IRC07:42
*** Miouge has joined #openstack-cinder07:43
*** chlong has quit IRC07:43
*** markvoelker has joined #openstack-cinder07:44
*** cnesa has joined #openstack-cinder07:44
openstackgerritEric Brown proposed openstack/cinder: Password config options should be marked secret
*** markvoelker has quit IRC07:50
*** afazekas_ has joined #openstack-cinder07:50
*** pradipm has joined #openstack-cinder07:56
*** lpetrut has joined #openstack-cinder08:00
*** e0ne has joined #openstack-cinder08:03
*** 21WABGMI1 has joined #openstack-cinder08:05
*** 21WABGMI1 has quit IRC08:05
*** ebalduf has joined #openstack-cinder08:10
*** ebalduf has quit IRC08:14
*** ronis has joined #openstack-cinder08:16
*** dulek has joined #openstack-cinder08:22
*** nshaikh has joined #openstack-cinder08:23
*** sgotliv has joined #openstack-cinder08:29
*** pradipm has quit IRC08:30
*** pradipm has joined #openstack-cinder08:43
openstackgerritrakesh mishra proposed openstack/cinder: Implement making volume snapshots public
*** karimb has joined #openstack-cinder08:49
*** rushiagr_away is now known as rushiagr08:49
*** TobiasE1 has joined #openstack-cinder08:53
*** jistr has joined #openstack-cinder08:53
openstackgerritwanghao proposed openstack/cinder: Add config option to override url for versions
*** jordanP has joined #openstack-cinder08:54
*** TobiasE has quit IRC08:56
*** sgotliv_ has joined #openstack-cinder08:57
*** dulek has quit IRC08:57
*** dulek has joined #openstack-cinder08:58
*** ndipanov has joined #openstack-cinder09:00
*** sgotliv has quit IRC09:00
*** dulek_ has joined #openstack-cinder09:04
*** dulek has quit IRC09:04
*** sgotliv has joined #openstack-cinder09:05
*** takedakn has joined #openstack-cinder09:05
*** sgotliv_ has quit IRC09:07
*** PaulCuzner has quit IRC09:08
*** TobiasE1 has quit IRC09:09
*** MasterPiece has quit IRC09:10
*** lpetrut has quit IRC09:10
*** ebalduf has joined #openstack-cinder09:11
*** yuriy_n17 has joined #openstack-cinder09:11
*** ebalduf has quit IRC09:15
*** pschaef has joined #openstack-cinder09:23
*** avishay_ has joined #openstack-cinder09:24
*** avishay has quit IRC09:24
*** abhishek_cb has joined #openstack-cinder09:25
*** bswartz has quit IRC09:30
*** dulek_ has quit IRC09:31
*** pradipm has left #openstack-cinder09:36
*** pradipm has joined #openstack-cinder09:44
*** markvoelker has joined #openstack-cinder09:47
*** davechen has quit IRC09:50
*** markvoelker has quit IRC09:52
*** Mandell has quit IRC09:53
*** EmilienM|afk is now known as EmilienM10:11
*** ebalduf has joined #openstack-cinder10:11
*** lpetrut has joined #openstack-cinder10:14
*** ebalduf has quit IRC10:16
openstackgerritwanghao proposed openstack/cinder: Add config option to override url for versions
*** takedakn has quit IRC10:20
*** pradipm has quit IRC10:27
*** pradipm has joined #openstack-cinder10:29
*** TobiasE has joined #openstack-cinder10:29
*** abhishek_cb has quit IRC10:33
openstackgerritShay Halsband proposed openstack/cinder: XtreamIO version 4.0 support
*** TobiasE1 has joined #openstack-cinder10:38
*** nshaikh has quit IRC10:38
*** TobiasE has quit IRC10:38
*** pradipm has quit IRC10:41
openstackgerritShay Halsband proposed openstack/cinder: XtreamIO version 4.0 support
*** eharney has quit IRC10:49
*** tbarron has joined #openstack-cinder10:53
*** eharney has joined #openstack-cinder11:02
*** asselin_ has joined #openstack-cinder11:06
*** nellysmitt has joined #openstack-cinder11:10
*** asselin_ has quit IRC11:11
*** ebalduf has joined #openstack-cinder11:12
*** ebalduf has quit IRC11:17
*** pcaruana has joined #openstack-cinder11:19
*** pradipm has joined #openstack-cinder11:27
*** rushiagr is now known as rushiagr_away11:33
*** abhishek_cb has joined #openstack-cinder11:36
*** sgotliv has quit IRC11:38
*** sgotliv has joined #openstack-cinder11:40
*** kallebe has joined #openstack-cinder11:42
*** chlong has joined #openstack-cinder11:51
*** chlong_ has joined #openstack-cinder11:51
*** bill_az has joined #openstack-cinder11:54
*** abhishek_cb has quit IRC11:58
*** deepakcs has quit IRC11:59
*** timcl has joined #openstack-cinder11:59
*** erlon has joined #openstack-cinder12:10
*** ganso_ has joined #openstack-cinder12:11
*** ebalduf has joined #openstack-cinder12:13
*** ebalduf has quit IRC12:18
*** pradipm has quit IRC12:24
*** pradipm has joined #openstack-cinder12:24
*** pradipm has quit IRC12:26
*** pradipm has joined #openstack-cinder12:27
*** pradipm has quit IRC12:28
*** pradipm has joined #openstack-cinder12:29
*** hodos has joined #openstack-cinder12:32
*** mtanino has joined #openstack-cinder12:33
*** akrog has joined #openstack-cinder12:33
*** annegentle has quit IRC12:33
mtaninoavishay_: ping12:34
*** avishay_ has quit IRC12:37
*** akrog is now known as geguileo12:37
*** diegows has joined #openstack-cinder12:41
*** takedakn has joined #openstack-cinder12:42
*** pradipm has quit IRC12:43
*** pradipm has joined #openstack-cinder12:43
*** dulek has joined #openstack-cinder12:44
*** IanGovett has joined #openstack-cinder12:44
*** pradipm has quit IRC12:45
*** pradipm has joined #openstack-cinder12:45
*** avishay_ has joined #openstack-cinder12:45
*** dulek_ has joined #openstack-cinder12:46
*** ho has quit IRC12:48
*** markvoelker has joined #openstack-cinder12:50
*** Ilja has quit IRC12:51
*** markvoelker has quit IRC12:56
*** sgotliv has quit IRC12:56
openstackgerritPetrut Lucian proposed openstack/cinder: SMBFS: Fix retrieving total allocated size
*** pradipm has quit IRC12:58
*** jkraj has joined #openstack-cinder13:05
*** afazekas_ has quit IRC13:05
*** avishay_ has quit IRC13:06
*** avishay__ has joined #openstack-cinder13:06
*** esker has joined #openstack-cinder13:07
*** BharatK has quit IRC13:09
*** sgotliv has joined #openstack-cinder13:09
*** lihit has joined #openstack-cinder13:13
*** aix has joined #openstack-cinder13:13
*** ebalduf has joined #openstack-cinder13:14
*** lihit has quit IRC13:17
*** bkopilov has quit IRC13:17
*** primechuck has quit IRC13:17
*** ebalduf has quit IRC13:18
*** akerr has joined #openstack-cinder13:20
*** chlong has quit IRC13:20
*** chlong_ has quit IRC13:20
*** vnagpal has joined #openstack-cinder13:22
*** rushiagr_away is now known as rushiagr13:22
*** vnagpal has quit IRC13:23
*** akerr_ has joined #openstack-cinder13:23
*** akerr has quit IRC13:26
openstackgerritAnton Arefiev proposed openstack/cinder: Add entry create and cast tasks to manage workflow
openstackgerritAnton Arefiev proposed openstack/cinder: Fix lvm manage existing volume
yuriy_n17DuncanT: Hi13:28
DuncanTyuriy_n17: Hi13:28
yuriy_n17Duncan: I'd like to discuss your comments about
yuriy_n17Duncan: What do you mean by saying 'It breaks client' yesterday?13:30
yuriy_n17Duncan: Just remember you it's about time zone in Cinder13:32
yuriy_n17DuncanT: What do you suggest?13:36
*** Yogi1 has joined #openstack-cinder13:36
*** takedakn has quit IRC13:37
DuncanTyuriy_n17: Clients that parse the time field currently break when extra info (like the timezone) is added13:39
DuncanTyuriy_n17: Depending on how their parser is written13:39
yuriy_n17DuncanT: Please give me an example to see this situation.13:42
*** baffle has quit IRC13:42
*** casusbelli has quit IRC13:43
*** mtanino has quit IRC13:44
DuncanTyuriy_n17: Do you have a system set up with your patch? I've not currently got one, it will take me a while to spin one up13:46
yuriy_n17DuncanT: ok13:47
DuncanTDo "cinder --debug show <some volume>" and look for the RESP BODY that comes back, and post up the exact value in the created_at field13:49
*** bswartz has joined #openstack-cinder13:50
geguileoHi, I'm new to Cinder and I have a quick question, can anybody help me?13:52
*** markvoelker has joined #openstack-cinder13:53
e0negeguileo: hi. what question do you have?13:54
yuriy_n17DuncanT: I've got a "created_at" with timezone13:54
geguileoe0ne: It's regarding iscsi_ip_address with tgt and lio helpers13:54
*** BharatK has joined #openstack-cinder13:54
geguileoe0ne: Shouldn't cinder only create a portal for that address instead of as it is doing now?13:54
yuriy_n17DuncanT: What's a problem?13:55
*** kaufer has joined #openstack-cinder13:55
DuncanTyuriy_n17:  So before your patch you get something that looks like "2014-03-05T15:37:06.000000"  - any parser written for that format in a client is broken by your patch13:55
DuncanTyuriy_n17: You've changed what the field contains13:55
DuncanTyuriy_n17: So you've broken the API contract cinder provides13:56
DuncanTyuriy_n17: Does that make sense?13:58
*** markvoelker has quit IRC13:59
*** sigmavirus24_awa is now known as sigmavirus2413:59
*** eharney has quit IRC14:00
e0neDuncanT: i've got you point. you mean that 3rd party clients won't work with these changes, do you?14:00
*** TobiasE has joined #openstack-cinder14:00
DuncanTe0ne: Exactly14:00
DuncanTe0ne: We can't change the format or semantics of an existing field14:01
*** TobiasE1 has quit IRC14:01
DuncanTe0ne: The easiest fix I can think of is to add new fields to the output for each time field, create_at_tz or whatever, with the new format14:02
*** pschaef has quit IRC14:02
*** bswartz has quit IRC14:02
*** annegentle has joined #openstack-cinder14:02
e0neDuncanT: oh.. personaly i don't like idea to have several time fields14:02
e0neDuncanT: but your consern about backward compatibility makes sense for me14:03
DuncanTe0ne: Either that or put just the timezone in the new field. Changing the format of the existing field is not acceptable14:03
*** asselin_ has joined #openstack-cinder14:03
*** pradipm has joined #openstack-cinder14:05
*** eharney has joined #openstack-cinder14:06
*** eharney has quit IRC14:06
*** eharney has joined #openstack-cinder14:06
e0neDuncanT: so, we have no chanсe to make error during api development14:08
DuncanTe0ne: Not true, we just have to fix and errors in a back compatible way, or leave them broken until the next major API revision14:09
e0neDuncanT: good point. i should to agree with you14:09
*** markvoelker has joined #openstack-cinder14:10
DuncanTe0ne: Back compatibility is one of those things it is hard to care about until you start to be the guy that customers shout at when it breaks... I have been that guy (and probably still am)14:10
smcginnisDuncanT. e0ne: What about a separate distinct call to get the server's timezone.14:11
e0neDuncanT: +114:11
smcginnisThen the client can determine the difference based off of that.14:11
smcginnisAnd translate other results if needed.14:11
*** anuragpalsule has quit IRC14:12
DuncanTsmcginnis: That would work, yeah, or we can include that as a new field in every response with a time in it14:12
DuncanTsmcginnis: Save having to make a distinct call14:12
DuncanTsmcginnis: Adding new fields is fine14:12
smcginnisDuncanT: I just worry about having the overhead on every call versus just on one call for those that actually care.14:13
smcginnisDuncanT: Granted small, but still extra data going back and forth.14:13
e0nesmcginnis: separate request for timezone doesn't help us while we parse logs14:14
DuncanTsmcginnis: The overhead should be tiny, and the alternative is the client, horizon etc making an extra REST call regularly, which is way worse. We can probably just cache the TZ at startup, and the overhead is too small to measure then14:14
smcginnise0ne: True14:14
e0netimezone info should be logged14:14
*** xyang has joined #openstack-cinder14:14
*** ebalduf has joined #openstack-cinder14:14
smcginnise0ne: OK, I agree.14:15
smcginnisJust thinking of alternatives that won't break anything.14:15
DuncanTyuriy_n17: Does that make sense to you?14:15
e0nesmcginnis: thanks:)14:15
smcginnise0ne: ;)14:15
yuriy_n17DuncanT: Thanks for the discussion.  Let's see what would be said today in API WG.14:16
e0neDuncanT, yuriy_n17 what timizone should be there? a sever timezone?14:17
*** changbl has quit IRC14:17
*** bswartz has joined #openstack-cinder14:17
DuncanTe0ne: Whatever timezone is being used in the time/date fields cinder is reporting14:17
*** ebalduf has quit IRC14:19
*** asselin_ has quit IRC14:19
e0neDuncanT: after going deeper into the sources, i'm not sure about any changes14:21
e0neDuncanT: by default, utcnow() is used everywhere14:21
DuncanTSo all our results are always in UTC? In which case is there any bug, other than a need to document that?14:22
e0neDuncanT: yep, if i didn't miss anything14:23
*** primechuck has joined #openstack-cinder14:26
*** cnesa has quit IRC14:27
DuncanTe0ne: Oh, hmmm, might just be worth putting that as a comment in the bug then?14:29
e0neDuncanT: i found usage only in db purge and two migrations14:29
DuncanTe0ne: Those might technically be bugs then14:30
*** afazekas_ has joined #openstack-cinder14:31
openstackgerritTom Barron proposed openstack/cinder: NFS backup driver
openstackgerritTom Barron proposed openstack/cinder: Refactor Swift backup driver and introduce chunking driver
*** lcurtis has joined #openstack-cinder14:31
e0neDuncanT: agree. i propose to: 1) fix that places to use utc ( in tests doesn't metter for it), 2) mark bug as invalid with a valid comment and 3) update api docs to notice that all dates are in utc tz14:31
e0neyuriy_n17: ^^14:32
yuriy_n17eOne: I see14:32
DuncanTe0ne: I'd fix the tests too, personally, and add a hacking check for to stop it accidentally being added somewhere important... better to go belt-and-braces14:32
*** TobiasE has quit IRC14:33
*** TobiasE1 has joined #openstack-cinder14:33
e0neDuncanT: i like your opinion about hacking rules :)14:33
DuncanTe0ne: I am an imperfect reviewer. Hacking rules don't get bored14:34
openstackgerritSteven Kaufer proposed openstack/cinder: Replication status periodic task optimization
e0neDuncanT: that's why i'm fixing hacking rules and other checks time to tome14:34
DuncanTe0ne: I should spend more time on it, to be honest, but convincing other people to fix them up is a close second best14:35
*** jungleboyj has quit IRC14:36
*** Yogi11 has joined #openstack-cinder14:38
*** xyang1 has joined #openstack-cinder14:39
*** Yogi1 has quit IRC14:40
e0neDuncanT: according to automation, how we decide what scanarios to add for rally gate in cinder? afaik, jgriffith was interesting on improvement it14:43
e0nei mean
*** j_king_ is now known as j_king14:43
*** deepakcs has joined #openstack-cinder14:44
DuncanTe0ne: Anybody can submit scenarios via the usual review process... it is in our repo, we control it14:45
*** IanGovett has quit IRC14:45
e0neDuncanT: sure. but may be we've got something like wishlist for it14:45
lpetrutHi, I'd like to deprecate a dependency version, specifically, I'd like the SMB driver to require version 1.7 of qemu-img (having a check, raising an exception otherwise) in order to remove some hacky workarounds, will this require a blueprint? Or is having DocImpact flag enough?14:45
*** tbarron has quit IRC14:46
*** kaisers has quit IRC14:47
*** cnesa has joined #openstack-cinder14:48
*** mriedem has joined #openstack-cinder14:49
*** tbarron has joined #openstack-cinder14:49
DuncanTe0ne: I'm not aware of one. Raising a wish-list / feature request bug, maybe?14:50
DuncanTlpetrut: I've no idea, sorry. Do those same havks apply to the NFS code too?14:50
DuncanTlpetrut: s/havks/hacks/14:52
lpetrutDuncanT: No, it's mainly targeting vhd/x images because of bugs/missing features on older versions of qemu-img.14:52
DuncanTlpetrut: I think that lowers the impact/risk then... still not sure of the process. I suggest putting a review up with a very clear commit message, and see if anybody complains?14:53
*** cnesa1 has joined #openstack-cinder14:53
*** cnesa has quit IRC14:53
lpetrutDuncanT: sure, thanks!14:53
geguileoe0ne: Do you have time to help me with my question? :)14:57
*** markstur_ has joined #openstack-cinder14:57
e0negeguileo: i'm sorry, i don't know answer right now14:57
geguileoe0ne: Oh, ok, thanks  :)14:58
*** markstur_ has left #openstack-cinder14:58
openstackgerritAnton Arefiev proposed openstack/cinder: Fix lvm manage existing volume
geguileoe0ne: I'll try bothering someone else later ;)14:58
e0negeguileo: afaik, iscsi_ip_address is configurable via cinder.conf14:59
*** bswartz has quit IRC14:59
geguileoe0ne: Yes, the thing is that as far as I can see it is not used for iSCSI with helpers lio and tgt, and it looks like a bug to me15:00
geguileoe0ne: But wanted confirmation15:00
*** bswartz has joined #openstack-cinder15:00
*** annashen has joined #openstack-cinder15:00
openstackgerritAnton Arefiev proposed openstack/cinder: Fix QoSSpecManageApiTest to work in parallel
*** kvidvans has joined #openstack-cinder15:08
*** e0ne is now known as e0ne_15:09
*** BharatK has quit IRC15:11
*** annashen has quit IRC15:12
*** bkopilov has joined #openstack-cinder15:12
*** ebalduf has joined #openstack-cinder15:15
*** annegentle has quit IRC15:16
*** nellysmitt has quit IRC15:19
*** e0ne_ has quit IRC15:19
*** ebalduf has quit IRC15:20
*** pradipm has quit IRC15:20
*** Yogi11 has quit IRC15:21
*** markvoelker has quit IRC15:21
*** markvoelker has joined #openstack-cinder15:21
*** markvoelker has quit IRC15:26
*** TobiasE1 has quit IRC15:27
openstackgerritabhiram moturi proposed openstack/cinder: Enhance unit tests for zfssa drivers
openstackgerritPetrut Lucian proposed openstack/cinder: SMBFS: Fix retrieving total allocated size
*** Mandell has joined #openstack-cinder15:28
*** kallebe has quit IRC15:28
*** pdeore_ has quit IRC15:36
*** jungleboyj has joined #openstack-cinder15:37
*** Yogi1 has joined #openstack-cinder15:38
*** e0ne has joined #openstack-cinder15:39
*** jcru has joined #openstack-cinder15:40
*** scottda has joined #openstack-cinder15:42
*** jcru has quit IRC15:45
*** jcru has joined #openstack-cinder15:46
*** rushil has joined #openstack-cinder15:48
*** Tross has left #openstack-cinder15:49
*** markvoelker has joined #openstack-cinder15:51
jungleboyjhemnafk: When you have a chance.  This had to be rebased:
*** kallebe has joined #openstack-cinder15:58
*** ronis has quit IRC16:02
openstackgerritRichard Hagarty proposed openstack/cinder: Add flash cache policy to 3PAR driver
*** mikedillion has joined #openstack-cinder16:03
*** anshul has quit IRC16:06
*** timcl has quit IRC16:07
*** sigmavirus24 is now known as Slackwarebot16:07
*** Slackwarebot is now known as sigmavirus2416:07
*** timcl has joined #openstack-cinder16:08
*** changbl has joined #openstack-cinder16:09
*** nkrinner has quit IRC16:10
*** mikedillion has quit IRC16:11
*** timcl1 has joined #openstack-cinder16:11
*** timcl has quit IRC16:12
*** ebalduf has joined #openstack-cinder16:16
*** annegentle has joined #openstack-cinder16:16
*** TobiasE has joined #openstack-cinder16:17
openstackgerritPetrut Lucian proposed openstack/cinder: Fix retrieving the volume path and format when using SMBFS drivers
*** ebalduf has quit IRC16:20
*** rmesta has joined #openstack-cinder16:21
*** sparr has quit IRC16:21
*** sparr has joined #openstack-cinder16:21
*** annegentle has quit IRC16:22
*** avishay__ has quit IRC16:22
*** avishay has joined #openstack-cinder16:22
openstackgerritMichal Dulko proposed openstack/cinder: Retry requests in scheduler to wait for metrics
*** crose has joined #openstack-cinder16:24
dulek_DuncanT, winston-d_zZZ: Finally I've updated patch for bug 1409012 to implement DuncanT's idea of retrying.16:24
openstackbug 1409012 in Cinder "Volume becomes in 'error' state after scheduler starts" [High,In progress] - Assigned to Michal Dulko (michal-dulko-f)16:24
dulek_I've had multiple problems with test_pure globally monkey-patching retry decorator.16:24
*** dannywilson has joined #openstack-cinder16:25
*** ronis has joined #openstack-cinder16:26
*** deepakcs has quit IRC16:27
*** thangp has joined #openstack-cinder16:27
*** bill_az has quit IRC16:28
*** tsekiyama has joined #openstack-cinder16:28
*** TobiasE has quit IRC16:28
*** IanGovett has joined #openstack-cinder16:29
*** Mandell has quit IRC16:33
*** jdurgin1 has quit IRC16:34
*** david-lyle_afk is now known as david-lyle16:36
nikesh_vedamsi have a couple of questions16:36
*** timcl1 has quit IRC16:37
*** IanGovett has quit IRC16:38
*** gary-smith_ has quit IRC16:39
*** gary-smith_ has joined #openstack-cinder16:40
*** rushiagr is now known as rushiagr_away16:40
*** IanGovett has joined #openstack-cinder16:41
dannywilsonnikesh_vedams: Good question.  I don't know the answer but someone over in openstack-infra channel might be able to help16:42
*** Apoorva has joined #openstack-cinder16:44
*** hemnafk is now known as hemna16:45
*** sgotliv has quit IRC16:46
*** rushiagr_away is now known as rushiagr16:47
*** pcaruana has quit IRC16:50
*** rwsu-afk is now known as rwsu16:50
hemnawinston-d_zZZ, DuncanT
*** takedakn has joined #openstack-cinder16:53
*** annashen has joined #openstack-cinder16:55
*** lpetrut has quit IRC16:56
*** ronis has quit IRC16:59
*** afazekas_ is now known as afazekas17:00
*** jistr has quit IRC17:02
DuncanT@hemna Looks like it does what I want... I can test it in the morning17:06
openstackgerritBob Callaway proposed openstack/cinder: NetApp eseries implementation for manage/unmanage
*** Ilja has joined #openstack-cinder17:08
*** dulek has quit IRC17:10
*** dulek_ has quit IRC17:10
openstackgerritBob Callaway proposed openstack/cinder: Add Manage/Unmanage support to NetApp NFS drivers
*** IanGovett has quit IRC17:10
hemnaDuncanT, ok cool.17:11
hemnaDuncanT, I tested it against LVM prior to adding the capability to LVM17:11
hemnaand the scheduler said "no hosts found"17:11
hemnawhich was correct17:11
hemnathen added the capability to the lvm stats, and the scheduler found the lvm host.17:11
hemnaso I think it does what you wanted.17:12
hemnaso the question is, do we just not update any drivers for K to add the multiattach capability in their stats?   this would prevent all creates with --allow-multiattach from finding a host17:12
*** timcl has joined #openstack-cinder17:12
hemnathe other way to look at it is, we could still add multiattach=True to a few drivers17:13
hemnacreates would work17:13
hemnaand a single attach would work17:13
hemnathen when L ships, those volumes can attach a second time, once the nova code lands.17:13
openstackgerritxing-yang proposed openstack/cinder: Fixed a concurrency issue in VMAX driver
*** xyang has quit IRC17:14
openstackgerritxing-yang proposed openstack/cinder: Fixed a concurrency issue in VMAX driver
*** annashen has quit IRC17:16
*** ebalduf has joined #openstack-cinder17:17
*** timcl has quit IRC17:17
*** annashen has joined #openstack-cinder17:17
*** avishay has quit IRC17:17
*** annegentle has joined #openstack-cinder17:18
*** nellysmitt has joined #openstack-cinder17:20
openstackgerritxing-yang proposed openstack/cinder: Fixed a problem in recreating masking view in VMAX
jungleboyjhemna: xyang1 If you two are doing reviews here are two that should be ready to go in:
*** ebalduf has quit IRC17:21
hemnaok just a sec17:21
jungleboyjPart of the priority list for K-3.17:21
*** _cjones_ has joined #openstack-cinder17:21
xyang1jungleboyj: Will take a look17:21
jungleboyjxyang1: Thank you.17:22
*** annegentle has quit IRC17:23
*** nellysmitt has quit IRC17:25
openstackgerritxing-yang proposed openstack/cinder: Fixed errors in docstrings in the VMAX driver
hemnaDuncanT, would you like me to add that gist as a follow up patch to my existing?17:26
hemnaDuncanT, it might make it easier to review.17:26
hemnaor I can add it to my existing patch.17:26
DuncanThemna: Followup is fine by me, whatever is easier for you. I'm happy enough with most of the code17:27
openstackgerritxing-yang proposed openstack/cinder: Check license before clone in VMAX driver
*** leeantho has joined #openstack-cinder17:30
xyang1jungleboyj: Why are V1 changes needed?17:32
*** mriedem is now known as mriedem_away17:34
hemnaDuncanT, ok17:35
hemnaI'll post a follow up17:35
*** Yogi1 has quit IRC17:35
jungleboyjxyang1: Hold on.17:35
*** IanGovett has joined #openstack-cinder17:36
jungleboyjI know I asked Steve that before.17:36
*** tries__ has joined #openstack-cinder17:36
*** tries__ has joined #openstack-cinder17:36
hemnajungleboyj, we are seeing some odd behavior with one of our drivers after an oslo update17:37
jungleboyjxyang1: Because the function signature changed sort_key vs. sort_keys17:37
jungleboyjhemna: Ruh roh.17:38
*** karimb has quit IRC17:39
jungleboyjhemna: I need to run.  Back shortly and we can chat.17:39
hemnajungleboyj, ok, I'm going to test something locally17:39
hemnaCI is now failing on every patch for lefthand17:39
kauferxyang1: The function used to only accept a single sort key, now it's a list so I changed to sort_keys17:39
hemnaand we haven't changed the lefthand driver, nor the client it uses.17:40
*** hodos|2 has joined #openstack-cinder17:41
*** nellysmitt has joined #openstack-cinder17:41
xyang1kaufer: So that is a common function used by both v1 and v2?17:41
kauferxyang1: correct, both v1 and v2 invoke the volume api get_all() function17:42
*** karimb has joined #openstack-cinder17:43
*** yuriy_n17 has quit IRC17:43
kauferxyang1: And that is were the kwarg in the signature changed17:43
xyang1kaufer: Ok17:43
*** karimb has quit IRC17:43
*** hodos has quit IRC17:44
*** sseverson has joined #openstack-cinder17:49
*** lpetrut has joined #openstack-cinder17:50
*** sgotliv has joined #openstack-cinder17:51
xyang1kaufer: jungleboyj is the v1 change backward compatible?17:54
kauferxyang1: Yes, it does not change the API contract, sorting is not supported in v117:55
kauferxyang1: The sort key/direction is still created_at/desc and that is now passed to the volume API as a list instead of a string17:56
xyang1kaufer: ok17:56
*** cpallares has joined #openstack-cinder17:57
*** jordanP has quit IRC18:01
*** kbyrne has quit IRC18:03
*** xyang has joined #openstack-cinder18:03
*** emagana has joined #openstack-cinder18:03
*** hodos|2 has quit IRC18:03
*** Mandell has joined #openstack-cinder18:04
e0nexyang1, kaufer: just 2c from mi side. i've tested it with cinder client and it works for me18:05
*** kbyrne has joined #openstack-cinder18:05
*** emagana has quit IRC18:07
xyang1e0ne: Great!  Thanks18:09
*** vilobhmm has joined #openstack-cinder18:13
e0nexyang1: np. i had to verify that it not breaks api v118:13
*** vilobhmm1 has joined #openstack-cinder18:14
*** afazekas has quit IRC18:16
*** ebalduf has joined #openstack-cinder18:17
*** vilobhmm has quit IRC18:18
*** ronis has joined #openstack-cinder18:19
*** annegentle has joined #openstack-cinder18:19
*** ebalduf has quit IRC18:22
*** annegentle has quit IRC18:24
*** bswartz has quit IRC18:26
*** patrickeast has joined #openstack-cinder18:29
jungleboyjhemna: I am back.  What did you break?18:35
hemnagetting very odd behavior with imports18:35
hemnaI can't explain it18:36
jungleboyjhemna: Not what I was expecting you to say.18:37
*** Miouge has quit IRC18:39
*** mtanino has joined #openstack-cinder18:40
*** cpallares has quit IRC18:42
*** bswartz has joined #openstack-cinder18:43
*** jgravel has quit IRC18:43
*** Miouge has joined #openstack-cinder18:45
jungleboyjhemna: Anything I can do to help right now?18:47
hemnajungleboyj, have a magic wand?18:47
jungleboyjhemna: He he he.18:48
jungleboyjhemna: You really want me to answer that?18:48
hemnahey now18:48
*** avishay has joined #openstack-cinder18:48
*** coolsvap is now known as coolsvap_18:48
*** bswartz1 has joined #openstack-cinder18:48
jungleboyjhemna: I don't think there was anything that got pulled in by the oslo changes that should cause weird import behavior.18:49
*** bswartz has quit IRC18:49
*** bswartz1 is now known as bswartz18:49
*** Mandell has quit IRC18:51
* jungleboyj gets out my red marker for tbarron 's NFS patch. :-)18:53
tbarron?? Woohoo!18:53
* jungleboyj knew thtat would make tbarron Happy.18:53
*** tries__ has quit IRC18:54
* tbarron wasn't calling jungleboyj names.18:54
*** Thelo has quit IRC18:54
jungleboyjOh, it starts with most of this code can be removed in the commit message.  Uh oh.18:54
tbarronjungleboyj: The eventual plan is to put a posix filesystem driver between the chunked driver and the nfs driver.18:55
tbarronBut I am not, for obscure NetApp legal reasons, allowed to write posix driver code.18:56
jungleboyjtbarron: I know, the commit message explains.  I was just giving you a hard time.18:56
tbarronSo I am posting an NFS driver that inherits directly from posix.18:56
tbarronjungleboyj: kk18:56
jungleboyjI remember talking about that at the meet-up and everyone was onboard with the plan.18:56
tbarronjungleboyj: cool18:57
jungleboyjJust thought it was funny as I get ready to review seeing that.18:57
*** Mandell has joined #openstack-cinder18:57
*** IanGovett has quit IRC18:58
*** avishay has quit IRC18:59
*** Mandell has quit IRC18:59
*** jgravel has joined #openstack-cinder18:59
*** e0ne is now known as e0ne_19:00
jungleboyjhemna: You talking about this:19:00
jungleboyj./cinder/volume/drivers/san/hp/ H302  import only modules.'from hp3parclient import client' does not import a module19:00
hemnajungleboyj, so kinda yah, we are looking at the lefthand driver19:01
smcginnisjungleboyj: That looks like the pep8 patch that just went in.19:01
jungleboyjYeah.  Ok, that wasn't me!19:01
smcginnisjungleboyj: Get out of jail free card. :)19:02
hemnathey changed the imports19:02
hemnaand removed the from hplefthandclient import client as hp_lh_client19:02
hemnaand are trying to access the package's module w/o importing it19:02
hemnathe CI for the patch FAILED19:02
*** Mandell has joined #openstack-cinder19:02
hemnaas it should19:02
smcginnishemna: Odd that wasn't caught by Jenkins...19:03
hemnasmcginnis, so in jenkins the import fails and falls back to the except19:04
hemnaand uses the fake19:04
hemnawhich causes the access to hplefthandclient.client to work, because hplefthandclient is a mock19:04
smcginnishemna: Oh, duh. I see.19:04
hemnaso I'll revert this patch for the lefthand driver19:05
smcginnishemna: Yeah, that could be a problem. :]19:05
hemnaso the lefthand 3rd party CI correctly fails on that patch19:05
hemnaand it was ignored.19:06
hemnaso everything worked as it should19:06
hemnaand we still missed it19:06
*** timcl has joined #openstack-cinder19:06
hemna'worked as it should' meaning the 3rd party CI failed correctly19:06
*** tries__ has joined #openstack-cinder19:06
*** tries__ has joined #openstack-cinder19:06
smcginnishemna: I think right now CI failures are so unreliable.19:06
smcginnishemna: It will be nice when they are.19:07
hemnasmcginnis, yah that's the other side of it19:07
hemnasince CI isn't really reliable, we don't vote19:07
smcginnishemna: Kind of crying wolf too much right now.19:07
*** jungleboyj has quit IRC19:07
hemnabut it does raise the issue19:07
smcginnishemna: And to be clear, I'm not referring to HP's CI but all.19:07
smcginnishemna: I agree, it's a problem.19:07
hemnaif folks aren't even looking at the 3rd party CI failures when they touch other people's drivers, then why are we doing 3rd party CI at all ?19:08
openstackhemna: Error: "!" is not a valid command.19:08
smcginnisThat is a really good point.19:08
*** jungleboyj has joined #openstack-cinder19:08
smcginnisIf someone touches someone elses driver, really they should investigate any third party CI failures.19:08
smcginnisI know I will from now on.19:08
akerr_ci is the bane of my life right now19:09
smcginnisakerr_: You aren't the only one. :)19:09
*** bswartz has quit IRC19:09
sigmavirus24akerr_: take that up with Ra's al Ghul19:09
sigmavirus24(bane... batman... nevermind)19:09
jungleboyjphew, so it wasn't me that broke things?19:09
hemnasmcginnis, yah, especially since they aren't voting19:09
smcginnisI completely agree with the need for CI (and the lefthand issue reinforces that) but it's a huge PITA until things mature.19:09
*** bswartz has joined #openstack-cinder19:09
hemnajungleboyj, yah it's not oslo19:10
smcginnisjungleboyj: We'll still blame you. :)19:10
hemnawe thought it might have been because of the timing of the failures19:10
hemnawrt another oslo review19:10
jungleboyjhemna: Yay!  Back to my regularly scheduled programming.19:10
* jungleboyj would only touch HPs drivers with the greatest care and respet.19:10
smcginnishemna: I did review that patch, so for my part I apologize for not catching that.19:11
*** jdurgin has quit IRC19:13
*** tries__ has quit IRC19:14
*** timcl1 has joined #openstack-cinder19:15
*** timcl has quit IRC19:16
*** ebalduf has joined #openstack-cinder19:18
jungleboyjtbarron: What is 0o770 for a file permissions setting?19:20
hemnasmcginnis, no worries.  the author should have looked at the 3rd party CI failures19:20
hemnabut yah, reviewers should take note:19:20
jungleboyjtbarron: Never seen an 'o' in there before.19:21
hemnaif a patch touches a driver, check the 3rd party CI for those drivers.19:21
smcginnishemna: agreed19:21
*** coolsvap_ is now known as coolsvap19:21
jungleboyjhemna: +119:21
hemnano thingee today ?19:21
jungleboyjhemna: Haven't seen him around.  He has had some personal stuff going on.19:22
*** Mandell has quit IRC19:22
*** ebalduf has quit IRC19:23
smcginnisSpeaking of CIs...19:24
smcginnisI'm trying to get our FC tests going.19:24
smcginnisI'm getting random test failures that time out.19:25
smcginnisI'm tracing one19:25
smcginnisAnd all within the same second a volume creation request is initiated, it runs, everything looks good.19:25
smcginnisI trace through the req ID and it goes along for a few milliseconds, then nothing.19:25
smcginnisAnd the test shows it timed out.19:26
smcginnisAny tips on how to triage something like this?19:26
*** Mandell has joined #openstack-cinder19:26
*** Mandell has quit IRC19:26
*** Mandell has joined #openstack-cinder19:27
patrickeastdoes it only repo on the CI?19:27
smcginnispatrickeast: Yeah. :[19:27
smcginnisWe can run through the tests all day long on our development machines and it's fine.19:27
smcginnispatrickeast: At least so far.19:27
hemnaso it's getting lost in the queue ?19:28
tbarronjungleboyj: you're talking chmod, I guess.19:28
jungleboyjtbarron: Yes.19:28
smcginnisJust inconsistently consistent enough to drive me crazy.19:28
tbarronit's reportedly the way to make sure you are using octal and be python3 compatible19:28
*** anshul has joined #openstack-cinder19:29
smcginnisOr consistently inconsistent. One of the two. :)19:29
*** tries__ has joined #openstack-cinder19:29
*** tries__ has joined #openstack-cinder19:29
*** asmith_brcd has joined #openstack-cinder19:29
jungleboyjtbarron: Ok, so 0o660 == 066019:29
patrickeastsmcginnis: we ran into some random failures like that (still hvae them on our backlog of mysteries to look into) but they don’t repro if you run the tests with concurrency=119:29
patrickeastwe looked into it and were unable to figure out yet if the issue was cinder, our driver, our backend, or something in tempest19:30
smcginnispatrickeast: I have seen that help other weird issues.19:30
tbarronjungleboyj: right, if you want to put a placeholder remark I'll chase down the link that documents this.19:30
tbarronthat way if I've misinterpreted or something we can fix it.19:30
*** jdurgin has joined #openstack-cinder19:30
smcginnispatrickeast: Seems like over 4 and you're asking for trouble, but even 2-4 have possible issues.19:30
smcginnispatrickeast: Really would hate to limit to one on the CI.19:30
jungleboyjtbarron: No, all good,. I googled and found plenty of examples of it being used.  Just no explanation as to why.19:30
tbarronI'd definitely appreciate more than my own two eyes on that19:30
smcginnisMight be a good test though.19:30
patrickeasti saw with 2-4 workers it was like 90% success rate19:30
patrickeastmaybe more19:30
jungleboyjpatrickeast: Do you have the latest comments from you spec integrated here:
smcginnispatrickeast: That's what I'm seeing with our iSCSI (which is odd that there's a difference).19:31
patrickeastjungleboyj: yep, its all up to date19:31
* jungleboyj takes my red pen over to patrickeast 's patch. :-)19:31
patrickeastsmcginnis: i’ve noticed issues with timing with udev stuff in the past, i suspect there might be more of them19:31
*** sgotliv has quit IRC19:31
patrickeastsmcginnis: with iscsi*19:31
smcginnispatrickeast: Guess I'll keep digging. But thanks for the concurrency = 1 suggestion. I think I'll try that once.19:33
smcginnispatrickeast: Do you have the syntax off hand? I think I always mess it up and have to dig for it. :)19:33
*** Yogi1 has joined #openstack-cinder19:33
patrickeastfunny enough i also found if we run with less resources (4 cores instead of 8, 8gb ram instead of 16) it fails less frequenty per test too… which kind of points to concurrency issues19:33
patrickeastfor tox i think its —CONCURRENCY=119:33
smcginnispatrickeast: I think unless you specify it will be concurrency=[# of cores]19:33
patrickeastmaybe not all caps19:33
*** Yogi11 has joined #openstack-cinder19:33
patrickeasti forget19:33
smcginnispatrickeast: Cool, thanks.19:34
smcginnispatrickeast: That's OK, I want to yell at it anyway so I'll try caps. ;)19:34
*** diemt has joined #openstack-cinder19:36
openstackgerritErlon R. Cruz proposed openstack/cinder: Adds pool aware scheduling for HNAS drivers
openstackgerritErlon R. Cruz proposed openstack/cinder: Adds SSH communication to HNAS drivers
openstackgerritErlon R. Cruz proposed openstack/cinder: Fix HNAS driver parsing errors
*** Yogi1 has quit IRC19:37
*** crose has quit IRC19:38
*** jgravel has quit IRC19:38
*** geguileo has quit IRC19:39
*** crose has joined #openstack-cinder19:40
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder: Fixes the import for Lefthand driver
*** anshul has quit IRC19:46
jungleboyjhemna: Typo in the commit message ^^19:46
openstackgerritJay Bryant proposed openstack/cinder: Sync middleware module from oslo-incubator
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder: Fixes the import for Lefthand driver
*** asmith_brcd has quit IRC19:50
*** e0ne_ is now known as e0ne19:52
*** jkraj has quit IRC19:55
*** sgotliv has joined #openstack-cinder19:56
*** Saba has joined #openstack-cinder19:56
*** kallebe has left #openstack-cinder19:56
SabaHi, how to get volume usage ??19:56
*** Saba is now known as Guest7862119:57
kauferjungleboyj: For my sort reviews do we need to ensure that the delta to the spec is merged?  One of the reviews is +A but my delta to the spec (for the new sorting syntax) is not:
*** jungleboyj_ has joined #openstack-cinder20:01
jungleboyj_hemna: Want me to nija merge 159586 ?20:02
*** jungleboyj has quit IRC20:03
hemnayes please20:05
hemnaonce jenkins +120:05
jungleboyj_hemna: Will do20:05
hemnaI'm writing up a ML post about this.20:06
patrickeastjenkins and the hp ci, right? :p20:06
hemnaheh yah20:06
*** vilobhmm1 has quit IRC20:07
*** vilobhmm has joined #openstack-cinder20:07
kauferjungleboyj_: For my sort reviews do we need to ensure that the delta to the spec is merged?  One of the reviews is +A but my delta to the spec (for the new sorting syntax) is not:
*** nellysmitt has quit IRC20:09
*** rlucio has joined #openstack-cinder20:10
jungleboyj_kaufer: So, it is updated to match the code but hasn't merged yet?20:11
openstackgerritMike Perez proposed openstack/cinder: Update v1 deprecation warnings
kauferjungleboyj_: Yeah, once all of the cross project and API WG sorting stuff was mereged then the spec was out of date, so I submitted a delta to the spec20:12
*** Mandell has quit IRC20:12
kauferjungleboyj_: the code reflects the updated spec20:12
jungleboyj_kaufer: Ok, it has a +2 from Mike.  I would just ping jgriffith and DuncanT and see if you can get one of them to merge .20:13
Guest78621Hi, Does anyone know how to compute volume usage in cinder ?20:14
kauferjungleboyj_: Thx20:14
jungleboyj_kaufer: Welcome.20:15
guitarzanGuest78621: cinder quota-usage-show will show you I think20:17
*** ebalduf has joined #openstack-cinder20:19
*** aix has quit IRC20:22
*** ebalduf has quit IRC20:23
*** sigmavirus24 is now known as sigmavirus24_awa20:32
Guest78621cinder quota-usage-show doesnt show the usage of volume20:33
*** tbarron has quit IRC20:33
*** tbarron has joined #openstack-cinder20:33
guitarzanGuest78621: I guess I don't understand what you're after20:35
guitarzanwhat is "usage of volume"20:35
Guest78621My aim is to get real usage of the Storage (not allocated)20:36
guitarzanah, that's been discussed, but I don't know where it ended up20:36
guitarzanthe easy answer is "ask your storage device" :)20:37
*** bill_az has joined #openstack-cinder20:38
Guest78621I want to read via meter for our monitoring software20:39
*** Mandell has joined #openstack-cinder20:40
*** mriedem_away is now known as mriedem20:47
openstackgerritIvan Kolodyazhny proposed openstack/cinder: Allow archiving deleted rows to shadow tables, for performance
openstackgerritIvan Kolodyazhny proposed openstack/cinder: Allow archiving deleted rows to shadow tables, for performance
openstackgerritSean McGinnis proposed openstack/cinder: Remove use of contextlib.nested
*** asmith_brcd has joined #openstack-cinder20:50
*** dustins has joined #openstack-cinder20:54
openstackgerritSean McGinnis proposed openstack/cinder: Remove use of contextlib.nested
*** tbarron1 has joined #openstack-cinder20:56
*** morganfainberg is now known as needscoffee20:57
*** tbarron has quit IRC20:59
e0nejungleboyj_: hi! do you think that we need new oslo.middleware from incubator?21:00
e0nehemna: i'm sorry that we broke you driver:(21:01
*** lcurtis has quit IRC21:05
*** rushiagr is now known as rushiagr_away21:06
jungleboyj_e0ne: Yes, so that we can stop getting the oslo.middleware warning every time flake8 is run.21:09
jungleboyj_It is a small change and as Mike pointed out it is good to pull in the library requirement.21:09
e0nejungleboyj_: but it will be replaced with oslo_middleware soon21:10
jgriffithhemna: that's a bummer on that hacking change21:11
*** jgravel has joined #openstack-cinder21:11
jgriffithhemna: but honestly, kinda cool that the issue was found and fixed in less than a day no?21:11
jungleboyj_e0ne: Ok, so do you think we should just remove oslo.middleware from incubator for this change?21:12
hemnajgriffith, yah it's all good.21:12
jgriffithhemna: uh oh... the sounds like "it's all good, but it's not" response21:13
e0nejungleboyj_: yes. we need to force moving to a stand-alone libs instead of incubator21:13
jungleboyj_jgriffith: Good point.  CI working.21:13
jungleboyj_Ok, let me try making that change.21:13
jungleboyj_Go ahead and -1 that.21:13
*** cnesa1 has quit IRC21:13
e0nejungleboyj_: now you are trying to do the change what will be removed with a next commit21:13
hemnahonestly, I just waited to raise the issue on the ML21:14
hemnato get folks thinking about it21:14
hemnawe aren't upset at all21:14
jungleboyj_e0ne: Good point.  I will push a new patch up.21:14
jgriffithhemna: ok, cool.  and yeah, it's a super valid point21:14
hemnaand in fact are happy that the CI did bring it to light and we were able to fix it.21:14
e0nejungleboyj_:  did you see my comment there?21:14
jgriffithhemna: I just don't know about the voiting thing at least yet.21:15
*** coolsvap is now known as coolsvap_21:15
jgriffithSurely could/should be better about at least "looking" though21:15
hemnayah I don't think voting makes sense right now either21:15
jgriffithhemna: don't disagree there21:15
hemnaI was chatting with asselin at lunch about a health dashboard21:15
*** lpetrut has quit IRC21:16
hemnait would be nice to gauge the health of individual CI over time21:16
jgriffithhemna: I keep wanting that too21:16
patrickeastthere was something in the works for that21:16
jgriffithI know anteaya has referenced a project under way multiple times since FTC meetup21:16
patrickeastnot sure what happened to it21:16
jgriffithpatrickeast: +121:16
erlone0ne: Hi Ivan, not sure what you mean in HNAS ssh change21:16
e0nejungleboyj_:  and this request
anteayawhat did I do?21:16
jgriffithanteaya: it's all your fault :)21:16
anteayaof course it is21:16
* anteaya dons her armour21:16
jgriffithanteaya: the topic of the third party CI health/dashboard came up again21:17
jgriffithanteaya: and no, nothing is your fault (this time) :)21:17
anteayaoh whew21:17
patrickeast^ this thing21:17
hemnaanteaya, fwiw, I posted on the cinder ML about a 3rd party CI issue.21:17
e0neerlon: as i anderstand your code, you  try to re-raise exception, do you?21:18
anteayahemna: I read with interest, good post21:18
anteayajgriffith: and nice reply21:18
hemnaanteaya, in the end as jgriffith commented, worked out well.21:18
anteayait did21:18
anteayaand was good to let folks know what is happening21:18
anteayaso thank you21:18
erlone0ne: yes, I just catch it, so I can log, and then I have to re-raise it so cinder can catch21:19
e0neerlon: from the python docs "If no expressions are present, raise re-raises the last exception that was active in the current scope. I"21:19
jungleboyj_e0ne: Oh, didn't know we had a patch for that out there.21:19
e0neerlon: so you need only 'raise' statement21:19
erlone0ne: do know why the change lost all reviews? it had a +2 IIRC21:19
mtaninojungleboyj_: Hi, Can I get your red pen for this?
jungleboyj_I need to drop off for a bit.  I will look at the patch you sent and mine.21:19
jungleboyj_e0ne: ^^21:20
jungleboyj_mtanino: Sure.21:20
mtaninojungleboyj_: I got an agreement from Avishay.21:20
*** ebalduf has joined #openstack-cinder21:20
smcginnise0ne, erlon: One other data point, if you don't just "raise" you lose the original traceback.21:20
jgriffithpatrickeast: anteaya oh dear, that's a victim of scope creep isn't it21:20
mtaninojungleboyj_: thank you :)21:20
smcginnisSo could make troubleshooting a little harder.21:20
jgriffithI would just like a simple dashboard showing, total-runs, passed, failed21:21
anteayajgriffith: yes, would be good to have some feedback on how to get it to a managble size21:21
jgriffiththat alone would be aweomse21:21
anteayajgriffith: and might possibly be small enough to exist21:21
smcginnisanteaya: +121:21
anteayajgriffith: do share your thoughts to that effect on the spec21:21
*** lcurtis has joined #openstack-cinder21:21
anteayaso weigh in21:21
e0neerlon: what change did lost them?21:21
anteayasince i don't know anyone who doesn't want a dashboard21:21
*** changbl has quit IRC21:21
erlone0ne: so, using only one exception I would have to Log inside the exeption21:21
anteayawe just have to have enough behind something to get something up21:22
smcginniserlon: No, you can still log. Then just "raise" alone reraises it.21:22
erlone0ne: the othere ones (in the dependencies) ware also rebased and keep the reviews21:23
e0neerlon: i'm talking about such case:
*** ebalduf has quit IRC21:24
smcginnise0ne, erlon: Just commented.21:25
*** tsekiyama has quit IRC21:26
smcginniserlon: Oh, also forgot to point out six.text_type(e) is not necessary there.21:26
*** tsekiyama has joined #openstack-cinder21:26
smcginnisThe logger handles that correctly.21:26
e0nesmcginnis: thanks!21:26
smcginnise0ne: np :)21:26
*** jungleboyj_ has quit IRC21:27
erlone0ne: smcginnis: o_O, it wasn't that 3 PS ago, someone asked to use21:27
xyangjgriffith: can you take a look of this small change?
smcginniserlon: Yeah, sorry, there was some confusion about that. We did some testing to validate and six is not needed.21:28
e0neerlon, smcginnis: maybe it was me:(, it's not required there but it makes this more excplicit. i won't -1 for six.text_type(exception)21:28
smcginnise0ne: I agree. But it is redundant because the logger ultimately ends up doing it too.21:29
erlone0ne: thanks for the paste, I was trying to understand the raise re-raise thing21:29
*** timcl1 has quit IRC21:30
erlone0ne: IIUC, just calling raise, will throw the last exeception21:30
e0neerlon: you're welcome! here the python docs:
e0neerlon: yes, it will21:30
erlone0ne: good, Ill fix that21:31
jgriffithxyang: way too much code there, I can't possibly review that today :)21:34
*** timcl has joined #openstack-cinder21:34
xyangjgriffith: ok, add to your review list then:)21:35
jgriffithxyang: LOL21:35
jgriffithxyang: approved... I am surprised though, I would've never thought that thin would gak over it only being 221:35
jgriffithxyang: in fact I would've thought you could set it to 121:35
jgriffithxyang: or is that specific to some "other" driver?21:36
xyangjgriffith: 1.0 was what I had originally.  you said it should be at least 2.0?:)21:36
jgriffithxyang: sounds like me :)21:36
*** jcru has quit IRC21:36
xyangjgriffith: so I changed to 2.021:36
jgriffithxyang: what i mean is your description says "that value is too low"21:36
xyangjgrffith: now everyone else told me 2 is too low21:36
jgriffithxyang: OHHH21:36
jgriffiththis is feedback/bike-shedding change, not a technical change21:37
jgriffithfair enough21:37
xyangjgriffith: so xtremio driver tried to use a different default value of 20.  pure also starts the default as 2021:37
jgriffithxyang: I was under the impression that you meant 2 was too small and the driver would reject it21:37
xyangjgriffith: that's why I thought may be 2 doesn't make sense in the real world21:37
jgriffithxyang: that's what I thought was weird and why I mentioned I thought even 1 should work21:38
*** timcl1 has joined #openstack-cinder21:38
jgriffithxyang: not that I think 1 should be the default (that's NO overprov)21:38
xyangjgriffith: it should work.  it is just that no one likes 2.0 as the default21:38
jgriffithxyang: I don't agree with people about 20 but I don't care either :)21:38
jgriffithxyang: it's a completely arbitrary "thing" if you ask me21:38
xyangjgriffith: you can change it in cinder.conf any way21:38
jgriffithxyang: yeah.. I follow ya now21:38
jgriffithxyang: right :)21:39
jgriffithxyang: thus my "bike-shedding" comment :)21:39
xyangjgriffith: it turns out you can use CONF.override to set the value, not it can't change the default21:39
*** timcl has quit IRC21:39
patrickeastfwiw the 20 for us isn’t a default per-say, it is only set to that when we would otherwise calculate it to NaN or infinite21:39
patrickeastit is only set when the array is otherwise empty21:39
jgriffithOHHHH... it's patrickeast 's fault!!!!21:40
* patrickeast hides21:40
jgriffithI should have known!!  Dang you!!!21:40
xyangjgriffith: right:)21:40
mtaninojgriffith: quick question. After refactoring SCSI target, anyone doesn't use "iscsi_num_targets" config option now.21:40
openstackgerritErlon R. Cruz proposed openstack/cinder: Adds pool aware scheduling for HNAS drivers
openstackgerritErlon R. Cruz proposed openstack/cinder: Adds SSH communication to HNAS drivers
mtaninojgriffith: should we deprecate this option or should we add some logic to use this?21:41
jgriffithmtanino: if it wasn't put back in IET I don't know of anyplace it's being used21:41
jgriffithmtanino: but lemme look21:41
*** SamYaple is now known as I-Am-Sam21:41
rhagartyxyang, question regarding your review of my patchset -
mtaninojgriffith: my proposed IET doesn't use the config21:41
*** I-Am-Sam is now known as SamYaple21:42
jgriffithmtanino: should be cool then21:42
xyangrhagarty: sure.21:42
jgriffithmtanino: honestly I barely even remember why/when we added that21:42
rhagartyyou suggest I get rid of continuation line '\' by using '()' instead. so just to be sure I'm doing it right, here are the lines in question -21:42
rhagartymock_client.getWsApiVersion.return_value = \21:42
rhagarty    wsapi_version21:42
rhagartyI should instead do this?21:42
rhagartymock_client.getWsApiVersion.return_value = (21:42
rhagarty    wsapi_version)21:42
mtaninojgriffith: me too.21:43
rhagartyok - just wanted to verify. Thanks21:43
jgriffithmtanino: actually looks like cxt is using it21:43
*** lcurtis has quit IRC21:44
mtaninojgriffith: That method does not use anywhere. just copied from tgt21:44
jgriffithmtanino: ohh.. that's just lovely!21:45
jgriffithdead method21:45
mtaninojgriffith: yes!!!21:45
jgriffithmtanino: yeah, let's deprecate it and remove that dead code from teh cxt code21:45
*** takedakn has quit IRC21:45
mtaninojgriffith: I got it.21:45
jgriffithmtanino: awesome, thanks!21:45
*** enterprisedc has quit IRC21:45
mtaninowill propose a patch.21:45
jgriffithmtanino: you want to do both or you want me to take one of them?21:46
mtaninojgriffith: np. I'll take both of them.21:46
jgriffithmtanino: okie dokie21:46
jgriffithmtanino: thanks21:46
*** Miouge has quit IRC21:46
*** enterprisedc has joined #openstack-cinder21:47
tsekiyamajgriffith: hemna: xyang: I've unified the data structure for alternative iSCSI targets with multipath. Could you check it if you have a chance?  The functional code is now just 2 lines in ...21:47
jgriffithtsekiyama: nice!21:47
jgriffithtsekiyama: I'm confused though21:49
jgriffithtsekiyama: you still have to different DS's (line 1323 of
jgriffithtsekiyama: s/to/two/21:49
*** tries__ has quit IRC21:50
jgriffithtsekiyama: I thought we talked about something like "additional-xxx": []21:50
jgriffithtsekiyama: maybe I'm on the wrong page21:50
*** markvoelker has quit IRC21:51
jgriffithtsekiyama: you're commit message is kind of along those lines21:51
*** markvoelker has joined #openstack-cinder21:51
jgriffithtsekiyama: but the comment in the code doesn't match21:51
openstackgerritRichard Hagarty proposed openstack/cinder: Add flash cache policy to 3PAR driver
tsekiyamajgriffith: ah, based on last IRC discussion, "target_portals" can be used to both multiatth and failover21:52
erlone0ne: smcginnis:  done, can you review when you have a chance?21:52
patrickeastoh nice, i like it a lot more now that it matches the multipath changes… i was playing around with that patch and having a big if/else to fill in the same stuff was kind of weird21:52
jgriffithtsekiyama: sure...21:52
tsekiyamajgriffith: but for old nova, it should return both single and multiple keys21:53
smcginniserlon: I have it up right now. :)21:53
jgriffithtsekiyama: what I am getting at is in your commit message21:53
e0neerlon: i'll try do it today after CI passed21:53
jgriffithtsekiyama: you ahve connection_info:{data:'target_portal'...., 'target_portals":[]....21:53
erlonsmcginnis: thanks! :)21:53
jgriffithwhich is cool by me21:53
jgriffithbut in the comment section of there's two different data structures called out?21:54
erlone0ne: ok21:54
openstackgerritIvan Kolodyazhny proposed openstack/cinder: Allow archiving deleted rows to shadow tables, for performance
*** e0ne has quit IRC21:54
tsekiyamajgriffith: when multipath=True is specified, it means that nova / brick is new and can understand 'target_portals'21:54
tsekiyamajgriffith: so we can omit 'target_portal'21:55
jgriffithtsekiyama: so check out my comment in the patch21:56
jgriffithsee if that makes any sense at all21:56
tsekiyamajgriffith: but multipath=False, we should contain both of them for old and new21:56
*** markvoelker has quit IRC21:56
erlonsmcginnis:  I have a patch with 3 -1 and one 2+, what happens?21:56
jgriffithtsekiyama: I'm proposing we just always contain both21:56
jgriffithtsekiyama: so we work no matter what version/capability nova or cinder backend has21:56
erlonsmcginnis: I read somewhere that a patch can not have any -121:56
*** markvoelker has joined #openstack-cinder21:57
jgriffithtsekiyama: and use the secondary/list entries to denote mutlipath or multi-attach etc21:57
jgriffithtsekiyama: don't know if that actually makes any sense or not21:57
tsekiyamajgriffith: Ah, ok21:57
openstackgerritThang Pham proposed openstack/cinder: VMware: Add vmware:adapter_type extra spec
jgriffithtsekiyama: but it solves the ugliness I hated in the nova patch of "if this DS, else that"21:57
jgriffithtsekiyama: I'm of course assuming others think that's a good idea and works21:58
*** karimb has joined #openstack-cinder21:58
guitarzanerlon: -1 patches can get merged21:58
smcginniserlon: You can have a -1 on a patch. It's just less likely that you'll get two +2's. :)21:58
guitarzanI don't know how often it happens21:58
* jgriffith merges his -1 patches all the time ;)21:58
tsekiyamajgriffith: I believe it works if both keys always contained.21:59
jgriffithtsekiyama: so I think that is "cleaner"21:59
jgriffithtsekiyama: but I'm curious if you agree or not?21:59
guitarzanerlon: I think that -2 is not actually there21:59
guitarzaner, +221:59
guitarzanoh, maybe it is22:00
guitarzanI see the checkmark now22:00
tsekiyamajgriffith: I just tried to not affecting drivers already implementing the new 'target_portals' etc.22:00
smcginniserlon: Interesting. Not sure why those -1's didn't get cleared on new patchsets...22:00
erlonsmcginnis: I havent seen taobai, for weeks and winston-d_zZZ and avishay keeps ignoring me :)22:01
jgriffithtsekiyama: right, and that's awesome!22:01
smcginniserlon: Just looking now, but were their concerns addressed? Or not actually valid?22:01
jgriffithtsekiyama: we can do a if key exists until things catch up too22:01
guitarzanerlon: they're just asking you to write unit tests, right?22:01
jgriffithbut once we go that route maybe my idea just doesn't matter22:02
*** Yogi11 has quit IRC22:02
*** takedakn has joined #openstack-cinder22:02
erlonguitarzan: smcginnis: yes, they concern was about unit tests, I implemented then in another patch, which where merged already, BTW22:02
guitarzanerlon: how does your other unit test pass for code that isn't merged?22:03
smcginniserlon: Can you put in a comment with a link to the patch that added the unit tests?22:03
guitarzanerlon: ah, there were no tests at all for that driver22:04
erlonguitarzan: actually the code passsed with out unit tests22:04
*** rushil has quit IRC22:04
erlonguitarzan: just this class22:04
smcginniserlon: Can you put that in this review so others can see there are now tests. :)22:04
tsekiyamajgriffith: ok i'll revise the patch so it always returns both 'target_portal' and 'target_portals'.22:04
guitarzanerlon: so it seems like you should be able to rebase and add a unit test for the code you changed22:04
erlonsmcginnis: is there already22:04
jgriffithtsekiyama: let's get other input22:04
smcginniserlon: Emmm... I would use "the code passes without unit tests" as an argument.22:04
guitarzansmcginnis: haha, good luck :)22:05
jgriffithtsekiyama: maybe I'm wrong on this22:05
guitarzanI do like that line though22:05
*** e0ne has joined #openstack-cinder22:05
smcginniserlon: I'm looking through the comments right now. I don't see that link there.22:05
*** rwsu has quit IRC22:06
smcginniserlon: Oh, nm. I see it now.22:06
erlonsmcginnis: guitarzan the new code I added was just small changes, covered by the added tests, so no need to new tests in the changes22:06
guitarzanerlon: that's not possible right?22:06
guitarzanyour tests *can't* cover the small change you added22:07
guitarzansince that small change isn't merged yet22:07
* erlon thinking22:07
smcginniserlon, guitarzan: I think I see what you mean.22:07
smcginniserlon: There are tests for these specific method calls.22:08
erlonguitarzan: is a parser error, so yes22:08
smcginniserlon: I don't see tests that would specifically cause it to follow the code path to exercise your changes.22:08
guitarzanright, what smcginnis said :)22:08
guitarzanyou introduced an if check22:09
smcginniserlon: But that said, I don't think we can say that every code path in the rest of the code base has coverage either.22:09
guitarzanadd a test that follows the new path22:09
guitarzansmcginnis: true22:09
smcginnisguitarzan: But no reason to not try to get more coverage, right?22:09
guitarzancertainly not22:09
guitarzanespecially if it's easy22:09
guitarzanthen winston and avishay would have no reason at all for a -122:09
*** takedakn has quit IRC22:11
erlonguitarzan: smcginnis: not all path of code are covered anyway(is that feasible? I have jgriffith discussing about unit tests coverage somewhere), but Ill add that tests to this path too22:11
*** akerr_ has quit IRC22:12
guitarzanerlon: I don't think anyone is saying that22:12
guitarzanerlon: I'm saying that if you want to address winston and avishay's -1, you could add a test that exercises the path you added22:12
guitarzanobviously thingee thinks it's ok now22:12
*** needscoffee is now known as needsmostcoffee22:12
smcginniserlon: I think if you can just add at least one more test that would cover the single_evs = True and single_EVS = False then coverage would be pretty good IMO.22:12
jgriffitherlon: what did I do?22:13
smcginnisjgriffith: It's all your fault.22:13
erlonguitarzan: hmmm, I tried to reach wiston to get feedback22:13
*** needsmostcoffee is now known as morganfainberg22:13
erlonjgriffith: :)22:13
smcginnisjgriffith: Nothing to worry about. Just the recurring unit test coverage discussion.22:13
jgriffithsmcginnis: ahh22:14
erlonjgriffith: once you opened a bug to discuss a acceptable coverage of unit tests22:14
smcginnisjgriffith: Or the "unit testing is crap" discussion. Take your pick. ;)22:14
jgriffitherlon: oh, I think that was Winston22:14
jgriffiththe "crap" one was mine :)22:14
*** jungleboyj_ has joined #openstack-cinder22:16
openstackLaunchpad bug 1404339 in Cinder "Unit tests for zfssa driver don't actually check anything" [Undecided,In progress] - Assigned to abhiram moturi (abhiram-moturi)22:16
erlonjgriffith: see, your faultt :)22:16
*** timcl1 has left #openstack-cinder22:19
*** takedakn has joined #openstack-cinder22:19
*** Mandell has quit IRC22:20
*** erlon is now known as erlon_away22:21
*** ebalduf has joined #openstack-cinder22:21
*** vilobhmm has quit IRC22:22
*** ebalduf has quit IRC22:25
jungleboyj_mtanino: Patch looks good.  Thanks.  xyang Can you take a look as well?22:27
jungleboyj_+A if you have no concerns.22:28
*** ronis has quit IRC22:28
mtaninojungleboyj_: thank you for the review.22:28
jungleboyj_mtanino: Thank you for working up a patch for that!22:28
mtaninojungleboyj_: do we need backpot it for juno-stable?22:29
*** rwsu has joined #openstack-cinder22:30
jungleboyj_mtanino: That is a good idea.  You know how to do that?22:30
mtaninojungleboyj_: I haven't do it. are there good wiki page for it?22:30
xyangjungleboyj_: in a meeting now. which one is this?  I'll take a look later22:30
smcginnisjungleboyj_: Isn't Juno security only now?22:31
mtaninosmcginnis: oh really?22:31
smcginnismtanino: I thought so.22:31
jungleboyj_smcginnis: No, I said that earlier, but that was a mistake.22:32
smcginnismtanino: I think we had a bug in one of our drivers that we wanted to backport but it was rejected because it wasn't a security issue.22:32
*** vilobhmm has joined #openstack-cinder22:32
*** vilobhmm has quit IRC22:32
jungleboyj_We shouldn't be shoving everything back, but this would be good.22:32
smcginnisjungleboyj_: I heard it from someone else I thought.22:32
jungleboyj_smcginnis: Yeah, that was my bad.  I corrected that.22:32
*** vilobhmm has joined #openstack-cinder22:33
jungleboyj_mtanino: Basically all you need to do is a 'git cherrypick -X <commit hash> into stable/juno'22:33
smcginnisjungleboyj_: Oh, I see your last comment now.22:34
jungleboyj_xyang: That is the change to deal with the migration_status if migration fails.22:34
mtaninojungleboyj_: I see. let me try after the patch will be merged.22:34
jungleboyj_xyang: Avishay and I have +2 on it, want someone more detached from the issue to look as well.22:34
smcginnisthingee: You're not online now, but if you see this can you take a look at
jungleboyj_mtanino: +2  I can help you through it if there are issues.22:34
xyangjungleboyj_: a link will help:)22:35
mtaninoxyang: sure.
*** takedakn has quit IRC22:36
*** takedakn has joined #openstack-cinder22:36
xyangmtanino: ok, thanks.22:36
jungleboyj_xyang: Sorry, thought I sent that.  :-p22:38
*** Guest78621 has quit IRC22:39
jgriffithmtanino: there maybe be an updated version of that around somewhere but that's what I use22:39
*** thangp has quit IRC22:40
mtaninojgriffith: oh, nice! script for making a backport patch22:40
*** takedakn has quit IRC22:43
*** Mandell has joined #openstack-cinder22:44
*** ganso_ has quit IRC22:47
*** annashen has quit IRC22:54
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder: Add volume multi attach support
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder: Add multiattach capabilities injection
openstackgerritJay Bryant proposed openstack/cinder: Use oslo.middleware
*** cnesa has joined #openstack-cinder23:02
*** Apoorva_ has joined #openstack-cinder23:03
*** primechuck has quit IRC23:03
*** vilobhmm has quit IRC23:04
*** vilobhmm has joined #openstack-cinder23:05
*** vilobhmm has quit IRC23:05
*** vilobhmm has joined #openstack-cinder23:06
*** Apoorva has quit IRC23:07
*** primechuck has joined #openstack-cinder23:09
*** takedakn has joined #openstack-cinder23:10
*** jdurgin has quit IRC23:11
*** kaufer has quit IRC23:20
*** diemt has quit IRC23:21
*** ebalduf has joined #openstack-cinder23:21
*** Ilja has quit IRC23:24
*** jdurgin has joined #openstack-cinder23:25
*** ebalduf has quit IRC23:26
jungleboyj_smcginnis: You still around?23:30
*** mriedem has quit IRC23:36
*** jungleboyj_ has quit IRC23:47
*** EmilienM is now known as EmilienM|afk23:49
*** Mandell has quit IRC23:53
*** jungleboyj_ has joined #openstack-cinder23:53
*** chlong has joined #openstack-cinder23:53
*** asmith_brcd has quit IRC23:55
*** jdurgin has quit IRC23:55
*** jdurgin has joined #openstack-cinder23:55
*** IanGovett has joined #openstack-cinder23:58

Generated by 2.14.0 by Marius Gedminas - find it at!