Thursday, 2015-09-03

openstackgerritVincent Hou proposed openstack/cinder: LVM: add the exception handling to volume copy
*** takedakn1 has joined #openstack-cinder00:03
*** akerr has joined #openstack-cinder00:07
*** takedakn has quit IRC00:07
*** takedakn1 is now known as takedakn00:07
*** akerr_ has joined #openstack-cinder00:08
*** takedakn1 has joined #openstack-cinder00:10
*** akerr has quit IRC00:11
*** takedakn has quit IRC00:11
*** takedakn1 is now known as takedakn00:11
*** rushil_ has quit IRC00:13
*** rushil has joined #openstack-cinder00:14
*** tobasco_ has quit IRC00:15
*** tobasco has joined #openstack-cinder00:16
*** garthb__ has quit IRC00:18
openstackgerritMatt Riedemann proposed openstack/python-cinderclient: Update path to subunit2html in post_test_hook
mriedemyou're welcome! ^00:19
*** goodygum has quit IRC00:19
*** mtanino_ has quit IRC00:21
*** goodygum has joined #openstack-cinder00:21
*** leeantho has quit IRC00:33
*** bill_az_ has quit IRC00:36
*** hodos has quit IRC00:36
*** takedakn1 has joined #openstack-cinder00:40
*** p0rtal_ has quit IRC00:43
*** takedakn has quit IRC00:44
*** p0rtal has joined #openstack-cinder00:44
*** ganso_ has quit IRC00:45
*** IanGovett has quit IRC00:45
openstackgerritXinXiaohui proposed openstack/cinder: Calculate virtual free capacity and notify
*** takedakn1 has quit IRC00:46
*** p0rtal has quit IRC00:48
*** rushil has quit IRC01:02
*** julim has joined #openstack-cinder01:05
*** zul has quit IRC01:14
jgriffithmriedem: ?01:15
*** julim has quit IRC01:15
jgriffithmriedem: of course thank you for the bug and the shiny patch to go with it01:15
mriedemapparently novaclient did a thing and other clients copied it01:16
mriedemand we didn't see the ML thread about it going to break01:16
openstackgerritTeruaki Ishizaki proposed openstack/cinder: Sheepdog: Improve image operations
*** zul has joined #openstack-cinder01:17
jgriffithmriedem: yeah, your reference is the first I saw the thread as well :(01:17
jgriffithmriedem: thanks for being on top of it and fixing it up01:17
*** julim has joined #openstack-cinder01:20
*** BharatK has joined #openstack-cinder01:20
openstackgerritThang Pham proposed openstack/cinder: Conversion to volume object
*** rbradbury has quit IRC01:26
*** ganso_ has joined #openstack-cinder01:27
*** Lee1092 has joined #openstack-cinder01:30
openstackgerritThang Pham proposed openstack/cinder: Register RPC and object versions
*** julim has quit IRC01:47
*** boris-42 has quit IRC01:50
*** sghanekar has quit IRC01:53
*** akerr_ has quit IRC02:23
openstackgerritYuriy Nesenenko proposed openstack/python-cinderclient: Remove duplicate code in functional tests
*** julim has joined #openstack-cinder02:26
*** akerr has joined #openstack-cinder02:27
*** akerr is now known as akerr_away02:27
*** akerr_away has quit IRC02:29
*** haomaiwang has quit IRC02:36
*** haomaiwang has joined #openstack-cinder02:38
openstackgerritTeruaki Ishizaki proposed openstack/cinder: Sheepdog: Improve image operations
*** hakimo_ has quit IRC02:54
openstackgerritJohn Griffith proposed openstack/cinder: Add placholder for migration backports in Liberty
*** mudassirlatif has quit IRC02:56
*** haomaiwang has quit IRC02:57
*** hakimo has joined #openstack-cinder02:57
*** haomaiwa_ has joined #openstack-cinder03:00
*** mriedem has quit IRC03:00
*** xiaohui has quit IRC03:02
*** xiaohui has joined #openstack-cinder03:02
*** bkopilov has quit IRC03:05
openstackgerritOpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements
*** rushil has joined #openstack-cinder03:07
openstackgerritOpenStack Proposal Bot proposed openstack/python-cinderclient: Updated from global requirements
*** rushil has quit IRC03:17
*** guest1 has joined #openstack-cinder03:19
openstackgerritGerald McBrearty proposed openstack/cinder: Switch SVC driver to use lsportfc to determine FC target WWPNS
*** erhudy1 has quit IRC03:29
*** guest1 has quit IRC03:30
*** guest1 has joined #openstack-cinder03:31
*** dims has joined #openstack-cinder03:35
openstackgerritJohn Griffith proposed openstack/cinder: Use oslo-config-generator
*** Zhongjun_ has joined #openstack-cinder03:48
*** julim has quit IRC03:54
*** gouthamr has quit IRC03:59
*** boris-42 has joined #openstack-cinder03:59
*** dims has quit IRC04:00
*** links has joined #openstack-cinder04:02
*** BharatK has quit IRC04:04
*** baojg has joined #openstack-cinder04:11
*** svilgelm has joined #openstack-cinder04:14
*** guest1 has quit IRC04:30
*** baojg has quit IRC04:31
*** p0rtal has joined #openstack-cinder04:41
*** p0rtal has quit IRC04:41
*** p0rtal has joined #openstack-cinder04:42
*** ganso_ has quit IRC04:45
*** shyama has joined #openstack-cinder04:47
*** yrabl has quit IRC04:49
*** svilgelm has quit IRC04:51
*** dannywilson has joined #openstack-cinder04:52
*** bkopilov has joined #openstack-cinder04:57
*** dims has joined #openstack-cinder04:59
*** BharatK has joined #openstack-cinder04:59
*** afazekas has joined #openstack-cinder05:04
*** BharatK has quit IRC05:06
*** dims has quit IRC05:06
openstackgerritTina Tang proposed openstack/cinder: Clone cg support in VNX driver
*** BharatK has joined #openstack-cinder05:13
*** shausy has joined #openstack-cinder05:16
*** deepakcs has joined #openstack-cinder05:18
*** p0rtal has quit IRC05:21
openstackgerritWilson Liu proposed openstack/cinder: Add hypermetro support for Huawei driver
*** p0rtal has joined #openstack-cinder05:25
*** p0rtal has quit IRC05:25
*** p0rtal has joined #openstack-cinder05:26
*** shausy has quit IRC05:27
*** lpetrut has joined #openstack-cinder05:29
*** p0rtal has quit IRC05:35
*** dannywilson has quit IRC05:41
openstackgerritTina Tang proposed openstack/cinder: Clone cg support in VNX driver
*** ishant has joined #openstack-cinder05:43
*** yrabl has joined #openstack-cinder05:46
*** nkrinner has joined #openstack-cinder06:02
*** yrabl has quit IRC06:08
*** yrabl has joined #openstack-cinder06:10
openstackgerritOpenStack Proposal Bot proposed openstack/cinder: Imported Translations from Transifex
*** shyama has quit IRC06:21
*** agarciam has joined #openstack-cinder06:23
openstackgerritVincent Hou proposed openstack/cinder: LVM: add the exception handling to volume copy
*** sgotliv has joined #openstack-cinder06:30
*** kaisers1 has quit IRC06:34
*** lpetrut has quit IRC06:37
*** stefan_amann has joined #openstack-cinder06:57
*** shyama has joined #openstack-cinder06:58
*** svilgelm has joined #openstack-cinder07:13
*** amoturi has quit IRC07:19
*** amoturi has joined #openstack-cinder07:20
*** guest1 has joined #openstack-cinder07:21
*** kvidvans has quit IRC07:22
*** guest1 has quit IRC07:22
*** ronis has joined #openstack-cinder07:26
*** anshul has joined #openstack-cinder07:31
*** anshul has quit IRC07:33
*** anshul has joined #openstack-cinder07:33
*** afazekas__ has joined #openstack-cinder07:33
*** ociuhandu has quit IRC07:41
*** Zhongjun_ has quit IRC07:47
*** lpetrut has joined #openstack-cinder07:47
*** ociuhandu has joined #openstack-cinder07:52
*** vbala has joined #openstack-cinder07:52
openstackgerritVincent Hou proposed openstack/cinder: Remove the destination volume check in delete_volume
*** links has quit IRC08:02
*** zhipeng has joined #openstack-cinder08:03
*** ndipanov has joined #openstack-cinder08:06
*** vbala has left #openstack-cinder08:12
*** zhipeng has quit IRC08:14
*** zhipeng has joined #openstack-cinder08:15
*** alexschm has joined #openstack-cinder08:17
*** zhipeng_ has joined #openstack-cinder08:23
*** hodos has joined #openstack-cinder08:25
*** zhipeng has quit IRC08:25
*** zhipeng_ has quit IRC08:27
*** zhipeng has joined #openstack-cinder08:29
*** jistr has joined #openstack-cinder08:37
*** zhipeng has quit IRC08:39
*** zhipeng has joined #openstack-cinder08:41
*** guest1 has joined #openstack-cinder08:45
*** guest1 has quit IRC08:46
*** vgridnev has joined #openstack-cinder08:58
*** d0ugal has left #openstack-cinder09:02
*** vgridnev has quit IRC09:02
*** zhipeng has quit IRC09:02
*** vgridnev has joined #openstack-cinder09:03
*** vgridnev has quit IRC09:03
*** vgridnev has joined #openstack-cinder09:03
*** bluex-pl has joined #openstack-cinder09:16
*** bluex-pl has quit IRC09:16
*** bluex-pl has joined #openstack-cinder09:16
openstackgerritDaniel Tadrzak proposed openstack/cinder: CGSnapshot Object
*** belmoreira has joined #openstack-cinder09:26
*** e0ne has joined #openstack-cinder09:32
duleke0ne: Hi! Can you please take a look at my comment on to confirm I'm not talking nonsense?09:55
*** IanGovett has joined #openstack-cinder10:02
*** e0ne has quit IRC10:04
*** vbala_ has joined #openstack-cinder10:05
*** vbala_ has quit IRC10:06
*** e0ne has joined #openstack-cinder10:08
*** kbyrne has quit IRC10:12
DuncanTdulek: This python stuff is all nonsense.10:13
*** kbyrne has joined #openstack-cinder10:15
*** vbala has joined #openstack-cinder10:16
DuncanTdulek: On a more serious note, I think it is going to be a worthwhile summit session to define exactly what AZ means to different people, and what AZ usecases we want to support. My understanding of an AZ means that the idea of a fallback AZ makes little sense, but clearly other people have a different idea of how AZs should work10:16
e0neDuncanT: +1. we need more feedback from operators10:18
DuncanTe0ne: I'm just emailing the submitter of that bug now. At the very least I can write up how we want AZs to work, and hopefully get details of how they want AZs to work too, then try to ensure cinder can do both10:19
DuncanTe0ne: Currently it's a bit organic, which I don't like - there's no good advice on how to actually build a resilient system10:19
DuncanTe0ne: If you have an idea of how AZs should work, please drop me an email (and anybody else who thinks they have an understanding of what they want from AZs - regardless of what we currently have)10:20
e0neDuncanT: personal mail for you or to openstack-dev?10:22
DuncanTe0ne: To me directly to start with please, I'll try to summarise before we take it to the list10:23
e0nedulek: i'm agree with DuncanT's comment10:23
e0nedulek: tbh, i don't like this patch but won't block it if operators want such strange feature10:24
dulekDuncanT: Haven't all 3 threads on openstack-dev on the topic explained what operators expect?10:24
e0neDuncanT: ok. I will mail you and try to get feedback from our ops10:24
e0nedulek: we still don't know WHY they want it10:25
DuncanTdulek: Not clearly, no. I will go through those as well though, and try to get diagrams of the (at least) two different models that are being deployed. We can then look at the failure cases for each model and see if it makes sense10:25
e0nedulek: it was an issue in Nova. and know they say: we don't want that issue to be fixed. it's good enough for us.10:26
DuncanTdulek: Having a good clear statement of what each group is trying to achieve would help me a lot for certain, and I doubt I'm alone. For my mental model of AZs, what is being asked for in that patch makes no sense, and providing a volume in anything other than the requested AZ is a bug10:27
dulekDuncanT, e0ne: An explanation why an operator want it:
dulekDuncanT: And I agree that diagrams and analysis of possible failures would be beneficial.10:28
e0nedulek, DuncanT: I saw it.  that's why I proposed an option AZ per backend10:28
duleke0ne, DuncanT: Sorry, got to go to a meeting, I'll answer you in some time.10:28
DuncanTdulek: But that explaination makes little sense unless they're willing to accept that cinder node failures will take down multiple AZs, and they are allowing nova to mount cross-AZ - if that is the case then fine, but I want to make that clear10:29
e0nedulek: but DuncanT's concern about 'fake AZ' looks right for some cases10:29
e0neDuncanT: +110:29
DuncanTdulek: I can then document how to configure that setup, and document what the failure modes are, and the advantages and disadvantages of doing so10:29
DuncanTdulek: I want all of the supported configurations to be documented10:30
*** links has joined #openstack-cinder10:31
*** yrabl has quit IRC10:33
*** aix has quit IRC10:46
*** deepakcs has quit IRC10:54
*** ganso_ has joined #openstack-cinder10:55
*** aix has joined #openstack-cinder11:15
*** ishant has quit IRC11:19
*** pschaef has joined #openstack-cinder11:19
*** eharney has joined #openstack-cinder11:21
*** BharatK has quit IRC11:22
*** takedakn has joined #openstack-cinder11:25
*** sayali has quit IRC11:26
*** timcl has joined #openstack-cinder11:27
nikeshmDotHill CI is passed on the bug-fix by hemna, its need one more +211:36
*** ociuhandu has quit IRC11:36
*** takedakn has quit IRC11:36
*** BharatK has joined #openstack-cinder11:36
*** takedakn has joined #openstack-cinder11:36
*** julim has joined #openstack-cinder11:37
*** Yogi1 has joined #openstack-cinder11:37
*** svilgelm has quit IRC11:39
*** takedakn1 has joined #openstack-cinder11:41
*** Yogi1 has quit IRC11:42
*** takedakn has quit IRC11:44
*** takedakn1 is now known as takedakn11:44
*** bkopilov has quit IRC11:51
*** breitz has quit IRC11:51
*** breitz has joined #openstack-cinder11:51
*** takedakn has quit IRC11:52
*** ociuhandu has joined #openstack-cinder11:55
*** diogogmt has joined #openstack-cinder11:56
*** alexpilotti_ has joined #openstack-cinder11:58
*** guest1 has joined #openstack-cinder12:11
*** pschaef has quit IRC12:11
DuncanTnikeshm: Approved. Any idea why the CI failed? Something you fixed or an intermittent issue?12:12
nikeshmDuncanT: Vedams CI is different from recheck-HP Storage CI12:14
nikeshmDuncanT: vedams CI was not failed on this patch12:14
DuncanTnikeshm: Oh, sorry, I misread the history. Thanks.12:15
nikeshmDuncanT: looks like recheck-HP Storage CI was failed12:15
nikeshmDuncanT: thanks for +2, :) , so what weeked plan (friday and saturday)12:16
DuncanTnikeshm: Beer. Lots of beer. Some dancing. Maybe a trip down to the desert.12:17
*** guest1 has quit IRC12:17
*** raildo-afk is now known as raildo12:18
DuncanT@nikeshm I should probably put up my casual review friday list before the end of the day, too12:19
nikeshmDuncanT: happy weekend after long tiring week12:20
DuncanTnikeshm: Cheers :-)12:21
*** JoseMello has joined #openstack-cinder12:22
*** martyturner has joined #openstack-cinder12:23
*** svilgelm has joined #openstack-cinder12:24
*** xyang1 has joined #openstack-cinder12:24
*** shyama has quit IRC12:25
*** akerr has joined #openstack-cinder12:27
*** diogogmt has quit IRC12:27
*** akerr has quit IRC12:28
*** svilgelm has quit IRC12:28
*** akerr has joined #openstack-cinder12:29
*** Yogi1 has joined #openstack-cinder12:31
*** diogogmt has joined #openstack-cinder12:31
*** svilgelm has joined #openstack-cinder12:34
*** stefan_amann has quit IRC12:34
*** shyama has joined #openstack-cinder12:34
*** stefan_amann has joined #openstack-cinder12:35
*** porrua has joined #openstack-cinder12:36
openstackgerritGerald McBrearty proposed openstack/cinder: Switch SVC driver to use lsportfc to determine FC target WWPNS
*** diogogmt has quit IRC12:40
*** dims has joined #openstack-cinder12:41
*** Yogi1 has quit IRC12:41
geguileodulek: ping - rolling upgrades with VOs12:42
*** sayali has joined #openstack-cinder12:42
dulekgeguileo: Hello!12:44
geguileodulek: Hi12:44
dulekgeguileo: What's up? :)12:44
geguileodulek: New worries, that's what's up  :-(12:45
openstackgerritabhiram moturi proposed openstack/cinder: ZFSSA driver to return project 'available' space
geguileodulek: I'm worried about rolling migrations with VOs12:45
geguileodulek: Because I believe we don't have the structure to support it12:45
dulekgeguileo: Okay, can you elaborate?12:45
geguileodulek: For upgrades we may have DB migrations12:46
* geguileo hoping he's wrong and dulek can remove his worries12:46
*** diogogmt has joined #openstack-cinder12:46
DuncanTVOs? What is a VO?12:47
geguileoDuncanT: Versioned Objects12:47
geguileodulek: So how do we manage DB migrations with the rolling upgrades?12:47
DuncanTgeguileo: We don't, yet12:48
DuncanTgeguileo: Some of the infrastructure is there, but far from all12:48
dulekgeguileo: Give me a minute to rethink if I'm not missing something.12:48
*** anshul has quit IRC12:49
geguileoDuncanT: My worry is that our current architecture does not allow us to do rolling upgrades12:49
DuncanTgeguileo: Well, given it isn't finished, no it doesn't. It will do though - thang had a blueprint for how it will fit together12:49
geguileodulek: Thanks, though I'm really hopping that it's me the one that is missing something12:49
geguileoDuncanT: I'll have a look, but I think it's missing the part that worries me12:50
DuncanTgeguileo: Basically we tell all services to pin the max object version they send over the wire (downgrading before send as necessary) until every service is upgraded, then we remove the pin12:50
geguileoDuncanT: Yeah, that's the easy stuff12:51
geguileoDuncanT: What about schema migrations?12:51
dulekgeguileo: Okay, so Nova wants to address it with two-phase DB migrations - first phase adds new columns and second removes them when all services are upgraded12:51
geguileodulek: Nova has a Conductor12:51
DuncanTgeguileo: I'm sure we had an answer, but I need to go back and re-read plans12:51
dulekgeguileo: I don't see how conductor changes two-phase migrations idea.12:52
*** sayali has quit IRC12:52
geguileoDuncanT: I'll have a look at the spec12:52
DuncanTgeguileo: I think the basic answer is objects need to be able to talk to both schema versions12:52
DuncanTgeguileo: I'm not sure it got as far as a concrete design, now you come to mention it :-(12:52
geguileodulek: So all new added colums will forcefully have default values or allow NULL, right?12:53
dulekgeguileo: (Neuron spec is a little better defined than Nova, and Neutron has no conductor)12:53
geguileodulek: Wrong12:53
geguileodulek: I spoke with a Neutron guy yesterday12:53
dulekgeguileo: Actually you cannot add new column to the DB without def value or allowing NULL.12:54
geguileodulek: And they have a kind of conductor12:54
geguileodulek: No, I mean, that new entries must have default value (not existing ones)12:54
geguileodulek: So neutron passes all DB access through their kind of conductor and then manage messages to agents12:55
*** chutwig has joined #openstack-cinder12:55
*** sayali has joined #openstack-cinder12:56
xekgeguileo, the code that actually talks with the db is sqlalchemy12:56
geguileoXD XD XD12:56
geguileoxek: That's not what we are talking about12:57
geguileoxek: That's a library for the abstraction layer12:57
xekgeguileo, so in a case of an upgrade, there may be two different sqlalchemy models, and that's why two-phase migrations are a good idea12:58
dulekgeguileo, DuncanT: xek is my team member who's a little more into oslo.VO and generally upgrades in various projects.12:58
geguileoxek: Yep, we would have different models on different nodes12:58
dulekgeguileo: About conductor stuff. I'm not that familiar with Neutron, so let's get back to Nova. It isn't true that all DB accesses in Nova's services go through conductor.12:59
*** BharatK has quit IRC13:00
geguileodulek: Well, they are "supposed to"   XD13:00
*** tpsilva has joined #openstack-cinder13:01
dulekgeguileo: I was always confused by what was the real goal of conductor, but you may be right. ;)13:01
xekdulek, geguileo it depends on the configuration13:01
xekwhat is always done by the conductor is backporting of objects for the computes which run with an older version13:02
openstackgerritOleksii Butenko proposed openstack/python-cinderclient: Add functional tests for python-cinderclient
geguileoxek: So compute nodes send requests directly to DB if DB version is the same as their VOs and through conductor if they are not?13:03
* geguileo must be understanding it wrong, because that sounds weird13:03
xekgeguileo, it depends on the  CONF.conductor.use_local setting13:05
* dulek is going to talk through a common version of what's happening with xek. ;)13:05
geguileodulek: What do you mean by that ^?13:05
openstackgerritKendall Nelson proposed openstack/cinder: WIP: Dynamically create cinder.conf.sample
*** timcl has quit IRC13:08
openstackgerritVipin Balachandran proposed openstack/cinder: VMware: Fix exception messages
dulekgeguileo: I've needed to make sure that we understand each other correctly, and that's easier to do in person and in my native language. ;)13:11
*** diogogmt_ has joined #openstack-cinder13:12
dulekgeguileo, DuncanT: Now I think that to handle DB migrations we also need online DB schema migrations like Nova or Neutron.13:12
dulekgeguileo, DuncanT: Given that we're actually removing or renaming columns through migrations, which I believe isn't happening.13:12
*** diogogmt has quit IRC13:13
*** diogogmt_ is now known as diogogmt13:13
* dulek is glancing through Liberty migration scripts.13:13
*** edmondsw has joined #openstack-cinder13:13
geguileodulek: But we may want to in the future13:13
*** dustins has joined #openstack-cinder13:15
*** sayali has quit IRC13:16
*** BharatK has joined #openstack-cinder13:16
dulekgeguileo: I agree, we should have that possibility.13:16
dulekgeguileo: From what I see we don't have any incompatible DB schema changes in Liberty, but that doesn't mean there won't be any in the future.13:17
*** Yogi11 has joined #openstack-cinder13:18
*** gouthamr has joined #openstack-cinder13:18
dulekgeguileo, DuncanT: Is two-phase DB schema migrations a solution for that in your opinion?13:18
duleks/Is/Are ;)13:18
geguileodulek: I'm going to give it some thought, as it isn't a trivial matter13:19
geguileodulek: But the first thing that comes to mind is that it could work13:19
DuncanTdulek: I don't know. I think we need a dummy migration as a demo... a migration that adds a column to one table and removes one from another, and shows the objects doing a version adaption13:19
DuncanTdulek: I guess adding and removing a column from the same table is the third test case13:20
dulekDuncanT: I don't think that's how it is supposed to work.13:20
DuncanTdulek: oh?13:20
dulekDuncanT: It's more like - add columns before upgrades - everything works fine, older versions just ignore the column (given it has default in the DB or is nullable).13:21
dulekDuncanT: After everything is upgraded - run second phase that will remove the unused columns.13:21
*** sayali has joined #openstack-cinder13:22
dulekDuncanT: A harder case would be how to handle conversions of column values. I'm not sure on that, need to take a look into Nova's and Neutron's specs.13:22
*** sayali has quit IRC13:24
dulekDuncanT: But Nova needs also to handle such cases - there can be a lot of conductors and these may not be upgraded at once.13:24
*** BharatK has quit IRC13:26
*** sayali has joined #openstack-cinder13:27
e0nevolume attachment API is broken in nova:(. only non-voting job in cinder found this issue:(13:35
e0ne*non-voting rally job13:35
*** jgregor has joined #openstack-cinder13:36
geguileoe0ne: With the new tests you added?13:37
e0negeguileo: no. volume attach scenario was merged a while ago13:38
e0nenobody cares that python-somethingclient works13:38
e0nenobody cares about non-voting job failed13:38
geguileoe0ne: Which client is broken?13:39
e0negeguileo: actually, it's not nova-client. it's nova api13:39
e0negeguileo: but simple test for client could find such issue13:39
geguileoe0ne: And their CI didn't catch it?13:39
e0negeguileo: no.13:39
openstackLaunchpad bug 1491842 in Cinder "Can't attach volume without specifying device name" [High,Confirmed]13:40
*** merooney has joined #openstack-cinder13:40
* geguileo has finally lost his faith in OpenStack's CIs13:40
*** links has quit IRC13:40
jgriffithgeguileo: I'm impressed that you ever had any13:40
jgriffithgeguileo: more != better13:41
geguileojgriffith: Well, it only lasted 6 months so...  ;-)13:41
jgriffithgeguileo: HA13:41
e0nejgriffith: :)13:42
e0nemy first lesson with openstack was: do not read the documentation. read the code!13:43
jgriffithe0ne: so I'm curious however on that bug13:43
geguileoe0ne: Yep, I had that one early on as well13:44
e0nejgriffith: imo,  root cause is: we don't test our clients with API well. tempest tests only API13:46
jgriffithe0ne: so that IS one of the reasons we were supposed to start moving functional in to Cinder remember13:47
jgriffithalthough this case I don't understand yet and it wouldn't have found it13:47
e0nejgriffith: absolutly agree13:47
jgriffithe0ne: I'm trying to understand the bug, and what exactly rally does that exposes it?13:48
e0nejgriffith: what do you mean?13:48
jgriffithI'm doing nova volume-attach all day long here13:48
e0nejgriffith: rally calls something like "nova volume-attach <instance_id> <volume_id>"13:48
*** eharney has quit IRC13:49
e0nejgriffith: but for now, nova requires something like "nova volume-attach <instance_id> <volume_id> <device_name>"13:49
*** asselin_ has joined #openstack-cinder13:50
jgriffithe0ne: yeah, so I'm trying to understand what's the "cinder" bug13:50
jgriffithsounds like a Rally bug13:50
* jgriffith is confused13:50
e0nejgriffith: it's a nova bug13:50
e0nejgriffith: my concern was: only Cinder CI found it13:50
jgriffithe0ne: although dberrange (I think) had proposed about dropping dev altogether13:50
jgriffithe0ne: oh... I see13:51
*** anshul has joined #openstack-cinder13:51
jgriffithe0ne: well, kinda :)13:51
johnthetubaguylooking to tag a new python-novaclient soon, thats very related13:51
jgriffithkinda not13:51
jgriffithjohnthetubaguy: :)13:51
jgriffithjohnthetubaguy: I knew there was some discussion I caught around this, just couldn't remember details :)13:52
johnthetubaguythe device_name is just not properly honoured, so we are trying to remove that (in the eventual sense)13:52
johnthetubaguybut thats a different discussion I guess13:52
jgriffithjohnthetubaguy: yeah, I know Vish fought it for years, sounds like it's time to just punt on it and be done :)13:52
jgriffithjohnthetubaguy: but I'm confused, because apparantly e0ne is expecting it to work today without the "dev" arg13:53
*** alexpilotti_ is now known as alexpilotti13:53
jgriffithwhich I never thought was something that was done13:53
jgriffithor supported yet, just talked about13:53
e0nejgriffith, johnthetubaguy: "device" argument should be optional13:53
johnthetubaguyI kinda thought thats been optional for a while, but I would have to go dig to say when, I could be wrong13:54
jgriffithe0ne: I must be wrong, I'll look; I thought the "option" was "dev=auto"13:54
johnthetubaguyoptional is a bit racey, as it just picks the "next" one13:54
jgriffithbut honestly haven't looked in quite a while so I should likely be ignored :)13:54
johnthetubaguyjgriffith: oh, possibly...13:54
johnthetubaguyyeah, not dug recently either13:55
e0nejohnthetubaguy: the same for the python-novaclient13:56
*** sayali has quit IRC13:57
jgriffithhmm... well then13:57
*** timcl has joined #openstack-cinder13:58
e0nelooks like optional params are required now13:58
*** mriedem has joined #openstack-cinder13:59
*** sayali has joined #openstack-cinder13:59
*** Yogi11 has quit IRC14:02
*** bnemec has quit IRC14:02
*** zhongjun has quit IRC14:02
*** bnemec has joined #openstack-cinder14:02
*** rushil has joined #openstack-cinder14:04
openstackgerritMarian Horban proposed openstack/cinder: Remove wsgi functionality
*** svilgelm has quit IRC14:07
*** thangp has joined #openstack-cinder14:08
*** stefan_amann has quit IRC14:08
*** rushil has quit IRC14:09
openstackgerritPeter Penchev proposed openstack/cinder: Reintroduce the StorPool driver.
*** stefan_amann has joined #openstack-cinder14:12
openstackgerritPeter Penchev proposed openstack/os-brick: Add the StorPool brick connector.
*** eharney has joined #openstack-cinder14:19
dulekjgriffith: - is my understanding presented in the comments there valid?14:28
*** rushil_ has joined #openstack-cinder14:28
*** vbala has quit IRC14:30
*** amoturi has quit IRC14:30
*** bkopilov has joined #openstack-cinder14:32
*** Yogi1 has joined #openstack-cinder14:32
*** devlaps has quit IRC14:33
*** eharney has quit IRC14:34
*** mtanino has joined #openstack-cinder14:35
*** chutwig has quit IRC14:36
jgriffithdulek: ahh... indeed14:37
*** markstur has quit IRC14:38
*** theanalyst has quit IRC14:39
*** amoturi has joined #openstack-cinder14:39
*** sayali has quit IRC14:40
jgriffithdulek: Edmund made some valid points as well.  But the gist of it is that "yes" we want to log the resource we're creating.  We don't have any handle to that in this case though14:40
*** asselin__ has joined #openstack-cinder14:40
jgriffithdulek: My opinion would be to merge this and file a bug, rather than drag Edmund through another change on it14:40
jgriffithdulek: mostly because he added it at my request ;)14:40
*** garthb has joined #openstack-cinder14:40
*** theanalyst has joined #openstack-cinder14:41
dulekjgriffith: Okay, I'm getting my -1 back.14:42
*** asselin_ has quit IRC14:42
*** markstur has joined #openstack-cinder14:43
jgriffithdulek: thanks for pointing that out, I updated my vote and comment14:44
nikeshmwhich nova service to restart after changing in devstack14:44
jgriffithnikeshm: api usually I tihnk14:45
dulekjgriffith: Ah, we're going this way. Okay, it's fine for me. Thanks for help!14:45
jgriffithdulek: ?14:45
openstackgerritTom Barron proposed openstack/cinder: NetApp E-Series over-subscription support
jgriffithdulek: No, I'm saying your -1 was correct/valid... I'm agreeing with you :)14:45
dulekjgriffith: "My opinion would be to merge this and file a bug" and then you get your +2 back.14:46
dulekjgriffith: That got me confused, but I agree getting that in RC is proper option.14:46
*** diogogmt_ has joined #openstack-cinder14:46
jgriffithdulek: yeah, I don't communicate well :)14:46
*** diogogmt has quit IRC14:47
*** diogogmt_ is now known as diogogmt14:47
dulekjgriffith: It's fine. ;)14:47
*** shyama has quit IRC14:48
*** eharney has joined #openstack-cinder14:48
*** asselin_ has joined #openstack-cinder14:49
*** amoturi has quit IRC14:49
*** asselin__ has quit IRC14:52
*** mtanino has quit IRC14:53
*** anshul has quit IRC14:53
*** jungleboyj has joined #openstack-cinder14:56
*** erhudy has joined #openstack-cinder14:59
*** Zhongjun_ has joined #openstack-cinder15:00
*** edtubill has joined #openstack-cinder15:00
*** jistr is now known as jistr|call15:02
*** ronis has quit IRC15:02
*** rushil_ has quit IRC15:04
*** sayali has joined #openstack-cinder15:05
*** markvoelker has joined #openstack-cinder15:08
*** esker has joined #openstack-cinder15:10
*** sayali has quit IRC15:10
*** vgridnev has quit IRC15:12
*** chutwig has joined #openstack-cinder15:14
*** esker has quit IRC15:14
*** alexpilotti_ has joined #openstack-cinder15:16
*** alexpilotti has quit IRC15:17
*** alexpilotti_ is now known as alexpilotti15:17
*** shyama has joined #openstack-cinder15:17
openstackgerritEdmund Rhudy proposed openstack/cinder: Adds allow_availability_zone_fallback option to Cinder
*** mriedem has quit IRC15:21
*** annegentle has joined #openstack-cinder15:21
*** sayali has joined #openstack-cinder15:22
*** p0rtal has joined #openstack-cinder15:26
*** dims has quit IRC15:26
*** p0rtal has quit IRC15:26
*** p0rtal has joined #openstack-cinder15:26
*** hemnafk is now known as hemna15:28
*** thangp has quit IRC15:29
*** harlowja_at_home has joined #openstack-cinder15:29
*** jdurgin1 has joined #openstack-cinder15:31
*** dannywilson has joined #openstack-cinder15:31
*** sayali has quit IRC15:32
*** nkrinner has quit IRC15:34
*** stefan_amann has quit IRC15:35
*** dims has joined #openstack-cinder15:36
*** sayali has joined #openstack-cinder15:36
*** mtanino has joined #openstack-cinder15:36
openstackgerritJohn Griffith proposed openstack/cinder: Use oslo-config-generator
*** markvoelker_ has joined #openstack-cinder15:42
*** thangp has joined #openstack-cinder15:42
*** jistr|call is now known as jistr15:44
openstackgerritJohn Griffith proposed openstack/cinder: Use oslo-config-generator
*** markvoelker has quit IRC15:45
*** harlowja_at_home has quit IRC15:48
*** breitz has quit IRC15:49
*** jdurgin1 has quit IRC15:50
*** rushil has joined #openstack-cinder15:56
*** bkopilov has quit IRC15:58
*** afazekas__ has quit IRC15:59
*** p0rtal_ has joined #openstack-cinder16:00
*** p0rtal has quit IRC16:03
*** zhongjun has joined #openstack-cinder16:03
*** p0rtal_ has quit IRC16:04
*** jistr has quit IRC16:06
*** jaypipes has joined #openstack-cinder16:07
*** bkopilov has joined #openstack-cinder16:10
*** p0rtal has joined #openstack-cinder16:11
*** p0rtal has quit IRC16:11
*** leeantho has joined #openstack-cinder16:11
*** aix has quit IRC16:12
*** p0rtal has joined #openstack-cinder16:12
*** crose has joined #openstack-cinder16:14
*** ndipanov has quit IRC16:21
*** shyama has quit IRC16:25
*** chutwig has quit IRC16:25
*** sayali has quit IRC16:26
*** pots1 has joined #openstack-cinder16:27
*** chutwig has joined #openstack-cinder16:29
*** pots has quit IRC16:29
*** mudassirlatif has joined #openstack-cinder16:31
*** vgridnev has joined #openstack-cinder16:32
*** shyama has joined #openstack-cinder16:32
*** ronis has joined #openstack-cinder16:34
*** e0ne has quit IRC16:34
*** sayali has joined #openstack-cinder16:35
*** annegentle has quit IRC16:36
*** dflorea has joined #openstack-cinder16:37
*** annegentle has joined #openstack-cinder16:37
*** esker has joined #openstack-cinder16:38
*** afazekas__ has joined #openstack-cinder16:40
*** dflorea has quit IRC16:41
*** mriedem has joined #openstack-cinder16:42
*** p0rtal_ has joined #openstack-cinder16:43
*** sghanekar has joined #openstack-cinder16:45
*** dflorea has joined #openstack-cinder16:45
*** p0rtal has quit IRC16:46
*** p0rtal_ has quit IRC16:47
*** lpetrut has quit IRC16:49
*** yuriy_n17 has quit IRC16:53
*** martyturner has quit IRC16:57
*** haomaiw__ has joined #openstack-cinder17:02
*** timcl has quit IRC17:02
*** Lee1092 has quit IRC17:03
*** haomaiwa_ has quit IRC17:03
*** bluex-pl has quit IRC17:04
*** p0rtal has joined #openstack-cinder17:04
*** p0rtal has quit IRC17:05
*** martyturner has joined #openstack-cinder17:05
*** belmoreira has quit IRC17:07
*** afazekas__ has quit IRC17:08
*** afazekas__ has joined #openstack-cinder17:12
openstackgerritDerrick Wippler proposed openstack/cinder: Hacking log format arg check
*** timcl has joined #openstack-cinder17:14
*** markvoelker_ has quit IRC17:16
*** erhudy has quit IRC17:19
*** julim has quit IRC17:21
*** afazekas__ has quit IRC17:21
*** kvidvans has joined #openstack-cinder17:23
*** esker has quit IRC17:23
*** zul has quit IRC17:26
*** zul has joined #openstack-cinder17:27
*** zul has quit IRC17:27
*** zul has joined #openstack-cinder17:28
*** martyturner has quit IRC17:30
jgriffithdiablo_rojo: hey17:31
*** annegentle has quit IRC17:33
*** hogepodge has joined #openstack-cinder17:33
hogepodgehi o/17:33
hogepodgeis there a good place to see the current list of drivers that are being tested in cinder CI?17:33
hogepodgeI've looked at the wiki, but that seems to be updated manually.17:33
jgriffithhogepodge: yah, that's a fun game :)17:34
jgriffithhogepodge: so two things I can think of:17:34
asselinhogepodge, this is a new one that I like:
jgriffithasselin: :)17:34
jgriffiththat's the first one :)17:34
hogepodgeasselin: oh, that's exactly what I need in the format I need it in.17:35
asselinmakes IBM GPFS CI look bad though17:35
asselinjungleboyj, ^^17:35
*** timcl has quit IRC17:36
jgriffithasselin: theyre flash one too17:36
jgriffithfunny how "nothing" ever fails17:36
jgriffithguess that's one way to do it :)17:36
asselinglad to hear that random failures are good to have :)17:37
*** timcl has joined #openstack-cinder17:37
*** alexschm has quit IRC17:37
SwansonWhile that page makes one of the dell drivers look really bad in our defense someone stole the SAN.17:38
nikeshmhemna: hi, thanks17:38
*** timcl1 has joined #openstack-cinder17:39
nikeshmhemna: is this also need to change
asselinSwanson, actually that's what I like about it. I can quickly see that I should ignore those failure results17:39
nikeshmor already some patch is there to change17:39
nikeshmhemna: thanks17:41
*** timcl has quit IRC17:41
*** annegentle has joined #openstack-cinder17:42
*** lpetrut has joined #openstack-cinder17:43
jgriffithSwanson: LMAO17:44
*** e0ne has joined #openstack-cinder17:46
*** p0rtal has joined #openstack-cinder17:47
*** sayali has quit IRC17:50
*** sayali has joined #openstack-cinder17:53
*** Zhongjun_ has quit IRC17:57
*** dflorea has quit IRC17:58
*** julim has joined #openstack-cinder17:58
*** lpetrut has quit IRC18:06
*** sayali has quit IRC18:06
*** martyturner has joined #openstack-cinder18:10
*** esker has joined #openstack-cinder18:17
*** esker has quit IRC18:19
*** garthb has quit IRC18:19
*** esker has joined #openstack-cinder18:21
openstackgerritVipin Balachandran proposed openstack/cinder: VMware: Fix exception messages
*** agarciam has quit IRC18:26
*** agarciam has joined #openstack-cinder18:26
*** rushil has quit IRC18:29
* hemna says the panic about volume migration sets in.....18:31
*** harlowja has quit IRC18:31
*** annegentle has quit IRC18:31
Swansonhemna: panic?18:31
* patrickeast panics18:32
*** harlowja has joined #openstack-cinder18:32
openstackgerritVipin Balachandran proposed openstack/cinder: VMware: Fix exception messages
e0newhat did i miss?18:33
patrickeastew yea i saw that18:34
hemnawell, for whatever reason I didn't know about update_migrated_volume and the mess that not implementing that creates.18:34
*** martyturner has quit IRC18:34
hemnawhat a freaking disaster18:34
patrickeastwe should definitely have to documented somewhere if it isn't already18:35
Swansonhemna: I had to implement update_migrated_volume to keep from losing track of my volumes.  Was horrible.18:35
hemnaI'm pretty sure our drivers are broken if someone migrates18:36
patrickeastif i understand correctly i'm safe by using 'name' with mine... but now i really need to go test it more thoroughly... again18:36
hemnawe use id18:36
hemnawah wah wahhhh18:37
patrickeasthaha now i'm glad you blocked my change to try and switch from name to id18:37
SwansonI was told to use id. I still use id.  I just have update_migrated_volume to fix things up.18:37
* hemna fires up a new vagrant to test migrating from lvm -> 3PAR/LeftHand18:37
*** esker has quit IRC18:38
hemnaI can taste the exception stack trace now.....18:38
SwansonSpeaking of migration did live migration work in kilo?18:38
hemnathis also seems like it should be a test case handled in 3rd party CI18:39
hemnanova live migration?18:39
*** esker has joined #openstack-cinder18:39
hemnaheh, Swanson I'd say no*18:39
*** esker has quit IRC18:39
hemnafor some drivers it might have.18:39
*** esker has joined #openstack-cinder18:40
SwansonI have a change in that fixes multiple calls to initialize_connection but I'd heard it was broken beyond that.  But I might be thinking of something else.18:40
SwansonNeed my full test environment working again....18:40
hemnaFC was probably borked with multipath enabled18:40
hemnathere is a bunch of code in nova that tries to preserve the multipath id18:41
hemnait's a mess18:41
hemnaI have plans to clean that out in M, now that nova uses brick and brick doesn't even use multipath_id anymore.18:41
*** timcl1 has quit IRC18:41
SwansonI think the issue is with iscsi.  Timezones are not helping me.18:42
openstackgerritGorka Eguileor proposed openstack/cinder: Remove API races from delete methods
openstackgerritGorka Eguileor proposed openstack/cinder: Add atomic conditional updates to objects
openstackgerritGorka Eguileor proposed openstack/cinder: WIP: Remove more API races
openstackgerritGorka Eguileor proposed openstack/cinder: Move get_by_id to CinderObject
openstackgerritGorka Eguileor proposed openstack/cinder: Improve metadata update operations
openstackgerritGorka Eguileor proposed openstack/cinder: Remove API races from attach and detach methods
openstackgerritPatrick East proposed openstack/cinder: Use consolidated update for failover_replication
*** kvidvans has quit IRC18:52
*** eharney has quit IRC18:52
*** lpetrut has joined #openstack-cinder18:53
*** jwcroppe_ is now known as jwcroppe18:58
*** martyturner has joined #openstack-cinder18:59
*** annegentle has joined #openstack-cinder19:01
openstackgerritGerald McBrearty proposed openstack/cinder: Switch SVC driver to use lsportfc to determine FC target WWPNS
*** eharney has joined #openstack-cinder19:07
*** timcl has joined #openstack-cinder19:08
*** agarciam has quit IRC19:12
openstackgerritDerrick Wippler proposed openstack/cinder: Hacking log format arg check
diablo_rojojgriffith: Hey. Sorry I couldn't reply right away;  I was in meetings.19:19
*** vgridnev has quit IRC19:21
*** alexpilotti has quit IRC19:26
*** lpetrut has quit IRC19:27
*** _cjones_ has joined #openstack-cinder19:29
*** lpetrut has joined #openstack-cinder19:30
*** changbl has quit IRC19:30
*** akerr has quit IRC19:31
*** _cjones_ has quit IRC19:31
diablo_rojojgriffith: I was confused by your pushing up a patch based heavily on Sergey's given your and the community's distaste with that approach.  I pushed up my WIP patch in the hopes that we could work together on it and was surprised to see that you had pushed up your own today returning to the non-dynamic approach. I would love to collaborate with you particularly since you seem to have a good understanding of the oslo-config-generat19:31
*** _cjones_ has joined #openstack-cinder19:32
jgriffithdiablo_rojo: couple things:19:32
jgriffithMy *distaste* doesn't really matter :)  and my distaste was in reference to something slightly different.  The reality is this is how it's being implemented currently.19:32
jgriffithdiablo_rojo: also, back in July I thought we'd have something done by Sept :)19:33
jgriffithdiablo_rojo: finally... it turns out that even the "auto generated" version seems to have a good bit of hard coding in it as well19:33
diablo_rojoRight, but a dynamic way of implementing it is what the community wanted I thought19:33
jgriffithdiablo_rojo: Yeah... and I want a pony :)19:33
diablo_rojoAnd no one had attempted it previously becuase it has been incredibly difficult19:34
jgriffithdiablo_rojo: my point is we need to fix it, if that's with your version, great19:34
diablo_rojoGotta give me a little credit with how new I am :)19:34
jgriffithif it's with Sergey and my version (like every other openstack project) that's cool too19:34
jgriffithor if it's step 1, then step 2 that's even beter19:34
jgriffithdiablo_rojo: OHHH19:34
jgriffithdon't take it the wrong way19:34
jgriffithNot a knock on you or your work at all!!!19:35
jgriffithThat's not the point19:35
jgriffithdiablo_rojo: sorry if it came across that way, not my intent19:35
jgriffithbottom line, it's a bug, it needs fixed... that's all19:35
diablo_rojoIf you and he are doing all the oslo-config-stuff right thats awesome. We can put my script to generate the cinder opts with what you have19:35
jgriffithdiablo_rojo: well... about that ;)19:35
jgriffithdiablo_rojo: a few things about your script we should look at19:36
*** lpetrut has quit IRC19:36
jgriffithdiablo_rojo: so there's no real reason to do the "import xxxx as some-name", we can nuke that step out of there19:36
*** _cjones_ has quit IRC19:36
jgriffithdiablo_rojo: I'm not sure I follow why a bunch of it is hard-coded?19:36
diablo_rojoI was doing that so that later when we reference the opts there is a name to get them from.19:37
diablo_rojoWhich hard coded parts are you referring to?19:37
diablo_rojoinitial setup?19:37
jgriffithkymgr, BRCD,CISCO BRCD....19:38
diablo_rojoThats the different sections19:38
diablo_rojothats not specific opt lists being hard coded19:38
diablo_rojoin Nova and other projects they have the lists of opts that have been registered broken up into sections19:38
jgriffithyeah... ?19:39
jgriffithand ?19:39
diablo_rojoso like CONF.register_opts(lost_o_opts, group = keymgr)19:39
diablo_rojolost_o_opts shouldn't go into the default, it needs to get sorted into the proper section19:39
diablo_rojowhich is what that is19:39
jgriffithdiablo_rojo: yes, I'm familiar, not following though19:39
*** rmetcalf has joined #openstack-cinder19:40
jgriffithdiablo_rojo: my point is "it's not an auto-generated" opts file if you hvae to hard code the sections in it :)19:40
diablo_rojoYou had said that I was hardcoding things?19:40
jgriffithsee what I mean?19:40
jgriffithdiablo_rojo: so look... like I said... not knocking what you're doing19:40
jgriffithdiablo_rojo: more than happy to just leave it up to you19:40
jgriffithdiablo_rojo: if you want me to start reviewing what you have and making suggestions happy to do it19:41
jgriffithdiablo_rojo: just say the word19:41
jgriffithbut we do need this and db migration fixed... :)19:41
diablo_rojoI get it. I can work on coming up with a way of doing that and push it up in another patch later on. I think the important thing right now is to get what you have and what I have together and merged.19:41
rmetcalfI need some Cinder code reading advice. With REST APIs how do I determine the code path for a GET and PUT when the URL is the same (/v2/tenant/type, list or create volume types)? Thanks in advance.19:41
jgriffithdiablo_rojo: sure, whichever works for you.  I'll leave it up to you  :)19:42
diablo_rojojgriffith: How should we put those halves together? Or shall I continue on the route I am currently on?19:42
jgriffithdiablo_rojo: also... you might consider something like oswalk all the py files, and read in what you need and process it as you go19:42
jgriffithdiablo_rojo: we don't have to put them together right now19:43
diablo_rojoThats kinda what happens now.19:43
jgriffithdiablo_rojo: "kinda" being the key word there :)19:43
diablo_rojoThe greps for all the CONF.register_opts(xxxxx)19:43
diablo_rojoand pipes that into my script19:43
jgriffithdiablo_rojo: right, what I'm suggesting is that instead you have just call "generate-opts" that just build all of that mess, don't mess with piping back and forth.  You'll get way better performance19:44
jgriffithdiablo_rojo: and you'll get more reuse out of your tools that you're building19:44
jgriffithdiablo_rojo: but anyway... completely up to you19:45
*** jecarey has joined #openstack-cinder19:45
jgriffithI'll buzz off and leave you to what you're doing19:45
diablo_rojoI had thought about that originally, it just made more sense to me to process all of that in a separate script19:45
jgriffithdiablo_rojo: cool!  Then go with it :)19:45
diablo_rojojgriffith: Okay, works for me :)19:45
jgriffithdiablo_rojo: let me know when you get an update, and thanks!19:46
diablo_rojoif you wanna make comments on what I have out there go ahead, I do appreciate the constuctive criticism.19:46
jgriffithdiablo_rojo: it is one of the trickier things to be trying to work on here I think19:46
diablo_rojoWill do!19:46
diablo_rojoTell me about it ;)19:46
diablo_rojoI didn't know the beast I was taking on when I started this.19:46
jgriffithdiablo_rojo: that's how we usually trick people into doing things around here ;)19:47
diablo_rojoClever, very very clever.19:47
*** sghanekar has quit IRC19:50
*** eharney has quit IRC19:55
*** eharney has joined #openstack-cinder20:00
chutwigDuncanT: here is my first draft of the cinder AZ stuff:
*** chutwig is now known as erhud120:04
*** erhud1 is now known as erhudy120:04
erhudy1let me know if it's what you were looking for20:05
*** sghanekar has joined #openstack-cinder20:09
*** pots has joined #openstack-cinder20:10
*** dflorea has joined #openstack-cinder20:10
*** pots has left #openstack-cinder20:11
*** e0ne has quit IRC20:12
*** vgridnev has joined #openstack-cinder20:14
*** pots has joined #openstack-cinder20:15
*** Yogi1 has quit IRC20:16
*** julim has quit IRC20:18
*** lpetrut has joined #openstack-cinder20:21
*** akerr has joined #openstack-cinder20:21
*** julim has joined #openstack-cinder20:21
*** kvidvans has joined #openstack-cinder20:22
*** ronis has quit IRC20:24
*** e0ne has joined #openstack-cinder20:25
*** jungleboyj has quit IRC20:25
*** akerr has quit IRC20:26
*** e0ne has quit IRC20:26
*** eharney has quit IRC20:27
*** tpsilva has quit IRC20:27
*** IanGovett has quit IRC20:29
*** julim has quit IRC20:33
*** thangp has quit IRC20:33
*** akerr has joined #openstack-cinder20:33
*** lpetrut has quit IRC20:38
*** eharney has joined #openstack-cinder20:42
*** julim has joined #openstack-cinder20:42
*** julim has quit IRC20:54
*** akerr is now known as akerr_away20:59
*** julim has joined #openstack-cinder21:02
*** jwcroppe_ has joined #openstack-cinder21:03
*** vgridnev has quit IRC21:05
*** dustins has quit IRC21:05
*** jwcroppe has quit IRC21:06
*** raildo is now known as raildo-afk21:08
*** mudassirlatif has quit IRC21:08
*** mudassirlatif has joined #openstack-cinder21:10
*** greghaynes has quit IRC21:10
*** akerr_away has quit IRC21:16
*** crose has quit IRC21:18
*** gouthamr has quit IRC21:18
*** jwcroppe_ has quit IRC21:21
*** jwcroppe has joined #openstack-cinder21:22
*** Yogi1 has joined #openstack-cinder21:25
*** eharney has quit IRC21:26
*** diablo_rojo has left #openstack-cinder21:27
*** porrua has quit IRC21:27
*** jwcroppe_ has joined #openstack-cinder21:29
*** jwcroppe has quit IRC21:31
*** IlyaG has joined #openstack-cinder21:36
*** annegentle has quit IRC21:38
*** merooney has quit IRC21:40
*** niknakpaddywak has quit IRC21:42
*** dflorea has quit IRC21:42
*** jwcroppe has joined #openstack-cinder21:43
*** rmetcalf has quit IRC21:45
*** jwcroppe_ has quit IRC21:45
*** jgregor has quit IRC21:45
*** jwcroppe_ has joined #openstack-cinder21:48
*** mriedem has quit IRC21:49
*** jwcroppe has quit IRC21:50
*** cbader has quit IRC21:50
*** haomaiwa_ has joined #openstack-cinder21:54
*** haomaiw__ has quit IRC21:56
*** dflorea has joined #openstack-cinder21:56
*** annegentle has joined #openstack-cinder21:57
*** annegentle has quit IRC22:03
*** mudassirlatif has quit IRC22:04
*** sghanekar_ has joined #openstack-cinder22:04
openstackgerritJohn Griffith proposed openstack/cinder: Add bug number to skip in VMware test
*** sghanekar has quit IRC22:07
*** JoseMello has quit IRC22:09
openstackgerritJohn Griffith proposed openstack/cinder: Add mechanism to update snapshot provider_id
*** sghanekar has joined #openstack-cinder22:11
*** sghanekar_ has quit IRC22:13
*** martyturner has quit IRC22:15
*** edtubill has quit IRC22:17
*** jwcroppe has joined #openstack-cinder22:20
*** jwcroppe_ has quit IRC22:22
*** Yogi1 has quit IRC22:23
*** jwcroppe_ has joined #openstack-cinder22:24
*** jwcroppe has quit IRC22:26
*** jwcroppe has joined #openstack-cinder22:27
*** jwcroppe_ has quit IRC22:29
*** IlyaG has quit IRC22:30
*** greghaynes has joined #openstack-cinder22:33
DuncanTerhudy1: So you if I'm reading that correctly, you want a single cinder AZ with one or more nova AZs?22:34
erhudy1yes, that's how we have things set up now22:35
*** sghanekar has quit IRC22:35
erhudy1we don't want cinder to care because dealing with data availability is ceph's responsibility22:35
DuncanTerhudy1: Thanks for the writeup. That config makes perfect sense and is definitely one I think we should support.22:36
erhudy1no problem22:37
jgriffithDuncanT: FWIW, that used to be our default :)22:44
jgriffithDuncanT: but some "do gooder" saw it as a bug and fixed it :(22:44
DuncanTjgriffith: We can un-fix it again. Fortunately easier to do with code than dogs22:45
*** sghanekar has joined #openstack-cinder22:45
jgriffitherhudy1: DuncanT one place to consider this falling apart is people that want to actually do locality scheduling based on AZ's (they're out there).  BTW, not disagreeing with the change22:45
jgriffithjust some background22:45
jgriffithand the fact that it's configurable makes "most" people happy22:46
erhudy1i recognize that down the road as our clusters continue to expand we'll probably reach a point where a single crush map spanning the entire cluster no longer works22:47
*** adrianofr has quit IRC22:47
erhudy1but i'd have to think about the implications of that and how we would try to deal with ceph in that scenario22:51
erhudy1and probably a lot of other assumptions we've made would break down before ceph did22:53
jgriffitherhudy1: well... the other thing to consider is that Ceph isn't the only supported backend for Cinder ;)22:53
erhudy1i'm just ruminating22:53
*** edmondsw has quit IRC22:53
erhudy1like a cow22:53
jgriffitherhudy1: I do the same.. drives people nuts :)22:53
erhudy1or maybe a COW22:53
openstackgerritJohn Griffith proposed openstack/cinder: Add placholder for migration backports in Liberty
*** mriedem has joined #openstack-cinder22:58
*** jecarey has quit IRC22:59
*** mdenny has quit IRC23:05
*** MIDENN_ has quit IRC23:05
*** dflorea has quit IRC23:12
*** markvoelker has joined #openstack-cinder23:13
*** markvoelker has quit IRC23:17
*** akerr has joined #openstack-cinder23:17
*** tbarron is now known as tbarron_afk23:19
*** dannywilson has quit IRC23:23
*** markvoelker has joined #openstack-cinder23:25
*** akerr is now known as akerr_away23:27
*** annegentle has joined #openstack-cinder23:29
*** dflorea_ has joined #openstack-cinder23:35
*** akerr_away has quit IRC23:39
*** dflorea_ has quit IRC23:41
*** dims has quit IRC23:44
*** hodos has quit IRC23:44
*** dflorea has joined #openstack-cinder23:45
*** dflorea has quit IRC23:48
*** changbl has joined #openstack-cinder23:48
*** pots1 has quit IRC23:52
*** dims has joined #openstack-cinder23:53
*** dflorea has joined #openstack-cinder23:57
*** aix has joined #openstack-cinder23:58

Generated by 2.14.0 by Marius Gedminas - find it at!