Monday, 2015-04-20

*** lpetrut has quit IRC00:08
*** primechuck has quit IRC00:31
*** heyun has joined #openstack-cinder00:52
*** zhenguo has joined #openstack-cinder00:54
*** primechuck has joined #openstack-cinder01:24
*** mriedem has quit IRC01:26
*** jungleboyj has joined #openstack-cinder01:31
*** Yogi1 has joined #openstack-cinder01:34
*** julim has quit IRC01:47
*** julim has joined #openstack-cinder01:48
*** Lee1092 has joined #openstack-cinder01:48
*** julim has quit IRC01:48
*** vilobhmm1 has joined #openstack-cinder01:56
*** vilobhmm1 has quit IRC02:13
*** vilobhmm1 has joined #openstack-cinder02:14
*** vilobhmm1 has quit IRC02:14
*** vilobhmm1 has joined #openstack-cinder02:14
*** annashen has joined #openstack-cinder02:16
openstackgerritJeegn Chen proposed openstack/cinder: Make VNX Cinder Driver aware of VNX Pool Full Threshold
*** annashen has quit IRC02:38
*** davechen has joined #openstack-cinder02:38
*** kfox1111 has quit IRC02:39
*** vilobhmm1 has quit IRC02:39
*** davechen1 has joined #openstack-cinder02:44
*** davechen2 has joined #openstack-cinder02:46
*** davechen has quit IRC02:46
*** davechen1 has quit IRC02:48
*** Yogi1 has quit IRC02:49
openstackgerritQiangGuan proposed openstack/cinder: Fix NoneType AttributeError problem
*** david-lyle has quit IRC02:55
*** annashen has joined #openstack-cinder02:57
*** annashen has quit IRC02:57
*** jamielennox is now known as jamielennox|away02:58
*** jamielennox|away is now known as jamielennox03:03
*** annashen has joined #openstack-cinder03:15
*** annashen has quit IRC03:20
*** annashen has joined #openstack-cinder03:21
*** annashen has quit IRC03:22
*** annashen has joined #openstack-cinder03:23
*** annashen has quit IRC03:27
*** coolsvap|afk is now known as coolsvap03:39
*** vilobhmm1 has joined #openstack-cinder03:54
*** ishant has joined #openstack-cinder04:04
*** rushiagr_away is now known as rushiagr04:20
*** annashen has joined #openstack-cinder04:31
*** aswadr has joined #openstack-cinder04:32
*** Longgeek has joined #openstack-cinder04:35
*** Longgeek has quit IRC04:35
*** Longgeek has joined #openstack-cinder04:36
*** BharatK has joined #openstack-cinder04:43
*** sks has joined #openstack-cinder04:48
*** tshefi has joined #openstack-cinder04:48
*** btran has joined #openstack-cinder04:50
*** annashen has quit IRC04:50
*** annashen has joined #openstack-cinder04:51
*** annashen has quit IRC04:56
*** boris-42 has quit IRC04:58
*** annashen has joined #openstack-cinder04:59
*** sgotliv has joined #openstack-cinder05:05
*** deepakcs has joined #openstack-cinder05:10
*** ishant has quit IRC05:13
*** nlevinki has joined #openstack-cinder05:14
*** nkrinner has joined #openstack-cinder05:23
*** ishant has joined #openstack-cinder05:38
*** annashen has quit IRC06:01
*** annashen has joined #openstack-cinder06:03
*** ankit_ag has joined #openstack-cinder06:04
*** afazekas_ has joined #openstack-cinder06:10
*** annashen has quit IRC06:12
*** anshul has joined #openstack-cinder06:14
*** annashen has joined #openstack-cinder06:14
*** annashen has quit IRC06:16
*** annashen has joined #openstack-cinder06:16
*** annashen has quit IRC06:20
*** anshul has quit IRC06:27
*** sgotliv has quit IRC06:28
*** annashen has joined #openstack-cinder06:39
*** anshul has joined #openstack-cinder06:40
*** lpetrut has joined #openstack-cinder06:44
*** anshul has quit IRC06:50
*** jamielennox is now known as jamielennox|away06:52
*** annashen has quit IRC07:02
*** anshul has joined #openstack-cinder07:04
*** ociuhandu has joined #openstack-cinder07:27
*** chlong has quit IRC07:29
*** e0ne has joined #openstack-cinder07:31
*** bkopilov has quit IRC07:32
*** bkopilov has joined #openstack-cinder07:32
*** e0ne has quit IRC07:34
*** e0ne has joined #openstack-cinder07:38
*** vilobhmm1 has quit IRC07:39
*** e0ne has quit IRC07:40
openstackgerritwanghao proposed openstack/cinder-specs: Support query volume filter by glance metadata
*** kaisers has joined #openstack-cinder07:42
openstackgerritwanghao proposed openstack/cinder-specs: Support query volume filter by glance metadata
*** ociuhandu has quit IRC07:45
*** pschaef has joined #openstack-cinder07:45
*** e0ne has joined #openstack-cinder07:45
*** wanghao has joined #openstack-cinder07:46
*** jistr has joined #openstack-cinder07:47
wanghaohi~ guys, there is a bug in nova side:  But it may need some suggestion came from cinder.  Is anyone can help me with this?07:50
*** sgotliv has joined #openstack-cinder07:51
wanghaoNova can't detach volume in init host routine07:51
wanghao1. create a vm using bootable volume.07:51
wanghao2. delete this vm07:51
wanghao3. restart service nova-compute when vm's task state is deleting.07:51
wanghaoWhen nova-compute is up, vm became deleted successful, but the bootable07:51
wanghaovolume is still in-use state and can't delete it using cinder delete volume.07:51
wanghaosolve method:07:51
wanghaoAdd init=True in _delete_instance when init_host, and raise exception07:51
wanghaowhen EndpointNotFound exists. It will set vm's status to error and07:51
wanghaomake user can re-issue a delete.07:51
wanghaoSome detail about this bug.07:52
*** annashen has joined #openstack-cinder08:03
*** jordanP has joined #openstack-cinder08:04
*** e0ne has quit IRC08:04
*** e0ne has joined #openstack-cinder08:08
*** annashen has quit IRC08:08
*** e0ne is now known as e0ne_08:09
*** e0ne_ is now known as e0ne08:11
*** e0ne has quit IRC08:17
openstackgerritTina Tang proposed openstack/cinder: Create consistgroup from cgsnapshot support in VNX driver
*** ndipanov has joined #openstack-cinder08:28
deepakcswanghao, so why can't nova init_host re-issue a delete if the instance state is 'deleting' instead of raising excp as you suggest ?08:30
*** pschaef has quit IRC08:30
*** pschaef has joined #openstack-cinder08:31
deepakcswanghao, or mark instance as 'error' during init_host itself if state is 'deleting'. When do we get EndpointNotFound excp ?08:31
*** anshul has quit IRC08:32
*** anshul has joined #openstack-cinder08:46
*** Yogi1 has joined #openstack-cinder08:50
*** alecv has joined #openstack-cinder08:50
*** chlong has joined #openstack-cinder08:50
*** kbyrne has joined #openstack-cinder08:54
*** Yogi1 has quit IRC08:54
*** e0ne has joined #openstack-cinder08:59
*** annashen has joined #openstack-cinder09:04
*** annashen has quit IRC09:09
wanghao@deepakcs: Before this patch, When nova get EndpointNotFound excp in init_host routine, it delete vm successful in nova side and just log a message about this exception. But in cinder side, this bootable volume still is in-use and can't detach it by user.09:13
*** e0ne is now known as e0ne_09:13
*** e0ne_ is now known as e0ne09:13
deepakcswanghao, whats the VM state after u restart n-cpu ?  I am trying to understand the problem first, not clear to me09:15
openstackgerritBharat Kumar Kobagana proposed openstack/cinder: GlusterFS: Support over subscription in thin provisioning
*** alexpilotti has joined #openstack-cinder09:17
wanghaoBefore restart n-cpu, VM state is deleting, After restart, it is deleted.  But when init_host in restart process, nova use context.get_admin_context to obtain context using it to send message to cinder. The context don't have auth_token, so when call terminate_connection to cinder , it will fail.09:19
*** chlong has quit IRC09:20
wanghaoBut in nova side, it just log a message about this exception, and go on with deleting vm successful.09:20
wanghaoSo, that's why VM is deleted after restart n-cpu.09:21
*** kaisers has quit IRC09:22
*** pcaruana has quit IRC09:27
*** ishant has quit IRC09:28
*** nlevinki_ has joined #openstack-cinder09:29
*** zerda has joined #openstack-cinder09:29
*** pcaruana has joined #openstack-cinder09:30
*** nlevinki has quit IRC09:33
*** chlong has joined #openstack-cinder09:33
*** lpetrut has quit IRC09:33
openstackgerritYuriy Nesenenko proposed openstack/cinder-specs: Checking the existence of volume
openstackgerritAnton Arefiev proposed openstack/cinder: Add missing backups entry to default quota class
*** primechuck has quit IRC09:47
*** davechen2 has left #openstack-cinder09:50
*** coolsvap is now known as coolsvap|afk09:51
*** boris-42 has joined #openstack-cinder09:51
openstackgerritwanghao proposed openstack/cinder: Implement function of import/export snapshots
*** aix has joined #openstack-cinder10:01
*** haomaiwa_ has quit IRC10:03
*** haomaiwa_ has joined #openstack-cinder10:04
*** kaisers has joined #openstack-cinder10:05
*** haomaiwa_ has quit IRC10:09
*** kaisers has quit IRC10:14
*** lpetrut has joined #openstack-cinder10:16
*** kaisers has joined #openstack-cinder10:18
*** haomaiwang has joined #openstack-cinder10:18
*** heyun has quit IRC10:18
*** nlevinki_ has quit IRC10:25
*** nlevinki has joined #openstack-cinder10:26
*** IanGovett has joined #openstack-cinder10:28
*** kaisers has quit IRC10:34
*** rushiagr is now known as rushiagr_away10:38
*** anshul has quit IRC10:42
*** zerda has quit IRC10:53
*** anshul has joined #openstack-cinder10:55
*** e0ne is now known as e0ne_10:56
*** e0ne_ is now known as e0ne10:58
*** sgotliv_ has joined #openstack-cinder11:05
*** sgotliv has quit IRC11:05
*** anshul has quit IRC11:14
*** e0ne is now known as e0ne_11:16
*** astellwa1 has joined #openstack-cinder11:23
*** astellwa1 has quit IRC11:23
*** afazekas_ has quit IRC11:24
*** e0ne_ has quit IRC11:26
*** TobiasE has joined #openstack-cinder11:26
*** anshul has joined #openstack-cinder11:27
*** nlevinki_ has joined #openstack-cinder11:28
*** afazekas_ has joined #openstack-cinder11:28
*** nlevinki has quit IRC11:30
*** akerr has joined #openstack-cinder11:51
*** deepakcs has quit IRC11:51
*** jistr has quit IRC11:57
*** chlong has quit IRC11:58
*** afazekas_ has quit IRC11:58
*** kaisers has joined #openstack-cinder11:59
*** haomaiwang has quit IRC12:01
*** jistr has joined #openstack-cinder12:03
*** zhenguo has quit IRC12:07
*** bswartz has quit IRC12:08
*** kaisers has quit IRC12:10
*** nlevinki_ has quit IRC12:10
*** ronis has joined #openstack-cinder12:11
*** lpetrut has quit IRC12:13
*** lpetrut has joined #openstack-cinder12:15
*** nlevinki_ has joined #openstack-cinder12:23
*** e0ne has joined #openstack-cinder12:26
*** timcl has joined #openstack-cinder12:30
*** jamielennox|away is now known as jamielennox12:31
*** jamielennox is now known as jamielennox|away12:32
*** jaypipes has joined #openstack-cinder12:33
*** erlon has joined #openstack-cinder12:34
e0neDuncanT: hi! should i add A/A HA issue description to the etherpad only or file a bug too?12:36
*** BharatK has quit IRC12:36
DuncanTe0ne: The etherpad for now, please. It'll get turned to bugs once there's some agreement that we've got a good set12:42
e0neDuncanT: ok, thanks12:43
aarefievDuncanT: hi, have a min?12:45
DuncanTaarefiev: Sure12:46
aarefievDuncanT: have a question about options for quotas12:46
DuncanTaarefiev: shoot12:46
aarefievDuncanT: I don't see where it is deprecated12:47
DuncanTgive me a minute12:47
aarefievDuncanT: or you mean it will i L12:47
aarefiev*in L12:47
DuncanTI thought it had been marked depricated some time ago, give me a minute12:47
DuncanTHmmm, there was a comment about it, which has since been lost, and it looks like the CFG options were never marked depricated :-(12:51
DuncanTOk, I'll remove the -1 and knock up a patch to mark them as depricated12:52
DuncanTWe meant to move to setting all the quotas in the db ages ago, looks like that work never got finished :-(12:52
*** esker has quit IRC12:52
aarefievDuncanT: ok thaks, where do you propose to store default values for quotas?12:53
DuncanTThe initial defaults I'm not sure about TBH12:54
DuncanTIf there's no default in the db then coping them in from the conf file, like you did, is the right thing to do12:54
DuncanTI'm just not sure what those migrations should do once the config files have been removed12:55
DuncanTProbably hard-code the values in the migration?12:55
e0neDuncanT: +1, hardcode current values from config12:55
DuncanTIt'll be a job for whomever does the deprications now anyway12:55
*** bswartz has joined #openstack-cinder12:55
DuncanTe0ne: Thanks12:55
e0neand add this to the release notes12:55
DuncanTEven if we depricate them now, we can't remove them until M, right?12:56
*** mwichmann has joined #openstack-cinder12:56
aarefievand we have a few migration with this options12:57
*** mwichmann has left #openstack-cinder12:57
aarefiev21 and 2612:58
smcginnisDuncanT: Still free? Got time to give your opinion on something?13:00
*** Yogi1 has joined #openstack-cinder13:00
*** pschaef has quit IRC13:00
*** sks has quit IRC13:02
*** lifeless has quit IRC13:05
*** annashen has joined #openstack-cinder13:08
*** annashen has quit IRC13:13
*** nlevinki_ has quit IRC13:14
*** sks has joined #openstack-cinder13:16
jgriffithsmcginnis: Not sure what's up wit those gate failures yet.  The one I have a patch up for, the others I can't reproduce13:19
jgriffithsmcginnis: aha!  Vincent figured it out (at least one of them indirectly)13:24
jgriffithsmcginnis: so the added reference get from the db13:24
smcginnisjgriffith: Just saw his comment.13:26
*** zerda has joined #openstack-cinder13:26
smcginnisjgriffith: So between your patch and that, looks like that covers it?13:26
jgriffithsmcginnis: maybe13:26
jgriffithsmcginnis: the second has a bunch of failures I can't reproduce13:27
smcginnisjgriffith: I'll keep my fingers crossed. ;)13:27
jgriffithsmcginnis: still trying to figure that one out13:27
jgriffithsmcginnis: :)13:27
*** bkopilov has quit IRC13:27
smcginnisjgriffith: Yeah, nothing directly obvious in any of this to me.13:27
smcginnisjgriffith: Glad you were able to find something at least.13:27
jgriffithsmcginnis: I can't take credit, it was Vincent who put the two together13:27
smcginnisjgriffith: He's not here right now so you can take all the credit. :D13:28
e0nejgriffith: morning13:29
*** crose has joined #openstack-cinder13:29
jgriffithsmcginnis: LOL13:29
jgriffithe0ne: hey there13:29
e0nejgriffith: i didn't put -1 for but hope my question won't be ignored13:29
jgriffithe0ne: I could probably change that13:30
smcginnise0ne: That is a good point.13:30
jgriffithe0ne: but...13:30
jgriffithe0ne: we're kinda stupid in what we're doing with those custom built dicts13:30
jgriffithe0ne: I'm of the opinion that if we pull a ref from the db in one layer we should pass the whole thing, not just the pieces we 'think' are important13:31
*** superdan is now known as dansmith13:31
jgriffithe0ne: I can certainly remove those13:31
e0nejgriffith: sounds reasonable to pass original objects throw all flows13:31
jgriffithe0ne: it's a valid concern... just use the "resource={'type': 'volume', 'id': volume_id} format13:31
jgriffithe0ne: yeah, but that should probably be follow up13:32
jgriffithe0ne: I'll respin those patches and remove the db get calls if that makes everybody "happier"13:32
jgriffithe0ne: frankly though I think we have some mistakes by not having the full reference13:33
e0nejgriffith: ok. i'm vote +++ and many + for unifiing logging and make it better. but performance should be goog too13:33
jgriffithe0ne: :)13:33
jgriffithe0ne: ok, thanks for the feedback13:33
e0nejgriffith: thank you for understanging13:33
* e0ne doesn't want to be a nerdy in reviews13:34
*** bkopilov has joined #openstack-cinder13:35
DuncanTsmcginnis: Sorry, was at standup. What do you want me to look at?13:35
*** annegentle has joined #openstack-cinder13:35
*** jistr_ has joined #openstack-cinder13:35
e0nejgriffith: just 2c more: imo, moving to objects will help us to pass the same dicts over the code13:37
jgriffithe0ne: agreed13:37
jgriffithe0ne: trouble is the cinder.api.xxxx code is kinda of a mess13:37
jgriffithe0ne: that's where it's broken in most places13:38
jgriffithe0ne: that is the next section I want to tackle for logging, it's rather ugly13:38
DuncanTjgriffith: I'm a big fan of the logging standardisation, stuff, I looked at the review, couldn't see anything fundamental wrong but your spelling hurt13:38
*** jistr has quit IRC13:38
jgriffithDuncanT: haha!13:39
*** jungleboyj has quit IRC13:39
jgriffithDuncanT: so I did a bash search and replace at Mitsuhiro's point that I screwed up the spelling on successfull13:39
e0nethere are 2 areas of improvement that will be forever: logging and testing (unit, functional, tempest, etc)13:40
jgriffithDuncanT: but the regex was wrong and put an extra 's' inbetween13:40
jgriffithI didn't even look13:40
jgriffithautomation is awesome!!!13:40
*** esker has joined #openstack-cinder13:40
jgriffithGreat way to make the same mistake over and over and over and over13:40
*** russellb has quit IRC13:40
*** bkopilov has quit IRC13:40
e0ne - sudo make me a sandwich13:41
e0neit's about automation:)13:41
*** julim has joined #openstack-cinder13:41
*** sigmavirus24_awa is now known as sigmavirus2413:42
*** ankit_ag has quit IRC13:42
*** sigmavirus24 has left #openstack-cinder13:42
*** bkopilov has joined #openstack-cinder13:42
jgriffithDuncanT: LOL!13:42
*** jistr_ has quit IRC13:44
*** jistr has joined #openstack-cinder13:44
*** zhenguo has joined #openstack-cinder13:44
smcginnisDuncanT: Sorry, now I was off getting coffee.13:45
*** russellb has joined #openstack-cinder13:45
smcginnisDuncanT: Was just wondering - I remembered you had agreed in a discussion earlier13:45
smcginnisDuncanT: about not using exact strings in unit tests.13:46
smcginnisDuncanT: So the question is the correct way to verify logging.13:46
*** mriedem has joined #openstack-cinder13:46
DuncanTsmcginnis: That was specifically for log strings, I believe, but the principle can apply in many places13:46
smcginnisDuncanT: Wondering if you agree with my response on line 277 here:
*** lpetrut has quit IRC13:47
DuncanTsmcginnis: I'll reply in the review, but I agree with you13:48
*** lpetrut has joined #openstack-cinder13:48
*** bkopilov has quit IRC13:49
smcginnisDuncanT: Cool, just wanted to get another opinion. Thanks!13:49
*** eharney has joined #openstack-cinder13:50
DuncanTReplied inline13:51
*** timcl has quit IRC13:52
*** e0ne is now known as e0ne_13:53
smcginnisDuncanT: Thanks!13:57
*** kaisers has joined #openstack-cinder13:58
*** rushil has joined #openstack-cinder13:59
*** timcl has joined #openstack-cinder13:59
*** Yogi11 has joined #openstack-cinder13:59
*** Yogi1 has quit IRC14:02
*** e0ne_ is now known as e0ne14:02
jgriffithe0ne: ok, so there was only "one" call that introduced a new get to the db14:04
jgriffithe0ne: I modified that one14:04
jgriffithe0ne: the rest all just moved the call up the chain14:04
jgriffithe0ne: cleaning those up running some tests will resubmit in a few14:04
*** bkopilov has joined #openstack-cinder14:04
jgriffithsmcginnis: I put the _LE back on that one14:04
jgriffithsmcginnis: should probably make sure it's the one you were referring to14:04
e0nejgriffith: thanks!14:05
jgriffithtbarron: frankly if I miss some items in replication and migration that's life14:05
smcginnisjgriffith: Which line was that. I can take a look.14:05
jgriffithtbarron: maybe you could create a patch to fix them up :)14:05
tbarronjgriffith: will do, I'll swithc my vote now.  ty.14:05
jgriffithtbarron: might as well wait til I put the new set up :)14:06
tbarronagain :-)  I'm a flip, flopper.14:06
tbarronjgriffith: kk, I will look for it.14:06
jgriffithsmcginnis: the Detach volume call around line 85714:06
jgriffithsmcginnis: was "msg = _("Volume %(id)s is attached to more than one instance"14:07
smcginnisjgriffith: Did I say to change that?14:08
smcginnisLooks like that one is used in the logger and the exception, so _() would be correct.14:08
smcginnisNot _LE().14:08
*** mtanino has joined #openstack-cinder14:08
smcginnisjgriffith: Sorry if I had said otherwise. :\14:08
*** sks has quit IRC14:08
* smcginnis looks back through comments14:09
jgriffithsmcginnis: you made a comment that Exception and Error are the same translation14:10
jgriffithsmcginnis: comment at 6:2114:10
smcginnisjgriffith: Oh, I think maybe I didn't state that right.14:10
smcginnisjgriffith: LOG.error and LOG.exception are the same.14:10
smcginnisjgriffith: But translation is different between LOG.[error|exception] and Exception(msg)14:11
smcginnisjgriffith: That make any sense?14:11
jgriffithsmcginnis: totally14:11
tbarronsmcginnis: jgriffith: so I think that implies there's no point in doing LOG.error(...) followed immediately by LOG.exception(ex)14:11
jgriffithsmcginnis: but I don't know what youre comment referred to then :)14:11
smcginnistbarron: Exactly!14:11
jgriffithtbarron: ah14:11
jgriffithtbarron: well FYI, I don't think I *added* those14:11
smcginnisjgriffith: Yeah, sorry. That comment was supposed to be a reply to tbarron.14:12
jgriffithguess I can try and scrub those out14:12
tbarronjgriffith: yes, you didn't add and we can change them in another patch.14:12
tbarronjgriffith: but I'd really love to have a file that does it all right that I can look at.14:12
smcginnistbarron: I agree.14:12
tbarronI learn by example, not by reading the logging rules.14:12
smcginnistbarron: I've been working on a series of patches to clean all this up.14:12
jgriffithsmcginnis: tbarron I think the difference was putting the exception in the log14:13
smcginnistbarron: Hoping so they do become good examples.14:13
jgriffithsmcginnis: tbarron the error just has an "error" message, the ex is logged in the log.exception only14:13
tbarronsmcginnis: that would be awesome.  I can't believe the amount of energy this project diverts to getting logging right.14:13
*** nkrinner has quit IRC14:13
jgriffithtbarron: that and spelling in comments :)14:14
tbarronI don't in any way mean to say that getting it right is not terribly important.14:14
*** nkrinner has joined #openstack-cinder14:14
jgriffithwe're KING!!!14:14
smcginnistbarron: I know. And I feel kind of bad having all these patches for something seemly so trivial, but I think it will be good in the end. I hope.14:14
tbarronbut I just want to have an easier way to do it.14:14
tbarronsmcginnis: +114:14
jgriffithsmcginnis: so what's your preference?  Fix it up now, or leave it for follow up?14:14
*** nkrinner has quit IRC14:14
*** mdenny has joined #openstack-cinder14:14
smcginnisjgriffith: I'm fine leaving it.14:14
smcginnisjgriffith: Ankit has a patch that is addressing some of them under cinder/volume.14:15
smcginnisjgriffith: And I am working my way through with hacking check support.14:15
*** nkrinner has joined #openstack-cinder14:15
smcginnisjgriffith: I've been leaving cinder/volume for last both because Ankit is addressing most and because it's going to be a huge PITA. :)14:15
DuncanTThe resource bits of this patch are extremely useful... I spent some time trying to write better logstash filters for our log messages, and it was a nightmare, much easier now14:16
smcginnistbarron: If you care to look:
openstackgerritJohn Griffith proposed openstack/cinder: Add resource tag to logging in
smcginnisjgriffith: /sould/should/14:18
*** afazekas_ has joined #openstack-cinder14:18
openstackgerritJohn Griffith proposed openstack/cinder: Add resource tag to logging in
tbarronsmcginnis: wow, you are going to be the logging go-to guy!  Though, to your point, if we have most of the "old" code fixed, there will be less confusion and less need to go-to.14:19
tbarronsmcginnis: so thank you.14:19
smcginnistbarron: That's what I'm hoping.14:19
smcginnisThere just seemed to be a lot of cases in code reviews where something was copy/pasted using wrong things.14:20
*** wormwood has quit IRC14:21
jgriffithsmcginnis: +114:22
e0nesmcginnis: +114:23
*** zerda has quit IRC14:23
*** jungleboyj has joined #openstack-cinder14:24
*** Tross1 has joined #openstack-cinder14:25
*** Tross has quit IRC14:25
tbarronDuncanT: just followed your spell-checker link.  Hadn't seen that before, and enjoyed it.14:26
jgriffithif anybody has a few cycles...14:26
smcginnisjgriffith, e0ne: Thanks! Glad to know folks aren't shaking their heads thinking what the hell is that guy think he's doing. :)14:26
jgriffithit woudl be AWESOME if you could download: 17505414:26
jgriffithand see if you can get it to fail14:26
jgriffithnever mind14:26
jgriffithI got it :)14:26
*** afazekas_ has quit IRC14:26
smcginnisjgriffith: Eureka moment?14:27
jgriffithsmcginnis: yeah :)14:27
*** tshefi has quit IRC14:28
*** emagana has joined #openstack-cinder14:28
*** afazekas_ has joined #openstack-cinder14:30
*** lifeless has joined #openstack-cinder14:32
*** afazekas_ has quit IRC14:36
jgriffithsmcginnis: nahh... the cleanup stuff is a PITA but it's godo14:37
smcginnisjgriffith: The volume directory is going to be painful. Especially if it takes a while to get through. I forsee a lot of rebasing.14:38
jgriffithsmcginnis: yeah, not sure how I'm going to tackle that one yet14:39
jgriffithsmcginnis: I'd love to see each Vendor do their driver on their own :)14:39
smcginnisjgriffith: That would be ideal. Not sure that's going to happen though. :)14:40
tbarronjgriffith: in, cinder/volume/, why are you using the weaker (volume_id) resource tag?  You got the volume itself at line 841.14:41
jgriffithtbarron: cuz IRDumb14:43
jgriffithtbarron: oversight14:43
tbarronjgriffith: kk (agreeing with the 2nd proposition, not the first)14:44
tbarronThat's all I need for the +114:44
*** emagana has quit IRC14:44
*** emagana has joined #openstack-cinder14:45
*** afazekas_ has joined #openstack-cinder14:47
*** russellb has quit IRC14:48
*** geguileo has quit IRC14:49
*** asselin_ has joined #openstack-cinder14:50
*** hemnafk is now known as hemna14:50
ameadehappy monday hemna14:51
hemnaw00t! Mon..day14:51
hemnaok Coffee helps14:51
*** geguileo has joined #openstack-cinder14:51
ameadeheh i am working from home this AM just so I can have a whole pot next to me14:52
*** russellb has joined #openstack-cinder14:53
jungleboyjhemna: Another Manic Monday.  :-)14:54
* jungleboyj has his thermos of coffee next to him.14:54
openstackgerritJohn Griffith proposed openstack/cinder: Add resource tag to logging in
jgriffithtbarron: ^^  and that's the last one!! :)14:54
smcginnisYou guys are way too chipper for a Monday morning. :)14:54
* e0ne does'nt undertand what is goint on. it's almost 6pm :)14:55
*** anshul has quit IRC14:56
jungleboyje0ne: :-)14:56
jungleboyje0ne: Did you have this conversation 8 hours ago?14:56
e0nejungleboyj: i did:)14:57
*** timcl has quit IRC14:57
*** annegentle has quit IRC14:57
*** strictlyb has quit IRC14:58
*** tsekiyama has joined #openstack-cinder14:58
*** crose has quit IRC14:59
openstackgerritJohn Griffith proposed openstack/cinder: Standardize logging in
*** annegentle has joined #openstack-cinder14:59
*** timcl has joined #openstack-cinder15:01
*** BharatK has joined #openstack-cinder15:01
*** rwsu has joined #openstack-cinder15:02
openstackgerritWalter A. Boring IV (hemna) proposed openstack/os-brick: Brick: Fix race in removing iSCSI device
*** annegentle has quit IRC15:04
*** annegentle has joined #openstack-cinder15:05
*** annegentle has quit IRC15:06
*** afazekas_ has quit IRC15:08
*** annegentle has joined #openstack-cinder15:10
*** annashen has joined #openstack-cinder15:10
openstackgerritJohn Griffith proposed openstack/python-cinderclient: Add ability to specify path var to testr
*** beekneemech is now known as bnemec15:10
*** e0ne is now known as e0ne_15:10
*** nkrinner has quit IRC15:12
*** jistr has quit IRC15:14
*** annashen has quit IRC15:15
*** strictlyb has joined #openstack-cinder15:16
*** e0ne_ is now known as e0ne15:17
SwansonDid the gerrit email toilet unclog over the weekend?  Seriously out of order messages in my inbox...15:17
*** annegentle has quit IRC15:18
*** afazekas_ has joined #openstack-cinder15:21
*** Yogi1 has joined #openstack-cinder15:23
*** flip214 has quit IRC15:24
*** harlowja_at_home has joined #openstack-cinder15:24
*** afazekas_ has quit IRC15:26
*** Yogi11 has quit IRC15:26
*** sgotliv_ has quit IRC15:27
*** jistr has joined #openstack-cinder15:28
*** Yogi1 has quit IRC15:28
*** Yogi1 has joined #openstack-cinder15:29
smcginnisSwanson: I had emails come in this morning from last Wednesday. Not sure if it's gerrit or something messed up with our mail server.15:29
smcginnisAnyone else see issues like that/15:29
thingeehemna: I wish my contributions to cinder could just be making coffee for everyone and approving changes here and there.15:30
thingeecoffee is important stuff15:30
*** lpabon has joined #openstack-cinder15:31
*** jdurgin1 has joined #openstack-cinder15:31
jungleboyjthingee: +115:31
jungleboyjsmcginnis: I didn't see anything like that.15:32
Swansonsmcginnis: I was seriously confused until I had about 1/2 a cup of coffee and then clarity descended15:32
smcginnisjungleboyj: Yeah, probably just us. Funky email scanner or something.15:32
smcginnisMight just have to switch over to gmail.15:32
jungleboyjsmcginnis: +215:33
SwansonEventually our firewall will be a physical wall crushing the incoming fibre.15:33
*** Yogi1 has quit IRC15:33
hemnathingee, +1 :)15:37
*** Yogi1 has joined #openstack-cinder15:37
*** afazekas_ has joined #openstack-cinder15:37
jungleboyjSwanson: Ahh, physical barrier, huh?  What are you, a government lab?15:38
Swansonjungleboyj: The great firewall of austin.15:39
thingeesmcginnis: you mind bringing me up to speed on the gate failures?15:39
*** dannywilson has joined #openstack-cinder15:39
Swansonno such option: config_file.  Joy.  Love debugging stacking failures...15:41
*** annashen has joined #openstack-cinder15:41
smcginnisthingee: I think jgriffith has a good understanding of what's going on. And a fix if I understood right.15:42
smcginnisthingee: Something was not getting faked correctly IIRC.15:42
*** dannywilson has quit IRC15:42
openstackgerritAlex Meade proposed openstack/cinder: NetApp E-Series: Fix instance live-migration with attached volumes
smcginnisthingee: That in a combination of some setup not being done in the right order was causing some py27 test failures.15:43
*** dannywilson has joined #openstack-cinder15:43
thingeesmcginnis: oh, yeah I already started approving that change15:43
thingeethanks jgriffith per the usual15:43
smcginnisthingee: Yeah, that was part of it.15:43
thingeewe have jgriffith who is captain gate and jungleboyj who is captain oslo15:43
smcginnisthingee: Hah! True.15:44
thingeeI only add additional exception catches to turn off gate issues :)15:44
openstackgerritJohn Griffith proposed openstack/python-cinderclient: Remove print statement in unit test
smcginnisIf you ignore something long enough it eventually goes away. ;)15:45
jungleboyj:-)  Keep this up and we will need to have Sargents, Admirals and Generals15:45
smcginnisjungleboyj: You'll need to be promoted at some point.15:46
hemnaphew, that's pretty innocuous15:46
hemnagood thing you don't see my debuging print statements.15:46
jungleboyjsmcginnis: Hmmm, who will promote me or can I self promote?15:46
hemnaLOG.debug("ASS = %s", connector)15:46
smcginnisjungleboyj: Hmm, wonder where this came from:
smcginnisjungleboyj: I think after a few beers at the summit you can promote yourself. :)15:47
*** afazekas_ has quit IRC15:47
jungleboyjsmcginnis: WTF, how did that ever get in!?!15:47
thingeejgriffith, jungleboyj: fastest approval evar15:48
smcginnisjungleboyj: Darn reviewers not doing their jobs.15:48
thingeejgriffith, jungleboyj: my version discovery patch actually removes that15:48
thingeeguess that's going to need to be rebased now15:48
jgriffiththingee: no it won't15:49
jgriffiththingee: it'll just merge fine15:49
jgriffiththingee: master changed to match you, should be fine I think15:49
thingeejgriffith: Oh I didn't mean just by that change...just I haven't checked on it in a while. Looks like it's just failing jenkins from a another gate issue15:49
openstackgerritMike Perez proposed openstack/python-cinderclient: Enable version discovery
jgriffiththingee: oh... hehe15:50
thingeewould be good to get this merged soon so we can gate on it15:50
thingeewould love to get this code removed soon :(15:50
thingeealso if anyone has refactoring suggestions ^15:51
*** dannywilson has quit IRC15:51
*** ronis has quit IRC15:52
*** mwichmann has joined #openstack-cinder15:52
*** mwichmann has left #openstack-cinder15:52
*** jistr_ has joined #openstack-cinder15:53
*** russellb has quit IRC15:53
*** pradipta has joined #openstack-cinder15:54
*** jistr has quit IRC15:56
kmartin_smcginnis, jungleboyj, jgriffith, rhe00: hemna posted a draft presentation and linked it to the etherpad...
*** _cjones_ has joined #openstack-cinder15:57
smcginniskmartin_: Yeah, looks pretty good for a start.15:57
kmartin_can you guys tack a look and plan on contributing your vendor slide15:58
*** russellb has joined #openstack-cinder15:58
smcginnisSo each of us just need to add a slide for our own storage piece.15:58
smcginniskmartin_: I'm hoping to get ours this week.15:58
kmartin_smcginnis, correct15:58
smcginnishemna: Just send updates to you?15:58
smcginnishemna: Thanks for putting that together!15:59
hemnais everyone ok with the names I put on the slides and the speaker order, etc  ?15:59
smcginnishemna: +115:59
*** _cjones_ has quit IRC15:59
*** _cjones_ has joined #openstack-cinder15:59
*** sseverson has joined #openstack-cinder15:59
smcginnisOr can I actually +2 on this one? :)15:59
hemnayou can!15:59
hemnaI'm going to try and get my slides done this week15:59
*** BharatK has quit IRC16:00
*** Longgeek has quit IRC16:00
thingeesmcginnis, jgriffith: think it would be good to start discussing summit panel stuff soon. sent an email to everyone.16:00
* jungleboyj is looking.16:01
smcginnisthingee: Agreed16:02
smcginnishemna: Do we also need a short blurb (title, company, etc.) for the Presenters slide?16:02
smcginnisDon't want to put too much there. Maybe just our names and the company we work for?16:03
* thingee remember his last panel where he was a red x because he didn't submit a picture. :)16:03
hemnasmcginnis, we could use the bio's from the preso page ?16:03
smcginnisI'd be fine with a red X. :)16:03
hemnaheh yah16:03
smcginnishemna: Yeah, that's fine.16:04
*** jistr_ is now known as jistr16:05
jungleboyjhemna: kmartin_ Looks good.  I will work on putting some slides together for the IBM piece.16:06
rhe00hemna: kmartin_: I will work on something from X-IO. Are we still thinking one slide per vendor?16:08
hemnarhe00, yah I think so16:08
rhe00hemna: ok, sounds good16:08
*** xyang1 has joined #openstack-cinder16:09
jgriffithjust wondering... is this a "look at me I'm vendor-xyz" panel or a Cinder team panel?16:09
jgriffithnot to be a jerk, but....  I didn't know this kinda feels off to me16:10
hemnajgriffith, I think this is for the FC preso16:10
hemnano ?16:10
jgriffithhemna: yeah16:10
hemnawe have 7 presenters...which is kinda silly, but it is what it is.16:10
jgriffithhemna: so typically we try and avoid the "vendorized" slides on these16:11
jgriffithhemna: you're missing my point16:11
*** rmesta has joined #openstack-cinder16:11
hemnaI'm sure I am.16:11
jgriffithhemna: ok... never mind16:11
smcginnisjgriffith: I think we just wanted to mention everyone involved.16:11
jgriffithsmcginnis: yeah, def16:11
smcginnisjgriffith: Primarily as a show of force that FC is a supported option.16:11
kmartin_jgriffith, 1 slide per vendor is too much?16:11
jgriffithsmcginnis: but Vendor branded slides for each section is what i'm getting at16:11
smcginnisjgriffith: Oh, not for each slide.16:12
hemnaso the other option is to just have everyone up on stage during the Q&A and not do vendor slides ?16:12
jgriffithkmartin_: not at all... I'm saying if we're doing slides they should probably be OpenStack template slides and consistent16:12
smcginnisjgriffith: Just to have a quick spot for one slide to show I'm vendor X and I support FC.16:12
hemnaI'm open to whatever.  I was just trying to accommodate everyone.16:12
smcginnisjgriffith: That's how I saw it.16:13
hemnajgriffith, I think that's fine.  I put up a template slide deck that we can use16:13
hemnait's in the etherpad16:13
smcginnisOpenstack branded slide with my company name and a couple bullet items about what we support. Next.16:13
hemnajgriffith, 3rd link in the etherpad is what I put together last week16:13
kmartin_jgriffith, has all the branding stuff removed16:14
jgriffithsmcginnis: so what I was getting at was OpenStack slide template, an intro slide with panelists and their affiliation16:14
hemnaas a starting point for the outline for the preso, according to the etherpad outline.16:14
jgriffithkmartin_: the dropbox one?16:14
smcginnisjgriffith: And that's it. Just the list of names and the companies?16:14
smcginnisjgriffith: I'm not opposed to that.16:15
jgriffithhemna: kmartin_ yeah... perfect16:15
smcginnisjgriffith: But if we have time and space just thinking it would be nice to have at least 30 seconds to add more.16:15
jgriffithhemna: kmartin_ sorry, there are way too many slide sets up there ;)16:15
kmartin_jgriffith, :)16:15
thingeexyang1: do we have docs for incremental backup support?16:16
*** jdurgin1 has quit IRC16:16
xyang1thingee: Oh, not yet.  The CLI part is not merged yet16:16
xyang1thingee: In the doc, I use CLI to illustrate16:16
thingeexyang1: mind linking me to the patch?16:17
thingeeboth docs and cli16:17
xyang1thingee: That error is in every cli patch16:19
xyang1thingee: I don't have a doc link yet.  I can create one and send to you later16:20
thingeexyang1: cool. would like to include it in the Kilo release notes, but would be good if we had something16:20
thingeeas is, I wouldn't include it just because it's not self explanatory16:20
thingeetbarron: hey, do we have any docs on nfs backup driver?16:21
xyang1thingee: Ok, I'll work on the doc.  When can we do the cli release?16:21
thingeexyang1: I think we're going to be doing a big cinderclient release soonish..would like to get the version discovery stuff in16:21
*** jcru has joined #openstack-cinder16:22
xyang1thingee: Ok16:22
openstackgerritYuriy Nesenenko proposed openstack/cinder-specs: Checking the existence of volume
thingeexyang1: do we have anything for configuring oversubscription in thin provisioning?16:23
rhe00does anyone here know the best way to get a hold of the documentation group? are they active on IRC or is the mailing list a safer bet?16:24
smcginnisjgriffith, hemna, kmartin_, jungleboyj, rhe00, xyang1: Do we have consensus? Vendor slide or no vendor slide? I'm fine either way I guess, just don't want to put on together if we don't want it.16:24
xyang1thingee: That is a flag in cinder.conf16:24
thingeerhe00: #openstack-docs or something?16:24
xyang1thingee: I'll take a look of the existing doc16:24
smcginnisrhe00: Anne Gentle is around usually.16:24
*** jistr has quit IRC16:24
xyang1thingee: And see if i should add anything16:24
rhe00seems like the documentation for our volume driver fell of the planet over the weekend16:24
kmartin_smcginnis, one vendor slide16:24
thingeexyang1: might be good to have some explanation. If I remember correctly, the names were a little overlapping, so it would be good to be clear.16:24
jgriffithsmcginnis: so my point was really just one slide deck and making it all "flow"16:25
xyang1thingee: Sure16:25
jgriffithsmcginnis: so like a google-doc that everyone can collaborate on16:25
thingeexyang1: thanks, include me on any reviews and I'll +116:25
thingeexyang1: think these would be excellent for the release notes16:25
jgriffithsmcginnis: but it's hemna and kmartin_ 's prez really, so it's up to them I think16:25
xyang1thingee: Sure.  Thanks16:25
thingeexyang1: it's awesome that we have stuff to include in the release notes for consistency groups!!16:25
smcginnisjgriffith: Sure, I get what you're saying.16:26
kmartin_smcginnis, jgriffith was looking at the wrong deck that had all the HP branding stuff, so I think he is ok with one slide per vendor using the OpenStack side template16:26
xyang1thingee: I need to update cg as well for 2 new apis16:26
smcginniskmartin_: Ah, I could see why. :)16:26
xyang1thingee: The doc16:26
xyang1smcginnis: I am ok with or without vendor slide.16:27
thingeetsekiyama: you mind updating the release notes about the multi path patches you worked on. There is overlap on these and I always forget which does what.
*** annashen has quit IRC16:27
DuncanTSo I don't want flights to Vancouver, BC, do I?16:28
smcginnisxyang1: I'll make one and if we pull it - oh well. :)16:28
tsekiyamathingee: okey, let me see16:28
smcginnisDuncanT: You could fly in to Seattle, but that would be more direct.16:28
thingeetsekiyama: thanks16:28
jgriffithDuncanT: yeah, you do16:29
smcginnisDuncanT: Shuttle or taxi looked like it was only about 10-15 away from the conference center.16:29
smcginnisOr in this case centre. :)16:29
* thingee still needs to get a hotel16:29
jgriffithsmcginnis: you sure about that?16:29
DuncanT"Vancouver, BC YVR"? That's the right airport?16:29
smcginnisDuncanT: Yep.16:29
smcginnisjgriffith: Which?16:29
DuncanTThere was some discussion earlier about there being more than one Vancouver16:30
jgriffithSeattle to Vancouver?  That's like 2.5 hours I think?16:30
openstackgerritHuang Zhiteng proposed openstack/cinder: RBD: Add missing Ceph customized cluster name support
smcginnisjgriffith: I meant you can fly in to Seattle and go up, but if you fly into Vancouver then it's really close.16:30
jgriffithsmcginnis: oh, yeah!  :)16:30
*** leeantho has joined #openstack-cinder16:30
smcginnisOnly reason I considered Seattle was to see some friends on the way.16:31
smcginnisAnd to avoid company travel being flagged as "international" :)16:31
jgriffithI was worried because DuncanT said "So I don't want flights to Vancouver, BC, do I?"16:31
jgriffithsmcginnis: was worried you were going to send him to the wrong country :)16:32
jungleboyjsmcginnis: I am ok with Vendor slides or we could do a list of all the drivers with FC support.  Don't have to have it company branded.16:32
thingeetsekiyama: also information on the volume copy bandwidth limit would be good. Might be good to propose something to
smcginnisjgriffith: That would be pretty cruel. ;)16:32
tsekiyamathingee: sounds good. Will try it.16:33
thingeetsekiyama: thanks again16:34
DuncanTjgriffith: Don't give people ideas!16:34
smcginnisDuncanT: You heard the next summit got moved to Siberia, right?16:36
tbarronthingee: was at lunch16:37
tbarronthingee: we don't have any specific docs on nfs backup driver, but I  can do some updates16:37
*** Yogi11 has joined #openstack-cinder16:38
thingeetbarron: thanks that would be great.16:38
thingeethe backup doc in general needs help16:38
thingeeDuncanT: ^16:38
thingeedoesn't really help with specific backends16:39
*** Yogi1 has quit IRC16:40
*** harlowja_at_home has quit IRC16:41
thingeetbarron: I would love to include the nfs backup driver in the release notes, but there's no explanation with setting things up from Cinder's perspective.16:42
*** alecv has quit IRC16:43
thingeee0ne, jgriffith: does it matter if we notify about this dep change? It's already installed by other projects.16:45
thingeesince it's already in the global requirements16:45
jgriffiththingee: absolutely yes16:45
thingeeok glad I asked then16:45
jgriffiththingee: the issue is that packagers aren't packaging that for Cinder currently16:46
jgriffiththingee: that's why it's kinda sketchy16:46
thingeeah ha, good point16:46
e0nethingee:  -1 was for commit message:(16:46
e0nejgriffith: +1. some vendors probably will need to update their deb/rpm/etc to add new requirement16:47
thingeeso what do we do now?16:47
jgriffithe0ne: thingee that's why it may not be valid to merge that16:47
jgriffiththingee: it's technically backporting "new" functionality16:48
thingeegot it. so is there any other way for us to backport the fix for this bug?16:48
thingeethat is depending on this16:49
e0nei didn't verify what projects already use it but it could impact cinder packaging and deployment16:49
jgriffiththingee: see my notes in that patch... the alternative is posted as a gist16:49
thingeejgriffith: sorry missed that, thanks16:49
jgriffiththingee: s/notes/comments/16:49
thingeejgriffith: yeah that makes sense to me16:50
thingeewrt to the gist16:50
thingeesorry for initially leading you down this path :)16:50
jgriffiththingee: nahh... no worries16:51
jgriffiththingee: so you want me to propose the alternative?16:51
thingeejgriffith: yes please16:51
jgriffiththingee: k... I'll do that later today16:51
*** annashen has joined #openstack-cinder16:51
thingeejgriffith: before juno is security backports only =/16:51
e0nefyi, retrying is used by glance so i believe that everyone has it in thers rpm/deb repo16:53
tbarronthingee: I'm reading over the release notes now to get context for your remark.  I did look at the backup section in the admin guide and do see that it is way too thin.16:53
thingeeoverall I'm really happy with the doc references from the release notes. good job to everyone making that happen :)16:53
e0nejgriffith, thingee: btw, should we add slots for db migrations for Kilo16:54
*** russellb has quit IRC16:55
*** vilobhmm1 has joined #openstack-cinder16:55
thingeeadd slots?16:55
jgriffithe0ne: yeah16:55
jgriffiththingee: we need to reserve slots for bakcports16:55
*** vilobhmm11 has joined #openstack-cinder16:55
*** haomaiwang has joined #openstack-cinder16:55
jgriffiththingee: e0ne but I always get messed up on the "when" for that :)16:55
e0nethingee: like this one
thingeeoh yeah16:56
e0nejgriffith: after FF and RC i believe we won't have many migrations so i'll go on with it16:57
ameadeDuncanT: rebased
jgriffithe0ne: great, thanks!16:57
ameadeit was near trivial16:57
jgriffithe0ne: it was on my list... but I'm obviously behind a bit :(16:57
e0nejgriffith: i can leave it for you16:57
tbarronthingee: I don't know how the doc process works, so bear with me.  Release notes will come out April 30.  How long do we have for updates between now and then?16:58
thingeee0ne, jgriffith: I can also do it. I know jgriffith has already a bit on his plate.16:58
jgriffithe0ne: thingee done16:58
*** russellb has joined #openstack-cinder16:58
thingeetbarron: you have from now until april 30? :)16:59
*** asselin__ has joined #openstack-cinder16:59
*** bswartz has quit IRC16:59
*** akerr has quit IRC16:59
tbarronthingee: I was thinking we could have a bullet that says: * Ability to backup cinder volumes to an NFS repository. **Read the docs for more info**16:59
*** vilobhmm1 has quit IRC16:59
thingeetbarron: sure. we need docs though  :)17:00
tbarronBut then we have to backfill
tbarronthingee: yup.17:00
tbarronthingee: so I could add a brief section that says: By default, backup/restore uses Swift to provide the backup repository.  If you choose to use NFS instead, configure your cinder.conf as follows. ...17:01
thingeethat might be fine for now17:01
e0nejbernard: fyi, i'm testing my fix for i'll publish patch once i'll get it tested not only with devstack all-in-one env17:02
openstackLaunchpad bug 1401335 in Cinder "rbd calls block eventlet threads" [High,Confirmed] - Assigned to Alan Jiang (ajiang)17:02
thingeetbarron: depending as this grows, we may split them up in different pages. can always change that later though17:02
tbarronthingee: but  I'm not qualified to speak to Ceph  or TSM17:02
thingeetbarron: yeah that's fine. I just meant having docs for nfs backup driver would be great for the release notes.17:02
tbarronthingee: OK, I will get that done this week.17:02
*** asselin_ has quit IRC17:02
*** akerr has joined #openstack-cinder17:03
*** jms has joined #openstack-cinder17:04
openstackgerritJohn Griffith proposed openstack/cinder: Reserve 5 migrations for Kilo backports
thingeejgriffith: thanks17:06
jgriffiththingee: sure17:06
jgriffithe0ne: thanks for the reminder! :)17:06
thingeewas just about to push something up :) ... wasn't sure if done meant "done! sold to thingee to do it" :)17:06
*** annashen has quit IRC17:06
*** lpetrut has quit IRC17:07
thingeee0ne: do you not have +2 on cinderclient?
*** Yogi1 has joined #openstack-cinder17:09
e0nethingee: oops. i've put +1 because my previous vote was +117:09
*** Yogi11 has quit IRC17:09
hemnathingee, so, we just have one more os-brick patch to land17:10
hemnaand I think it's good for a first drop to pypi17:10
* thingee checks17:10
hemnawhenever you have time.17:10
hemnano hurry17:10
e0nethingee: i'll test it once more today before +217:10
*** patrickeast has joined #openstack-cinder17:10
e0nehemna: thanks for addressing my comments17:12
hemnae0ne, np17:12
*** annashen has joined #openstack-cinder17:12
e0nejgriffith: one more thing: cherry-pick to stable/kilo17:14
*** ronis has joined #openstack-cinder17:14
jgriffithe0ne: actualy they don't need to go back17:15
jgriffithe0ne: they're just "place-holders" incase17:16
*** Guest36304 is now known as mgagne17:16
*** mgagne has joined #openstack-cinder17:16
e0nejgriffith: ok, makes sence17:16
jgriffithe0ne: so we just reserve the spots, don't need to actually move them17:16
jgriffithe0ne: yeah, I used to think it was the same way too... but I was educated :)17:17
*** xyang has joined #openstack-cinder17:17
*** deepakcs has joined #openstack-cinder17:17
jms(Let me know if this would be better directed elsewhere...)  Is there a way during debug to get the cinder files to do a trace on what call it's in when it outputs messages?17:17
e0nejgriffith: thanks for sharing knoweledge:)17:17
jgriffithjms: that should already be there17:17
jgriffithjms: if it's missing, you might look at your logging config in cinder.conf17:18
jgriffithjms: oh... wait, a full trace?17:18
jordanPguys, just to let you know that I silent the scality-ci for a couple of hours. Ipv4 vs ipv6 problem while running apt-get17:18
jgriffithjms: hmmm... best way I know is a breakpoint, but I'm sure there's a way to auto do that17:18
*** timcl has quit IRC17:20
jmsjgriffith: Yeah, a full trace ... I was looking at the pythong trace module, but that doesn't seem to do exactly what I'm needing (ofcourse, I don't know python.... I'm just working on this!). :/17:22
jmsjgriffith: That, or maybe manually starting cinder-volume with -m pdb ... pdg ... Whatever the python debugger module is?17:23
jmsI'm _pretty sure_ what I'm looking for it called from -volume anyway17:25
jms*is called17:25
jgriffithjms: yeah, restart with that may work... not sure.  I've always done it via breakpoints manually :)17:26
jmsjgriffith: I'm learning python as I go so... Whoo!17:27
*** zhenguo has quit IRC17:27
jgriffithjms: best way to do it :)17:28
jms... Yeah, not for those of us that ignore white space when reading code ... Makes it just a tad more difficult. ;)17:30
*** aswadr has quit IRC17:32
jgriffithmtreinish: you ever run into "git review" not recognizing the "Change-ID:" line?17:33
mtreinishjgriffith: not normally17:34
mtreinishjgriffith: the hook it installs normally makes it pretty error free17:35
deepakcseharney, ping17:35
*** lpetrut has joined #openstack-cinder17:35
jgriffithmtreinish: yeah, I have the hook, and it "works"17:35
eharneydeepakcs: hi17:35
jgriffithmtreinish: just for some reason this particular set pukes17:35
deepakcseharney, gettign this error while nova tries to do update_snapshot_status call to cinder17:36
mtreinishjgriffith: well it looks like one of the commits in the middle is missing it17:36
*** russellb has quit IRC17:36
jgriffithmtreinish: oh.. hmmm17:36
jgriffithmtreinish: ahhh!  You're a genius :)17:36
mtreinishand I don't know if it's able to add it or not in that situation17:36
deepakcseharney, my cinder/api/contrib/ function (_update_snapshot_status_) is not getting called17:36
jgriffithI know what happened17:36
deepakcseharney, instead wsgi.server directly returns 400 back to Nova17:36
jgriffithmtreinish: the one in the middle should be squashed actually17:36
jgriffithmtreinish: ok... thanks!!17:36
mtreinishjgriffith: heh, sure np17:37
mtreinishbtw, really like that tree :)17:37
jgriffithmtreinish: :)17:38
*** melwitt has joined #openstack-cinder17:38
*** harlowja_away is now known as harlowja17:38
*** kmartin_ has quit IRC17:39
eharneydeepakcs: Cinder doesn't throw an error somewhere?17:39
*** Yogi1 has quit IRC17:39
*** russellb has joined #openstack-cinder17:40
deepakcseharney, nope17:41
*** dustins has joined #openstack-cinder17:41
deepakcseharney, i am seeing c-api, c-sch, c-vol17:41
deepakcseharney,  as you can see, nova's update is reachign cinder (Action body: debug msg is present in c-api) but before it reaches, eventlet.wsgi.server is returning 400 back Nova17:41
deepakcseharney, is there a way to look at wsgi server logs to figure why it returns back 400 ?17:41
eharneydeepakcs: wsgi issues should appear in the api service log17:42
deepakcseharney, u mean c-api.log ?17:43
eharneydeepakcs: yes17:43
deepakcseharney, Hmm i don't see any, only the INFO msg from wsgi17:43
deepakcseharney, is my conclusion above on whats happening correct ?17:43
eharneydeepakcs: i think so but i don't know why it's issuing 40017:45
*** vilobhmm11 has quit IRC17:45
deepakcseharney, any idea on how to proceed further, have spent > 1 day, without luck :(17:45
*** lcurtis has joined #openstack-cinder17:46
eharneydeepakcs: make sure the log shows that it loaded the os-snapshot-actions extension at startup17:46
*** vilobhmm1 has joined #openstack-cinder17:47
*** russellb has quit IRC17:47
deepakcseharney, i don't see it being loaded, what needs to be done to have it loaded ?17:48
eharneydeepakcs: generally it should just happen automatically and show with all the other extensions loading17:49
openstackgerritJohn Griffith proposed openstack/python-cinderclient: Introduce functional testing to cinderlcient
*** bswartz has joined #openstack-cinder17:50
*** jordanP has quit IRC17:50
*** e0ne has quit IRC17:51
*** akerr has quit IRC17:51
*** timcl has joined #openstack-cinder17:52
*** russellb has joined #openstack-cinder17:52
deepakcseharney, 3171:2015-04-20 17:15:12.127 DEBUG cinder.api.extensions [-] Loading extension cinder.api.contrib.snapshot_actions.Snapshot_actions from (pid=5018) load_extension /opt/stack/cinder/cinder/api/
*** akerr has joined #openstack-cinder17:52
deepakcseharney, I was looking for the word "Loaded" and it didn't show up, but the above Loading msg does mean its loaded, right ?17:53
eharneydeepakcs: yes17:53
deepakcseharney, It was working a week before, until i added code to support old and new nova & cinder and then it broke17:54
deepakcseharney, looking at the changes I haven't yet figured what could cause the nova - cinder api call to break17:55
*** akerr has quit IRC17:57
*** Yogi1 has joined #openstack-cinder17:58
*** akerr has joined #openstack-cinder17:59
*** annegentle has joined #openstack-cinder18:01
deepakcseharney, any other ideas ? :)18:02
eharneydeepakcs: see if nova can talk to cinder in general with something like "nova volume-create"18:03
*** vilobhmm1 has quit IRC18:03
*** vilobhmm1 has joined #openstack-cinder18:05
*** vilobhmm1 has quit IRC18:05
*** vilobhmm1 has joined #openstack-cinder18:05
deepakcseharney, ok, does nova calling cinder's v1 Vs v2 API can cause this to happen ? I don't see any cinder API version in nova.conf, just wondering18:06
eharneydeepakcs: it shouldn't... Nova uses v1, but the extension code is shared between the two iirc18:07
*** rushil has quit IRC18:10
openstackgerritJohn Griffith proposed openstack/python-cinderclient: Add functional post test_hook
jgriffithmtreinish: cinder is next... then to start moving tests18:16
*** nikesh has joined #openstack-cinder18:18
rmstardo you guys know if i'd leave orphans around if i were to force delete a cinder volume while it is downloading?18:21
jgriffithrmstar: it *shouldn't*, it should handle things like aborting the download and cleaning up after itself18:22
jgriffithrmstar: that's not to say it does... but it *should* :)18:22
rmstarjgriffith: thanks.  i understand :)18:22
DuncanTrmstar: Currently, yes it leaves orphan volumes at times, orphan iscsi connections, and all sorts of other mess18:25
*** sgotliv has joined #openstack-cinder18:29
*** vilobhmm1 has quit IRC18:29
*** e0ne has joined #openstack-cinder18:29
*** vilobhmm1 has joined #openstack-cinder18:31
*** annashen has quit IRC18:37
rmstarDuncanT: do you know if there is anyway to clear off the "mess"?18:38
*** jcru has quit IRC18:40
mtreinishthingee: have you seen this:
*** jcru has joined #openstack-cinder18:41
mtreinishthingee: that seems very similar to the issue you are hitting in your ci right?18:41
thingeemtreinish: thanks, checking18:41
DuncanTrmstar: nope, sorry, you can script it but it really needs to be build into cinder I think. We're starting to work on it for liberty18:44
*** garthb_ has joined #openstack-cinder18:45
*** garthb has joined #openstack-cinder18:45
*** e0ne is now known as e0ne_18:46
*** deepakcs has quit IRC18:47
*** rushil has joined #openstack-cinder18:49
*** annashen has joined #openstack-cinder18:50
*** dannywilson has joined #openstack-cinder18:52
*** e0ne_ has quit IRC18:54
*** ebalduf has joined #openstack-cinder18:54
*** pradipta has quit IRC18:54
*** dannywilson has quit IRC18:56
*** e0ne has joined #openstack-cinder18:58
*** annashen has quit IRC19:03
*** annashen has joined #openstack-cinder19:03
e0nejgriffith: pep8 failed again on migrations placeholders19:07
jgriffithe0ne: hmm.... looking19:09
jgriffithe0ne: haha... stupid blank lines19:09
jgriffithe0ne: fixing now19:09
*** diemt has joined #openstack-cinder19:10
*** deepakcs has joined #openstack-cinder19:14
jmsWhere does driver_api come from (as used in cinder.scheduler.flow.create_volume.ScheduleCreateVolumeTask)? Specifically execute() ... I *think* that's where there error I'm getting is comming from ("No weighted host found") ... but I don't know _why_ it's getting that error. If I could figure that out it may help a bit. ;)19:21
jmsCould also have something to do with volume.filters ... or in addition to. :/19:21
jgriffithjungleboyj: hey... what the heck does this do?
jungleboyjjgriffith: Looking.19:26
jgriffithjungleboyj: hoping you know :) Not sure if you introduced or not19:27
jgriffithjungleboyj: reason I ask is:
jgriffithjungleboyj: I have no idea why that's failing19:28
jungleboyjI didn't, but it tests this:
jungleboyjChecking to make sure that no one is using assert_called_once_with()19:28
jungleboyjThat was pulled over from Nova.19:28
jgriffithjungleboyj: well... yeah, but checking *what*19:28
jgriffithjungleboyj: it's a unit test, checking "something"19:29
jgriffithbut I don't know what19:29
*** aix has quit IRC19:29
jgriffithjungleboyj: ie ?19:29
jungleboyjjgriffith: Oh, ok.19:29
jgriffithjungleboyj: ahhh... hm, I think I see what it's doing19:29
jgriffithit's a hard coded to cinder/tests in the hacking check19:30
jungleboyjjgriffith: So, the first one is making sure that assert_called_with(including a filename) doesn't griffer a fail in the check.19:30
jgriffithwhich is used by the unit test19:30
jungleboyjIt isn't really using that at all.19:30
jgriffithjungleboyj: wait... what?19:30
jungleboyjIt is just a string as that is what is expected as the second argument.  Some filename.19:30
jgriffithjungleboyj: so I'm confused.... the second arg is a "filename"19:32
jungleboyjjgriffith: ".assert_called_with(", "cinder/tests/") is just an example of a logical line that could go through the check.19:32
jgriffithjungleboyj: sure, but we're assertEqual'ing on it and it's failing :)19:32
jungleboyjcheck.check_assert_called_once( uses that function.19:32
jgriffithjungleboyj: yeah, I get that19:32
jungleboyjWhere is that happening?19:32
jgriffithtrying to figure out why I'm getting this:
jgriffithjungleboyj: see what I mean?19:33
jungleboyjThat is bizarre .  I am assuming you didn't change anything there?19:34
smcginnisThe check's unit test itself is failing. That's new. :{19:34
jgriffithsmcginnis: jungleboyj so I changed a bunch of stuff :)19:34
smcginnisjgriffith: I blame you then. :P19:34
jgriffithsmcginnis: jungleboyj I'm in the process of moving tests to test/unit/19:34
jgriffithsmcginnis: there's obviously something I need to change, but no diea what19:35
smcginnisjgriffith: Ah, did you update the exlusion path in
*** vilobhmm1 has quit IRC19:35
jgriffithNope :)19:35
jungleboyjTry changing that to cinder/tests/unit/test1.py19:36
*** annegentle has quit IRC19:36
smcginnisjungleboyj: Shouldn't this line actually be testing assert_called_once?
smcginnisjungleboyj: But that's besides the point.19:36
jgriffithjungleboyj: in the unit test?  thought of that, but the problem is that the other tests pass and I haven't modified them19:36
*** bill_az has joined #openstack-cinder19:36
*** crose has joined #openstack-cinder19:37
smcginnisThis is really odd. If the check didn't change and the check's unit test didn't change, then it should still work as expected regardless of where actual tests are located.19:37
jungleboyjLet me try something.19:37
jgriffithsmcginnis: one would think19:37
jgriffithsmcginnis: but honestly I don't understand how this test works anyway19:37
jungleboyjGive me a minute.19:37
smcginnisjgriffith: No git diff between these two ( and
smcginnisThe check or the check check? :)19:38
jgriffithsmcginnis: jungleboyj never mind... I get it!19:39
smcginnisLight bulb!19:39
jgriffithsmcginnis: kinda19:39
* jungleboyj must not have explained well?19:39
jgriffithjungleboyj: think maybe it's what you were trying to say19:40
smcginnisjgriffith: Inadvertent change to the check?19:40
* jungleboyj waits with baited breat19:41
jgriffithjungleboyj: so the issue is that the input is cinder.unit/tests19:41
jgriffithjungleboyj: so it wasn't cinder/unit/tests....19:41
jungleboyjjgriffith: Yep, that was what I was trying to say.  :-)19:41
jgriffithjungleboyj: ok... good job then :)19:42
jgriffithjungleboyj: and thanks!19:42
jungleboyjjgriffith: Welcome!19:42
rmstarDuncanT: ok.  thanks for your answer19:44
*** ronis has quit IRC19:45
*** lpabon has quit IRC19:48
*** vilobhmm1 has joined #openstack-cinder19:50
*** vilobhmm1 has quit IRC19:50
*** vilobhmm1 has joined #openstack-cinder19:50
openstackgerritJohn Griffith proposed openstack/cinder: Reserve 5 migrations for Kilo backports
*** kaisers has quit IRC19:56
openstackgerritIvan Kolodyazhny proposed openstack/cinder: Move RBD calls to a separate threads
e0nejungleboyj: hi! could you take a look on patch ^^ today if you'll get a time?19:57
jungleboyje0ne: I just opened it.  :-)19:57
*** julim has quit IRC19:58
e0nejungleboyj: cool. thanks19:58
e0neit's an alternative for
*** liusheng has quit IRC19:58
*** liusheng has joined #openstack-cinder19:59
*** jdurgin1 has joined #openstack-cinder19:59
*** Rockyg has joined #openstack-cinder19:59
*** annashen has quit IRC20:02
*** jdurgin has quit IRC20:03
jungleboyje0ne: Just commented.  The commit message needs to be cleaned up a bit and then it will be ok.20:04
e0nejungleboyj: thanks. fixing20:05
openstackgerritIvan Kolodyazhny proposed openstack/cinder: Move RBD calls to a separate threads
openstackgerritIvan Kolodyazhny proposed openstack/cinder: Move RBD calls to a separate threads
jungleboyjOk, that makes more sense.  Thank you.20:11
*** jdurgin1 has quit IRC20:12
e0nejungleboyj: thanks!20:14
jungleboyje0ne: You are welcome.20:15
e0nejungleboyj: i hate such fixes. 2 days tried to find env for reproduce it, half-a-day reading manuals/sources and result is 2 lines of code20:15
jungleboyje0ne: It can't all be easy.  You learned a lot though, right?20:16
e0nejungleboyj: yes, you're absolutely right.20:16
jungleboyje0ne: So, here is a bright side.  :-)20:17
e0nejungleboyj: :)20:17
*** crose has quit IRC20:20
*** Yogi1 has quit IRC20:23
mtaninojungleboyj: Thank you for trying the patch. I need to add more test cases but I'd like to agree the way to fix beforehand :)20:26
e0nethingee: I verified and change my vote for
*** jdurgin has joined #openstack-cinder20:26
jungleboyjmtanino: Welcome.  Yeah, I tried a few times to recreate the failure and didn't see it fail.20:27
jungleboyjYou want me to take my +1 off of it if you are planning to do more test cases?20:27
mtaninojungleboyj: nice!20:27
mtaninojungleboyj: yes. now you can lift off you +1 until I will add test cases.20:27
jungleboyjmtanino: Yeah, thank you for figuring that out!20:27
jungleboyjOk, back in a bit.20:28
*** timcl has quit IRC20:30
*** Yogi1 has joined #openstack-cinder20:30
*** akerr has quit IRC20:30
ameadewinston-d: hey if you get a chance, can you let me know if you are good with this patch?
* e0ne bb tomorrow20:31
*** jungleboyj has quit IRC20:34
*** Yogi1 has quit IRC20:34
*** deepakcs has quit IRC20:37
*** Yogi1 has joined #openstack-cinder20:43
*** Lee1092 has quit IRC20:51
*** eharney has quit IRC20:54
*** e0ne has quit IRC20:55
*** annegent_ has joined #openstack-cinder20:55
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder: WIP: Remove Brick from cinder codebase
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder: WIP: Remove Brick from cinder codebase
smcginnishemna: Just noticed something. One of the emails I got for the summit said the presentation should be formatted 16:9.20:59
smcginnishemna: How to change:
smcginnishemna: Widescreen only I guess.20:59
hemnawhere are the requirements to switch it to 16:9 ?21:00
smcginnishemna: Did you get an email titled "OpenStack Speaker Information"?21:01
hemnaI'm sure I did21:01
hemnaI'm looking21:01
smcginnishemna: "please come prepared with your VGA or DVI compatible laptop with your 16:9 presentation loaded and ready to present."21:02
hemnahrmm yah I see that21:02
hemnaI was actually going to start hacking on my slides21:02
smcginnisDoesn't look like it messes up the format too badly.21:03
hemnayah I just tried switching21:03
hemnajust kinda lost some vertical resolution21:04
hemnabut fine otherwise ?21:04
*** Yogi1 has quit IRC21:05
openstackgerritTom Barron proposed openstack/cinder: Fix range check for NFS used ratio
*** bswartz has quit IRC21:09
*** xyang has quit IRC21:17
thingeemtreinish: trying that patch now21:18
mtreinishthingee: cool, hopefully it'll work. I have no idea tbh21:25
mtreinishI just saw it go by somewhere and it looked on point21:25
*** jaypipes has quit IRC21:26
*** e0ne has joined #openstack-cinder21:30
*** mwichmann has joined #openstack-cinder21:30
thingeemtreinish: no dice =/21:33
mtreinishoh well, it was worth a shot at least21:34
openstackgerritWalter A. Boring IV (hemna) proposed openstack/python-cinderclient: Add volume multi attach support
*** e0ne has quit IRC21:35
*** lpetrut has quit IRC21:35
*** bswartz has joined #openstack-cinder21:38
openstackgerritJohn Griffith proposed openstack/cinder: Move unit tests into dedicated directory
jgriffiththingee: DuncanT hemna ^^21:40
jgriffithIt would be *best* to rip the band-aid and merge that sooner rather than later21:40
hemnajgriffith, is this in prep for creating cinder/tests/tempest ?21:40
*** mwichmann has left #openstack-cinder21:40
jgriffithBetter to let each patch in the queue rebase, as opposed to try and refactor that monster over and over and over until I dide21:41
*** erlon has quit IRC21:41
mtreinishhemna: wait, what is cinder/tests/tempest ?21:41
thingeehemna: in general, it's something we've been wanting to do for some time.21:41
jgriffithmtreinish: maybe that's something hemna is workign on?21:42
thingeehemna: splitting out unit tests21:42
jgriffithhemna: functional vs unit21:42
hemnaI thought folks were looking at pulling in tempest tests into each project instead them living in the tempest project ?21:42
hemnamaybe i'm drunk. dunno21:42
jgriffithhemna: drunk!!!21:43
jgriffithhemna: it's 2:43!!!21:43
mtreinishhemna: well that's something tons of people are confused about. Because there is a push for projects to spin up functional testing21:43
jgriffith14:43 to be more clear21:43
hemnait's 5:43 somewhere21:43
jgriffithhemna: you're close... it's spin up functional tests in the project using tempest-lib21:43
mtreinishbut that's doesn't necessarily mean pulling tempest tests into each project21:43
hemnaI see21:43
*** jungleboyj has joined #openstack-cinder21:43
mtreinishjgriffith: well functional test don't necessarily have to use tempest-lib either. Only if it makes sense for what you're trying to test21:44
mtreinishlike the nova functional tests have a good pattern for functional testing pieces without needing tempest-lib21:44
jgriffithmtreinish: fair enough21:44
jgriffithhemna: just one typo in teh commit message?  That's my only mistake WooHooooo!!!!21:45
mtreinishjgriffith: btw that change is gonna be total rebase hell for everyone else :)21:46
jgriffithmtreinish: I know21:46
jgriffithmtreinish: it's them or me21:46
hemnamtreinish, hence the hope that it lands first!21:46
jgriffithmtreinish: I don't know of another way to go about it, do you?21:46
*** dustins has quit IRC21:47
jgriffithmtreinish: and there's significantly higher risk of *me* breaking all the onesie twosies21:47
jgriffithmtreinish: and I know we'll never get everybody to cut the chord and build off of this21:47
jgriffithso I'm throwing the chaos grenade in21:47
mtreinishjgriffith: well it really comes down to how much you want to be stat padding21:48
jgriffithmtreinish: LOL21:48
mtreinishbecause you know stackalytics brings in the money21:48
jgriffithmtreinish: not at my company :(21:48
jgriffithmtreinish: but I've heard21:48
hemnadidn't thingee win that one in Kilo21:49
hemnawith all the driver removal patches ?21:49
jgriffithmtreinish: I get paid the same if it's one line or a billion, or ZERO commits vs 100021:49
jgriffithmtreinish: since they're all renames they don't mean squat anyway21:49
jgriffithmtreinish: besides... I thought you wanted me to get cracking on moving some tests around :)21:50
mtreinishjgriffith: sure but does your company get paid more if solidfire is contrib #30 instead of #31 :)21:50
mtreinishoh or think of the headlines21:50
mtreinishjgriffith: you could get something like:
mtreinishwhich has the best comment about sdague at samsung21:51
jgriffithHoly shit, WTF is that!21:51
jgriffithDdue, that's just sad21:52
jgriffithreally F'ing sad21:52
*** jcru has quit IRC21:52
*** mriedem is now known as mriedem_away21:53
* jgriffith needs to find a new project to work on... no more room here21:53
mtreinishjgriffith: hehe, I've got a couple that no one cares about. I'd love contribs on
jgriffithmtreinish: so does that mean I have to leave tha twork to someobdy from Redhat, HP, Mirantis or IBM?21:55
jgriffithmtreinish: tisk-tisk... no README21:56
hemnais this Dune ?21:56
mtreinishtbh, I only ever worked on it for a little bit. I would like to pick it up again soon21:56
mtreinishhemna: that would probably be cooler than what it really is21:57
mtreinishI just wrote it in anger after I started using the circuit simulation visualization tools at IBM which were the best 1982 had to offer21:58
jgriffithmtreinish: it's funny, after moving away from EE work I though for sure I'd be doing SPICE or FPGA etc21:59
jgriffithhaven't done a ckt analysis in forever... closest I came was buildign my barn and doing the power budget21:59
mtreinishjgriffith: up until about 1.5 years ago I thought that I was going to switch over to a mainframe SRAM design group at ibm22:02
mtreinishbut they never found the budget to transfer me for reals22:02
mtreinishso I haven't really touched ckt stuff since then either22:03
mtreinishunless you count rewiring my car's audio system :)22:03
*** melwitt has quit IRC22:04
*** jms has quit IRC22:07
patrickeasti recall seeing folks discussing issues with test_volume_boot_pattern, is that still in progress?22:14
patrickeastwasn’t sure which bug was related to it22:14
smcginnisthingee: ^^ That was you, wasn't it?22:21
smcginnispatrickeast: I seem to recall it being an issue with nova calling things out of sequence.22:21
thingeepatrickeast: I'm hitting an issue with my ci22:21
patrickeastthingee: same :(22:22
patrickeastlooking into intermittent issues this week22:22
patrickeastwanted to see if it was the same issue or not22:22
patrickeastthingee: do you have a bug open for yours?22:23
thingeepatrickeast: no22:24
thingeepatrickeast: here's an example of mine though with volumebootpatternv2...
thingeesame request, creates a volume, then eventually does a remove export22:25
thingeewhile the convert is happening22:25
thingeemtreinish: ^22:25
*** rushil has quit IRC22:25
thingeeonce that remove export happens, then nova complains there's no session22:26
patrickeastthingee: hmm interesting, i’ll have to go dig through my vol log and see if that happens for me too22:27
patrickeastthingee: i end up getting a timeout in the actual test as the failure reason, do you see that too?22:27
thingeeactually this maybe correct behavior for me. it's suppose to eventually detach the volume after copy volume to image22:27
mtreinishthingee: would that be related to:
mtreinishwhich I'm still not sure whether it's a valid change or not22:27
mtreinishI wasn't able to reproduce that reported bug22:28
patrickeasti saw which describes my symptoms exactly, but it kind of looks like this is a catch-all for something broke and the test timed out22:28
openstackLaunchpad bug 1298472 in tempest "SSHTimeout in tempest trying to verify that computes are actually functioning" [Critical,New] - Assigned to Matt Riedemann (mriedem)22:28
mtreinishpatrickeast: yeah that's pretty much what it is. Debugging why tempest fails to ssh into a guest 196 secs after nova says it's available is very hard22:30
mtreinishpatrickeast: oh but in that case it wasn't the ssh which caused the timeout (well it is but sort of)22:32
mtreinishpatrickeast: that test hit the per test timeout to ensure we don't hang forever22:32
mtreinishpatrickeast: look at "20:59:55 Captured traceback:" in that console log22:32
thingeepatrickeast: ah, yeah I'm not having that problem22:32
mtreinishthat's the timeout fixture which will kill a test if OS_TEST_TIMEOUT is elapsed (it defaults to 500 in .testr.conf in tempest)22:33
patrickeastmtreinish: ooo yea its the fixture timeout22:33
patrickeastiirc i set that extra high…22:33
* patrickeast checks the config22:33
mtreinishpatrickeast: yeah you did22:34
mtreinishlook right below that22:34
mtreinishit killed it after 1209.126 secs22:34
patrickeastah yep22:34
mtreinishso something is really slow in your env22:34
patrickeasti set it to 120022:34
patrickeastso what i’ve noticed is that when it works, its like 160 tops22:35
patrickeasti set it to 1200 to see if maybe it needed more time22:35
patrickeastbut its either ~2 min or never seems to finish22:35
patrickeastalthough maybe i’m just overloading this initiator22:36
*** jaypipes has joined #openstack-cinder22:37
thingeepatrickeast: when did your issue start happening?22:44
SwansonI've seen something like that every so often.  I haven't checked the latest round from our ci to know if it is still happening.22:45
patrickeastthingee: its been occurring over the last week or two since i got my system moved into a DMZ and back online, not sure exactly when something would have been introduced since it was off for about 3 weeks waiting until we fixed the security issues22:45
patrickeastterrible timing with the release cycle… but it is what it is22:45
*** _cjones_ has quit IRC22:46
mtreinishpatrickeast: yeah I can't really say one way or the other22:46
mtreinishnothing really stands out in the logs as causing a failure from a quick glance22:46
mtreinishI would have expected something to timeout somewhere after waiting so long22:46
mtreinishmaybe on your san (or whatever hardware you're testing with)22:47
patrickeastyea i’m almost thinking it might be a network issue with my openstack provider nodepool is using to spin these devstack vm’s up on22:47
thingeepatrickeast: I haven't figured out yet if mine is related to the multi-attach change. It's hard to tell, because my ci passes if I use a patch before multi-attach, but also the multi-attach patch caused a series of changes in nova and tempest22:47
openstackgerritJohn Griffith proposed openstack/cinder: Move unit tests into dedicated directory
patrickeasthmm interesting, that would make sense that it could have caused some issues22:48
patrickeasttouched a lot of code paths22:48
patrickeastmtreinish: yea i’ve looked into issues with our data paths, but the iscsi stuff all seems ok22:48
thingeebut I have raised an issue where cinder is returning an exception that nova isn't expecting.22:48
thingeethat was introduced in multi-attach22:49
thingeemtreinish: which by the way, catching that exception did not help22:49
*** _cjones_ has joined #openstack-cinder22:49
mtreinishthingee: really, did it just fail later on?22:49
jgriffiththingee: patrickeast if each of you want to send me a datera cluster and pure array I'll be happy to help troubleshoot :)22:50
jgriffiththingee: patrickeast make sure you opt for the insurance on the shipping :)22:50
mtreinishjgriffith: heh, that's a good idea, I wish I had thought of it. My home array could always use a boost22:51
jgriffithmtreinish: "Griffith resellers"22:52
thingeejust got a recruiter email with just buzz words. it's literally devopsclouddockertestdrivendevelopmentcontinuousintegration22:52
jgriffiththingee: you should jump all over that one!  NOT22:52
thingeejgriffith: ha!22:52
thingeegriffith resellers22:52
thingeeI'm imagining walking on a farm and then inside a barn there are server racks.22:53
hemnapowered by horse manure22:53
* thingee used to imagine that's what jgriffith's "lab" was22:53
thingeelaugh all you want, but you city slickers with your drought and fancy silicon valley will be coming back to old mcdonald.22:54
hemnadried out, it can burn the boil water turn a turbine.22:54
*** annegent_ has quit IRC22:57
thingeexyang1: besides jenkins complaining, the incremental doc makes sense to me. thanks22:59
thingeeincremental backup*22:59
*** rmesta has quit IRC23:01
*** annashen has joined #openstack-cinder23:03
*** hemna is now known as hemnafk23:07
*** annashen has quit IRC23:08
openstackgerritJay Bryant proposed openstack/cinder: Add hacking check for str and unicode in exceptions
patrickeastthingee: if you have a chance can you target some bp’s for me? i’ve got a few for L-123:11
patrickeastlemmie know if its easiest to email a list or post links here23:13
*** jungleboyj has quit IRC23:15
*** mdenny has quit IRC23:16
thingeejgriffith, mtreinish: so what I've observed now is that in the volumebootpattern, the second instance it brings up with the volume it originally created, it never does an export. I see the volume is in-use in cinder, but I don't see an export happening, so it's like initialize connection is not happening.23:18
thingeethe export definitely is there for the first instance though, but is gone after the instance is deleted and the volume is detached23:22
*** lcurtis has quit IRC23:22
*** IanGovett has quit IRC23:26
*** esker has quit IRC23:27
thingeetake that back, initialize_connection is definitely happening23:28
openstackgerritJohn Griffith proposed openstack/python-cinderclient: Introduce functional testing to cinderlcient
openstackgerritJohn Griffith proposed openstack/cinder: Move unit tests into dedicated directory
thingeejgriffith: spelled cinderclient wrong ^23:30
jgriffiththingee: must not be a problem since you knew it was cinderclient :)23:31
jgriffiththingee: send me a new keyboard :)  Mine sucks :)23:31
thingeejgriffith: I'll send you a das keyboard with no labels!23:31
jgriffithCouldn't be any worse23:31
jgriffithmtreinish: ok... now, trying to remember what our priorities were for test moves?23:32
thingeewith the cherry blue key switches. your co-workers and wife will love you.23:32
jgriffiththingee: I have my old IBM keyboard from 1993... it's clicky-clacky enough23:32
thingeeoh nice! I like those23:33
jgriffiththingee: and my Gold Touch... which I'm not allowed to use in the house any more :)23:33
mtreinishjgriffith: I win Model M from 92 :)23:33
jgriffithmtreinish: ha!  Nice23:33
jgriffithmtreinish: mine is just on display on my desk at work now23:33
thingeemtreinish: that adapter to USB is slowing down your keystrokes23:33
openstackthingee: Error: "!" is not a valid command.23:33
jgriffithmtreinish: every once in a while someobdy walks by and goes "Ohhh wow!"23:34
mtreinishI have one older than me, but I didn't want to deal with the adapter to get it working on my computer23:34
*** arif-ali has quit IRC23:34
*** chlong has joined #openstack-cinder23:34
mtreinishjgriffith: priorities for test moves?23:35
*** ebalduf has quit IRC23:35
jgriffithmtreinish: yeah... tempest--->cinder23:37
jgriffithmtreinish: there were some inparticular that you guys were sick of dealing with I thought :)23:37
mtreinishjgriffith: heh, is that a trick question? :)23:39
jgriffithmtreinish: LOL... not trick, but loaded ;)23:40
mtreinishjgriffith: but if you're thinking about it, I would look at: first23:40
*** Rockyg has quit IRC23:41
jgriffithmtreinish: That's exactly what I was looking for23:41
mtreinishjgriffith: fwiw, I think the tests which are valuable are those which are really testing an internal cinder function indirectly23:42
mtreinishvaluable to move23:42
mtreinishalso I expect a bunch of the negative tests are valid to migrate too23:43
jgriffithmtreinish: which there are a lot of IMO23:43
mtreinishbecause most of them just test the api layer and maybe some db23:43
jgriffithmtreinish: yeah, I just thought I saw something in my logs between you and sdague about something being "gone soon"23:44
*** david-ly_ has joined #openstack-cinder23:44
mtreinishjgriffith: oh we might have just been talking about the cli tests :)23:44
jgriffithmtreinish: yeah, looking back that's what it was23:44
mtreinishjgriffith: but sdague's nova flavor functional tests are a really good example of what we should be doing for functional testing:
*** jamielennox|away is now known as jamielennox23:45
mtreinishwell not self approving with only 1 +223:45
*** david-ly_ is now known as david-lyle23:46
*** mtanino has quit IRC23:46
openstackgerritJohn Griffith proposed openstack/python-cinderclient: Add functional post test_hook
*** garthb_ has quit IRC23:53
*** garthb has quit IRC23:53
*** annegentle has joined #openstack-cinder23:57
*** openstackgerrit has quit IRC23:58
*** openstackgerrit has joined #openstack-cinder23:58

Generated by 2.14.0 by Marius Gedminas - find it at!