Tuesday, 2015-07-28

*** jwang has quit IRC00:00
*** jwang has joined #openstack-cinder00:00
patrickeasttsekiyama: i think if we change https://github.com/openstack/cinder/blob/master/cinder/image/glance.py#L326 to be more like http://paste.openstack.org/show/405989/ it should do what we expect00:01
patrickeastwhat i’m not sure of though is if anyone actually needs the metadata00:01
patrickeastlooks like upload_volume in image_utils.py doesn’t actually use the return value anyway00:02
tsekiyamapatrickeast: Confirmed that it works as expectedly00:03
tsekiyamapatrickeast: but I'm not sure why it actually need metadata..00:04
patrickeasttsekiyama: i dont think it does all the places in here where its called we ignore the return value https://github.com/openstack/cinder/blob/master/cinder/image/image_utils.py#L33900:04
* patrickeast is searching around other places to verify00:04
patrickeastunfortunately searching for .update( finds a lot of stuff : (00:05
tsekiyamapatrickeast: yeah.. maybe we could just remove the returning values and run unittests to check what will fail..00:06
patrickeasttsekiyama: good idea00:06
patrickeasttsekiyama: in theory if it breaks anyone too bad it should show up on their 3rd party ci too00:07
*** timcl has quit IRC00:07
thingeeis there any documentation on glance artifacts?00:07
thingeeI don't have the patience for the video presentation I found from the paris summit.00:08
patrickeastthingee: i haven’t found any good sources00:08
*** markvoelker has joined #openstack-cinder00:09
tsekiyamapatrickeast: haven't seen anything yet either00:09
*** Guest10__ has joined #openstack-cinder00:09
patrickeastthingee: the best docs i could find were scattered around in here http://docs.openstack.org/developer/glance/00:10
thingeepatrickeast: I did a search in that, and all I got was the metadata catalog00:11
thingeeOn how I can search on artifacts.00:11
tsekiyamaor glance specs? http://specs.openstack.org/openstack/glance-specs/specs/kilo/artifact-repository.html00:11
thingeeI still don't know what a dang artifact is.00:12
thingeeah I see00:13
thingeetsekiyama: thanks00:13
thingeewow, now that's a spec00:14
thingeetsekiyama: how well are artifacts working out in glance, you know?00:14
patrickeasthaha yea, thats the most detailed one i’ve ever seen00:14
*** andreykurilin__ has quit IRC00:14
tsekiyamathingee: I don't know much about it00:15
thingeetsekiyama: ok so I'm assuming you're looking into this bug already?00:16
* thingee checks if there's already an open bug00:16
tsekiyamathingee: yep, and confirmed that patrickeast's patch fixes it00:17
thingeeCouldn't find a bug. I'll report it00:17
thingeetsekiyama, patrickeast: not the first time this has happened https://bugs.launchpad.net/cinder/+bug/121046700:19
openstackLaunchpad bug 1210467 in Cinder "cinder upload-to-image broken with v2 glance api" [Undecided,Fix released] - Assigned to Edward Hope-Morley (hopem)00:19
patrickeastmebbe we should add some check/gate tests that use v2 glance00:20
patrickeastalthough i suppose thats easier said than done00:20
thingeetsekiyama, patrickeast: anyone able to verify kilo?00:21
*** jasondotstar has quit IRC00:21
*** angela-s has quit IRC00:21
thingeetsekiyama: https://bugs.launchpad.net/cinder/+bug/147873700:22
openstackLaunchpad bug 1478737 in Cinder "Cinder upload-to-image with glance v2 fails" [Medium,In progress] - Assigned to Tomoki Sekiyama (tsekiyama)00:22
tsekiyamathingee: ok I'll try it on Kilo too00:22
*** jasondotstar has joined #openstack-cinder00:22
*** dims has quit IRC00:22
patrickeasti don’t have a kilo env ready right now, could spin one up if needed00:23
patrickeastthingee: tsekiyama: although a quick check of the code appears that it should repro on kilo too00:23
tsekiyamamy kilo env is killed by OOM lol00:25
*** mtanino_ has joined #openstack-cinder00:25
*** mtanino has quit IRC00:26
tsekiyamamm sorry i'll check it tomorrow00:27
*** tsekiyam_ has joined #openstack-cinder00:27
patrickeasti’ve got one stacking right now… take a few min and i can verify00:29
*** xyang has joined #openstack-cinder00:30
*** tsekiyama has quit IRC00:31
*** dims has joined #openstack-cinder00:32
*** xyang has quit IRC00:33
*** salv-orlando has joined #openstack-cinder00:33
*** DericHorn-HP has quit IRC00:37
*** salv-orlando has quit IRC00:37
*** laughterwym has joined #openstack-cinder00:39
*** jasondotstar is now known as jasondotstar|afk00:43
thingeepatrickeast: so what's the plan now with your caching work and this cinder glance store being revived?00:45
thingeeI guess you still need both00:45
patrickeastthingee: heh yea good question00:45
patrickeastso i’ve though about this a bunch00:45
patrickeasti really don’t like the idea of having two code paths doing such similar things00:46
patrickeastbut i can’t figure out a good way around it00:46
*** heyun has joined #openstack-cinder00:46
patrickeastthe biggest problem is the image formats00:46
patrickeastthe cinder backend stuff only helps out a bunch if they are raw00:46
*** openstackgerrit has quit IRC00:46
patrickeastor if we code the cinder backend to do something tricky like keep the original image data in one volume and a second copy as a raw one to clone off of (basically what the cache does)00:47
patrickeastbut there is not limiations for it00:47
patrickeastand you would need to duplicate it around on all your cinder backends00:47
*** openstackgerrit has joined #openstack-cinder00:47
patrickeastso i came back to the conclusion that a cache which glance doesn’t need to know about still makes sense00:47
thingeepatrickeast: so in the end, how are communicating to operators why you need one or the other?00:48
thingeetsekiyam_: ^00:48
patrickeastthingee: i don’t have a real solid plan for that yet00:49
patrickeastthey kind of have different use cases00:49
*** _cjones_ has quit IRC00:50
patrickeastthe cache isn’t really a viable solution if you just want to use cinder as the glance backend to store all the images, but if you still want some improvement and to keep your glance stuff using swift or whatever its a nice alternative00:50
patrickeastor for like our customers, i’m sure i would be looked at like an idiot if i told them the only way they got fast image operations is to put *all* the images on their tier 1 flash array00:51
thingeepatrickeast: just to make sure I understand myself, you use the glance store know about a single storage backend. Sometimes the copy will be quick we're cloning to the same storage backend. However, you cache work will just work across all backends and still involves glance keeping track of what volumes still represent an image?00:51
thingeewow lots of typos in that00:52
patrickeasthaha no worries00:52
patrickeastthingee: so the cache is per-backend00:52
patrickeastthingee: and glance doesn’t know about it00:52
patrickeastthingee: after the first copy you would see faster volume creation00:52
thingeewith you so far00:53
patrickeastthingee: this is similar to tsekiyam_ ’s, but the big difference is in the image-volume formats00:53
patrickeastthingee: so like for the cache, we download the image run it through qemu-whatever and get raw data on the volume00:53
patrickeastthingee: we then clone that and stick it in the cache00:54
*** diogogmt has joined #openstack-cinder00:54
patrickeastthingee: if you are the glance backend you might have a volume but the data is actually qcow or some other compressed info00:54
patrickeastthingee: so you can’t just directly clone it and run00:54
*** annashen has joined #openstack-cinder00:54
patrickeastthingee: in the case where it is a raw image with no container they function almost identical sans the size limiations and stuff the cache adds00:55
thingeepatrickeast: thanks for the explanation, that helps00:55
thingeepatrickeast: yeah I don't see a use case difference in there :P00:55
*** jbernard has quit IRC00:56
patrickeastthingee: haha00:56
*** jbernard has joined #openstack-cinder00:56
thingeepatrickeast, tsekiyam_: here's my worry about these two efforts. We're going to focus on supporting both, people deploy the different options which effective do the same. we now have to support both00:56
thingeepatrickeast, tsekiyam_ unless there is a clear reason to have both, I'm unsure why we would want to continue to go down this path00:57
thingeepatrickeast: is your patch ready to go for testing?00:57
patrickeastthingee: yea pretty much00:58
thingeepatrickeast: excellent going to spin up an environment now with it. I'm assuming it's fine for me to continue with lvm?00:59
*** annashen has quit IRC00:59
thingeepatrickeast: if you don't mind, I'm going to reference this conversation on the ML to see if others have thoughts.01:00
thingeepatrickeast: I think your explanation is excellent to get people up to speed01:00
tsekiyam_thingee: One thing I can add to patrickeast's explanation is that my approach can skip uploading volume to Glance01:03
thingeeah that's true01:04
tsekiyam_thingee: if the image format is raw. It just register the cloned volume url so it is quite fast even if the volume size if 1TB.01:04
patrickeastthingee: tsekiyam_: i was actually playing with that with the cache too01:04
patrickeastfor any image format we can jus take a clone and toss it in the cache01:04
patrickeastas a backup if the glance code tsekiyam_ added couldn’t handle it01:05
patrickeastalthough i’m removing it in the next update (which is going up in the next few min if i can figure out how i broke this unit test)01:07
patrickeastthere are some things that would need to be reworked to make it work correctly01:07
*** lixiaoy1 has joined #openstack-cinder01:07
thingeetsekiyam_, patrickeast http://lists.openstack.org/pipermail/openstack-dev/2015-July/070714.html01:10
*** krypton has quit IRC01:11
patrickeastmaybe a topic for the mid-cycle01:11
*** krypton has joined #openstack-cinder01:11
thingeepatrickeast: oh there's a merge conflict with your patch01:13
*** Apoorva has quit IRC01:14
patrickeastthingee: tsekiyam_: good news… that bug doesn’t repro on kilo, bad news… not sure why01:14
patrickeastthingee: oh uh… one sec01:14
patrickeasti’ve been rebasing the stupid thing for like the last 4 hours01:14
patrickeastkeep getting distracted01:14
openstackgerritPatrick East proposed openstack/cinder: Generic image-volume cache  https://review.openstack.org/19579501:15
thingeepatrickeast: haha, you'll never get things merged in OpenStack with that attitude! should be rebasing once an hour at least01:15
openstackgerritTomoki Sekiyama proposed openstack/cinder: Efficient image download and upload for Glance Cinder backend  https://review.openstack.org/20175401:15
patrickeastone more rebase01:15
tsekiyam_patrickeast: ^ hope this fixes the rebase mistakes01:15
openstackgerritPatrick East proposed openstack/cinder: Generic image-volume cache  https://review.openstack.org/19579501:15
*** Guest1___ has joined #openstack-cinder01:16
*** krypton has quit IRC01:16
patrickeasttsekiyam_: did you test it?01:16
patrickeasttsekiyam_: thats different from how i fixed it XD01:17
patrickeastthingee: that latest one should be ok now, assuming the issue with create_volume got fixed01:17
*** jasondotstar|afk has quit IRC01:17
thingeepatrickeast: patch set 21 still has cherry pick issues01:18
*** davechen has joined #openstack-cinder01:18
*** Guest10__ has quit IRC01:18
patrickeastthingee: are you grabbing the dependant change too?01:19
thingeepatrickeast: of course not01:19
*** haomaiwang has quit IRC01:22
thingeepatrickeast: alright environment is going up01:23
*** haomaiwang has joined #openstack-cinder01:23
thingeepatrickeast: there's a beer with my name on it somewhere. I'll bb on later tonight if anyone wants to gossip on reviews01:24
*** davechen1 has joined #openstack-cinder01:24
patrickeastthingee: haha word, let me know if you run into issues that patchset01:24
thingeealso people keep in mind https://etherpad.openstack.org/p/cinder-liberty-2-reviews01:25
thingeebunch of stuff merged. and some stuff moved out of liberty-2 that just doesn't stand a chance IMO. Very likely image caching, migration will be in there too01:25
*** davechen has quit IRC01:25
*** davechen1 is now known as davechen01:26
davechenthingee: hi boss,01:26
*** vilobhmm has quit IRC01:26
thingeethese are just going to liberty-3 of course01:26
thingeedavechen: don't call me boss :P01:26
davechenthingee: yeah, I believe it's true.01:26
davechenthingee: some question from me about the cinder client01:27
* thingee makes an angry look at jungleboyj01:27
davechenthingee: I am not sure whether it's correct, so want your answer.01:27
thingeedavechen: so simply, I just think people should be able to get the image metadata.01:27
thingeeit's not intuitive IMO to expect people to know to list the details of a volume to get image metadata01:28
davechenthingee: yes, currently, we have the API already in the extenstions.01:28
thingeedavechen: excellent, then we're done on that part :)01:28
davechenthingee: but it's used by cinder volume show/list command.01:28
davechenthingee: if we explictly want a API to do that, should we do some refacoring?01:29
*** jwcroppe has quit IRC01:30
*** Yogi1 has quit IRC01:30
davechensince, these API has been defined, and I assume it's been designed as this, not sure it's okay to that? change the meaning of these existing API or just expose somem new API to do that?01:30
*** jwcroppe has joined #openstack-cinder01:30
thingeeI'm just looking at things like cinder metadata-show and cinder snapshot-metadata-show. Those all have cinder metadata set or cinder snapshot-metadata set ... so it just seems like it would be expected we'd have a cinder image-metadata-show01:31
thingeeif we're effectively adding in cinder image-metadata set01:31
davechenthingee: so, we need some new API exposed in the cinder side?01:31
thingeedavechen: yes01:31
thingeeI don't know what other's think though01:32
thingeethat's just me thinking out loud of what's obvious to me01:32
thingeedavechen: I guess I'm saying take my advice, someone might disagree when they have time to review things :P01:33
davechenthat should be okay, I not sure other's opinion as well, so there will be two different APIs, one for cinder lists/show and one for cinder image-metadata stuff?01:33
thingeedavechen: but I'll argue my reason for having you do it.01:33
thingeedavechen: if you'll excuse me, I was about to step out for dinner.01:34
davechenthingee: okay, excellent, both of them should be in the some module, I meant both of the APIs should be defined in volume_image_metadata extension, right?01:34
davechenthingee: of course01:34
*** jwcroppe has quit IRC01:35
davechenthingee: I am going to propose a new patch both in the cinder cli and cinder API, it that okay?01:35
davechenthingee: so we can let that patch go at first?01:35
thingeeyeah it looks like you can define an index method in the volume_image_metadata extension01:37
thingeejust like how volume_metadata and snapshot_metadata define the index method01:37
thingeedavechen: https://github.com/openstack/cinder/blob/master/cinder/api/v2/volume_metadata.py#L4101:37
thingeebut yes, the volume_image_metadata extension change, and update to the cinder client will be correct01:38
thingeeI'll bbl01:38
thingeedavechen: thanks01:38
openstackgerritYuanHui Xu proposed openstack/cinder: Add mock cases for IBM FlashSystem  https://review.openstack.org/20537101:38
davechenthingee: okay, got you.01:38
davechenthingee: can we do these in a seperate review?01:39
davechenthingee: we need at least two patches for that.01:40
davechenthingee: one for cinder api change/ one for cinder client change, am i right?01:40
*** mylu has joined #openstack-cinder01:41
*** Lee1092 has joined #openstack-cinder01:42
*** leeantho has quit IRC01:43
*** salv-orlando has joined #openstack-cinder01:44
*** salv-orlando has quit IRC01:48
openstackgerritwanghao proposed openstack/cinder: Implement function to manage/unmanage snapshots  https://review.openstack.org/14459001:51
*** abhi has quit IRC01:52
*** abhi has joined #openstack-cinder01:53
*** annashen has joined #openstack-cinder01:55
*** mylu has quit IRC01:56
openstackgerritxing-yang proposed openstack/cinder: Add multiple pools support to VMAX driver  https://review.openstack.org/18999002:00
*** annashen has quit IRC02:00
*** tsekiyam_ has quit IRC02:03
*** mtanino_ has quit IRC02:06
*** raunak has quit IRC02:09
*** jasondotstar|afk has joined #openstack-cinder02:12
*** annashen has joined #openstack-cinder02:17
*** sghanekar_ has quit IRC02:17
*** zhenguo has joined #openstack-cinder02:19
*** bkopilov has quit IRC02:35
*** dims has quit IRC02:35
*** changbl has quit IRC02:37
*** Guest1___ has quit IRC02:38
*** zhonghua-lee has joined #openstack-cinder02:38
*** mragupat has joined #openstack-cinder02:40
*** mylu has joined #openstack-cinder02:44
*** mylu has quit IRC02:45
*** daneyon_ has quit IRC02:47
openstackgerritwanghao proposed openstack/cinder: Incremental backup improvements for L  https://review.openstack.org/18832802:50
*** mragupat has quit IRC02:51
*** hakimo_ has joined #openstack-cinder02:52
*** patrickeast has quit IRC02:54
*** hakimo has quit IRC02:54
*** annashen_ has joined #openstack-cinder02:56
openstackgerritwanghao proposed openstack/cinder: Fixing notify message of manager's manage_existing flow  https://review.openstack.org/20589602:58
*** annashen_ has quit IRC03:01
*** Apoorva has joined #openstack-cinder03:07
*** krypton has joined #openstack-cinder03:09
*** BharatK has joined #openstack-cinder03:19
*** jasondotstar|afk has quit IRC03:24
*** diogogmt has quit IRC03:24
*** DericHorn-HP has joined #openstack-cinder03:25
*** jwcroppe has joined #openstack-cinder03:30
*** edtubill has joined #openstack-cinder03:30
*** edtubill has left #openstack-cinder03:31
*** coolsvap|away is now known as coolsvap03:32
*** annashen has quit IRC03:45
*** julim has quit IRC03:46
*** avishay has joined #openstack-cinder03:53
*** vilobhmm has joined #openstack-cinder03:55
*** annashen has joined #openstack-cinder03:57
*** annashen has quit IRC04:01
*** bkopilov has joined #openstack-cinder04:02
*** jwcroppe has quit IRC04:02
*** patrickeast has joined #openstack-cinder04:03
*** jwcroppe has joined #openstack-cinder04:03
*** markvoelker has quit IRC04:04
*** jwcroppe_ has joined #openstack-cinder04:06
*** jwcroppe has quit IRC04:08
*** daneyon has joined #openstack-cinder04:09
*** BharatK has quit IRC04:15
*** jwcroppe_ has quit IRC04:16
*** jwcroppe has joined #openstack-cinder04:17
*** jwcroppe has quit IRC04:22
*** daneyon has quit IRC04:22
openstackgerritApoorva Deshpande proposed openstack/cinder: Fix order of arguments in assertEqual  https://review.openstack.org/20629604:25
*** Apoorva has quit IRC04:32
*** harlowja_at_home has joined #openstack-cinder04:32
*** BharatK has joined #openstack-cinder04:33
openstackgerritwanghao proposed openstack/cinder: Add pagination to backups  https://review.openstack.org/20449304:33
*** DericHorn-HP has quit IRC04:37
*** krypton has quit IRC04:37
*** BharatK has quit IRC04:38
*** sgotliv has joined #openstack-cinder04:38
*** DericHorn-HP has joined #openstack-cinder04:41
*** BharatK has joined #openstack-cinder04:43
*** DericHorn-HP has quit IRC04:46
*** BharatK has quit IRC04:47
*** BharatK has joined #openstack-cinder04:49
*** sgotliv has quit IRC04:50
*** abehl has joined #openstack-cinder04:50
*** deepakcs has joined #openstack-cinder04:51
*** shyama has joined #openstack-cinder04:54
*** krypton has joined #openstack-cinder04:54
*** patrickeast_ has joined #openstack-cinder04:54
*** jwcroppe has joined #openstack-cinder04:56
*** patrickeast has quit IRC04:56
*** patrickeast_ is now known as patrickeast04:56
*** jwcroppe has quit IRC04:58
*** annashen has joined #openstack-cinder04:58
*** jwcroppe has joined #openstack-cinder04:58
*** jwcroppe_ has joined #openstack-cinder05:01
*** annashen has quit IRC05:02
*** jwcroppe has quit IRC05:03
*** markvoelker has joined #openstack-cinder05:05
lixiaoy1: thingee: I answered your comments in https://review.openstack.org/#/c/199846/, please have a look05:05
*** ub has quit IRC05:06
*** markvoelker has quit IRC05:10
*** e0ne has joined #openstack-cinder05:14
*** DericHorn-HP has joined #openstack-cinder05:21
*** harlowja_at_home has quit IRC05:32
*** shyama has quit IRC05:35
*** e0ne has quit IRC05:42
*** DericHorn-HP has quit IRC05:43
*** e0ne has joined #openstack-cinder05:44
*** e0ne has quit IRC05:51
*** jamielennox is now known as jamielennox|away05:52
*** rushiagr_away is now known as rushiagr05:55
*** salv-orlando has joined #openstack-cinder05:55
*** e0ne has joined #openstack-cinder05:55
*** agarciam has joined #openstack-cinder05:56
*** jamielennox|away is now known as jamielennox05:56
*** gaurangt has joined #openstack-cinder05:58
*** annashen has joined #openstack-cinder05:59
*** albertom has quit IRC06:01
*** annashen has quit IRC06:03
*** harlowja_at_home has joined #openstack-cinder06:04
*** vlaza has joined #openstack-cinder06:06
*** krypton_ has joined #openstack-cinder06:08
*** albertom has joined #openstack-cinder06:09
*** krypton has quit IRC06:11
*** krypton_ has quit IRC06:12
*** vilobhmm has quit IRC06:13
*** e0ne has quit IRC06:13
*** shyama has joined #openstack-cinder06:14
*** chlong has quit IRC06:17
*** patrickeast has quit IRC06:22
*** harlowja_at_home has quit IRC06:23
*** belmoreira has joined #openstack-cinder06:25
deepakcsthingee: https://review.openstack.org/#/c/150511/4/specs/liberty/standard-capabilities.rst06:29
deepakcsthingee: reg. the Q on why snapshot is not considered as a capability, was curious to know your response to that.06:29
*** ub has joined #openstack-cinder06:33
*** nkrinner has joined #openstack-cinder06:42
*** afazekas has joined #openstack-cinder06:44
-openstackstatus- NOTICE: zuul is stuck and about to undergo an emergency restart, please be patient as job results may take a long time06:44
*** ChanServ changes topic to "zuul is stuck and about to undergo an emergency restart, please be patient as job results may take a long time"06:44
*** jwcroppe_ has quit IRC06:48
*** markus_z has joined #openstack-cinder06:59
*** salv-orlando has quit IRC06:59
*** annashen has joined #openstack-cinder07:00
*** annashen has quit IRC07:05
*** markvoelker has joined #openstack-cinder07:06
*** markvoelker has quit IRC07:11
openstackgerritDave Chen proposed openstack/cinder: Role based properties protection  https://review.openstack.org/16009007:12
openstackgerritliuxinguo proposed openstack/cinder: Add iSCSI multipath support for Huawei driver  https://review.openstack.org/18836507:19
*** ishant__ has joined #openstack-cinder07:21
*** sgotliv has joined #openstack-cinder07:23
openstackgerritDave Chen proposed openstack/cinder: Policies based properties protection  https://review.openstack.org/16014507:29
*** shyama has quit IRC07:34
*** lpetrut has joined #openstack-cinder07:34
*** e0ne has joined #openstack-cinder07:37
*** salv-orlando has joined #openstack-cinder07:50
*** coolsvap is now known as coolsvap|away07:57
*** agarciam has quit IRC07:57
openstackgerritliuxinguo proposed openstack/cinder: Add iSCSI multipath support for Huawei driver  https://review.openstack.org/18836507:57
*** sgotliv has quit IRC07:58
*** agarciam has joined #openstack-cinder08:00
*** annashen has joined #openstack-cinder08:01
*** abehl has quit IRC08:01
*** ChanServ changes topic to "The OpenStack Block Storage Project Cinder | The New Kids On the Block | https://wiki.openstack.org/wiki/Cinder"08:01
-openstackstatus- NOTICE: zuul has been restarted and queues restored. It may take some time to work through the backlog.08:01
*** jwcroppe has joined #openstack-cinder08:02
openstackgerritSzymon Wróblewski proposed openstack/cinder: POC: Tooz locks demo  https://review.openstack.org/18564608:02
openstackgerritSzymon Wróblewski proposed openstack/cinder: Tooz locks  https://review.openstack.org/18353708:02
*** jistr has joined #openstack-cinder08:03
*** annashen has quit IRC08:05
*** jordanP has joined #openstack-cinder08:07
*** kjelly has joined #openstack-cinder08:12
*** ndipanov has quit IRC08:12
kjellyHi guys, does anyone know why the condition is "project_only and is_user_context(context)" instead of "project_only"?   https://github.com/openstack/cinder/blame/master/cinder/db/sqlalchemy/api.py#L25708:13
*** sgotliv has joined #openstack-cinder08:20
*** shyama has joined #openstack-cinder08:22
*** bluex-pl has joined #openstack-cinder08:29
*** bluex-pl has quit IRC08:29
*** bluex-pl has joined #openstack-cinder08:30
*** alexpilotti has joined #openstack-cinder08:33
dulekkjelly: I guess it's for the situation when we have project_only and user isn't admin.08:33
dulekkjelly: User can see only resources from his own project.08:33
dulekkjelly: And admin can see all of the resources.08:34
dulekkjelly: I think that's why the filter is added here.08:34
*** agarciam has quit IRC08:35
*** agarciam has joined #openstack-cinder08:36
*** ub has quit IRC08:36
*** ub has joined #openstack-cinder08:39
openstackgerritwanghao proposed openstack/cinder: Enhance deletion efficiency when backup init host  https://review.openstack.org/20635908:41
*** aix has joined #openstack-cinder08:45
kjellyIf admin want to see all of the resources, we can just set project_only to false to get all data.08:45
*** laughterwym has quit IRC08:45
kjellydulek: If admin want to see all of the resources, we can just set project_only to false to get all data.08:45
kjellydulek: "project_only and is_user_context(context)" means that admin always get all resources.08:46
dulekkjelly: :param project_only: if present and context is user-type, then restrict query to match the context's project_id.08:47
dulekkjelly: I guess that's the point of project_only param.08:48
kjellydulek: I am working on the bug. https://bugs.launchpad.net/cinder/+bug/1477625 . the bug happened because cinder will add admin role to contxt , so user can show any volumes.08:52
openstackLaunchpad bug 1477625 in Cinder "DB access to show volumes may not be properly controlled" [Undecided,Confirmed] - Assigned to jelly (coding1314)08:52
kjellydulek: If the condition is "project_only" instead of "project_only and is_user_context(context)", we don't have the problem because we just get the volume in project.08:53
dulekkjelly: Ah, I get it. You're probably right...08:55
kjellydulek: If the condition is "project_only" and admin want to get all resources, we can just pass project_only to false to get all volumes.08:55
*** coolsvap|away is now known as coolsvap08:56
dulekkjelly: I think there's --all-tenants param.08:56
*** sgotliv has quit IRC08:59
kjellydulek: By the way, to solve the bug 1477625, maybe we can add "viewable_admin_meta" param to db.volume_get instead of using context.elevated.       https://github.com/openstack/cinder/commit/0505bb268942534ad5d6ecd5e34a4d9b0e7f5c04#diff-33011bbeb0dd51161827701307e6d033R26509:01
openstackbug 1477625 in Cinder "DB access to show volumes may not be properly controlled" [Undecided,Confirmed] https://launchpad.net/bugs/1477625 - Assigned to jelly (coding1314)09:01
*** annashen has joined #openstack-cinder09:01
*** jwcroppe has quit IRC09:02
openstackgerritPradeep Sathasivam proposed openstack/cinder: Adds VF support with HTTPS southbound connector  https://review.openstack.org/18479909:02
dulekkjelly: This also seems possible to me.09:04
*** ub has quit IRC09:06
*** annashen has quit IRC09:06
*** ub has joined #openstack-cinder09:07
*** markvoelker has joined #openstack-cinder09:07
*** boris-42 has quit IRC09:10
*** markvoelker has quit IRC09:11
*** ndipanov has joined #openstack-cinder09:12
*** jwcroppe has joined #openstack-cinder09:16
*** jwcroppe has quit IRC09:16
mkodererthingee: I put my comment on jgriffith rework (https://review.openstack.org/#/c/201812/)09:19
mkodererthingee: I planed to have a small design session with john.. but unfortunately currently I am a bit busy with other tasks09:20
*** agarciam has quit IRC09:22
*** agarciam has joined #openstack-cinder09:22
*** agarciam has quit IRC09:23
*** abhishekk has joined #openstack-cinder09:24
*** jasondotstar|afk has joined #openstack-cinder09:29
*** agarciam has joined #openstack-cinder09:35
*** sasi has joined #openstack-cinder09:36
*** sasi has left #openstack-cinder09:37
*** jwcroppe has joined #openstack-cinder09:38
*** jwcroppe has quit IRC09:42
*** lixiaoy1 has quit IRC09:44
*** davechen has left #openstack-cinder09:52
openstackgerritKuo-tung Kao proposed openstack/cinder: non-admin user don't get volumes in other project  https://review.openstack.org/20645009:52
*** kjelly has quit IRC09:53
*** kjelly has joined #openstack-cinder09:53
*** kjelly has quit IRC09:54
*** jasondotstar|afk has quit IRC09:54
*** aix has quit IRC09:55
*** Zhongjun has joined #openstack-cinder09:57
*** ronis has joined #openstack-cinder09:59
*** annashen has joined #openstack-cinder10:02
*** jwcroppe has joined #openstack-cinder10:04
*** annashen has quit IRC10:07
*** jasondotstar|afk has joined #openstack-cinder10:12
openstackgerritLei Zhang proposed openstack/cinder: Make concurrent backups for same volume impossible  https://review.openstack.org/20291410:16
*** jwcroppe has quit IRC10:19
*** aix has joined #openstack-cinder10:22
*** dims has joined #openstack-cinder10:31
*** salv-orl_ has joined #openstack-cinder10:34
*** salv-orlando has quit IRC10:37
*** heyun has quit IRC10:44
*** gaurangt has left #openstack-cinder10:44
*** salv-orl_ has quit IRC10:47
*** timcl has joined #openstack-cinder10:51
*** ankit_ag has joined #openstack-cinder10:55
*** avishay_ has joined #openstack-cinder10:56
*** avishay has quit IRC10:57
*** salv-orlando has joined #openstack-cinder11:00
*** salv-orl_ has joined #openstack-cinder11:01
*** annashen has joined #openstack-cinder11:03
*** salv-orlando has quit IRC11:05
*** ishant__ has quit IRC11:05
*** annashen has quit IRC11:07
*** markvoelker has joined #openstack-cinder11:08
*** markvoelker has quit IRC11:13
openstackgerritliuxinguo proposed openstack/cinder: Add iSCSI multipath support for Huawei driver  https://review.openstack.org/18836511:14
*** zhonghua-lee has quit IRC11:21
*** marcusvrn has joined #openstack-cinder11:21
*** sgotliv has joined #openstack-cinder11:22
*** breitz1 has joined #openstack-cinder11:25
*** breitz has quit IRC11:25
openstackgerritliuxinguo proposed openstack/cinder: Add iSCSI multipath support for Huawei driver  https://review.openstack.org/18836511:28
openstackgerritliuxinguo proposed openstack/cinder: Add iSCSI multipath support for Huawei driver  https://review.openstack.org/18836511:35
*** gaurangt has joined #openstack-cinder11:36
*** abehl has joined #openstack-cinder11:36
*** coolsvap is now known as coolsvap|away11:37
*** gaurangt has left #openstack-cinder11:39
openstackgerritBharat Kumar Kobagana (BharatK) proposed openstack/cinder: GlusterFS backup driver  https://review.openstack.org/18225411:40
openstackgerritBharat Kumar Kobagana (BharatK) proposed openstack/cinder: Posix backup driver  https://review.openstack.org/16364711:40
*** samueldmq has joined #openstack-cinder11:41
samueldmqI need to discover identity url from inside cinder11:41
samueldmqI have two options, i) look at the service catalog of the token or ii) look at the config encryption_auth_url11:42
samueldmqis that config always set properly ?11:42
*** bkopilov has quit IRC11:45
*** jasondotstar|afk has quit IRC11:46
*** jwcroppe has joined #openstack-cinder11:51
openstackgerritJordan Pittier proposed openstack/os-brick: Remotefs: add ScalityFS support  https://review.openstack.org/19663111:53
openstackgerritBharat Kumar Kobagana (BharatK) proposed openstack/cinder: GlusterFS: Using 'fallocate' instead of 'dd' to create raw volumes  https://review.openstack.org/16452711:55
*** annashen has joined #openstack-cinder12:04
*** agarciam has quit IRC12:08
*** agarciam has joined #openstack-cinder12:08
*** markvoelker has joined #openstack-cinder12:09
*** annashen has quit IRC12:09
*** Zhongjun has quit IRC12:12
*** markvoelker has quit IRC12:14
openstackgerritBharat Kumar Kobagana (BharatK) proposed openstack/cinder: GlusterFS: 'fallocate' instead of 'dd' to create raw volumes  https://review.openstack.org/16452712:18
*** jwcroppe has quit IRC12:19
*** jwcroppe has joined #openstack-cinder12:20
*** laughterwym has joined #openstack-cinder12:21
*** akerr has joined #openstack-cinder12:24
*** eharney has joined #openstack-cinder12:28
*** markvoelker has joined #openstack-cinder12:29
*** edmondsw has joined #openstack-cinder12:33
*** bswartz has quit IRC12:35
*** jasondotstar|afk has joined #openstack-cinder12:37
*** timcl has quit IRC12:38
DuncanTsamueldmq: no that property is not always set.12:38
*** timcl has joined #openstack-cinder12:39
DuncanTsamueldmq: config rather. You should look in the context catalogue12:39
*** BharatK has quit IRC12:42
*** zhenguo has quit IRC12:42
*** coolsvap|away is now known as coolsvap12:45
samueldmqDuncanT: nice, I was discussing with jamielennox in #keystone about that.. actually the ksmiddleware already instantiates the ksclient plugin to check if tokens are valid, etc12:48
samueldmqDuncanT: after that, it passes the plugin down as a env var, that could be re-used by cinder12:49
samueldmqDuncanT: nova already does that12:49
samueldmqDuncanT: this is how middleware passes it down  : https://github.com/openstack/keystonemiddleware/blob/master/keystonemiddleware/auth_token/__init__.py#L16412:49
samueldmqDuncanT: and nova code using it : https://github.com/openstack/nova/blob/master/nova/context.py#L3712:49
*** jasondotstar|afk is now known as jasondotstar12:49
samueldmqDuncanT: basically to not re-create the code for parsing the catalog, etc :)12:50
*** changbl has joined #openstack-cinder12:50
DuncanTsamueldmq: patches welcome. What do you want the keystone server for?12:52
DuncanTsamueldmq: backup code already looks in the catalogue to find swift btw12:53
samueldmqDuncanT: hmm, do you have a pointer to that code12:54
samueldmqDuncanT: actually we are implementing hierarchical quotas (cc ericksonsantos)12:54
samueldmqDuncanT: so we need to know the project hierarchy12:54
*** julim has joined #openstack-cinder12:55
samueldmqDuncanT: (asking keystone server)12:55
*** changbl has quit IRC12:55
DuncanTsamueldmq: got you. Give me two minutes to find that code12:56
*** BharatK has joined #openstack-cinder12:56
*** rushiagr is now known as rushiagr_away12:56
samueldmqDuncanT: sure :)12:57
*** merooney has joined #openstack-cinder12:57
openstackgerritSzymon Wróblewski proposed openstack/cinder: Tooz locks  https://review.openstack.org/18353713:00
*** afazekas has quit IRC13:02
*** annashen has joined #openstack-cinder13:05
DuncanTsamueldmq: backup/drivers/swift.py:13013:06
DuncanTsamueldmq: Fixing the barbican plugin so that is doesn't need the keystone url in the config (and also so that it can pick the barbican url out of the catalog) would be nice too13:07
openstackgerritEric Harney proposed openstack/cinder: Fix update_migrate_volume() exception handling  https://review.openstack.org/19873113:08
*** annashen has quit IRC13:09
openstackgerritEric Harney proposed openstack/cinder: Move update_migrated_volume() to BaseVD  https://review.openstack.org/19873113:10
*** xyang1 has joined #openstack-cinder13:12
openstackgerritEric Harney proposed openstack/cinder: Log which service is down  https://review.openstack.org/20621113:12
*** edmondsw has quit IRC13:14
*** xyang has joined #openstack-cinder13:14
*** ganso_ has joined #openstack-cinder13:15
*** bswartz has joined #openstack-cinder13:15
*** kjelly_ has joined #openstack-cinder13:17
*** kjelly_ is now known as kjelly13:17
*** xyang1 has quit IRC13:17
*** mylu has joined #openstack-cinder13:19
*** rushiagr_away is now known as rushiagr13:21
*** Zhongjun has joined #openstack-cinder13:21
openstackgerritSzymon Wróblewski proposed openstack/cinder: Tooz locks  https://review.openstack.org/18353713:22
*** Yogi1 has joined #openstack-cinder13:24
bluex-plgeguileo: hi13:26
DuncanTeharney: Thanks for a review I'm not in any way conflicted or confused about, I needed that13:26
geguileobluex-pl: hi13:26
eharneyDuncanT: glad to help :)13:26
bluex-pli just finished reading your blog post and looking through poc patches, very interesting13:28
*** xyang1 has joined #openstack-cinder13:28
bluex-plI would like to propose merging https://review.openstack.org/#/c/205840/ into https://review.openstack.org/#/c/183537/ and adding you as co-author13:29
geguileobluex-pl: Sounds good to me13:30
ericksonsantosDuncanT, thanks for point it out.  Since encryption_auth_url conf is not always set, I think we really need to get the keystone url from the service catalog.13:30
geguileobluex-pl: We would need to add tests for the code in my patch though13:30
bluex-plgeguileo: sure, consider it done13:30
geguileobluex-pl: Awesome!13:30
*** akshai has joined #openstack-cinder13:31
geguileobluex-pl: By the way, I really liked your Tooz patch  :-)13:31
bluex-plgeguileo: Thanks ^^13:31
DuncanTericksonsantos: Sounds like the right answer to me. I'm just working on a patch to remove the requirement for the barbican service URL from the plugin, I'll keep an eye out for your patch and consider using whatever technique you guys use to get the keystone url to remove the need for encryption_auth_url too13:32
dulekgeguileo: Do you think that a part of your approach can get into L?13:33
ericksonsantosDuncanT, nice :)13:33
geguileodulek: Depends on whether we decide is the right approach or not and who else would be collaborating13:34
dulekgeguileo: I'm not sure if you saw my spec - https://review.openstack.org/#/c/202615/13:34
geguileodulek: No, I didn't :(13:34
dulekgeguileo: I think it is compatible with your approach, just describes first steps - that is locks.13:34
geguileodulek: I have been in EuroPython13:34
*** abehl has quit IRC13:35
geguileodulek: I'll read your spec13:35
dulekgeguileo: Thanks :)13:35
geguileodulek: Sorry I didn't see it earlier and thanks for pointing it out  :)13:35
dulekgeguileo: I have also some questions on approach described in the blogpost.13:35
geguileodulek: Shoot13:36
smcginnishemna: Congrats on the nova os-brick patch!13:36
dulekgeguileo: For example - how do you plan to take LVM-like drivers?13:37
dulekgeguileo: When scheduler decides where the request will go.13:37
geguileoWell, in theory you cannot have A-A on LVM, right?13:37
geguileoSorry, stupid comment13:38
dulekgeguileo: Right. Sorry, I'll get back to you in a moment.13:39
geguileodulek: The scheduler will decide on a cluster, not a specific node, right?13:39
*** sgotliv_ has joined #openstack-cinder13:40
*** sgotliv has quit IRC13:41
samueldmqDuncanT: sorry I was holding another conversation in #keystone13:41
samueldmqDuncanT: thanks for the links :)13:41
DuncanTsamueldmq: You're welcome13:42
*** bkopilov has joined #openstack-cinder13:45
*** deepakcs has quit IRC13:47
*** abehl has joined #openstack-cinder13:48
*** thangp has joined #openstack-cinder13:50
*** belmoreira has quit IRC13:51
*** edmondsw has joined #openstack-cinder13:51
openstackgerritCurt Bruns proposed openstack/os-brick: WIP: Set executor to Putils  https://review.openstack.org/20654913:59
openstackgerritCurt Bruns proposed openstack/os-brick: WIP: Set executor to Putils  https://review.openstack.org/20654914:01
dulekgeguileo: I'm here again, sorry - it's always multiple people want something when I'm talking.14:03
geguileodulek: Jaja, no problem :)14:03
dulekgeguileo: We probably want to have scheduler to decide on particular node.14:03
dulekgeguileo: So it will use the filters.14:04
geguileodulek: I disagree14:04
dulekgeguileo: I mean in case of LVM.14:04
geguileodulek: In my opinion scheduler selects backend14:04
geguileodulek: And he doesn't care which node of the cluster does the operation14:04
geguileodulek: Since the backend is the same14:05
*** abhishekk has quit IRC14:05
geguileodulek: That's the whole idea of Active-Active, right?14:05
geguileodulek: Exact same node14:05
dulekgeguileo: That would be a big change to how LVM driver works now.14:05
geguileodulek: That can work simultaneously14:05
geguileodulek: No, I mean14:06
*** annashen has joined #openstack-cinder14:06
geguileodulek: If you are using LVM with local storage (as with devstack)14:06
geguileodulek: Then you cannot have HA Active-Active14:06
dulekgeguileo: I think that scheduler selects LVM hosts basing for example on information if host is able to fit the volume in.14:06
geguileodulek: Because the storage is local to the node14:06
geguileodulek: That is correct14:06
geguileodulek: But it's not that the host is able to fit the volume in14:06
geguileodulek: But that the storage is able to fit it14:07
geguileodulek: The node is just an intermediary14:07
*** ankit_ag has quit IRC14:08
dulekgeguileo: Then how can you tell if node has enough free space to create a volume?14:08
dulekgeguileo: Right now I think scheduler decides that.14:08
geguileodulek: Yes, but based on information sent from the nodes, right?14:09
dulekgeguileo: That's right.14:09
geguileodulek: Then what's the difference?14:09
geguileodulek: Nodes will keep updating that information14:09
geguileodulek: But it will receive N times as many refreshes as it does now (N being the number of active nodes in the cluster)14:10
*** boris-42 has joined #openstack-cinder14:10
geguileodulek: Since each node will be sending that information from the same backend14:10
dulekgeguileo: It would need to aggregate them...14:10
geguileodulek: Why?14:10
geguileodulek: The free space is not the sum of those14:10
*** annashen has quit IRC14:11
geguileodulek: It's the latest one that counts14:11
dulekgeguileo: Because every LVM node has only knowledge of it's own capabilities I think.14:11
geguileodulek: They are all going to the same storage backend14:11
dulekgeguileo: It doesn't have the view of the whole cluster.14:11
geguileodulek: The thing is that if LVM is set to local storage (you store data in local HD) you cannot have A-A14:11
dulekgeguileo: That's right, I just want to think how to integrate that. So I guess we just have LVMs as different hosts as we have now?14:12
geguileodulek: Yes14:13
geguileodulek: And they can be configured as A-P but not A-A14:13
dulekgeguileo: Okay, so we still have host option in the config, nothing changes there.14:13
dulekgeguileo: Cool.14:13
geguileodulek: Nothing changes, we use the same host configuration as we do now14:14
geguileodulek: But you set the same host for nodes from the same cluster14:14
dulekgeguileo: Got it!14:14
geguileodulek: And they must access the exact same backend14:14
dulekgeguileo: I've also took a look on how to detect if oslo.messaging disconnected from RabbitMQ.14:15
*** msno has joined #openstack-cinder14:15
dulekgeguileo: That's unfortunately hard in current state.14:15
geguileodulek: Really?  :(14:15
msnowhat should be vendor_name for cinder.volume.drivers.emc.emc_cli_iscsi.EMCCLIISCSIDriver for cinder in tempest.conf14:15
msnoi tried vendor_name=EMC but my test fails14:16
dulekgeguileo: I think so, but maybe deeper investigation should be done. Anyway just for you to know. ;)14:16
geguileodulek: Thanks for looking into it, it's good to know14:16
dulekgeguileo: And just a nit - I guess it's "for the sake of", and not "for the shake of". ;)14:18
geguileodulek: XD XD XD14:18
geguileodulek: Thanks14:18
geguileodulek: I'll fix it :)14:18
openstackgerritThang Pham proposed openstack/cinder: Conversion to volume object  https://review.openstack.org/17705414:19
dulekgeguileo: I wonder if we're able to go on with merging Tooz locks into Liberty, and then follow your approach in M.14:20
dulekgeguileo: BPs deadline is this Friday I think.14:20
geguileodulek: I think that's reasonable14:20
geguileodulek: And if we have locks for manager and drivers it will be big step14:21
*** Tross has quit IRC14:21
*** tsekiyama has joined #openstack-cinder14:21
dulekgeguileo: Ah, probably one more step is needed - reworking https://github.com/openstack/cinder/blob/master/cinder/volume/manager.py#L319-L381 to work in distributed version.14:23
dulekgeguileo: Am I right?14:23
geguileodulek: Yes, there are patches for that ;-)14:24
geguileodulek: https://review.openstack.org/#/c/205838/14:24
*** annegentle has joined #openstack-cinder14:26
dulekgeguileo: Is this strongly tied with other of your patches or it can be extracted and used on it's own?14:26
geguileodulek: That's a good question.  XD14:27
geguileodulek: I think that one has only 2 dependencies14:28
*** esker has joined #openstack-cinder14:28
geguileodulek: But I'd have to check if it would work without the other stuff14:28
openstackgerritEric Harney proposed openstack/cinder: Move retype() to BaseVD  https://review.openstack.org/20655914:29
*** jaypipes has joined #openstack-cinder14:29
dulekgeguileo: But the dependencies are all this queued statuses stuff, which may be a little harder to push through.14:29
eharneyi'd appreciate if folks could review https://review.openstack.org/#/c/206559/ and its dep -- this is really breaking some NFS-ish drivers14:30
geguileodulek: Well, the only reason I needed to change the init_host method in the first place is because of the change in the ACK14:30
*** Yogi11 has joined #openstack-cinder14:30
dulekgeguileo: Is it? In case of A/A when starting a c-vol we don't know if volume is processed by some other c-vol or orphaned.14:32
dulekgeguileo: Or am I wrong (hopefully?)?14:32
*** vlaza has quit IRC14:32
geguileodulek: You are right14:32
geguileodulek: But if you remove the stuck fixing from there, where do you fix them?14:34
*** Yogi1 has quit IRC14:34
geguileodulek: And how will you be able to tell apart a stuck resource from one that just got queued by the API?14:34
tbarrongeguileo: dulek: I want to read https://review.openstack.org/#/c/205838/ today myself, as I'm also looking to fix backup init_host for issues of the same kind.14:35
tbarronor similar at least.14:35
geguileotbarron: I'm not sure if that patch will be helpful for current code14:37
dulekgeguileo: Ah, right, we cannot just acquire the lock in init_host and do cleanups - action might been just scheduled by the API and is inside RabbitMQ, without any c-vol processing yet.14:38
tbarrongeguileo: I figured that might be the case, but it will help expand my "conceptual repetoire" at least :-)14:38
geguileodulek: Precisely!!14:38
geguileodulek: That's why I introduced those new statuses14:38
geguileotbarron: Cool14:39
dulekgeguileo: So this statuses actually mean "on the way to c-vol"?14:40
geguileodulek: Yes14:41
geguileodulek: It means the operation has been requested, and once they are received by a volume node and they have the locks they can change them to *ing state14:41
dulekgeguileo: Can the c-vol drop the operation intentionally and let it go back to RabbitMQ?14:43
dulekgeguileo: Or will it always lock it - which basically means queuing on greenthreads.14:43
geguileodulek: It cannot let it go back to the broker14:43
geguileodulek: Because for that to hapen it would need to drop the whole connection to the broker14:44
geguileodulek: RabbitMQ support requeuing, but other brokers don't14:44
dulekgeguileo: Okay, got it14:44
geguileodulek: We could always send a new request from volume node14:44
geguileodulek: But I think that's kind of "dirty"14:45
dulekgeguileo: Not too elegant idea.14:45
dulekgeguileo: exactly14:45
geguileodulek: I agree14:45
dulekgeguileo: This seems like it's really well thought, but it seems that minimal set of improvements is: Tooz locks + post-action acks + "about to be processed" statuses + cleanup done with locks.14:47
dulekgeguileo: Anything else?14:47
geguileodulek: Races in the API nodes14:48
geguileodulek: Small change to how we report service status to the DB14:49
geguileodulek: And maybe I should have a look at the scheduler interactions you mentioned to confirm everything is OK14:49
*** erlon has joined #openstack-cinder14:49
dulekgeguileo: It should be I think...14:49
dulekgeguileo: Okay, so it's like all-or-nothing here.14:50
geguileodulek: Well, I think tooz locks can be implemented regardless of everything else14:50
geguileodulek: Since we are going to need them anyway, regardless of the solution14:50
dulekgeguileo: That's true, but are they adding any value alone?14:50
geguileodulek: Well, for one we would add all locks that are missing14:51
*** Adriano__ has joined #openstack-cinder14:51
geguileodulek: Since we would be going through all operations14:51
geguileodulek: And if we get that in L it means it will better tested when we start working on the rest of the stuff14:52
dulekgeguileo: That's a value, but can be achieved without Tooz.14:52
*** jwcroppe has quit IRC14:52
dulekgeguileo: I'm looking for a rationale here.14:52
*** jwcroppe has joined #openstack-cinder14:52
geguileodulek: Yes, but if we add Tooz with just file locks (no DLM) we make it easier for the next phase14:52
geguileodulek: That's the rationale14:52
geguileodulek: For the final solution we will need Tooz14:53
geguileodulek: And using Tooz with file locks will not change current behavior14:53
dulekgeguileo: Yup... Okay, so locks cleanup + Tooz as first step seems fine.14:53
geguileodulek: I think so, since it's basically what we have now14:53
geguileodulek: Not much to think about :)14:54
*** daneyon has joined #openstack-cinder14:55
*** jwcroppe has quit IRC14:56
dulekgeguileo: Can you review the spec for A/A then? Feel free to update it if you find that easier.14:58
*** mdenny has joined #openstack-cinder14:58
geguileodulek: Yes, I'll review it now14:59
dulekgeguileo: Thanks!14:59
*** daneyon has quit IRC15:00
*** crose has joined #openstack-cinder15:01
openstackgerritxing-yang proposed openstack/cinder: Add multiple pools support to VMAX driver  https://review.openstack.org/18999015:02
kjellyhi guys, I am working on the bug 1477625 . there are many ways to solve the problem. the simplest way is like https://review.openstack.org/#/c/206450/1 . another way is that add new argument "viewable_admin_meta" to db.volume_get function. And don't use context.elevated() . https://github.com/openstack/cinder/blob/master/cinder/volume/api.py#L402  Any suggestion for the bug?15:02
openstackbug 1477625 in Cinder "DB access to show volumes may not be properly controlled" [Undecided,In progress] https://launchpad.net/bugs/1477625 - Assigned to jelly (coding1314)15:02
*** crose has quit IRC15:05
*** msno has quit IRC15:06
*** annashen has joined #openstack-cinder15:06
openstackgerritxing-yang proposed openstack/cinder: EMC VMAX Create CG from CG Snapshot  https://review.openstack.org/19528015:09
*** nkrinner has quit IRC15:10
*** daneyon has joined #openstack-cinder15:11
*** annashen has quit IRC15:11
*** mylu has quit IRC15:13
geguileodulek: You really don't mind if I update the spec?15:13
*** ronis has quit IRC15:13
dulekgeguileo: Definitely not, go on.15:14
*** agarciam has quit IRC15:14
geguileodulek: Ok, because the comments were a pita15:14
*** agarciam has joined #openstack-cinder15:14
*** edtubill has joined #openstack-cinder15:15
hemnaos-brick in nova just landed.15:15
smcginnishemna: Woot woot15:15
xyang1hemna: Congrats!15:15
hemnashould I panic now?15:15
smcginnishemna: Probably. :)15:15
avishay_hemna: mazal tov!15:15
hemnathanks for all the help guys.15:16
Swansonhemna: Congrats!  And it is definitely time to panic.  Or at least after liberty-2 gets into the wild.15:17
hemnaheh yah15:18
hemnathanks :)15:18
*** mtanino has joined #openstack-cinder15:18
*** mylu has joined #openstack-cinder15:18
*** dannywilson has joined #openstack-cinder15:22
*** markus_z has quit IRC15:23
*** igajsin_ has joined #openstack-cinder15:25
*** jasondotstar has quit IRC15:25
*** avishay_ is now known as avishay15:30
avishayjordanP: hey, what did you have in mind here?  https://review.openstack.org/#/c/205945/15:30
jordanPavishay, I posted a new comment to be more specific15:30
*** jdurgin1 has joined #openstack-cinder15:31
jordanPavishay, maybe I am a bit picky today... sorry15:31
avishayjordanP: i think so :)15:31
jordanPbut as os-brick could be one day also used by Nova, better safe than sorry15:32
avishayjordanP: the unit test is basically a copy of the code, don't think it helps much15:32
jordanPI don't think it's a copy of the code, the idea is to test that the exception is properly handled15:32
jordanPwe should also unit test the 'unhappy path' when things fail15:32
avishayjordanP: i don't see much value in it for this case, but i will add it if you insist15:33
avishayjordanP: thanks15:34
jordanPavishay, I insist :) It's should take 5 min at most... (15 min if you are not a a mock expert). You'll lose your +2 but I am sure hemna will give it back15:35
jordanPavishay, what do you think of the 'too broad except statement' comment I also left ?15:35
*** lcurtis has joined #openstack-cinder15:36
openstackgerritCindy Pallares proposed openstack/cinder: Fix multi-line docstrings to meet hacking rules  https://review.openstack.org/18482015:36
*** aix has quit IRC15:38
*** Apoorva has joined #openstack-cinder15:39
*** laughterwym has quit IRC15:40
avishayjordanP: what error are you worried about?  if somebody replaced the directory with a file and it returns ENOTDIR?  i think that's the only other error code for this call...15:40
geguileodulek: Have you looked into zookeepers lock mechanism in Tooz?15:40
jordanPavishay, you're saying that the only errno that os.listdir could raise is ENOENT and ENOTDIR ?15:42
dulekgeguileo: I have some knowledge on that, what's special about them?15:42
avishayjordanP: as far as i can see, yes15:42
geguileodulek: Expiration timeout of a lock15:42
jordanPI am impressed with you low level cPython knowledge :)15:42
jordanPavishay, what about errno.EIO or errno.ENOMEM ?15:43
geguileodulek: Or if you prefer, if a node dies, does its locks get released? how? when?15:43
jordanPif you list a super big directory ?15:43
dulekgeguileo: Ah, that thing. So I think Redis locks are dropped when service holding it dies.15:45
geguileodulek: Redis I know, it's Zookeeper that worries me15:45
dulekgeguileo: I'm not sure if I checked that explicitly, but that was my understanding.15:45
avishayjordanP: if you can attach enough devices to a system to make listing /dev/disk/by-path run out of memory i will give you $10015:45
geguileodulek: Redis locks have an expiration date that you update each time you call heartbeat15:45
dulekgeguileo: ZooKeeper is basically a storage.15:45
* jordanP is going to spawn his 8 MB of RAM VM....15:46
dulekgeguileo: It uses Paxos consensus algorithm to get the state right.15:46
avishayjordanP: go ahead, i'm sure listing the directory is the first thing that will fail15:46
dulekgeguileo: I think that when a node die other nodes will decide that they have quorum and will drop the lock.15:46
avishayjordanP: if you really want to nitpick my 2 line change to this point, go ahead, i will just abandon it and you can do it however you like.15:47
dulekharlowja: Isn't my explanation for geguileo right? ;)15:47
*** alexpilotti has quit IRC15:47
geguileodulek: I would feel better if we knew that for sure15:47
*** annashen has joined #openstack-cinder15:48
jordanPavishay, that's sad. I have 15 min of free time in front of me. Can I propose a new patchset and add myself as a coauthor  ?15:48
avishayjordanP: enjoy15:48
jordanPI'll add a unit test and I'll narrow the except statement.15:48
dulekgeguileo: ZooKeeper is Tooz's reference implementation, I wouldn't expect it to be worse that other.15:49
avishayjordanP: abandoned - you can do whatever you'd like15:49
geguileodulek: XD XD XD15:49
geguileodulek: No comment on that15:49
jordanPavishay, You can't abandoned it. Because then I can"t push a new patchset to it15:49
avishayjordanP: i will feel better knowing that i won't run out of memory listing that directory15:49
dulekgeguileo: If I recall correctly it even doesn't need explicit heartbeating, ZooKeeper takes care of that.15:49
jordanPavishay,  I need to keep the change ID to keep track of the conversation15:49
avishayjordanP: restored, enjoy15:50
jordanPavishay, thanks15:50
smcginnisThere is such thing as being too defensive in programming. Just saying.15:50
geguileodulek: Yes, I don't see a heartbeat15:50
*** agarciam has quit IRC15:51
geguileodulek: But I don't see anywhere that it mentions anything about how these are droped15:51
*** _cjones_ has joined #openstack-cinder15:51
*** annashen has quit IRC15:51
geguileodulek: And Redis is also recommended solution by the Tooz project: http://docs.openstack.org/admin-guide-cloud/content/section_telemetry-cetral-compute-agent-ha.html15:51
*** Yogi1 has joined #openstack-cinder15:52
dulekgeguileo: Ceilometer you mean?15:52
geguileodulek: Well, they say there that Tooz guys recommend both15:52
geguileodulek: Anyway, we should ask someone who knows how that stuff works, because I don't feel like digging into that myself15:52
*** changbl has joined #openstack-cinder15:53
geguileoI already did it for Redis15:53
*** Yogi11 has quit IRC15:53
*** jungleboyj has joined #openstack-cinder15:53
*** BharatK has quit IRC15:55
*** logan2 has quit IRC15:55
*** breitz1 is now known as breitz15:55
*** BharatK has joined #openstack-cinder15:56
openstackgerritApoorva Deshpande proposed openstack/cinder: Fix order of arguments in assertEqual  https://review.openstack.org/20629615:56
dulekgeguileo: http://zookeeper.apache.org/doc/trunk/recipes.html#sc_recipes_Locks15:57
dulekgeguileo: "The removal of a node will only cause one client to wake up"15:57
geguileodulek: Yes, that's talking about exclusion15:58
geguileodulek: And it's implying the release removes it15:58
dulekgeguileo: This suggest that it will drop it, but they aren't saying that explicitly. ;)15:58
geguileodulek: It doesn't say anything about a node dying15:58
geguileodulek: I don't think so15:58
geguileodulek: I mean, I don't think that's what that sentence is talking about15:59
dulekgeguileo: Isn't "removal of a node" equal to "node dies"?15:59
dulekgeguileo: Ah, no, it isn't.15:59
geguileodulek: They are talking about queue nodes15:59
*** kjelly has quit IRC16:00
geguileodulek: Ok, I found the answer, these are ephemeral nodes and will be removed automatically :)16:01
geguileodulek: Members of the group create ephemeral nodes under the group node. Nodes of the members that fail abnormally will be removed automatically when ZooKeeper detects the failure.16:01
dulekgeguileo: :)16:01
geguileodulek: So we are good  :)16:01
openstackgerritSzymon Wróblewski proposed openstack/cinder: Tooz locks  https://review.openstack.org/18353716:01
*** BharatK has quit IRC16:06
*** coolsvap is now known as coolsvap|dinner16:06
*** annashen has joined #openstack-cinder16:07
*** Tross has joined #openstack-cinder16:08
*** merooney has quit IRC16:09
*** bluex-pl has quit IRC16:11
*** annegentle has quit IRC16:11
*** annashen has quit IRC16:12
*** lpetrut has quit IRC16:14
*** ub2 has joined #openstack-cinder16:18
*** ub has quit IRC16:19
*** leeantho has joined #openstack-cinder16:19
*** e0ne has quit IRC16:20
*** Yogi1 has quit IRC16:20
*** Zhongjun has quit IRC16:22
*** jistr has quit IRC16:22
*** logan2 has joined #openstack-cinder16:24
*** Tross has quit IRC16:26
*** Apoorva has quit IRC16:27
openstackgerritGorka Eguileor proposed openstack/cinder-specs: Add Cinder Volume A/A HA lock spec  https://review.openstack.org/20261516:28
geguileodulek: ^16:28
dulekgeguileo: Thanks, I'll take a look.16:29
geguileodulek: you'll probably want to change some things  :-)16:29
openstackgerritJordan Pittier proposed openstack/os-brick: Fix race in check and access of /dev/disk/by-path/  https://review.openstack.org/20594516:30
SwansonIf I wanted to back port a driver from kilo to juno or icehouse are we talking significant differences?16:32
*** jdurgin1 has quit IRC16:32
*** jasondotstar has joined #openstack-cinder16:34
openstackgerritJordan Pittier proposed openstack/os-brick: Remotefs: add ScalityFS support  https://review.openstack.org/19663116:37
*** hodos has joined #openstack-cinder16:38
*** annashen has joined #openstack-cinder16:38
SwansonJust posted a fix to stable/kilo and two CI's gave me a pass in less than a minute!  Woo!16:39
*** thangp has quit IRC16:40
*** thangp has joined #openstack-cinder16:41
*** jordanP has quit IRC16:41
hemnaentire CI running in <1minute ?16:41
hemnaum....I'd say that's probably a false +16:41
avishayhemna: mock CI :)16:41
hemnamock the world CI ?16:42
hemnamock *.*16:42
Swansonhemna: I think you're right.  Either that or they are running their CI on Deep Thought which seems like overkill.16:42
*** lpetrut has joined #openstack-cinder16:42
avishayhemna: no chance of the driver getting booted for failing all the time :)16:42
avishaySwanson: which CIs?16:42
hemnawhat's the review ?16:43
SwansonIBM DB2 CI StorPool16:43
Swansonand should be between those two.16:43
*** sghanekar_ has joined #openstack-cinder16:43
Swansonhemna: https://review.openstack.org/206615  I have suspicions I'll be trimming that down to just the payload filter change.  But I have high hopes.16:44
hemnaoh DB2 CI16:45
hemnaso that might simply be running the DB migration scripts16:45
avishayi don't see that any CIs posted on that patch, and I think DB2 probably only does db-related tests16:45
*** ronis has joined #openstack-cinder16:45
SwansonGot two emails.16:46
SwansonWhich is fine, I guess, so long as they aren't posting on the review.16:46
*** mriedem has joined #openstack-cinder16:47
mriedemoh hello cinder peeps16:48
mriedemsome nova people, myself included, are wondering if this is a priority for cinder16:48
jungleboyjEveryone run!16:48
mriedemb/c it's DOA for our FF16:48
*** Yogi1 has joined #openstack-cinder16:48
hemnait would have been nice to get it in16:48
jungleboyjhemna: ^^16:48
jungleboyjVery nice.16:48
mriedemit has quite a few issues at this point16:49
hemnaI'd rather see it working :)16:49
hemnathan push it through.16:49
mriedemyeah, it has rpc and rest api impacts16:49
*** johnthetubaguy has joined #openstack-cinder16:49
mriedemand the code came in pretty late and low on the radar16:50
mriedemplus, random vendor pushing it16:50
mriedemrather than cinder core like the os-brick change16:50
mriedemanyway, i guess this is just a heads up16:50
hemnano worries.16:50
hemnathanks for the heads up16:50
mriedemwe've been trying to prioritize sub-teamy type things like virt drivers in nova, we should probably start talking about doing that with cross project stuff16:51
mriedemwe track some of that here https://etherpad.openstack.org/p/liberty-nova-priorities-tracking16:51
jungleboyjmriedem: Thanks for letting us know.16:51
johnthetubaguyyeah, a cinder or volume group in there would be awesome16:51
mriedemb/c last week someone popped up complaining about glance v2 not making liberty nova16:51
*** dims has quit IRC16:52
mriedembut if glance isn't pushing it and it aint broken, we aren't actively pursuing more work16:52
hemnamriedem, +116:52
mriedemhenceforth let it be know that jungleboyj is the nova/cinder liasion :)16:52
*** dims has joined #openstack-cinder16:52
*** dims has quit IRC16:52
*** dims has joined #openstack-cinder16:53
mriedemsheesh, joking tw16:53
johnthetubaguyseriously though, if someone fancies that, it would be great to add that in... https://wiki.openstack.org/wiki/CrossProjectLiaisons#Inter-project_Liaisons ;)16:53
*** jaypipes has quit IRC16:54
johnthetubaguy(although, I can't say they are working that well right now, I need to really try push that forward more...)16:54
*** jaypipes has joined #openstack-cinder16:54
mriedemi wasn't aware that table existed :)16:55
hemnayah interesting16:57
*** annashen has quit IRC16:57
hemnaI suppose I could sign up for it.  I have done lots of nova <--> Cinder stuffs16:57
johnthetubaguyhemna: that would be cool16:57
*** sghanekar_ has quit IRC16:57
hemnaI'm not sure what it entails16:58
hemnakinda feels like it should be a PTL thing16:58
hemnaso I'll defer to thingee for advice16:58
johnthetubaguyhemna: its really just trying to scale out the PTL roles I guess16:58
hemnaI'm sure he's overwhelmed with things to do16:58
johnthetubaguyhemna: I think the idea is to jump into nova meetings when there is something important you need doing16:58
johnthetubaguyhemna: and track the nova process, and be a point of contact16:59
johnthetubaguyhemna: honestly, we don't really have a good pattern to follow yet, so its more... aspirational16:59
*** annashen has joined #openstack-cinder16:59
*** ub2 has quit IRC17:00
hemnaok, well I'll defer to thingee at this point as the PTL.  I can help though17:00
hemnait's hard to be aware of everything all the time for sure.17:00
*** coolsvap|dinner is now known as coolsvap|mt17:01
*** coolsvap|mt is now known as coolsvap|mtg17:01
thingeefwiw mriedem, glance v2 is broken in Cinder today https://bugs.launchpad.net/cinder/+bug/147873717:02
openstackLaunchpad bug 1478737 in Cinder "Cinder upload-to-image with glance v2 fails" [Medium,In progress] - Assigned to Tomoki Sekiyama (tsekiyama)17:02
thingeetsekiyama: speaking of, are you pushing a patch for this ^17:02
mriedemthingee: i heard glance v2 doesn't work with glance v1 created shapshots either, so yeah, some issues with nova moving to glance v217:03
thingeemriedem: I was noting it on the ML http://lists.openstack.org/pipermail/openstack-dev/2015-July/070714.html17:03
*** mylu has quit IRC17:04
*** mylu has joined #openstack-cinder17:05
*** Tross has joined #openstack-cinder17:05
*** lpetrut has quit IRC17:06
*** vilobhmm has joined #openstack-cinder17:06
*** avishay has quit IRC17:08
smcginnisAnyone else seeing their test_volume_boot_pattern tests suddenly starting to fail?17:10
hemnasaw that last night as well I thinks17:11
smcginnishemna: Any idea what's going on there?17:11
*** ub has joined #openstack-cinder17:12
hemnanah I haven't dug into it yet today17:12
smcginnisI'll see if I can do a little digging.17:12
*** mylu has quit IRC17:13
mriedemssh failures or other?17:15
*** mylu has joined #openstack-cinder17:16
*** boris-42 has quit IRC17:20
*** jungler has quit IRC17:20
thingeebreitz: hi17:20
*** jungler has joined #openstack-cinder17:21
*** boris-42 has joined #openstack-cinder17:22
*** abehl has quit IRC17:22
thingeebreitz: unless there's a good reason for a PM, I'd rather keep the convo here.17:22
*** merooney has joined #openstack-cinder17:23
*** Apoorva has joined #openstack-cinder17:25
openstackgerritSonia Ghanekar proposed openstack/cinder: Set VERSION on the Nimble driver.  https://review.openstack.org/20623917:26
breitzah sure - did not want to bother everybody else with the first q17:27
breitzthingee:  was wondering if there was anybody that was willing to help mentor a new guy a bit.17:27
breitzand if not - I was looking for a suggestion for a bug or two to look at fixing17:28
*** patrickeast has joined #openstack-cinder17:29
*** annegentle has joined #openstack-cinder17:30
thingeebreitz: first off welcome to the team17:32
breitzheh - thanks!17:32
thingeebreitz: so I have a bug that I started fixing, but stopped that I could hand off to you. I usually just tell people to pair up with someone working on significant changes in Cinder though17:33
breitzthingee, that sounds good.  I can pick up the bug - and try to pair up with somebody.17:34
*** annegentle has quit IRC17:35
thingeebreitz: here's the bug though https://bugs.launchpad.net/nova/+bug/1437904 https://review.openstack.org/#/c/178860/17:35
openstackLaunchpad bug 1437904 in Cinder "generate_sample.sh uses MODULEPATH environment variable, conflicts with environment-modules" [Low,In progress] - Assigned to Mike Perez (thingee)17:35
breitzsuggestions of who I can pair up with would be most welcome17:35
thingeebreitz: vilobhmm is working on nested quota inside of Cinder. sure help would be appreciated there http://specs.openstack.org/openstack/cinder-specs/specs/liberty/cinder-nested-quota-driver.html17:37
breitzthingee: ok, I'll see if I can provide some help there.  thanks17:38
thingeebreitz: patrickeast is working a generic image caching feature. testing would be great here http://specs.openstack.org/openstack/cinder-specs/specs/liberty/image-volume-cache.html17:38
thingeebreitz: here's the patch for image caching https://review.openstack.org/#/c/195795/17:39
patrickeastoo yea, definitely would like to get more folks trying it out17:39
breitzk - this sounds like a good starting point to ramp up with.  awesome17:40
thingeepatrickeast: is there any documenting that's necessary for getting this working with new drivers that breitz could help with?17:40
patrickeastthingee: breitz: mmm not really, it doesn’t require any driver changes or anything, you only need to set the config options for the internal tenant and to enable the cache17:41
thingeebreitz: so I know some projects like to start people off with random bugs, but I lean more towards having people work with others on significant changes. So eventually you yourself will be working on significant changes :)17:42
thingeebreitz: unfortunately you're coming in when we're almost through a release cut, so things are sort of at testing phases, but testing out changes coming through the pipeline is very helpful17:42
*** annashen has quit IRC17:42
thingeepatrickeast: perhaps documentation for enabling it ;D17:43
thingeebreitz: ^17:43
patrickeastthingee: yep, definitely will want some17:43
breitztesting is fine - much learning to be had even with that.17:43
patrickeastthingee: breitz: an example of my local.conf i’m using to play around with it http://paste.openstack.org/show/406111/17:43
patrickeastnote the image_volume_cache_* and cinder_internal_tenant_* options17:44
*** mylu has quit IRC17:44
thingeeI'd like to see documentation for it somewhere in http://docs.openstack.org/admin-guide-cloud/content/managing-volumes.html17:44
thingeepatrickeast: I guess we still have this mess on the ML to figure out17:44
patrickeastthingee: haha yea, i figured i would wait for it to land before writing up too much documentation for it17:45
*** mylu has joined #openstack-cinder17:45
thingeepatrickeast: I appreciate your last email to thread. I still haven't decided what to do17:46
* thingee decided to make his x1 carbon work better with linux last night. he's now typing on a 2010 macbook pro.17:47
*** ronis has quit IRC17:47
thingeeI think I scared breitz off17:47
*** markvoelker has quit IRC17:47
patrickeastthingee: hopefully that email made sense… i felt like i wrote a small novel and still didn’t really hit all the points very well17:48
thingeepatrickeast: I do like short emails with bullet points, but I still appreciated the info17:48
breitzheh - scared yes - scared off - not yet.  ;-)17:48
breitzmuch to figure out17:48
*** ub has quit IRC17:49
*** crose has joined #openstack-cinder17:50
*** ub has joined #openstack-cinder17:51
*** haomaiwang has quit IRC17:51
*** annashen has joined #openstack-cinder17:52
thingeepatrickeast: Is tsekiyama not posting a fix for https://bugs.launchpad.net/bugs/1478737 ?17:52
openstackLaunchpad bug 1478737 in Cinder "Cinder upload-to-image with glance v2 fails" [Medium,In progress] - Assigned to Tomoki Sekiyama (tsekiyama)17:52
*** haomaiwang has joined #openstack-cinder17:52
patrickeastthingee: we had a working fix yesterday17:52
breitzI also had some code questions about stuff that is now moved into os-brick.17:53
thingeepatrickeast: that bug is targeted for l-217:53
thingeepatrickeast: guess I can move it to l-3 since I'm trying to do a cut today17:53
thingeeanyone from core want to give the last +2 on https://review.openstack.org/#/c/184820/ once jenkins is ready?17:53
thingeepretty simple hacking update ^17:54
thingeebreitz: hemna is brick masta17:54
thingeeis our*17:54
patrickeastthingee: i can put a fix up for it right now if you wanna sneak it into l-217:54
thingeepatrickeast: that'd be cool17:55
breitzk - basically much of what is in os_brick/remotefs/remotefs.py is specific to linux - mount command, parsing mount cmd output.  Was looking for a way to make it less so.17:56
*** annashen has quit IRC17:56
breitzor better put - was looking for suggestions about how that might be done17:56
*** changbl has quit IRC17:57
thingeejgriffith: https://review.openstack.org/#/c/200237/17:58
*** ronis has joined #openstack-cinder17:58
*** jasondotstar has quit IRC17:58
*** lpetrut has joined #openstack-cinder18:00
*** mriedem is now known as mriedem_meeting18:00
*** timcl has quit IRC18:00
*** timcl has joined #openstack-cinder18:00
*** abehl has joined #openstack-cinder18:01
mtaninojgriffith: Could you visit this fix for LVM? https://review.openstack.org/#/c/200237/18:02
tsekiyamathingee: patrickeast: oh, I'm writing a unit test for patrickeast's patch for the bug https://bugs.launchpad.net/bugs/1478737 . I'm going to post it soon.18:03
openstackLaunchpad bug 1478737 in Cinder "Cinder upload-to-image with glance v2 fails" [Medium,In progress] - Assigned to Patrick East (patrick-east)18:03
patrickeasttsekiyama: ah excellent18:04
*** jungler has quit IRC18:04
*** jungler has joined #openstack-cinder18:05
*** xyang1 has quit IRC18:05
*** sgotliv_ has quit IRC18:06
vilobhmmbreitz : hello18:07
vilobhmmi am working on the nested quota driver right now18:08
vilobhmmas part fo http://specs.openstack.org/openstack/cinder-specs/specs/liberty/cinder-nested-quota-driver.html18:08
vilobhmm2 patches have been merged and the 3rd patch for the driver is under review18:08
breitzvilobhmm:  hi18:10
vilobhmmbreitz : ^618:10
breitzmy first order of business is to get a working dev/test env setup.18:10
*** changbl has joined #openstack-cinder18:10
breitzafter that I'm interested in helping as much as I can.18:11
vilobhmmlet me know if you need help with that18:12
vilobhmmbreitz : 1] http://docs.openstack.org/developer/devstack/ 2] http://www.rushiagr.com/blog/2013/05/27/cinder-on-devstack-quick-start/18:12
breitzk.  thanks.18:13
vilobhmmshould be helpful for you to start with it18:13
breitzI think I've read both of those - but I'll re-read them again to be sure18:13
breitzbut thanks for the links - I appreciate it18:13
*** avishay has joined #openstack-cinder18:14
breitzI've messed with devstack on FC - but was going to switch to ubuntu for a platform instead.18:14
breitzI'm familiar with openstack and cinder (to a point).  Like many my introduction came with writing a cinder driver.18:16
*** angela-s has joined #openstack-cinder18:18
*** lpetrut has quit IRC18:20
openstackgerritTomoki Sekiyama proposed openstack/cinder: Fix exception on uploading a volume to image with glance v2 API  https://review.openstack.org/20664518:22
tsekiyamapatrickeast: thingee: ^ posted the patch with unit tests for glance api v2.18:24
*** agarciam has joined #openstack-cinder18:26
*** annashen has joined #openstack-cinder18:27
*** xyang1 has joined #openstack-cinder18:27
*** ub has quit IRC18:28
*** haomaiwang has quit IRC18:29
*** haomaiwang has joined #openstack-cinder18:30
*** mylu has quit IRC18:34
avishayhemna: would appreciate your input here to get this bug fixed - https://review.openstack.org/#/c/205945/18:35
hemnaavishay, so, I added a comment in the previous patch I think18:35
hemnaif that listdir fails, then we should return an empty list IMHO18:36
avishayhemna: in practice it doesn't matter, because no other error will ever occur.  it's ugly and sets bad precedent IMO but i just don't want it the patch to collect dust.18:36
hemnaso that code is only used to fetch the list of existing /dev/disk/by-path entries at iscsi disconnect time18:37
hemnaavishay, https://github.com/openstack/os-brick/blob/master/os_brick/initiator/connector.py#L50018:38
smcginnisavishay: I was fine with it as it was. If it hits something like ENOMEM then their are bigger problems.18:39
avishayhemna: correct18:39
avishaysmcginnis: yep18:39
hemnaif it fails to get that listing, then it basically means there are no volumes.18:39
*** mylu has joined #openstack-cinder18:39
hemnaso I dunno18:39
avishayhemna: i prefer without checking errno, but don't care much. your call18:40
hemnaso if that devices list comes back empty, it tries to do an iscsiadm logout basically.18:40
*** vokt has joined #openstack-cinder18:40
avishayhemna: yep18:40
hemnaso maybe what you have is fine I guess18:41
hemnaif we hit ENOMEM, we won't try and do a logout18:41
hemnaif we raise18:41
hemnawhich is probably safer.18:41
*** markvoelker has joined #openstack-cinder18:42
hemnawe can always change it later18:42
avishayhemna: if we hit ENOMEM, VMs will be killed, the system will be unusable, etc. it really won't matter.18:42
hemnayah, but we won't make it worse by nuking iscsi connections....or trying to :)18:42
avishayhemna: i think it's a bad precedent to check all kinds of things we shouldn't care about all over, but again, doesn't matter for this case18:42
hemnaENOENT is really the only real normal failure18:43
openstackgerritErlon R. Cruz proposed openstack/cinder: Fix concurrent attaches on HNAS iSCSI driver  https://review.openstack.org/20378018:43
openstackgerritErlon R. Cruz proposed openstack/cinder: Fix HNAS iSCSI 32 targets limitation error  https://review.openstack.org/20665118:43
patrickeastthat seems like a pretty reasonable change to me… we shouldn’t leave the iscsi portal open if we get ENOENT18:44
hemnayah that's what I think18:44
hemnaif we get other spurious errors, we raise and crap out18:44
avishayhemna: i stated my opinion, you're the core member, you make the call :)18:45
avishayhemna: i don't care as long as some version of it is merged :)18:45
hemnaI think it's good.18:46
hemnajust waiting on jenkins18:46
*** markvoelker_ has joined #openstack-cinder18:46
openstackgerritCurt Bruns proposed openstack/cinder: Add deactivate step to extend_lv  https://review.openstack.org/19881318:47
*** markvoelker has quit IRC18:48
avishayhemna: great, thanks18:48
*** mragupat has joined #openstack-cinder18:49
*** abehl has quit IRC18:49
*** ronis has quit IRC18:56
*** timcl1 has joined #openstack-cinder19:03
*** e0ne has joined #openstack-cinder19:03
*** akshai has quit IRC19:04
*** timcl has quit IRC19:05
*** akshai has joined #openstack-cinder19:06
*** ub has joined #openstack-cinder19:07
*** hemna is now known as hemnalunch19:07
*** avishay has quit IRC19:24
*** mriedem_meeting is now known as mriedem19:27
*** ub has quit IRC19:29
*** markvoelker_ has quit IRC19:29
*** akshai has quit IRC19:30
*** annashen has quit IRC19:30
*** ub has joined #openstack-cinder19:30
*** akshai has joined #openstack-cinder19:32
vilobhmmsmcginnis, jungleboyj : If you can review this https://review.openstack.org/#/c/206170/3 it would be helpful thanks :)19:32
smcginnisvilobhmm: SUre, I'll take a look.19:32
vilobhmmthanks smcginnis ! :)19:32
*** agarciam has quit IRC19:37
*** annegentle has joined #openstack-cinder19:40
*** ccesario has joined #openstack-cinder19:43
*** jasondotstar has joined #openstack-cinder19:44
ccesariohello... does someone have idea about fix this error ?  http://paste.openstack.org/show/aKjp8hSKG1Bu9oku6adt/19:45
*** annashen has joined #openstack-cinder19:53
*** julim has quit IRC19:55
*** annashen has quit IRC19:57
*** btran has joined #openstack-cinder20:01
*** alexpilotti has joined #openstack-cinder20:06
*** rushiagr is now known as rushiagr_away20:06
*** alexpilotti has quit IRC20:12
dannywilsonccesario: says "volume service is down", check cinder volume log contents to see why it crashed/didn't start20:13
*** ub2 has joined #openstack-cinder20:15
*** Apoorva has quit IRC20:15
*** ub has quit IRC20:17
ccesariodannywilson,  http://paste.openstack.org/show/406137/20:18
ccesarioit seems is started20:18
*** merooney has quit IRC20:19
*** e0ne has quit IRC20:19
*** annashen has joined #openstack-cinder20:20
dannywilsonccesario: check log file to be sure, if using devstack possibly at /opt/stack/logs/screen/screen-c-vol.log20:20
*** jaypipes has quit IRC20:20
*** e0ne has joined #openstack-cinder20:21
*** e0ne has quit IRC20:21
*** julim has joined #openstack-cinder20:22
*** mtanino has quit IRC20:23
ccesariodannywilson, http://paste.openstack.org/show/7FsXMrdq0NhMVHShDR5X/20:23
*** sgotliv_ has joined #openstack-cinder20:25
dannywilsonccesario: "Update driver status failed: (config name vsphere) is uninitialized." from log file, can you paste cinder.conf too?20:25
*** shyama has quit IRC20:28
ccesariosure .. a momment pls20:29
*** Apoorva has joined #openstack-cinder20:30
*** jaypipes has joined #openstack-cinder20:31
*** salv-orlando has joined #openstack-cinder20:31
ccesariodannywilson, http://paste.openstack.org/show/LdDWkc1PBvKeXe7EN7UZ/20:31
*** salv-orl_ has quit IRC20:34
dannywilsonccesario: Hmm "enabled_backends = vsphere" I am not familiar with this volume type but it seems like this might be the problem20:35
dannywilsonccesario: also not familiar with cinder.volume.drivers.vmware.vmdk.VMwareVcVmdkDriver driver, sorry20:35
dannywilsonccesario: also looks like you haven't set this up fully based on this line "vmware_host_ip = 192.168.X.X"20:36
*** Lee1092 has quit IRC20:37
SwansonI would have assumed there would have to be a volume backend of some kind listed.  Presumably with the vsphere name.20:38
*** ganso_ has quit IRC20:40
ccesariowell... I 'm  not a expert with this... I 'm testing openstack with vmware environment... and I follow some guides to try it with devstack :)20:40
*** Yogi1 has quit IRC20:41
*** mtanino has joined #openstack-cinder20:41
*** alexpilotti has joined #openstack-cinder20:41
*** Tross has quit IRC20:41
ccesariothen can exists errors in this config :/20:42
*** logan2 has quit IRC20:42
dannywilsonccesario: the process started but it failed to initialize a back end to create volumes on20:42
dannywilsonccesario: sorry, ignore that one20:43
dannywilsonccesario: yes, there are errors in the cinder.conf file, do a little googling on using the VMwareVcVmdkDriver and how to set up cinder.conf with it20:44
*** logan2 has joined #openstack-cinder20:45
ccesariolet me try do it :) .. thanks for the tips  dannywilson20:48
dannywilsonccesario: np, good luck20:48
ccesariothank you... I'll report you eheheheh20:48
*** alexpilotti has quit IRC20:52
*** IlyaG has joined #openstack-cinder20:53
*** akerr has quit IRC20:55
*** esker has quit IRC20:58
*** esker has joined #openstack-cinder21:00
*** jaypipes has quit IRC21:04
*** esker has quit IRC21:04
*** DericHorn-HP has joined #openstack-cinder21:06
*** aix has joined #openstack-cinder21:06
*** mylu has quit IRC21:07
*** bswartz has quit IRC21:07
*** DericHorn-HP has quit IRC21:09
SwansonI've never had every single test fail before.  Wow.21:10
smcginnisSwanson: Nice. You're just that good.21:10
*** jasondotstar has quit IRC21:11
*** delatte has quit IRC21:15
*** thangp has quit IRC21:15
Swansonsmcginnis: Looks like a sea of ci setup issues.21:16
SwansonHard to tell since the logs are not showing up.21:21
openstackgerritVilobh Meshram proposed openstack/cinder: Fix get default quota values for subprojects  https://review.openstack.org/20617021:26
*** esker has joined #openstack-cinder21:29
*** resker has joined #openstack-cinder21:31
*** esker_ has joined #openstack-cinder21:32
*** crose has quit IRC21:32
*** esker__ has joined #openstack-cinder21:33
*** esker___ has joined #openstack-cinder21:33
*** esker has quit IRC21:34
*** esker has joined #openstack-cinder21:34
*** resker has quit IRC21:35
*** resker has joined #openstack-cinder21:35
*** esker_ has quit IRC21:36
*** esker_ has joined #openstack-cinder21:37
*** esker__ has quit IRC21:37
*** esker___ has quit IRC21:38
*** esker has quit IRC21:39
*** esker has joined #openstack-cinder21:39
*** resker has quit IRC21:40
openstackgerritxing-yang proposed openstack/cinder: Clone CG  https://review.openstack.org/20347321:40
*** resker has joined #openstack-cinder21:40
*** esker_ has quit IRC21:42
*** dims has quit IRC21:42
*** esker has quit IRC21:44
openstackgerritxing-yang proposed openstack/cinder: Clone CG  https://review.openstack.org/20347321:44
*** resker has quit IRC21:45
*** eharney has quit IRC21:45
*** esker has joined #openstack-cinder21:46
openstackgerritxing-yang proposed openstack/cinder: Clone CG  https://review.openstack.org/20347321:46
*** resker has joined #openstack-cinder21:47
*** esker_ has joined #openstack-cinder21:48
*** esker has quit IRC21:50
*** esker has joined #openstack-cinder21:51
*** resker has quit IRC21:52
*** jwcroppe has joined #openstack-cinder21:52
*** esker_ has quit IRC21:52
*** changbl has quit IRC21:53
*** resker has joined #openstack-cinder21:53
*** annashen_ has joined #openstack-cinder21:53
*** edtubill has left #openstack-cinder21:54
*** esker_ has joined #openstack-cinder21:54
*** mriedem has quit IRC21:54
openstackgerritErickson Filipe Guedes dos Santos proposed openstack/cinder: Nested Quota Driver: Get Project Hierarchy  https://review.openstack.org/20617121:55
openstackgerritErickson Filipe Guedes dos Santos proposed openstack/cinder: WIP : Cinder Nested Quota Driver  https://review.openstack.org/20536921:55
*** esker has quit IRC21:56
*** esker has joined #openstack-cinder21:56
*** resker has quit IRC21:57
*** resker has joined #openstack-cinder21:58
*** annashen_ has quit IRC21:58
*** esker_ has quit IRC21:59
*** coolsvap|mtg is now known as coolsvap|away22:00
*** esker has quit IRC22:01
*** dims has joined #openstack-cinder22:02
*** resker has quit IRC22:02
*** jungleboyj has quit IRC22:06
*** hemnalunch is now known as hemna22:09
hemnapatrickeast, any idea what's up w/ the iscsi failures for pure here: https://review.openstack.org/#/c/205945/ ?22:09
hemnavolumes stuck in attaching22:09
hemnaI don't see errors in n-cpu or c-vol22:10
*** mdenny has quit IRC22:13
cebrunshemna: sorry for the delay in getting this brick bug going.  It's WIP as I wanted some feedback before updating all the testcode.  Thx! https://review.openstack.org/#/c/206549/22:14
hemnacebruns, cool thanks for working on it :)22:15
*** haomaiwang has quit IRC22:15
patrickeasthemna: uhh good question22:16
cebrunshemna: no prob.  When you have a chance, see if that's what you and jgriffith were envisioning. :)22:17
*** haomaiwang has joined #openstack-cinder22:17
*** mragupat has quit IRC22:17
patrickeasthemna: i’ve been having a lot of weirdness with my ci system the last week or so, i’m looking into but so far havne’t figure out exactly what the problem is22:18
patrickeasthemna: very strange that there are no errors22:18
*** annegentle has quit IRC22:18
hemnacebruns, yah basically.22:18
hemnacebruns, I think it should be putils all the time basically.22:19
hemnathere was a need early on to make it configurable, but I don't think that's the case now.22:19
cebrunshemna: Ok - the Class "Executor" might be misleading since we will be hard-coding putils as the executor.  :)22:20
hemnacebruns, https://github.com/openstack/cinder/blob/master/cinder/utils.py#L572-L58522:20
hemnacebruns, I think it's safer just to leave it in for now.22:21
hemnaand by it, I mean the Executor22:21
cebrunshemna: agree - no point in major surgery when a comment would suffice.22:21
cebrunshemna: In that blob you sent - my thought is to just remove the "execute=processutils.execute," parameter from the method alltogether and any other place there is "execute=execute" or similar.22:28
hemnacinder is passing it in currently to force it to processutils22:29
hemnaso we just have to be safe to not break cinder/nova now in every patch22:29
*** rhe00 has joined #openstack-cinder22:29
*** rhedlind has quit IRC22:30
*** bswartz has joined #openstack-cinder22:30
cebrunsOk - cool.  That's why I wanted to check-in and see if the direction was what you had in mind.22:33
*** jungleboyj has joined #openstack-cinder22:43
*** marcusvrn has quit IRC22:45
openstackgerritPatrick East proposed openstack/cinder: Generic image-volume cache  https://review.openstack.org/19579522:46
*** mriedem has joined #openstack-cinder22:47
jamielennoxhey everyone i've got a tempest failure coming from ceph in cinder-backup22:48
jamielennoxthe actual tempest run is in glance_store eg https://review.openstack.org/#/c/193422/22:48
jamielennoxfailing test is:  tempest.api.volume.admin.test_volumes_backup.VolumesBackupsV1Test.test_volume_backup_export_import[id-a99c54a1-dd80-4724-8a13-13bf58d4068d]22:49
jamielennoxerror message at: http://logs.openstack.org/22/193422/3/check/gate-tempest-dsvm-full-ceph-src-glance_store/098988f/logs/screen-c-bak.txt.gz#_2015-07-25_10_43_27_92422:49
jamielennoxWARNING cinder.backup.manager [req-41e2626d-5540-4ff0-a586-9183e0a98e43 tempest-VolumesBackupsV2Test-1882584612] Backup service cinder.backup.drivers.ceph does not support verify. Backup id 0cb52178-01d9-459c-b081-1eb3ba4972b5 is not verified. Skipping verify.22:49
*** garthb has joined #openstack-cinder22:50
jamielennox2015-07-25 10:43:30.704 5045 ERROR oslo_messaging.rpc.dispatcher ImageNotFound: error opening image volume-0000-0000-0000-0000.backup.0cb52178-01d9-459c-b081-1eb3ba4972b5 at snapshot None22:50
*** garthb_ has joined #openstack-cinder22:50
*** garthb_ has quit IRC22:50
jamielennoxi'm only guessing that's the cause but does anyone have any ideas22:51
jamielennoxmtreinish: have you seen ^ before?22:51
jgriffithjamielennox: known issue / bug22:52
jgriffithjamielennox: restoring an external backup to Ceph doesn't work.  That test was/is skipped for Ceph backends currently22:52
jamielennoxjgriffith: it's hitting glance_store fairly consistently, is there a workaround or should i disable the gate job?22:52
mtreinishjamielennox: is that on a ceph run22:52
jamielennoxmtreinish: yep22:52
mtreinishyep it is, there is a bug open for that22:52
jgriffithjamielennox: should skip it similar to the cinder-dsvm-ceph job IMO22:52
mtreinishjamielennox: https://bugs.launchpad.net/cinder/+bug/147641622:53
openstackLaunchpad bug 1476416 in Cinder "backup.api:import_record sets volume_id to '0000-0000-0000-0000'" [High,Confirmed] - Assigned to Ivan Kolodyazhny (e0ne)22:53
jamielennoxjgriffith, mtreinish: ok, i'll propose to skip it, thanks22:53
jamielennoxjgriffith: ceph jobs seem to be running on cinder: https://review.openstack.org/#/c/197075/22:56
jamielennoxor is that a different test22:56
jamielennoxgate-tempest-dsvm-full-cephSUCCESS in 46m 39s22:56
jgriffithjamielennox: Sorry... to be more specific, just skip "that job" on the cep run22:56
jgriffithjamielennox: you're correct, it's still running, but we added a skip to the projects file for that particular test22:56
* jgriffith grabbing the link....22:57
jamielennoxjgriffith: if you can show me that i'd appreciate it, i haven't seen skipping individual tests from projects file22:57
openstackgerritWalter A. Boring IV (hemna) proposed openstack/os-brick: WIP: remove the iscsi rescan during disconnect  https://review.openstack.org/20326522:57
earlephilhowermriedem: ping22:58
jgriffithjamielennox: try this out: https://review.openstack.org/#/c/203845/22:58
jgriffithjamielennox: but NOTE, I don't think the .*slow filter should be in there22:59
jamielennoxjgriffith: that's a fun regexp, yea i'll have to change it to the new job name22:59
jgriffithjamielennox: sure... good luck :)22:59
mriedemearlephilhower: pong22:59
*** lcurtis has quit IRC23:00
*** dannywilson has quit IRC23:00
earlephilhowermriedem:  Howdy.  I'm trying to update the new os-brick shim for HGST w/your breaking out of volume drivers in Nova bp/consolidate-libvirt-fs-volume-drivers.  But it looks like your branch was done before os-brick support pulled in.23:01
mriedemearlephilhower: yeah, so i'm in the process of rebasing that bastard today/tonight23:01
mriedemi'm a few changes off yet, but should have that rebased and up in a couple of hours23:01
*** haomaiwang has quit IRC23:01
earlephilhowermriedem:  Great, thanks.  Been beating my head around how to make it all line up23:01
*** haomaiwang has joined #openstack-cinder23:02
earlephilhowerI'll keep up on it and try again tomorrow if there's a new rebase...23:02
earlephilhowers/keep/give/ ^^23:03
jamielennoxmriedem: the above bug is what i was pinging you about yesterday, the glance guys said to ask you about it, i think we've got a solution23:03
*** dims has quit IRC23:05
*** edmondsw has quit IRC23:07
jamielennoxjgriffith, mtreinish: skip ceph test: https://review.openstack.org/20675923:14
jungleboyjscottda: You still around?23:14
jgriffithjamielennox: LGTM23:17
*** xek has quit IRC23:18
scottdajungleboyj: hi23:20
jungleboyjscottda: Hey, I was just chatting with cebruns too.23:20
jungleboyjAustin's has a private dinner Thursday already so the earliest they could get us in is 8 .  What do you think?23:20
*** xek has joined #openstack-cinder23:21
scottdaSeem's OK...It's the last "official" day of the mid-cycle, so people might want to go late....23:21
jungleboyjscottda: That is true.  If not we can grab a drink somewhere beforehand.23:22
jungleboyjDown at that pizza place maybe that was good.23:22
scottdaThat's what I was thinking, just wanted to phrase it in a way that didn't make me look like a lush :)23:22
scottdaPlenty of places around.23:22
jungleboyjcebruns: Beat you to it.  /me is planning plenty of beverage time.23:22
jungleboyjHey, I am trying to pick a place near the hotel.  Who is the lush?23:23
scottdaFort Collins is known for beer and bikes. They kinda go together.23:23
jungleboyjOk, I will reserve Austin's at 8 then.23:24
*** r-daneel has joined #openstack-cinder23:24
*** NobodyCam has quit IRC23:24
jungleboyjI will bring it up in the meeting tomorrow and make sure everyone is cool with that.23:24
cebrunsbeer+bikes go well together...up to a point.  :)23:24
*** r-daneel has quit IRC23:24
*** NobodyCam has joined #openstack-cinder23:24
jungleboyjcebruns: That was what I was thinking.  Add altitude ...23:24
*** r-daneel has joined #openstack-cinder23:25
thingeewinston-d_zZZ: you know what would really be good... some documentation on the filters. not sure if I'm blind23:26
thingeewinston-d_zZZ: filter schedulers to be exact23:26
mtaninojgriffith: Hi, Can I get a one more core review for this LVM patch? https://review.openstack.org/#/c/200237/23:30
jungleboyjcebruns: scottda Funny, I called back and he was glad I did.  There was a cancellation they hadn't gotten documented.23:30
jgriffithmtanino: hmm.. never thought of applying that to the secure delete op23:30
jungleboyjcebruns: scottda So, 6:30 on Thursday.23:31
scottdajungleboyj: uhhhh...doesn't that mess up the early drinking?23:32
jungleboyjscottda: It leave plenty of time for late drinking.  :-)23:33
cebrunsjungleboyj: Sweet!  You made him an offer he couldn't refuse, eh?23:33
*** rhagarty_ has quit IRC23:33
*** rhagarty has quit IRC23:33
jungleboyjOr we can do early drinking Friday instead.  :-)  Just thinking, for those who want free food but might be leaving early Friday.23:33
*** dims has joined #openstack-cinder23:34
*** esker has joined #openstack-cinder23:34
mtaninojgriffith: For clear_volume() path, sparse is always set to False for the secure delete op23:34
*** akshai has quit IRC23:35
mtaninojgriffith: Does this cover your concern?23:35
jungleboyjscottda: Didn't realize you liked the early drinking idea that much.  :-)23:35
jungleboyjcebruns: Yeah, guaranteed income.23:36
scottdaI never start before 5:00 PM....23:36
*** esker has quit IRC23:36
jungleboyjWell, we have an hour and a half then.23:36
scottdabut since my team is in Ireland, that ends being 10:00 AM23:36
*** esker has joined #openstack-cinder23:36
jungleboyjscottda: Even earlier if you start with DuncanT .  :-)23:37
jungleboyjscottda: Seriously though, 6:30 sound ok?23:37
*** merooney has joined #openstack-cinder23:37
*** rhagarty has joined #openstack-cinder23:38
*** rhagarty_ has joined #openstack-cinder23:39
mtaninojgriffith: As I mentined about it in the [NOTE] of my previous patch(https://review.openstack.org/#/c/182473/), do you want to add same NOTE for my current proposed patch?23:39
jgriffithmtanino: I wasn't really concerned about it TBH, I just didn't consider it :)23:39
scottdasure, 6:30 should be fine. It's only minutes away, and folks could stop in their hotels if they like.23:39
mtaninojgriffith: hmm,23:40
jungleboyjscottda: Yeah, that was what I was thinking.  I will see what people think in the meeting tomorrow.23:40
*** jasondotstar has joined #openstack-cinder23:41
jungleboyjWant to get a more concise number attending.  If everyone is planning to go they want to do a reduced menu.23:41
jgriffithmtanino: so the change LGTM23:41
jungleboyjIf we don't have to do that, it would be good.23:41
*** Yogi1 has joined #openstack-cinder23:41
jgriffithmtanino: I haven't tested it, but traditionally your'e never submitted anything that I found you didn't test out23:41
scottdacool. Are you going to post a generic 'mid-cycle logistics' topic to the meeting agenda?23:41
jgriffithand I'm assuming that's the case here as well ;)23:42
*** kaisers has quit IRC23:42
mtaninojgriffith: so, since the target of this patch is L-3, I'm fine to wait your enough test, and then get this to merge!23:43
*** erlon has quit IRC23:43
jgriffithmtanino: I already gave it my vote :)23:43
mtaninojgriffith: sure. thank you.23:44
*** kaisers has joined #openstack-cinder23:44
*** IlyaG has quit IRC23:44
*** albertom has quit IRC23:46
jungleboyjscottda: Oh, I can do that if you like?23:48
scottdaIt's ok, I'll put something up.23:49
*** albertom has joined #openstack-cinder23:49
jungleboyjOk, cool.23:49
scottdaBTW, are you able to connect to etherpad? I keep getting kicked off.23:49
jungleboyjYeah, I had no problem.23:49
jungleboyjWait, is there something wrong with your wifi?23:49
* scottda will not jinx himself with wifi jokes23:50
thingeehemna: getting back to you earlier discussion with johnthetubaguy and mriedem ... I think it would be good if scottda (if he has time) to be our liason in nova23:50
jungleboyjHe he he.23:50
thingeehemna: I think you already have a bit on your plate since have to drop multi-attach work23:51
thingeehemna: but if you're really interested and want to take on more, I won't stop you :) ... just a recommendation23:51
scottdascottda: I'm OK with that, although I feel quite ignorant compared to hemna23:52
*** salv-orlando has quit IRC23:52
thingeescottda: I think it's great work to step up to. hemna can help you still23:53
mriedemhemna: btw, i'm rebasing my refactor series for the libvirt volume drivers so if you see me adding post-mortem comments in https://review.openstack.org/#/c/175569/ they are just notes to myself so i can go back and cleanup later23:53
thingeeI just think hemna is doing great work elsewhere at the moment, even if does happen to touch on the nova side.23:53
scottdasure, he's a busy guy. I've the cycles and I'd be happy to do it.23:54
*** annashen_ has joined #openstack-cinder23:54
jungleboyjWow, a Nova liaison!23:55
*** DTadrzak has quit IRC23:55
*** annashen has quit IRC23:55
openstackgerritVilobh Meshram proposed openstack/cinder: Cinder Nested Quota Driver  https://review.openstack.org/20536923:56
*** DTadrzak has joined #openstack-cinder23:56
thingeeso scheduler_hints extension has no policy. is that on purpose?23:57
*** annashen_ has quit IRC23:59

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!