Wednesday, 2015-06-17

*** ijw has joined #openstack-nova00:00
*** igordcard_ has quit IRC00:01
*** ssurana1 has joined #openstack-nova00:02
*** haomai___ has joined #openstack-nova00:02
*** tjones2 has quit IRC00:02
*** tjones1 has joined #openstack-nova00:02
*** haomaiwa_ has quit IRC00:03
*** ozialien has quit IRC00:04
*** ssurana has quit IRC00:04
*** Nikolay_St has quit IRC00:08
*** armax has joined #openstack-nova00:09
*** takedakn has joined #openstack-nova00:11
*** ijw has quit IRC00:13
*** ijw has joined #openstack-nova00:14
*** takedakn has quit IRC00:16
*** Nikolay_St has joined #openstack-nova00:21
*** mtanino has quit IRC00:22
*** otter768 has joined #openstack-nova00:23
*** sdake has quit IRC00:24
*** isd has quit IRC00:26
*** yingjun has joined #openstack-nova00:28
*** ferest has quit IRC00:30
*** baoli has joined #openstack-nova00:32
*** haomai___ has quit IRC00:34
openstackgerritDavanum Srinivas (dims) proposed openstack/python-novaclient: cleanup openstack-common.conf and sync updated files  https://review.openstack.org/18911900:38
*** tpeoples has joined #openstack-nova00:38
*** ctrath has quit IRC00:38
*** jasondotstar has quit IRC00:42
*** oro has quit IRC00:42
*** tjones1 has quit IRC00:44
tonybDo we actually support python-novaclient windows? (well anything otehr than MacOS and Linux really)00:45
*** tjones1 has joined #openstack-nova00:47
*** agireud has quit IRC00:47
*** ssurana1 has quit IRC00:47
*** sdake has joined #openstack-nova00:48
openstackgerritwangxiyuan proposed openstack/nova: Fix wrong check when use image in local  https://review.openstack.org/18697600:50
*** mdrabe has quit IRC00:50
*** electrocucaracha has quit IRC00:52
openstackgerritLi Yingjun proposed openstack/nova: Raise an exception when adding an already associated security group  https://review.openstack.org/16068900:54
*** otter768 has quit IRC00:54
*** tjones1 has quit IRC00:55
*** sdake_ has joined #openstack-nova00:58
*** sdake has quit IRC01:01
*** tobe has joined #openstack-nova01:05
*** zz_jgrimm is now known as jgrimm01:06
*** jgrimm is now known as zz_jgrimm01:07
*** diogogmt has quit IRC01:07
*** penick has quit IRC01:08
*** ijw has quit IRC01:08
*** apuimedo has quit IRC01:09
*** ijw__ has joined #openstack-nova01:09
*** unicell has quit IRC01:09
*** ijw__ has quit IRC01:10
*** _liusheng has joined #openstack-nova01:10
*** ijw__ has joined #openstack-nova01:10
*** dsanders has quit IRC01:11
*** ijw__ has quit IRC01:11
*** _liusheng has quit IRC01:11
*** liusheng_ has joined #openstack-nova01:12
*** zxen has quit IRC01:12
*** patrickeast has joined #openstack-nova01:12
*** SourabhP has quit IRC01:12
*** patrickeast has left #openstack-nova01:19
openstackgerritwangxiyuan proposed openstack/nova: Fix bdm image_meta when boot from volume  https://review.openstack.org/18624701:20
*** bradjones has quit IRC01:21
*** patrickeast has joined #openstack-nova01:21
*** baoli has quit IRC01:24
*** claudiub has joined #openstack-nova01:26
*** bradjones has joined #openstack-nova01:29
*** bradjones has quit IRC01:29
*** bradjones has joined #openstack-nova01:29
*** Guest68549 is now known as dan01:30
*** patrickeast has quit IRC01:32
*** mtecer has quit IRC01:34
*** kiran-r has joined #openstack-nova01:37
tonybdims: Thanks for the support on: https://review.openstack.org/19047201:43
*** ociuhandu has quit IRC01:46
*** diogogmt has joined #openstack-nova01:49
*** pixelbeat has joined #openstack-nova01:49
*** ssurana has joined #openstack-nova01:49
*** gyee is now known as operator9901:49
*** baoli has joined #openstack-nova01:49
*** pixelb has quit IRC01:49
*** nic has quit IRC01:55
*** ssurana has quit IRC01:57
*** iamjarvo has joined #openstack-nova01:57
*** baoli has quit IRC01:59
*** mdrabe has joined #openstack-nova02:00
dimstonyb: my pleasure02:02
*** vilobhmm has joined #openstack-nova02:02
*** dims has quit IRC02:04
*** yingjun has quit IRC02:05
*** vilobhmm has quit IRC02:07
*** ypraveen has quit IRC02:07
*** dims has joined #openstack-nova02:08
*** jasondotstar has joined #openstack-nova02:09
*** jasondotstar has quit IRC02:09
*** jwcroppe has quit IRC02:10
*** jasondotstar has joined #openstack-nova02:10
*** kiran-r has quit IRC02:11
*** Marga_ has quit IRC02:14
*** Marga_ has joined #openstack-nova02:16
*** Marga_ has quit IRC02:16
*** Marga_ has joined #openstack-nova02:17
*** dims has quit IRC02:19
openstackgerritwangxiyuan proposed openstack/nova: Fix bdm image_meta when boot from volume  https://review.openstack.org/18624702:21
*** mdrabe has quit IRC02:23
*** jasondotstar has quit IRC02:24
*** claudiub has quit IRC02:29
*** changbl has joined #openstack-nova02:32
*** Marga_ has quit IRC02:33
*** Marga_ has joined #openstack-nova02:34
*** Marga_ has quit IRC02:37
*** haomaiwang has joined #openstack-nova02:38
*** liusheng_ has quit IRC02:48
*** liusheng has joined #openstack-nova02:48
*** dsanders has joined #openstack-nova02:50
openstackgerritheijlong proposed openstack/nova-specs: Add uuid field to security-groups for server show  https://review.openstack.org/18420102:50
*** baoli has joined #openstack-nova02:51
*** tjones1 has joined #openstack-nova02:54
*** otter768 has joined #openstack-nova02:55
*** pixelbeat has quit IRC02:55
*** dpaterson has joined #openstack-nova02:55
*** otter768 has quit IRC02:59
*** sdake_ is now known as sdake03:00
*** varunarya has joined #openstack-nova03:03
*** kiran-r has joined #openstack-nova03:05
*** unicell has joined #openstack-nova03:06
*** Marga_ has joined #openstack-nova03:09
*** marzif has joined #openstack-nova03:09
*** liusheng has quit IRC03:09
*** liusheng has joined #openstack-nova03:10
*** Marga_ has quit IRC03:14
*** prashup has quit IRC03:18
*** varunarya_ has joined #openstack-nova03:19
*** varunarya has quit IRC03:21
*** raildo has quit IRC03:21
*** samueldmq has quit IRC03:21
*** htruta has quit IRC03:22
*** afaranha has quit IRC03:22
*** ericksonsantos has quit IRC03:22
*** tellesnobrega has quit IRC03:22
*** baoli has quit IRC03:24
*** iamjarvo has quit IRC03:26
*** tpeoples has quit IRC03:26
*** iamjarvo has joined #openstack-nova03:29
*** baoli has joined #openstack-nova03:29
*** markvoelker has quit IRC03:31
*** coolsvap|afk is now known as coolsvap03:34
*** ijw has joined #openstack-nova03:37
*** tjones1 has quit IRC03:39
*** baoli has quit IRC03:40
*** aswadr has joined #openstack-nova03:40
*** Marga_ has joined #openstack-nova03:43
*** tjones1 has joined #openstack-nova03:44
*** Marga__ has joined #openstack-nova03:45
*** rajesht has joined #openstack-nova03:47
*** Marga_ has quit IRC03:47
*** Marga_ has joined #openstack-nova03:48
*** Marga__ has quit IRC03:49
*** vilobhmm has joined #openstack-nova03:53
*** ildikov has quit IRC03:58
*** iamjarvo has quit IRC04:00
*** prashup has joined #openstack-nova04:01
*** varunarya_ has quit IRC04:02
*** ozialien has joined #openstack-nova04:03
*** tjones1 has quit IRC04:03
*** kiran-r has quit IRC04:06
openstackgerritDiana Clarke proposed openstack/nova: Add unit tests for the exact filters  https://review.openstack.org/19251404:07
*** puniuee has joined #openstack-nova04:08
*** rook has quit IRC04:09
*** unicell has quit IRC04:10
*** rushiagr_away is now known as rushiagr04:10
lifelessnova / IBM peoples04:12
lifelessis ORD the right airport for the midcycle ?04:12
*** marzif has quit IRC04:13
lifelessno04:13
lifelessRST??04:13
*** dpaterson has quit IRC04:15
StevenKlifeless: RST is in Minnesota, not the other Rochester04:15
StevenKROC is the Rochester in NY04:15
*** patrickeast has joined #openstack-nova04:15
puniueeHello all, I need some help in figuring out why some of my CellsComputeAPITestCase are failing04:16
*** ijw_ has quit IRC04:19
*** ijw_ has joined #openstack-nova04:19
lifelessStevenK: https://wiki.openstack.org/wiki/Sprints/NovaLibertySprint04:22
*** ijw_ has quit IRC04:22
lifelessStevenK: says MN04:22
lifelessand RST04:22
lifeless / MSP04:22
*** achanda has joined #openstack-nova04:23
*** patrickeast has quit IRC04:23
StevenKlifeless: Blah, I keep forgetting that the US likes reusing city names04:24
*** puniuee has quit IRC04:26
*** josecastroleon has quit IRC04:31
*** beagles has quit IRC04:31
*** achanda has quit IRC04:32
*** markvoelker has joined #openstack-nova04:32
*** josecastroleon has joined #openstack-nova04:32
*** markvoelker has quit IRC04:37
*** varunarya_ has joined #openstack-nova04:38
*** ijw_ has joined #openstack-nova04:40
*** puniuee has joined #openstack-nova04:43
*** ijw has quit IRC04:44
tonyblifeless: No MSP04:46
tonyblifeless: https://wiki.openstack.org/wiki/Sprints/NovaLibertySprint04:46
tonybAnd now I read the whole scrollback04:47
openstackgerritFei Long Wang proposed openstack/nova: Refactor image's code to support both versions  https://review.openstack.org/14487504:47
*** b3nt_pin has joined #openstack-nova04:47
*** b3nt_pin is now known as Guest7107504:48
*** ijw has joined #openstack-nova04:50
*** fawadkhaliq has joined #openstack-nova04:55
*** ijw has quit IRC04:55
*** otter768 has joined #openstack-nova04:56
*** tojuvone has joined #openstack-nova04:58
*** irenab has joined #openstack-nova04:58
*** otter768 has quit IRC05:00
*** ozialien has quit IRC05:01
*** tojuvone has quit IRC05:02
*** achanda has joined #openstack-nova05:02
*** whenry has quit IRC05:03
*** fifieldt has joined #openstack-nova05:10
*** IanGovett has joined #openstack-nova05:12
*** nihilifer has joined #openstack-nova05:12
*** ypraveen has joined #openstack-nova05:13
*** haomaiwang has quit IRC05:14
*** Kennan has quit IRC05:14
*** ypraveen has quit IRC05:16
*** dsanders has quit IRC05:16
*** vishwana_ has joined #openstack-nova05:16
*** browne has quit IRC05:16
*** Maike has joined #openstack-nova05:18
*** vishwanathj has quit IRC05:19
*** Kennan has joined #openstack-nova05:20
*** ypraveen has joined #openstack-nova05:22
lifelessStevenK: *cough newcastle*05:23
lifelessStevenK: *cough perth*05:24
lifelessStevenK: *cough... need I continue ?05:24
*** Nikolay_St has quit IRC05:26
*** nkrinner has joined #openstack-nova05:31
*** achanda has quit IRC05:31
*** lpetrut has joined #openstack-nova05:33
*** Guest71075 has quit IRC05:34
*** b3nt_pin_ has joined #openstack-nova05:37
openstackgerritDheeraj Gupta proposed openstack/nova: Add HostMapping object  https://review.openstack.org/19209805:38
*** tobe has quit IRC05:39
*** arnaud___ has joined #openstack-nova05:41
*** sudipto has joined #openstack-nova05:43
*** belmoreira has joined #openstack-nova05:44
*** oomichi has quit IRC05:45
*** doron_afk has joined #openstack-nova05:45
*** b3nt_pin_ is now known as b3nt_pin05:48
*** ijw has joined #openstack-nova05:48
*** slong has quit IRC05:49
*** b3nt_pin is now known as Guest9419405:49
*** ijw has quit IRC05:50
*** tojuvone has joined #openstack-nova05:51
*** deepthi has joined #openstack-nova05:53
*** IanGovett has quit IRC05:55
*** markvoelker has joined #openstack-nova05:55
*** ypraveen has quit IRC05:56
*** ypraveen has joined #openstack-nova05:58
*** haomaiwa_ has joined #openstack-nova05:59
*** markvoelker has quit IRC06:00
*** Mike_D_laptop has joined #openstack-nova06:01
openstackgerritgaryk proposed openstack/nova: VMware: use min supported VC version in fake driver  https://review.openstack.org/19252406:01
*** kiran-r has joined #openstack-nova06:02
*** Mike_D_laptop has quit IRC06:02
*** browne has joined #openstack-nova06:03
*** garyk has quit IRC06:03
*** IanGovett has joined #openstack-nova06:05
*** destiny has joined #openstack-nova06:06
*** doron_afk is now known as doron_06:07
*** IanGovett has quit IRC06:09
*** moshele has joined #openstack-nova06:13
*** browne has quit IRC06:14
*** radek_ has joined #openstack-nova06:15
*** abhishekk has joined #openstack-nova06:15
*** doron_ is now known as doron_afk06:15
openstackgerritOpenStack Proposal Bot proposed openstack/nova: Imported Translations from Transifex  https://review.openstack.org/19003406:15
*** puniuee has quit IRC06:16
*** fawadkhaliq has quit IRC06:17
*** rajesht has quit IRC06:18
*** exploreshaifali has joined #openstack-nova06:19
*** ypraveen has quit IRC06:19
*** tobe has joined #openstack-nova06:20
*** unicell has joined #openstack-nova06:21
*** rajesht has joined #openstack-nova06:22
*** dulek has joined #openstack-nova06:22
*** Marga_ has quit IRC06:23
*** afazekas has joined #openstack-nova06:23
*** dulek has quit IRC06:25
*** park_ has joined #openstack-nova06:26
abhishekkhi all, is there any meeting for cinder nova volume discussion?06:26
*** mpaolino has joined #openstack-nova06:27
*** browne has joined #openstack-nova06:32
*** puniuee has joined #openstack-nova06:33
*** ianbrown has quit IRC06:39
*** salv-orl_ has joined #openstack-nova06:41
*** boris-42 has quit IRC06:42
*** salv-orlando has quit IRC06:44
*** sahid has joined #openstack-nova06:46
*** garyk has joined #openstack-nova06:46
*** dulek has joined #openstack-nova06:49
*** browne has quit IRC06:49
*** Longgeek has joined #openstack-nova06:49
openstackgerritAnkit Agrawal proposed openstack/nova: Test patch race condition in delete (do not merge)  https://review.openstack.org/19173206:50
*** lpetrut has quit IRC06:50
*** moshele has quit IRC06:51
*** arnaud___ has quit IRC06:52
*** Longgeek_ has joined #openstack-nova06:53
*** romainh has joined #openstack-nova06:56
*** otter768 has joined #openstack-nova06:56
*** Longgeek has quit IRC06:57
*** vilobhmm has quit IRC06:58
*** moshele has joined #openstack-nova07:01
*** otter768 has quit IRC07:01
*** armax has quit IRC07:03
*** pilgrimstack has joined #openstack-nova07:04
*** pilgrimstack has joined #openstack-nova07:04
*** pilgrimstack has left #openstack-nova07:04
openstackgerritRajesh Tailor proposed openstack/nova: Validate int using utils.validate_integer method  https://review.openstack.org/19253507:15
*** sdake has quit IRC07:20
*** puniuee has quit IRC07:21
*** Longgeek_ has quit IRC07:26
*** alex_klimov has joined #openstack-nova07:26
*** matrohon has joined #openstack-nova07:28
*** haypo has joined #openstack-nova07:28
*** ijw has joined #openstack-nova07:30
*** park_ has quit IRC07:34
*** cfriesen has quit IRC07:34
*** unicell has quit IRC07:35
*** unicell has joined #openstack-nova07:35
*** puniuee has joined #openstack-nova07:38
*** doude has joined #openstack-nova07:39
*** ociuhandu has joined #openstack-nova07:42
*** jlanoux has joined #openstack-nova07:43
*** ildikov has joined #openstack-nova07:44
*** markvoelker has joined #openstack-nova07:44
*** ociuhandu has quit IRC07:45
*** doude has quit IRC07:46
*** doude has joined #openstack-nova07:47
*** Longgeek has joined #openstack-nova07:47
*** park_ has joined #openstack-nova07:47
*** Longgeek has quit IRC07:48
*** Longgeek has joined #openstack-nova07:48
*** markvoelker has quit IRC07:49
*** Nikolay_St has joined #openstack-nova07:50
*** varunarya_ has quit IRC07:52
*** jyuso1 has joined #openstack-nova07:53
*** markus_z has joined #openstack-nova07:56
*** lucasagomes has joined #openstack-nova07:59
*** ZZelle has quit IRC08:02
*** ZZelle has joined #openstack-nova08:03
*** ociuhandu has joined #openstack-nova08:04
*** apuimedo has joined #openstack-nova08:07
openstackgerritBartosz Fic proposed openstack/nova: Live migration failure in API doesn't set vm_state to Error  https://review.openstack.org/16891608:08
*** park_ has quit IRC08:08
*** jcoufal has joined #openstack-nova08:08
*** jcoufal has quit IRC08:09
*** jcoufal has joined #openstack-nova08:10
*** derekh has joined #openstack-nova08:12
*** doron_afk has quit IRC08:13
*** dsanders has joined #openstack-nova08:15
*** dsanders has quit IRC08:16
*** aswadr has quit IRC08:18
*** Nikolay_St has quit IRC08:18
*** aswadr has joined #openstack-nova08:19
*** Nikolay_St has joined #openstack-nova08:20
openstackgerritAlex Xu proposed openstack/nova-specs: Allow specifying cpu model/features in flavor extra-specs or image metadata  https://review.openstack.org/16898208:20
park_heijlonghello johnthetubaguy, would you please help to take a review about https://review.openstack.org/#/c/184201/....I updated as suggestion...thank you ~08:21
*** park_ has joined #openstack-nova08:21
*** asrangne has joined #openstack-nova08:21
johnthetubaguypark_heijlong: just looking at your questions on there08:21
johnthetubaguypark_heijlong: I am saying we generate a uuid for all the nova security groups, and use the neutron ones like we do today08:22
johnthetubaguypark_heijlong: its possibly I am totally miss understanding the relationship there08:23
openstackgerritDeepak C Shetty (deepakcs) proposed openstack/nova: Nova part of volume snapshot improvements  https://review.openstack.org/17281308:23
park_heijlongjohnthetubaguy, yep, my latest version made the correction..08:23
*** pixelbeat has joined #openstack-nova08:24
park_heijlongjohnthetubaguy: the latest made the correction as your suggestion.08:24
*** yassine has joined #openstack-nova08:24
*** aswadr has quit IRC08:24
johnthetubaguypark_heijlong: OK, thank you08:25
bauzasmorning Nova08:25
park_heijlongjohnthetubaguy, really thanks for review :) feel free to leave comments08:25
johnthetubaguypark_heijlong: no problems, I am really happy to see someone fixing up this API :)08:25
*** doron_afk has joined #openstack-nova08:27
*** varunarya has joined #openstack-nova08:30
*** puniuee has quit IRC08:31
johnthetubaguypark_heijlong: so I think there is one extra bit, we need a note about all the add security group API, etc, that will need to now accept uuids instead of name or id?08:32
*** rotbeard has joined #openstack-nova08:32
* johnthetubaguy is having to drop online while he drives over to talk to some Xen folks08:33
*** belmoreira has quit IRC08:33
johnthetubaguyBobBall: see you in a bit!08:33
BobBallIndeed!08:33
*** belmoreira has joined #openstack-nova08:33
park_heijlongjohnthetubaguy: make sense, will update08:34
*** varunarya has quit IRC08:37
*** dguitarbite has quit IRC08:38
*** dguitarbite has joined #openstack-nova08:38
pczesnomoshele: Hi08:41
*** mdbooth has quit IRC08:42
*** mdbooth has joined #openstack-nova08:45
*** ociuhandu has quit IRC08:52
*** coolsvap is now known as coolsvap|afk08:52
garykanyone had the latest devstack and nova blow up on them - please see http://paste.openstack.org/show/297829/08:56
*** ociuhandu has joined #openstack-nova08:57
kashyapgaryk: I just did a `make unstack` away on an F22 machine. Will let you know if my next run goes well.08:57
*** otter768 has joined #openstack-nova08:57
garykkashyap: thanks08:58
*** tobe has quit IRC08:58
garyki am not sure if it is my env or the code … gut feeling is that it is the former08:58
*** tobe has joined #openstack-nova08:59
kashyapgaryk: Your gut feeling is right, it's always DevStack :-)09:00
*** pask81 has joined #openstack-nova09:01
*** tobe_ has joined #openstack-nova09:01
*** tobe_ has quit IRC09:01
*** danpb has joined #openstack-nova09:02
*** otter768 has quit IRC09:02
*** coolsvap|afk is now known as coolsvap09:03
*** tobe has quit IRC09:03
*** gszasz has joined #openstack-nova09:04
*** dguitarbite has quit IRC09:07
*** dguitarbite has joined #openstack-nova09:08
*** park_ has quit IRC09:09
*** varunarya has joined #openstack-nova09:09
*** dguitarbite has quit IRC09:10
openstackgerritStefan Amann proposed openstack/nova: Support host type specific block volume attachment  https://review.openstack.org/19255609:14
*** igordcard_ has joined #openstack-nova09:17
snikitinGuys, I have a question. Do we have any freeze period for nova specs?09:17
*** ndipanov has joined #openstack-nova09:21
*** tobe has joined #openstack-nova09:23
garykkashyap: seems to have been something in my env… thanks09:28
*** varunarya has quit IRC09:30
*** markvoelker has joined #openstack-nova09:33
*** markvoelker has quit IRC09:37
*** jyuso1 has quit IRC09:40
*** haomaiwa_ has quit IRC09:41
*** pask811 has joined #openstack-nova09:43
*** pask81 has quit IRC09:46
*** pask811 has quit IRC09:48
*** mpaolino has quit IRC09:49
*** arnaud___ has joined #openstack-nova09:53
*** Marga_ has joined #openstack-nova09:54
*** mpaolino has joined #openstack-nova09:56
*** tdurakov has quit IRC09:56
*** arnaud___ has quit IRC09:57
kaisersHi Nova! Can somebody please give me a hint for debugging a DHCP issue in a devstack/nova-network setup: Guests do send DHCP requests, dnsmasq receives the requests but the answers seem to be lost. Happens with cirros & Ubuntu 14.04 guest OS. Hints on where to look for reasons?09:58
*** tdurakov has joined #openstack-nova09:58
*** romainh has quit IRC09:59
kashyapgaryk: Yes, likewise here too.09:59
kaisersSettings are using FlatDHCPManager09:59
kaisersOh, whoopsie, this channel was only for dev, not help req, right? If yes: my apologies!10:01
openstackgerritPrashantUpadhyay proposed openstack/nova: adding check to see if volume size specifed by user is greater than the size of OS-image specified by user  https://review.openstack.org/19257110:05
*** dguitarbite has joined #openstack-nova10:06
*** oro has joined #openstack-nova10:07
*** ianbrown has joined #openstack-nova10:07
openstackgerritPrashantUpadhyay proposed openstack/nova: Adding check to see if volume size specifed by user is greater than the size of OS-image specified by user  https://review.openstack.org/19257110:07
openstackgerritPrashantUpadhyay proposed openstack/nova: Adding check to see if volume size specifed by user is greater than the size of OS-image specified by user  https://review.openstack.org/19257110:09
*** jmccrory has quit IRC10:10
*** lpetrut has joined #openstack-nova10:10
*** lpetrut has quit IRC10:10
*** varunarya has joined #openstack-nova10:10
openstackgerritPrashantUpadhyay proposed openstack/nova: Adding check to see if volume size specifed by user is greater than the size of OS-image  https://review.openstack.org/19257110:11
*** davidlenwell has quit IRC10:12
*** davidlenwell has joined #openstack-nova10:12
*** jmccrory has joined #openstack-nova10:12
*** alex_klimov has quit IRC10:13
openstackgerritPrashantUpadhyay proposed openstack/nova: check for volume size  https://review.openstack.org/19257110:13
*** oro has quit IRC10:14
*** varunarya has quit IRC10:16
*** dims has joined #openstack-nova10:17
*** tobe has quit IRC10:17
*** dims has quit IRC10:17
*** dims has joined #openstack-nova10:18
*** ianbrown has quit IRC10:19
openstackgerritPrashantUpadhyay proposed openstack/nova: check for volume size  https://review.openstack.org/19257110:20
*** lpetrut has joined #openstack-nova10:21
*** ianbrown has joined #openstack-nova10:26
*** haomaiwang has joined #openstack-nova10:28
*** claudiub has joined #openstack-nova10:31
openstackgerritChris Gacsal proposed openstack/nova: Allow multiple volume attachments  https://review.openstack.org/15303310:31
*** oomichi has joined #openstack-nova10:32
*** kiran-r has quit IRC10:39
*** vladikr has quit IRC10:41
*** jasondotstar has joined #openstack-nova10:41
*** varunarya has joined #openstack-nova10:45
*** mwagner_afk has quit IRC10:48
*** alex_klimov has joined #openstack-nova10:49
*** aix has joined #openstack-nova10:49
openstackgerritAndrzej Kuriata proposed openstack/nova-specs: Introduce Normalized Compute Units  https://review.openstack.org/19260910:52
*** chaochin has quit IRC10:53
garykdanpb: i have tried to address your comment with the VC API link.10:54
*** otter768 has joined #openstack-nova10:58
*** Marga_ has quit IRC11:02
*** otter768 has quit IRC11:03
*** sahid has quit IRC11:05
*** sahid has joined #openstack-nova11:06
*** jasondotstar has quit IRC11:06
danpbgaryk: ok, i'm afraid I think we might want to go the blueprint+spec route for this first11:08
danpbgaryk: i didn't -2 it though - wait to see whether other nova cores agree or not11:09
*** Marga_ has joined #openstack-nova11:09
garykdanpb: ok, thanks. i already have a BP for it - https://blueprints.launchpad.net/nova/+spec/vmware-limits11:10
garykthe CPU limits were approved in J and are used in production today.11:11
garykthey are supported by flavors11:11
garykdo you still think that I need a spec?11:11
garyki think tat we are exposing VC functionality like any other driver is exposing its ability to do resource allocations11:12
danpbusage against flavours is ok, as the admin is in control of that and they understand their hardware11:12
danpbi think usage against images though introduces the new problems i menton there11:12
danpbbecause you're pushing decisions to the tenant user who is not supposed to know about hardware11:13
danpbso i'm not sure how they'd meaningfully decide on cpu limits expressed in this manner11:13
garykok, understood.11:13
*** varunarya has quit IRC11:13
danpbthere's also the permissions issue that the cloud admin may not want to allow control over this11:13
danpbso we need to at least decide policy for usage of image vs flavour11:14
danpbto ensure the cloud admin can prevent usage on the image if desired11:14
openstackgerritPeter Penchev proposed openstack/nova-specs: Resubmit the StorPool libvirt attachment driver.  https://review.openstack.org/19261311:14
danpbfor numa we simply said flavour always takes priority if set11:14
danpbso cloud admin can block usage on flavor by simply setting numa policy on the image11:14
garykthe permissions are a good point. i will go to look at glance and see how these are enforced.11:14
danpbbut i'm not sure it is so easy to achieve that in this case11:15
garyki am not sure why a tenant would set meta properties as this is information that is is related to the infra and the code11:15
danpbas the cloud admin may wish to just have the default behaviour without setting anything on the flavor11:15
alex_xudanpb: hi, I submitted initial version for introduce api for cpu model https://review.openstack.org/168982 . would you like take a look at whether the direction is right, then I'm going to fill up more detail. thanks in advance!11:16
*** oro has joined #openstack-nova11:16
*** tojuvone has quit IRC11:16
garykdanpb: so why would we enable the user to set numa cpu's nodes etc. that should only be available to the admin11:17
danpbthere are use cases where performance critical applications have a need for very specific numa topologies11:18
garykthat is the same case here11:18
*** romainh has joined #openstack-nova11:18
danpbso to avoid an explosion in the number of flavours, its possible to set it on the image11:19
garyksomeone may need X resource for a specific image11:19
danpbbut any settings on the image are only honoured, if the flavour does not have a specific topology set11:19
danpbthat wa the cloud admin can prevent tenant user from overriding stuff if they desire11:19
garykbut the user can exploit that if she.he knew how to go about that11:19
garykdanpb: the glance metadata can be controlled by the policy.json file11:21
*** markvoelker has joined #openstack-nova11:22
*** coolsvap is now known as coolsvap|away11:22
*** ianbrown has quit IRC11:22
*** deepthi has quit IRC11:23
*** rushiagr is now known as rushiagr_away11:24
openstackgerritPeter Penchev proposed openstack/nova: Add the Nova libvirt StorPool attachment driver.  https://review.openstack.org/14073311:26
*** markvoelker has quit IRC11:27
openstackgerritShaoHe Feng proposed openstack/nova-specs: Attach/detach SR-IOV interface  https://review.openstack.org/13991011:31
*** shaohe_robot has joined #openstack-nova11:32
*** tellesnobrega has joined #openstack-nova11:33
openstackgerritPetrut Lucian proposed openstack/nova: Hyper-V: Fix serial port issue on Windows Threshold  https://review.openstack.org/19261711:34
*** samueldmq has joined #openstack-nova11:34
*** ericksonsantos has joined #openstack-nova11:36
openstackgerritAdelina Tuvenie proposed openstack/nova: Adds Hyper-V imagechace cleanup  https://review.openstack.org/19261811:36
*** mpaolino has quit IRC11:38
*** kiran-r has joined #openstack-nova11:38
openstackgerritShaoHe Feng proposed openstack/nova-specs: enable pci api  https://review.openstack.org/19262211:40
*** tellesnobrega has quit IRC11:40
*** ericksonsantos has quit IRC11:41
*** sdake has joined #openstack-nova11:41
*** samueldmq has quit IRC11:41
*** oomichi has quit IRC11:42
*** tellesnobrega has joined #openstack-nova11:43
*** prashup has quit IRC11:44
openstackgerritPetrut Lucian proposed openstack/nova: Hyper-V: Fix serial port issue on Windows Threshold  https://review.openstack.org/19261711:46
*** thangp has joined #openstack-nova11:46
*** lucasagomes is now known as lucas-hungry11:47
*** prashup has joined #openstack-nova11:48
openstackgerritDaniel Berrange proposed openstack/nova: ironic: convert driver to use nova.objects.ImageMeta  https://review.openstack.org/18936211:48
openstackgerritDaniel Berrange proposed openstack/nova: objects: convert config drive to use ImageMeta object  https://review.openstack.org/18932311:48
openstackgerritDaniel Berrange proposed openstack/nova: libvirt: convert driver to use nova.objects.ImageMeta  https://review.openstack.org/18811911:48
openstackgerritDaniel Berrange proposed openstack/nova: libvirt: make instance compulsory in blockinfo APIs  https://review.openstack.org/18811811:48
openstackgerritDaniel Berrange proposed openstack/nova: hyperv: convert driver to use nova.objects.ImageMeta  https://review.openstack.org/18930211:48
openstackgerritDaniel Berrange proposed openstack/nova: objects: add virtual 'image_meta' property to Instance object  https://review.openstack.org/19014511:48
openstackgerritDaniel Berrange proposed openstack/nova: virt: convert hardware module to use nova.objects.ImageMeta  https://review.openstack.org/18878911:48
openstackgerritDaniel Berrange proposed openstack/nova: test: add MatchType helper class as equivalent of mox.IsA  https://review.openstack.org/19014411:48
openstackgerritDaniel Berrange proposed openstack/nova: libvirt: convert to use instance.image_meta property  https://review.openstack.org/19014611:48
openstackgerritDaniel Berrange proposed openstack/nova: compute: convert manager to use nova.objects.ImageMeta  https://review.openstack.org/18968511:48
openstackgerritDaniel Berrange proposed openstack/nova: xenapi: convert driver to use nova.objects.ImageMeta  https://review.openstack.org/18879011:48
openstackgerritDaniel Berrange proposed openstack/nova: utils: ignore block device mapping in system metadata  https://review.openstack.org/19262711:48
openstackgerritDiana Clarke proposed openstack/nova: Add unit tests for the exact filters  https://review.openstack.org/19251411:49
*** dprince has joined #openstack-nova11:53
sudiptobauzas, Hi, I am doing a BP implementation to remove the scheduled_at from the instances table. Can you give any pointers on what might be required from the nova objects per say? I don't see any code in Nova currently working on that field - does it mean - that we still have to care about the objects?11:54
garykdanpb: i understand the issues with the image meta and need to figure out how we can address that. would it be possible to at least get the BP for the flavor stuff approved.11:55
sudiptosudipto, rather - does that mean - we can do away with the object changes? Also - since this field was meant/updated from the scheduler, do we need to bother about the conductor?11:55
danpbgaryk: oh is it not approved - i thught you'd already done the flavor stuff11:56
garykdanpb: the flavors for cpu was added in juno. now we are adding vif, disk and memeory11:57
*** dulek_ has joined #openstack-nova11:57
sudiptodansmith, if you could also give your thoughts on my question above, that'd really help...11:57
*** samueldmq has joined #openstack-nova12:01
*** dulek has quit IRC12:01
danpbgaryk: if its just a blueprint no spec for that, then i believe johnthetubaguy has to toggle the approved flag12:01
garykdanpb: ok, thanks. i'll reach out to him12:02
garykjohnthetubaguy: can you please take a look at https://blueprints.launchpad.net/nova/+spec/vmware-limits12:03
garykit would be nice if we could get this approved to help get the patches for the featre in.12:03
*** sdake has quit IRC12:04
*** Marga_ has quit IRC12:04
*** markvoelker has joined #openstack-nova12:05
*** htruta has joined #openstack-nova12:06
bauzassudipto: I don't understand your question, you mean how to remove scheduled_at from the object, right ?12:10
sudiptobauzas, yeah sort of.12:10
*** puniuee has joined #openstack-nova12:10
bauzassudipto: ok, then it's not possible12:10
sudiptobauzas, yeah - i mean it probably will exist there but we will ensure it's not written to the db?12:11
bauzassudipto: I mean, since the object needs to provide a backwards compat, you have to keep scheduled_at for the Instance object12:11
bauzassudipto: so, the field is nullable atm, right ?12:11
*** rushiagr_away is now known as rushiagr12:11
sudiptobauzas, yeah :)12:11
*** jlvillal has quit IRC12:11
sudiptobauzas, actually i see it initialized to NULL itself.12:11
bauzassudipto: okay, then you don't need to remove it, just leave a comment saying that it would be removed for a new major version for the Instance object12:12
bauzassudipto: but you can remove it from the DB12:12
bauzassudipto: and also remove it from the API, if it's given12:12
bauzassudipto: meaning that you won't have a scheduled_at field for the API and DB, but leave it for the object12:13
bauzassudipto: until we are sure that all the deployments are not using it, then we can remove it from the object12:13
sudiptobauzas, ok - do i need to handle a case - like if the instance object has scheduled_at then we make it NULL etc?12:13
*** ferest has joined #openstack-nova12:14
sudiptobauzas, i was thinking that's not needed, since we are initializing it to NULL atm...12:14
*** moshele has quit IRC12:14
bauzassudipto: the field is possibly updated by any object method, right ?12:15
bauzassudipto: so you perhaps need to work on leaving it unset12:15
bauzasI mean for the object12:16
*** Marga_ has joined #openstack-nova12:16
*** Guest94194 is now known as b3nt_pin12:16
*** b3nt_pin is now known as beagles12:16
bauzasremove it from the DB, make sure the field is left unset for the Instance object, and then remove it from the API12:16
bauzassudipto: do you have a BP for that work ?12:16
bauzassudipto: even a spec if you need a REST API change N12:16
bauzas?12:16
sudiptobauzas, don't see it being updated by any object method.12:16
sudiptobauzas, yeah - approved one here: https://blueprints.launchpad.net/nova/+spec/cleanup-scheduled-at12:17
sudiptobauzas, since none of the methods are updating it, i am guessing the unset requirement is automatically met?12:18
bauzassudipto: then yet12:20
bauzasyeah12:20
bauzassudipto: possibly then that you just have to remove it from the API12:20
*** VW has joined #openstack-nova12:20
bauzassudipto: ok, your spec is not mentioning a change in the REST API12:22
*** aysyanne has joined #openstack-nova12:22
bauzassudipto: is the field exposed by the API?12:22
*** lpetrut has quit IRC12:22
sudiptobauzas, great! Yeah - there isn't a REST API change needed IMHO...12:22
*** lpetrut has joined #openstack-nova12:22
sudiptobauzas, no afaik - on code grep.12:22
bauzassudipto: re: your question about objects12:22
bauzassudipto: so, what you need is to make sure that you have a compatibility section for the object saying if scheduled_at is provided by the DB or not12:23
*** oro has quit IRC12:23
*** doron_afk is now known as doron_12:23
bauzassudipto: since the object can be calling remotely the DB, you can't make sure that both DB and object are in sync12:23
*** jlvillal has joined #openstack-nova12:23
bauzassudipto: but that's basically it12:23
*** ferest has quit IRC12:24
sudiptobauzas, yeah so i guess what you are saying is - if the object contains scheduled_at then do xyz (that is probably make it NULL)?12:24
bauzassudipto: ie. given the version you'll bump, you'll provide the scheduled_at field from the previous versions, or not if that's the new versions12:24
*** mwagner_afk has joined #openstack-nova12:24
bauzassudipto: I'm just saying that if the field is set, you have to assume different behaviours depending on if the DB is including that field or not12:25
*** Marga_ has quit IRC12:25
*** raildo has joined #openstack-nova12:26
*** ociuhandu has quit IRC12:26
sudiptobauzas, yeah and by setting means - i understood - any object method working over it? which is not the case. Infact, there isn't anything that's setting or working on that field in the code base i greped on.12:26
bauzassudipto: the question I have is, any method is using scheduled_at or not ?12:26
*** tojuvone has joined #openstack-nova12:27
bauzassudipto: do you know the last time when that field has been set ?12:27
*** exploreshaifali has quit IRC12:27
bauzassudipto: oh you said that https://review.openstack.org/#/c/143725/12:27
sudiptobauzas, Afaik, no - this is not being set by any method. The reference, where it was removed is here: https://review.openstack.org/#/c/143725/12:28
sudiptobauzas, yeah12:28
bauzassudipto: ok, so previously that commit, it was set, righr ?12:29
sudiptobauzas, yeah previous to that commit it was being set.12:29
sudiptobauzas, which should be around Juno.12:31
*** kylek3h has joined #openstack-nova12:31
*** ivasev has joined #openstack-nova12:33
*** dane-fichter has joined #openstack-nova12:33
bauzassudipto: okay, so if the object field is unset, you probably have to find a way to set it, if some old nodes waiting for it ?12:34
bauzas(open question)12:34
*** dulek_ is now known as dulek12:34
sudiptobauzas, Isn't that something which should be dealt with even without this BP?12:35
*** oro has joined #openstack-nova12:36
*** dims has quit IRC12:36
*** dims has joined #openstack-nova12:36
*** kodoku has joined #openstack-nova12:39
*** dulek has quit IRC12:39
bauzassudipto: honestly, I dunno, I didn't code-review that stuff12:39
*** kodoku has left #openstack-nova12:40
*** shz has joined #openstack-nova12:40
*** dulek has joined #openstack-nova12:40
sudiptobauzas, ok - let me just get a patchset out - and then ask people for their review comments.12:40
bauzasfair12:40
*** shihanzhang has quit IRC12:43
*** kylek3h has quit IRC12:44
openstackgerritLi Yingjun proposed openstack/nova: Fixes TypeError when libvirt version is BAD_LIBVIRT_CPU_POLICY_VERSIONS  https://review.openstack.org/19264312:45
*** puniuee has quit IRC12:45
*** lucas-hungry is now known as lucasagomes12:45
*** jlanoux has quit IRC12:46
*** exploreshaifali has joined #openstack-nova12:46
*** moshele has joined #openstack-nova12:48
*** VW has quit IRC12:49
*** doron_ is now known as doron_afk12:49
*** prashup has quit IRC12:50
openstackgerritDavanum Srinivas (dims) proposed openstack/nova-specs: Allow list all instances for non-admin based on policy  https://review.openstack.org/19115112:53
*** VW has joined #openstack-nova12:54
dimsjohnthetubaguy: filed a spec in support of a bug that seems to need a microversion. please take a look when you get a chance ^12:54
*** baoli_ has joined #openstack-nova12:55
*** eharney has joined #openstack-nova12:55
*** moshele has quit IRC12:56
*** atuvenie has joined #openstack-nova12:56
*** IanGovett has joined #openstack-nova12:56
*** moshele has joined #openstack-nova12:56
*** ddutta has joined #openstack-nova12:57
*** edmondsw has joined #openstack-nova12:57
*** baoli_ has quit IRC12:58
*** baoli has joined #openstack-nova12:58
*** otter768 has joined #openstack-nova12:59
*** vladikr has joined #openstack-nova13:02
*** ddutta has quit IRC13:02
*** mjturek1 has joined #openstack-nova13:02
openstackgerritNikola Dipanov proposed openstack/nova: Add documentation for block device mapping  https://review.openstack.org/19265113:03
*** otter768 has quit IRC13:03
ndipanovjohnthetubaguy, sdague ^^ some docs13:04
sdaguendipanov: \o/ more docs always good13:04
openstackgerritNikola Dipanov proposed openstack/nova: Add documentation for block device mapping  https://review.openstack.org/19265113:05
ndipanovyeah I was embarassed enough about never writing those that I decided to write them :)13:05
*** mdrabe has joined #openstack-nova13:06
*** sudipto has quit IRC13:06
johnthetubaguyndipanov: thats awesome, thank you13:07
*** diogogmt has quit IRC13:07
johnthetubaguyndipanov: can I add that patch into the blueprint I am working on?13:07
johnthetubaguyndipanov: this one: https://blueprints.launchpad.net/nova/+spec/devref-refresh-liberty13:07
openstackgerritJohn Garbutt proposed openstack/nova: Add documentation for block device mapping  https://review.openstack.org/19265113:08
ndipanovsure - but I left it as a last patch on branch for a reason if that's ok13:08
* johnthetubaguy stops hiding now he is back home from visiting the Xen peoples13:08
johnthetubaguygaryk: I here there is a blueprint that needs approving?13:09
*** VW has quit IRC13:09
johnthetubaguys/here/hear/13:09
johnthetubaguyndipanov: I am totally cool with that13:09
kashyapndipanov: Nice work on block dev docs! Bookmarked to read :-)13:10
dane-fichterjohnthetubaguy: is nova moving to using python-glanceclient anytime soon?13:10
openstackgerritJim Rollenhagen proposed openstack/nova: Ironic: block storage support  https://review.openstack.org/18465213:11
johnthetubaguydane-fichter: so we still have to work out whats the right thing for those, its a little taffled up in the openstack SDK discussions13:12
*** sahid has quit IRC13:12
johnthetubaguydane-fichter: I think the real issue is having something a bit like brick we are getting from cinder for doing things with glance13:12
bauzasjohnthetubaguy: ndipanov: if you please, let me advert some spec I really want to get it approved for Liberty since I'm targeting to have it priotized https://review.openstack.org/#/c/179224/9/specs/liberty/approved/check-destination-on-migrations.rst,cm13:12
*** dpaterson has joined #openstack-nova13:13
ndipanovbauzas, ok will look but coffee first!13:13
bauzasjohnthetubaguy: ndipanov: took the last comments from you, guys13:13
garykjohnthetubaguy: hi, yes, if possible could we please get https://blueprints.launchpad.net/nova/+spec/vmware-limits approved13:13
johnthetubaguybauzas: so I think we had users asking for that, I was just a bit worried about the use case bit last time I looked, not had chance to revisit that yet13:13
bauzasndipanov: coffee is definitely a good option13:13
garykit is exposing specifc driver functionality via flavors13:13
bauzasjohnthetubaguy: yeah, we have customers waiting for it13:13
*** jamielennox is now known as jamielennox|away13:13
garykvia image properties requieres discussing and i need to post a spec and another bp13:13
*** jecarey has joined #openstack-nova13:13
*** ociuhandu has joined #openstack-nova13:13
bauzasjohnthetubaguy: and that's also caring about fixing how migrations are done to be consistent with regular boots13:14
*** amotoki_ has joined #openstack-nova13:14
*** nihilifer has quit IRC13:14
bauzasnot the story with lots of 'if' branches to explain to ops why the request didn't fail or failed because they specified a dest host or not13:14
johnthetubaguygaryk: is this all feature parity stuff, or are some of these new?13:14
*** jog0 has joined #openstack-nova13:15
johnthetubaguygaryk: we have done a spec for this one, but its was a really small one: http://specs.openstack.org/openstack/nova-specs/specs/liberty/approved/hyperv-storage-qos.html13:15
*** tojuvone has quit IRC13:15
johnthetubaguygaryk: if its just a "me too" blueprint, we can approve that at the next nova meeting (I added it to the agenda to make sure I don't forget)13:15
garykjohnthetubaguy: ok, thanks13:16
johnthetubaguygaryk: actually general question, do you folks, as a vmware subteam I mean, have any other blueprint "paperwork" thats needed for liberty at this point13:16
johnthetubaguygaryk: do let me know soon, and we should still be able to get it sorted in time, at this point13:17
johnthetubaguygaryk: I think the blueprint is just missing a few detail, btw, like the names of the image properties / flavor extra specs that are required to activate the code13:17
garykjohnthetubaguy: no, just this one and the console support (but that spec has already been approved)13:17
johnthetubaguygaryk: ah, thats good, so we are close13:17
dimsjohnthetubaguy: not sure if you saw my request above (https://review.openstack.org/191151) in support of a bug fix13:18
garykjohnthetubaguy: for the image properties dan want me to draft a seperate spec. i will do that tomorrow13:18
johnthetubaguygaryk: so if you could submit a brief spec for that vmware-limits blueprint, I think thats the easiest way13:18
johnthetubaguydims: sorry, znc missed that some how, we have been falling out13:18
garykjohnthetubaguy: ok, sure will do13:18
johnthetubaguygaryk: awesome, thank you13:18
dimsthanks just queueing it up on your list :)13:18
*** mpaolino has joined #openstack-nova13:18
*** kylek3h has joined #openstack-nova13:19
johnthetubaguydims: oh thats a good one to get fixed, not totally sure that actually need a spec, but probably easier to do it as a spec13:19
johnthetubaguydims: its just a policy change right?13:19
dane-fichterjohnthetubaguy: can i ask you a couple questions about your comments on the image signing / encryption spec?13:20
johnthetubaguydane-fichter: yes, sure13:20
*** jamielennox|away is now known as jamielennox13:21
dimsjohnthetubaguy: check the policy and if policy is ok then allow them to get the data they need13:21
johnthetubaguydims: ignore me, I understand why this needs a spec now, thanks13:21
dimsneeded a microversion13:21
johnthetubaguydims: yeah, sounds good... slight idea though, I will add it on the spec13:21
*** xyang1 has joined #openstack-nova13:23
dimsack thanks13:23
*** lucasagomes is now known as lucas-brb13:23
*** ddutta has joined #openstack-nova13:25
*** IanGovett has quit IRC13:25
alex_xudansmith: hi, I submitted initial version for new api of cpu model https://review.openstack.org/168982, hope you give suggestion if this is on the right direction when you have time, then I will fill up more detail. thanks!13:26
openstackgerritSylvain Bauza proposed openstack/nova-specs: Proposes to check destination on migrations  https://review.openstack.org/17922413:27
*** artom has joined #openstack-nova13:29
BobBallbauzas: FYI I'm looking at redeploying to pick up a new zuul in case there is another fix.  A redeploy likely won't be ready for a few days as I also need to switch accounts (to avoid redploying twice in quick succession)13:33
*** jamielennox is now known as jamielennox|away13:33
bauzasBobBall: no worries, my finger likes copy/pasting "xen: recheck" ;)13:33
bauzasBobBall: thanks for the update and the great investigation job you did13:33
BobBallAh good.  Well, maybe I'll leave it as it is then.  I wouldn't want to deprive your fingers of such joy ;)13:33
bauzasBobBall: eh :)13:34
ndipanovkashyap, do give feedback if something is unclear or missing on that doc pls!13:34
bauzasBobBall: seriously surprised to see I'm the only one yelling at it13:34
bauzasBobBall: probably because I'm French, I assume13:34
kashyapndipanov: Will do this week. Testing the selective block dev migration patches from libvirt git now.13:35
BobBallI assume it's something particular about your patchset that's triggering it13:35
bauzasBobBall: my git-review client, maybe ?13:35
bauzasBobBall: because my branch is just a regular git branch with a certain number of commits :)13:35
BobBallOr a strange interaction between the patches you've got and the current state of the repo?13:35
BobBallHave you rebased on master lately?13:36
bauzasBobBall: doing that very often13:36
*** erlon has joined #openstack-nova13:36
bauzasBobBall: I'm working on the nova objects, and those stuff don't let you being late on your branch13:36
BobBallOr perhaps the number of changes - if you change the same files several times, maybe there is something that exposes.  Dunno.13:36
bauzasBobBall: or perhaps I'm the only guy who sends more than one commit, meaning I'm the only guy who works here ? maybe13:37
BobBallhehe13:37
*** jskinner has joined #openstack-nova13:37
BobBallTrue.  I try to avoid working whenever I can.13:37
* BobBall looks around to check that his boss hasn't joined IRC.13:37
bauzaswe all do, that's why Frenchies have decent vacations13:37
* BobBall should move to France.13:38
bauzasthink moving in August or May, you should enjoy13:38
*** VW has joined #openstack-nova13:39
bauzasBobBall: anyway, thanks for your support, greatly appreciated :)13:39
BobBallAnytime13:40
openstackgerritChris Gacsal proposed openstack/nova: REST API changes to support multiple volume attach  https://review.openstack.org/15303813:40
*** burt has joined #openstack-nova13:41
*** lucas-brb is now known as lucasagomes13:42
*** ctrath has joined #openstack-nova13:42
*** jamielennox|away is now known as jamielennox13:42
*** mriedem_away is now known as mriedem13:42
*** moshele has quit IRC13:44
*** kbyrne has quit IRC13:46
*** Longgeek has quit IRC13:48
*** aswadr has joined #openstack-nova13:49
*** Longgeek has joined #openstack-nova13:49
mriedemgaryk: https://review.openstack.org/#/c/169732/ looks ok to me now if you want to get it rebased13:50
garykmriedem: thanks! i'll rebase13:51
*** asrangne has quit IRC13:52
openstackgerritgaryk proposed openstack/nova-specs: Propose VMware limits, reservation and shares  https://review.openstack.org/19267513:52
*** alex_klimov has quit IRC13:54
*** prashup has joined #openstack-nova13:55
*** moshele has joined #openstack-nova13:56
garykjohnthetubaguy: ^13:57
openstackgerritgaryk proposed openstack/nova: VMware: detach cinder volume when instance destroyed  https://review.openstack.org/16973213:58
garykmriedem: ^ rebased. thanks13:59
openstackgerritgaryk proposed openstack/nova: VMware: remove redundant check for block devices  https://review.openstack.org/17064114:00
*** cfriesen has joined #openstack-nova14:01
mriedemBobBall: i see that xenproject ci voted +1 on this immediately https://review.openstack.org/#/c/169732/ - is that due to filtering the change? i.e. since it's not xen or common code it's ignored by xenproject ci?14:03
BobBallcorrect mriedem14:03
* BobBall checks on the status of the zuul change to stop that14:03
mriedemBobBall: is there any way from it not giving the +1?14:03
openstackgerritgaryk proposed openstack/nova: VMware: Use actual VM state instead of using the instance vm_state  https://review.openstack.org/19062714:03
mriedemah ok14:03
dansmithah, I was wondering the same yesterday14:03
dansmithit's very misleading14:04
BobBallmriedem / dansmith: https://review.openstack.org/#/c/188383/14:04
dansmithhmm, seems like there should be another way to prevent that, but IANA3PCIA14:05
*** ijuwang has quit IRC14:06
BobBallThe only option we have ATM is to run a job - whether it is the real job or a noop.  "A result" will always be reported to a project that you're running tests for14:06
openstackgerritgaryk proposed openstack/nova: VMware: convert driver to use nova.objects.ImageMeta  https://review.openstack.org/18908514:06
*** tonytan4ever has joined #openstack-nova14:06
BobBallThe issue here is we use skipif to not run on changes that can't affect our CI - I guess mostly to reduce load14:07
BobBallSo from there there are no jobs :)14:07
dansmithBobBall: you don't have to explain, it just seems like you should be able to do frontend filtering, but I guess you're trying to use the (newer) skipif bits and that's buggy, right?14:07
BobBallfrontend filtering?14:07
BobBallI'm happy to filter any-which-way that makes sense.  I just assumed that skipif was the best way to do it14:07
*** ijuwang has joined #openstack-nova14:07
dansmithlike, never letting gerrit events get to you that don't apply14:07
dansmithBobBall: I'm sure it is14:08
BobBallAh... Unfortunately the gerrit stream doesn't contain any information on the changes in the patchset.  Only zuul gets to see them when it pulls the change14:08
dansmithhmm, okay14:09
dansmithbefore skipif,14:09
BobBallWould you prefer that the CI had a "I'm not testing this change" no op job?14:09
dansmithpeople were not running things against changes that didn't apply,14:09
BobBalle.g. "dummy-not-testing"?14:09
dansmithbut maybe they weren't using zuul14:09
dansmithBobBall: no, this fix is the right fix I'm sure, no need to change course, I'm just expressing confusion not proposing a change14:09
BobBallYup.  I added some custom code in the Citrix CI that downloaded the patchset and conditionally skipped a test.  But it was seriously ugly and quite error prone14:09
dansmithBobBall: if this was going to persist for a long time, some mitigation would be good, but I assume this isn't far off from landing14:10
*** artom has quit IRC14:10
BobBallI hope it's not, agreed.14:10
ndipanovbauzas, so what happens to the old way of forcing hosts once we introduce this?14:11
bauzasndipanov: you mean when migrating ?14:12
bauzasndipanov: or when booting ? sorry if unclear, but force_hosts is a hack for the boot command14:12
openstackgerritMatt Riedemann proposed openstack/nova: Support Network objects in set_network_host  https://review.openstack.org/19239614:12
*** mdrabe has quit IRC14:12
openstackgerritEd Leafe proposed openstack/nova: Add documentation for block device mapping  https://review.openstack.org/19265114:13
ndipanovright when booting14:13
edleafendipanov: great doc writeup on block device mapping. I had some grammar nits, so I just fixed them and pushed a new patch. ^14:13
openstackgerritgaryk proposed openstack/nova: Compute: enable admin to decide action for conflicting states  https://review.openstack.org/19004714:13
ndipanovedleafe, many thanks!14:14
edleafendipanov: glad to help!14:14
*** kbyrne has joined #openstack-nova14:15
ndipanovbauzas, when booting... will that have the same behavior now14:15
ndipanovor we leave that case as it was?14:15
*** zzzeek has joined #openstack-nova14:15
openstackgerritgaryk proposed openstack/nova-specs: Propose VMware limits, reservation and shares  https://review.openstack.org/19267514:17
*** nagyz has joined #openstack-nova14:17
*** doron_afk is now known as doron_14:18
bauzasndipanov: I would like to do this as a separate spec14:18
bauzasndipanov: because I dislike force_hosts but I would prefer to have a distinct way to ask for forcing a destination than just using --az fake_az:host14:19
ndipanovbauzas, I mean ok but how much more work is that on top of this.14:19
ndipanovup to you14:19
bauzasndipanov: MHO is that we should provide a new option when booting (ie. the same --force_destination) with an optional destination host14:20
mriedemdanpb: semi interesting convo in -infra right now about libvirt and bug 135393914:20
openstackbug 1353939 in OpenStack Compute (nova) "Rescue fails with 'Failed to terminate process: Device or resource busy' in the n-cpu log" [High,Fix committed] https://launchpad.net/bugs/1353939 - Assigned to Matt Riedemann (mriedem)14:20
mriedemdanpb: basically ceilometer has a compute agent that polls libvirt directly14:20
mriedemwhich theoretically could be causing that ebusy maybe14:20
bauzasndipanov: but I think it should be done at M14:20
*** jog0 has quit IRC14:20
bauzasndipanov: because just the current spec is enough large for Liberty only for me14:20
danpbmriedem: hmm, i'd be surprised if ceilometer could do something that would block SIGKILL from working in reasonable timeframe14:21
* mriedem graps at straws14:21
mriedem*grasps14:21
*** iamjarvo has joined #openstack-nova14:22
*** nkrinner has quit IRC14:23
*** exploreshaifali has quit IRC14:24
*** jog0 has joined #openstack-nova14:24
*** alex_klimov has joined #openstack-nova14:25
claudiubdanpb: hi14:26
*** Marga_ has joined #openstack-nova14:26
ndipanovbauzas, OK fair enough14:27
*** mdbooth has quit IRC14:27
ndipanovalso I see that we merged the RequestSpec spec with all the duplication...14:27
ndipanovok let's see how that goes once you get further in14:28
ndipanovmaybe it'll be ok14:28
ndipanovmaybe it'll be bad14:28
bauzasndipanov: otp14:28
*** sudipto has joined #openstack-nova14:28
*** mdrabe has joined #openstack-nova14:28
*** VW has quit IRC14:28
*** jlanoux has joined #openstack-nova14:29
*** sudipto has quit IRC14:30
*** zz_jgrimm is now known as jgrimm14:30
nagyzsup guys14:31
*** cdent has joined #openstack-nova14:31
*** xyang1 has quit IRC14:31
cdentmriedem: and...go!14:31
mriedemcdent: so for this libvirt pollster thing in ceilometer, was there ever a bp or spec in nova for doing something different with notifications from nova?14:32
sdaguemriedem: this stuff goes back a *long* ways14:32
cdentNot that I'm aware of, but that's not the same as no.14:32
mriedemthe get_instance_diagnostics stuff in nova now provides a versioned serializable structured object with the vm data14:32
cdentI think there's some historical fud that's been a blocker in the past.14:32
mriedemsdague: i get that, which predates the get_instance_diagnostics rewrite for v2.114:32
sdaguecdent: what's the default poll interval here?14:33
cdentsdague: 600 seconds14:33
openstackgerritStefan Amann proposed openstack/nova: Support host type specific block volume attachment  https://review.openstack.org/19255614:33
cdentwhich pretty much every operator says is far far far too low, and all the older devs say "you need to keep that as high as possible"14:33
mriedemso ceilometer could be using the v2.1 API to get instance diagnostics on that big of an interval, or nova could just have a periodic task in the compute manager that emits the notification14:33
cdents/low/long/14:33
mriedemkeep it high so you're not hammering the hypervisor?14:34
cdentlong term I want to kill all polling from ceilo and just hear notifications14:34
sdagueright, I think we moved that up because otherwise ceilometer was consuming over 50% of all resources on the gate test nodes14:34
mriedemi'm not sure how you'd trigger a notification in the diags case14:34
danpbyeah, the nova virt driver really should have a way to push performance data to ceilometer14:34
*** electrocucaracha has joined #openstack-nova14:34
danpbbecause libvirt can easily push out bulk stats data for every running guest in one go14:34
cdentsdague: there's two different sources of load: one is hammering the hypervisor, the other is asking the nova-api for resource metadata14:35
mriedemdanpb: the compute manager could already do this regardless of virt driver14:35
cdentthere is a spec in progress to vastly the improve the latter (we nope)14:35
danpbwhich will be waaaay more efficient than doing one rpc request on instance diagnostics for each guest14:35
mriedemdanpb: but i wonder if the other virt drivers can do that14:35
danpbmriedem: you still don't want the compute manager querying the virt driver for each individual guest14:35
mriedemespecially vmware/ironic14:36
danpbso it would at least need a new virt/driver.py  api to request bulk stats14:36
cdentresource metadata cachign: https://github.com/openstack/ceilometer-specs/blob/master/specs/liberty/resource-metadata-caching.rst14:36
sdaguemriedem: well, that's their own fault for being silly14:36
mriedemtrue14:36
*** sudipto has joined #openstack-nova14:36
cdentgetting the hypervisors to drop a regular notification onto messaging would be a huge win for any of the many things that like info, not just ceilo14:38
kashyapmriedem: Unintentionally I've just reproduced the  thing locally "error: Failed to terminate process 6426 with SIGKILL: Device or resource busy"14:39
sdaguecdent: agreed, so there was a failed attempt to get someone to write down what they wanted that schema to be a couple cycles ago when it looked like stacktac and ceilo would collab on it14:39
kashyapWhile testing migration related things as I wanted to kill off an unwanted guest.14:39
danpbanyway, insane architecture design decisions of the past aside, I really struggle to see how this ceilometer code would prevent the kenrel handling SIGKILL in a timely manner14:39
sdagueany idea if anyone is still trying to get there, because I think that's probably the missing next step14:39
mriedemi think cdent said he had nothing to work on and this was right up his alley :)14:39
cdentthat's exactly right mriedem :)14:40
sdaguedanpb: yeh, it was a shot in the dark. It was just surprising that we had 2 active clients.14:40
mriedemmaybe this jason myers on the ceilometer spec would be interested in helping on the nova side14:40
mriedemdefining the required spec14:40
cdentsdague: It's remained sort of latent on the ceilometer agent for the past two cycles14:40
mriedem*schema i mean14:40
danpbmriedem: did i recall you saying this only happened on certain OS versions ?14:40
*** kiran-r has quit IRC14:41
mriedemdanpb: the retry loop we added seems to work with newer libvirt/qemu in the centos6/fc21 jobs14:41
*** sambol has joined #openstack-nova14:41
cdentmriedem: jason's a good idea, this is right in his area of concern and interest14:41
mriedemdanpb: but those are also run far less often since they are experimental jobs14:41
sdaguemriedem: we run a lot less of those jobs, so are you sure it's statistically significant?14:41
mriedemsdague: jinx14:41
mriedemi'll jump over to -ceilometer and ping jasonamyers14:42
*** jasonamyers has joined #openstack-nova14:42
danpbmriedem: so you're saying the bulk of the failures are on the ubuntu images then14:43
cdentI momentarily had some traction with sandywalsh on notifications-related stuff but a) he left the building b) there was stronger interest in formalizing notification schema14:43
cdentb, to me, has felt rather like a distracting bike shed (after initial enthusiasm)14:43
*** VW has joined #openstack-nova14:44
sdaguedanpb: mriedem is saying that by absolute could, yes. But by absolute count we also run at least 100x more jobs on ubuntu images, so it's not really clear that means anything.14:44
sdaguefwiw, the most recent logstash fail is on f2114:44
sdaguehttp://logs.openstack.org/29/192329/2/check/check-tempest-dsvm-f21/6bd72b0/logs/screen-n-cpu.txt14:44
mriedemcdent: yeah (b) seems like it will never happen14:45
mriedemb includes versioned notifications i assume14:45
*** belmoreira has quit IRC14:45
cdentyeah, which, yeah great whatever, but first we need some notifications14:45
danpbsdague: yeah, agreed14:45
cdentSo it looks like jasonamyers and I are willing to do something here, can we define what that "something" is?14:46
sdagueso... actually14:46
sdaguehttp://logstash.openstack.org/#eyJmaWVsZHMiOltdLCJzZWFyY2giOiJtZXNzYWdlOihcImxpYnZpcnRFcnJvcjogRmFpbGVkIHRvIHRlcm1pbmF0ZSBwcm9jZXNzXCIgTk9UIFwiaW4gX3NodXRkb3duX2luc3RhbmNlXCIpIEFORCB0YWdzOlwic2NyZWVuLW4tY3B1LnR4dFwiIiwidGltZWZyYW1lIjoiNjA0ODAwIiwiZ3JhcGhtb2RlIjoiY291bnQiLCJvZmZzZXQiOjAsInRpbWUiOnsidXNlcl9pbnRlcnZhbCI6MH0sInN0YW1wIjoxNDM0NTUyMzQzNDE1LCJtb2RlIjoic2NvcmUiLCJhbmFseXplX2ZpZWxkIjoiYnVpbGRfbmFtZSJ914:46
sdaguethis is only faiing on f21, centos7, and the multinode jobs14:46
openstackgerritJeffrey Zhang proposed openstack/nova: Fix the wrong address ref when the fixed_ip is invalid  https://review.openstack.org/19270714:46
openstackgerritDan Smith proposed openstack/nova: WIP: Test removing flavors in sysmeta from compat_instance  https://review.openstack.org/19270814:46
sdaguemriedem: that's the right query, right?14:47
sdagueit was the one linked14:47
danpbwhat OS does the multinode job run on ?14:47
mriedemsdague: this is the e-r query http://goo.gl/kgNNKj14:47
mriedemdanpb: trusty14:48
sdaguemriedem: ok, the elastic recheck signature is stale then14:48
*** unicell has quit IRC14:48
*** exploreshaifali has joined #openstack-nova14:48
sdaguemriedem: we should get rid of that other er sig14:49
sdaguemriedem: http://goo.gl/kgNNKj is only 4 hits in 7 days?14:49
*** diogogmt has joined #openstack-nova14:50
sdague3 of which are centos714:50
*** unicell has joined #openstack-nova14:50
openstackgerritChris Gacsal proposed openstack/nova: REST API changes to support multiple volume attach  https://review.openstack.org/15303814:50
*** Jeffrey4l has joined #openstack-nova14:50
*** varunarya has joined #openstack-nova14:51
*** tjones1 has joined #openstack-nova14:52
Jeffrey4lbeagles, hi, I want to change the nova api current behavior. I was asked for some help from u. The detail can be found https://review.openstack.org/#/c/154068/11/nova/network/neutronv2/api.py . Could u give me some advice?14:52
*** sdake_ has joined #openstack-nova14:53
mriedemdanpb: fyi, this is what ceilometer would be looking to replace with a bulk diag notification: http://docs.openstack.org/admin-guide-cloud/content/section_telemetry-compute-meters.html - anything with 'pollster'14:53
mriedemsdague: looking14:53
*** dulek has quit IRC14:53
*** xyang2 has joined #openstack-nova14:53
openstackgerritJosh Durgin proposed openstack/nova-specs: RBD instance snapshots  https://review.openstack.org/18824414:53
beaglesJeffrey4l, looking14:54
*** moshele has quit IRC14:54
*** moshele has joined #openstack-nova14:54
Jeffrey4lthanks.14:54
*** moshele has quit IRC14:54
*** moshele has joined #openstack-nova14:55
danpbmriedem: those don't look too unreasonable14:55
*** tonytan_brb has joined #openstack-nova14:55
*** tonytan4ever has quit IRC14:55
*** xylan_kong has left #openstack-nova14:55
claudiubdanpb: hi14:56
*** tonytan_brb has quit IRC14:56
*** Jeffrey4l has quit IRC14:56
danpbclaudiub: hi ?14:56
*** tonytan4ever has joined #openstack-nova14:56
*** ssurana has joined #openstack-nova14:56
*** Jeffrey4l has joined #openstack-nova14:56
claudiubdanpb: So, from what I understand, you have experience in libvirt using shared storage, right?14:57
sdaguedanpb: so what would you envision as the flow to get all this stuff out as notifications?14:57
danpbclaudiub: i've used nfs once or two to test migration, if that counts as experiance14:57
*** jdurgin1 has joined #openstack-nova14:58
danpbsdague: i guess the nova compute manager would periodically query virt driver for bulk stats about all guests14:59
cdentsdague, danpb: A big question for me (re flow) is whether the notifications would come as a big huge drop about all the current instances on a node, or whether there are efficiencies gained by the data gathering being granular?14:59
danpband then push that data out on the message bus to some target ?14:59
mriedemdanpb: is there a page somewhere that shows what a bulk data dump from libvirt looks like?14:59
*** nihilifer has joined #openstack-nova15:00
claudiubdanpb: I see. Well, there is a thingy, that I was hoping you would have some advice on. So the scenario is that there are multiple compute nodes over the same shared storage. The issue is that they both report the same disk resources in this case, which can be problematic, if you spawn 1 instance on one one with 200GB storage, the other compute nodes will not see that 200 GB as used storage..15:00
danpbif we want to unicast the data on the bus, the act of deploying ceilometer would involve a nova.conf setting to tell it where to target the data15:00
claudiubdanpb: ... and report it as free.15:00
*** otter768 has joined #openstack-nova15:00
danpbclaudiub: yes, the resource manager handling of shared storage is utterly fubar currently15:01
*** browne has joined #openstack-nova15:01
sdaguecdent: good questions, I honestly expect it's more efficient to collect and send in bulk, but that's just a guess.15:01
jdurgin1danpb: dansmith: jaypipes: thanks for your comments on https://review.openstack.org/#/c/188244/ - ready  for a final review now15:01
danpbclaudiub: last time i looked at it, i came to the conclusion it couldn't be fixed in its current form and would need majorly redesigning15:01
jaypipesjdurgin: it's right at the top of my list of reviews to get to today.15:02
mriedemsdague: cdent: danpb: jasonamyers: ildikov: i'll sign up to take notes and get those into the ML, then start working a rough draft spec with help from jasonamyers on the ceilometer particulars (what ceilometer needs out of nova)15:02
mriedemctrath: ^15:02
danpbcdent: given that list of data items that ceilometer wants - if we queried libvirt for each guest, for each type of data, that would require on the order of  O(n*6)  libvirt API calls15:02
jdurgin1jaypipes: thanks!15:02
*** mpaolino has quit IRC15:03
danpband even though they're fast, that still doesn;t really scale nicely15:03
claudiubdanpb: I see... well, this is a problem I'm currently facing and I was hoping for some answers.. well, if by any chance I find a worthwhile solution, I'll let you know.15:03
cdentdanpb: I wasn't suggesting we considering doing one for each datapoint, just wondering what the options are15:03
danpbnewer libvirt has an API call that lets us query data for all running guests in a single api call15:03
sdaguemriedem: sounds great, thanks15:03
sdaguedanpb: when did that get added?15:03
danpbso my preference would be to design something that lets us take advantage of that15:03
jasonamyersthank you mriedem15:03
cdentdanpb: makes sense15:04
cdentmriedem++15:04
danpbsdague: about version 1.2.815:04
dansmithjaypipes: you're the primary reviewer that needs to approve that I think, so let me know when you're happy with it and I'll hit it again15:04
phschwartzHey all, can I get some more people reviewing https://review.openstack.org/#/c/154521/ I want to try to get it in as soon as possible so we can begin testing it in a new release version he in Rax.15:04
danpbsdague: but that's not a big deal - if we run against older libvirt we'd simply fallback to querying data the slow way15:04
mriedemdanpb: ok, so we'd have a version cap but could test this in the centos6/fc21 job at least15:04
jaypipesdansmith: will do.15:04
mriedemdanpb: ah, or that15:04
*** xyang1 has joined #openstack-nova15:04
sdaguedanpb: ok, so we'd still need a per instance query in the fallback15:04
danpbyep15:05
*** otter768 has quit IRC15:05
sdaguebut that seems still less expensive than the current way15:05
*** openstackgerrit has quit IRC15:05
danpbyeah, as you only have the overhead of the libvirt rpc system there which is low15:05
danpbthe real key is to avoid nova rpc overhead per instance15:05
*** openstackgerrit has joined #openstack-nova15:05
*** amotoki_ has quit IRC15:06
openstackgerritAnkit Agrawal proposed openstack/nova: Test patch race condition in delete (do not merge)  https://review.openstack.org/19173215:06
danpbfor reference, the libvirt api i'm referring to is   http://libvirt.org/html/libvirt-libvirt-domain.html#virConnectGetAllDomainStats15:06
*** mdrabe has quit IRC15:07
*** agireud has joined #openstack-nova15:08
*** xyang2 has quit IRC15:08
*** xyang2 has joined #openstack-nova15:08
*** ssurana has quit IRC15:08
*** amotoki_ has joined #openstack-nova15:08
dansmithphschwartz: what all have you changed in this since jerdfeldt's last version? Just rebasing?15:08
*** mdrabe has joined #openstack-nova15:09
openstackgerritDaniel Berrange proposed openstack/nova: objects: add virtual 'image_meta' property to Instance object  https://review.openstack.org/19014515:09
*** iamjarvo has quit IRC15:09
openstackgerritDaniel Berrange proposed openstack/nova: libvirt: convert to use instance.image_meta property  https://review.openstack.org/19014615:09
openstackgerritDaniel Berrange proposed openstack/nova: compute: convert manager to use nova.objects.ImageMeta  https://review.openstack.org/18968515:09
phschwartzdansmith: Rebase + a few changes related to code review, things like changes of use of iteritems and compressing of a nested if statement that didn't need to be nested.15:09
dansmithphschwartz: okay15:10
*** afazekas has quit IRC15:10
dansmithphschwartz: given the recent conversations and the fact that we know that contract is a little risky right now, I wonder if we should make the nova-manage contract command just return "not supported"?15:10
dansmithsince contract is completely optional15:11
danpbdansmith: i was wondering what/where  VersionedObjectSerializer is used ?15:11
dansmithdanpb: it's NovaObjectSerializer in our tree at the moment15:12
phschwartzdansmith: I discussed that with johnthetubaguy to see if he felt if we should do just that, his thoughts were since it was experimental that allowing to still function would be fine.15:12
dansmithdanpb: it's plugged into our RPC layers so that objects get serialized automatically over the wire15:12
phschwartzdansmith: I have no issues making it "not supported" at the moment though15:12
*** iamjarvo has joined #openstack-nova15:12
*** armax has joined #openstack-nova15:12
danpbi was doing some hacking on oslo.messaging that would invole olso messaging depending on something that used vesioned objects15:12
dansmithjohnthetubaguy: what about a --no-really-I-mean-it flag?15:12
*** haypo has left #openstack-nova15:12
*** iamjarvo has quit IRC15:12
danpbbut because versioned objects has this class, i ended up with a circular dependancy between olso messaging & versionedobjects :-(15:12
dansmithjohnthetubaguy: just because I'm not sure we're exposing it as experimental enough before things could be destroyed15:13
*** Nikolay_St has quit IRC15:13
*** iamjarvo has joined #openstack-nova15:13
*** xyang2 has quit IRC15:13
dansmithdanpb: yeah, not sure why messaging should know anything about objects.. it's a lower layer, right?15:13
phschwartzdansmith: technically it is a "no op" unless someone removes a column from a modle.15:13
phschwartz*model15:13
dansmithdanpb: the actual serializer interface is defined in oslo.messaging though15:13
dansmithphschwartz: yep15:14
danpbdansmith: well versioned objects aren't just used for things going over the message bus :-)15:14
*** xyang2 has joined #openstack-nova15:14
danpbdansmith: they're so useful i was trying to use them for instrumenting API calls and serializing them out to disk15:14
dansmithdanpb: sure, but... in oslo.messaging?15:14
danpband oslo messaging was one of the things i wanted to instrument15:14
danpbi guess i'll have to think about how to restructure things15:15
danpbso i can avoid creating a circular dep15:15
dansmithdanpb: well, I'd have to see an example I guess to help more15:15
dansmithdanpb: I'm both intrigued and terrified by what you're doing :D15:15
danpblol15:15
*** xyang2 has quit IRC15:16
*** doron_ is now known as doron_afk15:17
*** mtanino has joined #openstack-nova15:18
dansmithjaypipes: I think you should do the +W on this: https://review.openstack.org/#/c/154521/1915:19
dansmith(or the -1 of course)15:20
mriedemdansmith: dims: what was the 1:1 clustered hypervisor change again? that was to enforce a single compute node/host for vmware/ironic right?15:22
*** doron_afk has quit IRC15:22
*** xyang2 has joined #openstack-nova15:22
mriedemi think that might be required for this bulk stats thing to work for those driveres15:23
johnthetubaguydansmith: phschwartz: year. I think I requested an experimental flag at some point but forgot about that, --i-might-kill-your-cloud=True or whatever it is15:23
*** afazekas has joined #openstack-nova15:23
phschwartzjohnthetubaguy: If you would like, I can add that now15:23
dansmithjohnthetubaguy: I think that'd be a real good idea, yeah15:24
dimsmriedem: to be able to expose the vcenter+multiple-esx hierarchy in nova API and CLI - so i can say deploy straight to esx-3215:24
mriedemdims: do you remember the patch that was going to enforce that at all?15:24
mriedemi think it was abandoned in kilo15:24
dimsmriedem: never got started, so no code15:24
dansmithphschwartz: throw a -1 on there so we know you're working on that and don't land it before you do15:25
johnthetubaguyphschwartz: dansmith: +115:25
jaypipesdansmith: done15:25
mriedemdims: i remember there was some change that came out of the juno summit in ATL which was agreed to be a thing, basically deprecate or not allow multiple vcenters for a single compute service15:25
mriedemsomething like that15:25
dansmithdoh15:25
dansmithheh15:25
dansmithphschwartz: well, maybe let that land and follow up with that flag?15:25
dimsmriedem: that's in already. one can only specify one vcenter cluster with one nova-compute process15:26
phschwartzdansmith: yeah, he just beat me to it. I will wait for it to gate and then put up a follow up with the flag15:26
dimsearlier one could specify multpiple vcenter clusters using just one nova-compute15:26
dansmithphschwartz: k15:26
mriedemdims: is that this? https://review.openstack.org/#/c/163426/15:27
*** afazekas has quit IRC15:28
dimsmriedem: looking15:28
mriedemi should have -1ed that :) we should have updated the options help text since that's what goes in the docs, and we should have had a DocImpact/UpgradeImpact so we updated the release notes15:28
mriedemi can add that15:28
openstackgerritJordan Pittier proposed openstack/nova: libvirt:when deleting snapshot use qemu-img to blockRebase if domain is stopped  https://review.openstack.org/19273615:29
dimsmriedem: yep15:29
*** oomichi has joined #openstack-nova15:29
mriedemoh it's in kilo, i can update the release notes for that yet15:30
mriedemi'll still update the help text to say multiple is a no-no15:30
dimsright15:30
dimsgaryk: we need to revive https://review.openstack.org/#/c/103916/15:30
dimsmriedem: ^^ that's what we should get into liberty15:30
bauzasalex_xu: still around by a very late time ?15:31
*** garyk has quit IRC15:31
bauzasor maybe an API subteam member ?15:31
bauzasgilliard, sdague ?15:31
bauzasmy question is about https://review.openstack.org/#/c/179224/10/specs/liberty/approved/check-destination-on-migrations.rst,cm15:31
*** yamahata has joined #openstack-nova15:31
*** Maike has quit IRC15:32
bauzasand what should be the behaviour if someone is specifying a flag like --force_destination but would not provide a destination host15:32
bauzas(that's silly, but possible)15:32
bauzasshould we just drop the request by a 400 ?15:32
*** yamahata has quit IRC15:32
bauzasor ignore the flag ?15:32
bauzasgilliard: sdague: alex_xu: ^15:32
*** yamahata has joined #openstack-nova15:33
johnthetubaguymriedem: good spot on that one, thanks, I guess it was missing the upgradeimpact tag :(15:33
* bauzas is considering calling 555-API-WG15:34
bauzas:p15:34
dansmithtoo many numbers15:35
bauzashuh, we have 10 digits here15:35
dansmithwell, the 555 in the US is an exchange,15:35
dansmithwhich means only four digits after it15:35
edleafebauzas: yes, return a 40015:35
bauzasdansmith: I see :)15:36
edleafedansmith: unless it's an area code15:36
edleafedansmith: then it should be 555-API-WG5515:36
* bauzas just remembers the Last Action Hero movie15:36
dansmithedleafe: well, the conventional example of 555 numbers are using it as an exchange15:36
edleafedansmith: yeah, but he's French, so... :)15:37
dansmithedleafe: and 555 as an area code goes to actual things15:37
dansmithedleafe: right, but I assumed 555 had no special significance there15:37
dansmithassumed he was making a US-centric movie reference15:37
edleafedansmith: correct, of course15:37
dansmithwhich he was15:37
bauzashum, 555 even has a wikipage https://en.wikipedia.org/wiki/555_%28telephone_number%2915:38
bauzasdansmith: yeah, like I said, 555 is well known for us :)15:38
dansmithokay, that's sorted, what's next? :)15:38
*** jamielennox is now known as jamielennox|away15:38
*** rajesht has quit IRC15:39
*** jdurgin1 has quit IRC15:39
bauzasedleafe gave me his opinion, I take it :)15:39
bauzasthe magical number worked !15:39
*** ssurana has joined #openstack-nova15:40
edleafebauzas: there should be a magic API WG incantation you can chant15:40
dansmithwell, of course, because once you typed APIW, you were connected and the G was just superfluous15:40
*** vishwana_ has quit IRC15:40
*** vishwanathj has joined #openstack-nova15:41
*** lpetrut has quit IRC15:42
openstackgerritSylvain Bauza proposed openstack/nova-specs: Proposes to check destination on migrations  https://review.openstack.org/17922415:44
sdaguebauzas: it should be an explicit 400 if the provide non sensible parameters15:44
mriedemjasonamyers: ctrath: etherpad with notes on the thing https://etherpad.openstack.org/p/nova-hypervisor-bulk-stats-notify15:45
tdurakovdanpb: hi!15:45
bauzassdague: fair, ty15:45
*** ericksonsantos has joined #openstack-nova15:46
tdurakovdanpb:about  https://review.openstack.org/#/c/183057/3 VirtCPUModel  uses VirtCPUFeature for feature list, do you think we should change that in api?15:46
*** jamielennox|away is now known as jamielennox15:47
*** unicell1 has joined #openstack-nova15:47
*** hemnafk is now known as hemna15:47
*** unicell has quit IRC15:50
*** Jeffrey4l has quit IRC15:51
*** kiran-r has joined #openstack-nova15:52
*** jecarey has quit IRC15:53
*** exploreshaifali has quit IRC15:53
*** tjones1 has quit IRC15:54
*** tjones1 has joined #openstack-nova15:56
*** kiran-r has quit IRC15:56
*** cfriesen has quit IRC15:56
*** kiran-r has joined #openstack-nova15:57
*** jamielennox is now known as jamielennox|away15:58
ctrathmriedem: thanks15:58
*** tsekiyama has joined #openstack-nova15:58
*** Sukhdev has joined #openstack-nova15:58
*** tojuvone has joined #openstack-nova15:59
beaglesJeffrey4l, was there a specific thing you were looking for an assist on?15:59
*** krtaylor has quit IRC15:59
*** haypo has joined #openstack-nova16:00
*** SourabhP has joined #openstack-nova16:00
*** iamjarvo has quit IRC16:00
*** ypraveen has joined #openstack-nova16:01
mriedembtw, ctrath is new to nova and will be working on it, so feel free to dump everything on him16:01
*** dulek_home has joined #openstack-nova16:02
*** tonytan4ever has quit IRC16:03
*** kiran-r has quit IRC16:03
*** kiran-r has joined #openstack-nova16:03
*** exploreshaifali has joined #openstack-nova16:03
ctrathmriedem: thanks again. :-*16:04
*** jlanoux has quit IRC16:04
*** rotbeard has quit IRC16:06
*** jamielennox|away is now known as jamielennox16:06
*** arnaud___ has joined #openstack-nova16:06
*** vilobhmm has joined #openstack-nova16:06
*** tpeoples has joined #openstack-nova16:08
*** Jeffrey4l has joined #openstack-nova16:08
openstackgerritTimofey Durakov proposed openstack/nova-specs: nova-api hypervisor.cpu_info change from string to json object  https://review.openstack.org/18305716:08
*** jcoufal has quit IRC16:10
*** kiran-r has quit IRC16:10
*** dims_ has joined #openstack-nova16:11
*** krtaylor has joined #openstack-nova16:11
*** garyk has joined #openstack-nova16:11
*** achanda has joined #openstack-nova16:12
*** electroc_ has joined #openstack-nova16:13
*** dims has quit IRC16:13
*** jecarey has joined #openstack-nova16:13
*** unicell1 has quit IRC16:14
openstackgerritMatt Riedemann proposed openstack/nova: vmware: update help text for vmware.cluster_name option  https://review.openstack.org/19275316:15
*** nic has joined #openstack-nova16:16
*** electrocucaracha has quit IRC16:16
*** gszasz has quit IRC16:17
openstackgerritChris Gacsal proposed openstack/nova-specs: Add REST API impact to the multi-attach spec  https://review.openstack.org/19275516:18
*** rook has joined #openstack-nova16:18
*** ptm is now known as PaulMurray16:19
*** Mike_D_laptop has joined #openstack-nova16:19
*** rook is now known as rook-spec16:19
PaulMurrayndipanov,16:21
*** achanda_ has joined #openstack-nova16:21
*** gabor-t has joined #openstack-nova16:21
PaulMurrayndipanov, oops sorry  - autocomplete failure16:21
ndipanovPaulMurray, no worries16:21
*** neophy has joined #openstack-nova16:22
*** agireud has quit IRC16:23
lxslidansmith: before I go any further, is this the sort of thing you were expecting please? re: no-more-soft-delete16:23
lxslidansmith: http://paste.openstack.org/show/298813/16:23
*** achanda has quit IRC16:24
bauzasndipanov: around for discussing why I made https://review.openstack.org/#/c/188487/ this way ?16:24
dansmithlxsli: that looks awesome.. it *might* be too much given how many instances we have.. I wonder if you could cut out some of the middle layers.. you'd just have to use judgment on which ones make sense16:24
dansmithlxsli: maybe start with the fully verbose layout,16:25
*** sdake_ is now known as sdake16:25
dansmithand we can trim it down in the review too, if you'd prefer16:25
ndipanovbauzas, ok16:25
lxslidansmith: OK I'll slog on!16:25
bauzasndipanov: okay, lemme try to be clear16:25
bauzasndipanov: so, I'm okay with creating a new select_nodes(RequestSpec) method16:26
dansmithlxsli: let me edit that paste as an example16:26
bauzasndipanov: but this means I'll have to change all the southbound method signatures, right?16:26
ndipanovwhat's southbound?16:26
*** dsanders has joined #openstack-nova16:26
lxslidansmith: removing middle layers is easy, it's doing the trace that takes the time16:26
bauzasndipanov: the methods called by the API server methods16:27
*** markus_z has quit IRC16:27
bauzasndipanov: so for example, given select_nodes(RequestSpec)16:27
dansmithlxsli: http://paste.openstack.org/show/298814/16:27
bauzasndipanov: it will call FilterScheduler.select_nodes(Spec)16:27
dansmithlxsli: you can also probably stop at some layer in the trace once you know the semantics of what is going on16:27
lxslidansmith: ah! that's very helpful!16:27
bauzasndipanov: but I'll also have to create it select_nodes() for the ChanceScheduler16:28
*** afazekas has joined #openstack-nova16:28
bauzasndipanov: then, I'll also have to modify the signature of nova.filters16:28
lxslidansmith: great, that often becomes obvious relatively early, but chasing through the layers of API code gets tedious16:28
bauzas(or create a new method)16:28
ndipanovyes - you add a completely new RPC call and method16:28
*** vilobhmm has quit IRC16:28
dansmithlxsli: that parenthetical comment is exactly what I think we need to evaluate in each case... knowing what it is doing, why, and if we *really* need read_deleted, or if we just do it to avoid a trace16:28
ndipanovbauzas, yes - nova filters need to be changed in a single patch16:28
dansmithlxsli: yep16:28
bauzasndipanov: then I'll have to create host_passes2(RequestSpec)16:29
bauzasall of that will need to be done in a single patch, right?16:29
ndipanovwhy?16:29
ndipanovso select_destination_objectified() works with objects16:29
bauzasndipanov: I see, you're asking to duplicate all the calls16:29
openstackgerritSudipta Biswas proposed openstack/nova: The 'scheduled_at' field is a part of the Nova instances. Currently, the 'scheduled_at' field is no longer updated by the Nova scheduler, instead the 'launched_at' field is updated for the same purpose. The 'scheduled_at' column is now redundant and shoul  https://review.openstack.org/19276016:29
*** lpetrut has joined #openstack-nova16:30
ndipanovselect_destinations() gets dicts, forward ports them, and works with objects16:30
ndipanovboth of these only call filters which are converted in a single patch16:30
ndipanov(but anyway they would have to be converted in a single patch or have some kind of a silly decorator that does isinstance blah16:30
bauzasndipanov: I'll have to change all the filters at once, right?16:30
*** achanda_ has quit IRC16:31
ndipanovyeah but it's like renaming stuff16:31
openstackgerritSudipta Biswas proposed openstack/nova: Remove 'scheduled_at' - DB cleanup  https://review.openstack.org/19276016:31
ndipanovit's not like there's million of t hings to change16:32
ndipanovit's just filters16:32
ndipanov(and tests)16:32
ndipanov250 line patch at most16:32
*** matrohon has quit IRC16:32
ndipanovthat's stupid easy to review16:32
ndipanovand the win is that there is no backporting to old format which is hell16:32
ndipanovtrust me16:32
ndipanovdownside is taht you get to write less patches and your stackalytics suffer from that :P16:33
alex_xubauzas: just +1, I think 400 is ok for me16:33
*** openstackgerrit has quit IRC16:33
*** agireud has joined #openstack-nova16:34
*** openstackgerrit has joined #openstack-nova16:34
*** jamielennox is now known as jamielennox|away16:35
*** Marga_ has quit IRC16:35
*** oomichi has quit IRC16:36
ndipanovbauzas, fwiw - you could do all that without adding a new method16:36
ndipanovbut I see no reason not to16:36
*** jog0 has quit IRC16:37
*** krtaylor has quit IRC16:37
ndipanovyou would ghetto-ise all the cruft in hte old method and have a new shiny method to work with!16:37
*** agireud has quit IRC16:37
*** sudipto has quit IRC16:38
*** arnaud___ has quit IRC16:38
dansmith+1 for ghetto-ization16:38
*** artom has joined #openstack-nova16:38
*** neophy has quit IRC16:38
*** neophy has joined #openstack-nova16:39
*** ypraveen has quit IRC16:39
openstackgerritIldiko Vancsa proposed openstack/nova-specs: Add REST API impact to the multi-attach spec  https://review.openstack.org/19275516:41
*** kylek3h_ has joined #openstack-nova16:43
*** thangp_ has joined #openstack-nova16:43
*** apuimedo_ has joined #openstack-nova16:43
*** med_` has joined #openstack-nova16:43
*** oro_ has joined #openstack-nova16:43
*** gyee_ has joined #openstack-nova16:43
*** cdent_ has joined #openstack-nova16:43
*** pixelbeat_ has joined #openstack-nova16:43
*** vladikr_ has joined #openstack-nova16:43
*** changbl_ has joined #openstack-nova16:43
*** fifieldt_ has joined #openstack-nova16:44
*** wverdugo5001 has joined #openstack-nova16:44
*** mtanino has quit IRC16:45
*** kylek3h has quit IRC16:45
*** vladikr has quit IRC16:45
*** oro has quit IRC16:45
*** thangp has quit IRC16:45
*** cdent has quit IRC16:45
*** pixelbeat has quit IRC16:45
*** apuimedo has quit IRC16:45
*** fifieldt has quit IRC16:45
*** josecastroleon has quit IRC16:45
*** changbl has quit IRC16:45
*** operator99 has quit IRC16:45
*** med_ has quit IRC16:45
*** tcammann has quit IRC16:45
*** cdent_ is now known as cdent16:45
*** kylek3h_ is now known as kylek3h16:45
*** josecastroleon1 has joined #openstack-nova16:45
*** nic has quit IRC16:45
*** nic has joined #openstack-nova16:46
*** a| has joined #openstack-nova16:46
*** a| has quit IRC16:46
bauzasndipanov: sorry, was afk for a bit16:47
bauzasndipanov: had to fix something16:47
*** penick has joined #openstack-nova16:47
*** mfisch has quit IRC16:47
*** logan2 has joined #openstack-nova16:47
bauzasndipanov: so, you're saying, just changing the filters by hydrating the object there, then doing backwards ?16:48
*** tcammann has joined #openstack-nova16:48
*** sambol has quit IRC16:48
*** krtaylor has joined #openstack-nova16:48
*** mtanino has joined #openstack-nova16:49
*** jamielennox|away is now known as jamielennox16:49
*** mfisch has joined #openstack-nova16:50
*** mfisch has quit IRC16:50
*** mfisch has joined #openstack-nova16:50
*** kiran-r has joined #openstack-nova16:50
ndipanovno16:51
*** unicell has joined #openstack-nova16:51
*** Marga_ has joined #openstack-nova16:51
*** alexpilotti has joined #openstack-nova16:51
bauzasndipanov: sorry for being dumb, but I don't get your point :/16:52
*** romainh has left #openstack-nova16:52
*** exploreshaifali has quit IRC16:52
ndipanovyou have one patch that moves the whole scheduler to use the new Objects16:52
bauzasndipanov: I'm not that sure it would be a +200/-200 one then16:52
ndipanovmore or less16:52
*** prashup has quit IRC16:52
bauzasndipanov: just because of some cruftness in the HostManager for example16:52
ndipanovand then old method when it gets the dicts - forward ports them (that already exists)16:53
*** sudipto has joined #openstack-nova16:53
bauzasndipanov: okay, I can give a try16:54
*** yamahata has quit IRC16:54
ndipanovit literally does 3 lookups on filter properties16:54
bauzasndipanov: but my main worries is the size of the change for using the ReqSpec dot notation as a whole16:54
ndipanovnothing else touches it16:54
ndipanovmeh16:54
mriedembauzas: do dot notation changes in a separate patch16:54
mriedembulk dot notation changes that is16:55
*** yamahata has joined #openstack-nova16:55
ndipanovthat's better than having a ton of useles commits plus a bunch of code that is useless (backporting)16:55
mriedemconvert the unit tests to use objects rather than dicts and change to dot notation in the code at the same time to enforce that it's working16:55
ndipanovmriedem, it can't be done like that16:55
ndipanovit has to be one shot with the RPC change16:55
ndipanovto avoid backporting to dicts16:55
ndipanovwhich has been proven horrible in a number of cases16:55
mriedemsorry, came in late to a thing i'm not following16:56
bauzasmriedem: that was my initial series where ndipanov is -1 it16:56
ndipanovand we actually don't need it here - it';s nowhere near as bit as for example instance in the compute manager16:56
ndipanovit';s just filters!16:56
ndipanovliterally16:56
bauzasndipanov: and HostManager.consume_from_instance() again16:56
*** lucasagomes is now known as lucas-dinner16:56
jaypipesjohnthetubaguy, ndipanov, danpb: https://review.openstack.org/#/c/183057/ could use a re-review from you. thx in advance. I know tdurakov is eager to get started on that work.16:57
bauzasbut I can give a test and see16:57
mriedemndipanov: literally? https://twitter.com/awfulannouncing/status/52675435126482944016:57
*** belmoreira has joined #openstack-nova16:57
ndipanov:)16:57
*** achanda has joined #openstack-nova16:57
*** abhishekk has quit IRC16:58
openstackgerritSudipta Biswas proposed openstack/nova: Remove 'scheduled_at' - DB cleanup  https://review.openstack.org/19276016:58
bauzasndipanov: anyway, I got your views, lemme see how I can achieve it16:58
*** derekh has quit IRC16:58
*** vilobhmm has joined #openstack-nova16:58
bauzasalaski: cells meeting in 2mins right?16:58
alaskicorrect16:59
bauzascool16:59
*** kiran-r has quit IRC16:59
*** kiran-r has joined #openstack-nova16:59
*** haypo has quit IRC16:59
ndipanovbauzas, one more thing you'll need to do16:59
*** dulek_home has quit IRC17:00
bauzasndipanov: sure, any feedback is good to hear17:00
ndipanovmove all the junk from scheduler.utils to the object itself17:00
ndipanovbut we should all be super happy about that really17:00
bauzasndipanov: yeah, that's the target, again I just wanted some small patches17:00
ndipanovyeah taht can be a separate patch of course17:00
bauzasndipanov: my stackalytics stats are good, ty :p17:01
bauzasndipanov: even if my review ratio is decreasing dramatically because of that stuff :p17:01
ndipanovthe fact that you even look at that makes me think less of you :P17:01
*** otter768 has joined #openstack-nova17:01
openstackgerritSudipta Biswas proposed openstack/nova: Remove 'scheduled_at' - DB cleanup  https://review.openstack.org/19276017:01
*** atuvenie has quit IRC17:01
bauzasndipanov: meh, don't you do that every 1 hour ?17:01
bauzas:p17:01
*** Marga_ has quit IRC17:01
*** chaochin has joined #openstack-nova17:02
*** aswadr has quit IRC17:02
*** asrangne has joined #openstack-nova17:02
*** achanda has quit IRC17:02
*** dulek_home has joined #openstack-nova17:02
*** dulek_home has quit IRC17:03
*** exploreshaifali has joined #openstack-nova17:03
*** haomaiw__ has joined #openstack-nova17:04
*** rushiagr is now known as rushiagr_away17:04
*** haomaiwang has quit IRC17:04
*** asrangne__ has joined #openstack-nova17:05
*** achanda has joined #openstack-nova17:05
*** otter768 has quit IRC17:06
*** alex_klimov has quit IRC17:07
*** Marga_ has joined #openstack-nova17:08
*** yamahata has quit IRC17:08
ndipanovok bauzas need to run - but if you don't trust me ask dansmith... if we can get away with not doing the new -> legacy conversion and only always forwardport - it saves us a ton of headache17:08
bauzasndipanov: yeah, I got the rationale17:08
bauzasndipanov: which I appreciate17:08
*** asrangne has quit IRC17:08
bauzasndipanov: I just need to see the level of complexity that it implies, but maybe I ubercared too much17:09
openstackgerritSudipta Biswas proposed openstack/nova: Remove 'scheduled_at' - DB cleanup  https://review.openstack.org/19276017:09
bauzasoverthinking is like a second nature for me :(17:09
ndipanovI think the scheduler is the perfect place to do it17:09
*** moshele has quit IRC17:09
ndipanovwe trade a beefier (but easy) patch for less and way simpler code17:09
ndipanovobviously cannot be done for (for example) compute manager17:10
*** Marga_ has quit IRC17:11
ndipanovwhat might trip you up is that filter props get sent to the compute host... :(17:11
bauzasndipanov: I know17:11
*** achanda has quit IRC17:11
bauzas...17:11
ndipanovshite17:12
ndipanovah well17:12
ndipanovI'm off - talk tomorrow :)17:12
bauzassure thign17:12
bauzasfor that, I was just thinking to hydate the Spec but leave the existing filt_props passed to the compute manager17:12
*** ndipanov has quit IRC17:12
*** jamielennox is now known as jamielennox|away17:13
bauzasie. we hydrate the object right before calling the scheduler API, but we leave the existing dicts, until someone enough foolish is doing that for the compute side17:13
openstackgerritJosh Durgin proposed openstack/nova-specs: RBD instance snapshots  https://review.openstack.org/18824417:13
bauzascrap, nikola left :p17:13
openstackgerritQin Zhao proposed openstack/nova: Make resize api of compute manager to send flavor object  https://review.openstack.org/16186017:14
openstackgerritQin Zhao proposed openstack/nova: Support flavor object in migrate_disk_and_power_off  https://review.openstack.org/19278617:14
*** dims_ has quit IRC17:16
*** ypraveen has joined #openstack-nova17:16
*** dims has joined #openstack-nova17:17
*** SourabhP has quit IRC17:18
*** achanda has joined #openstack-nova17:18
*** cfriesen has joined #openstack-nova17:19
*** atuvenie has joined #openstack-nova17:21
*** achanda_ has joined #openstack-nova17:22
openstackgerritDaniel Berrange proposed openstack/nova: objects: add virtual 'image_meta' property to Instance object  https://review.openstack.org/19014517:25
openstackgerritDaniel Berrange proposed openstack/nova: libvirt: convert to use instance.image_meta property  https://review.openstack.org/19014617:25
openstackgerritDaniel Berrange proposed openstack/nova: compute: convert manager to use nova.objects.ImageMeta  https://review.openstack.org/18968517:25
*** achanda has quit IRC17:25
*** ddutta has quit IRC17:25
*** ildikov has quit IRC17:30
*** sdake_ has joined #openstack-nova17:30
*** ctrath has quit IRC17:31
*** varunarya has quit IRC17:31
*** VW has quit IRC17:32
openstackgerritEd Leafe proposed openstack/nova: Add logging when filtering returns nothing.  https://review.openstack.org/19218417:32
*** achanda_ has quit IRC17:33
*** sdake has quit IRC17:34
*** ctrath has joined #openstack-nova17:35
*** rajesht has joined #openstack-nova17:35
*** belmoreira has quit IRC17:38
vilobhmmjaypipes : ping17:38
*** yassine has quit IRC17:39
openstackgerritDavanum Srinivas (dims) proposed openstack/nova-specs: Allow list all instances for non-admin based on policy  https://review.openstack.org/19115117:39
*** rajesht has quit IRC17:39
dimsalex_xu: johnthetubaguy: thanks for the review, i added exactly what the current is and new behavior will be17:40
openstackgerritDan Smith proposed openstack/nova: Remove compat_instance()  https://review.openstack.org/19270817:40
openstackgerritDan Smith proposed openstack/nova: Fix testing object fields with missing instance rows  https://review.openstack.org/19280517:40
*** afazekas has quit IRC17:41
*** aix has quit IRC17:42
*** danpb has quit IRC17:42
openstackgerritWalter A. Boring IV (hemna) proposed openstack/nova: Switch to using os-brick  https://review.openstack.org/17556917:43
*** amotoki_ has quit IRC17:43
mriedemnova version 12 and keystone version 8 is going to be very confusing for awhile17:44
mgagneI think an opinion on https://github.com/openstack/nova/blob/master/nova/virt/libvirt/driver.py#L124817:44
mriedemcinder version 717:44
mgagneIf the revert fails, the user won't get an InterfaceAttachFailed exception, is it the desired behavior?17:45
mgagneAFAIK, he will get a 500 instead17:45
mriedemmgagne: what will they get? 500 for ProcessExecutionError?17:45
mriedemmgagne: put that in an excutils.save_and_reraise_exception context17:45
mriedemfor the unplug17:45
mriedemactually no, since that would raise the libvirterror17:46
mriedemjust a try/except should be good17:46
mgagneyes17:46
openstackgerritQin Zhao proposed openstack/nova: Support flavor object in migrate_disk_and_power_off  https://review.openstack.org/19278617:46
vilobhmmdansmith :  regarding https://review.openstack.org/#/c/190322/ need your insights on what will be a better approach for us going ahead….irc://irc.freenode.net:6667/#1. should the services information and the service liveliness information be stored in same datastore and let the common API fetch/update data from same source of truth OR irc://irc.freenode.net:6667/#2. keep the service information always in DB and service liveliness information in17:46
mriedemmgagne: remember to report a bug in case you want to backport that fix17:47
mgagnemriedem: should I log the original exception and just fallback to a common InterfaceAttachFailed?17:48
*** Longgeek has quit IRC17:48
mgagnelog (and mute)17:48
*** tdurakov_ has joined #openstack-nova17:50
*** Marga__ has joined #openstack-nova17:50
openstackgerritNicolas Simonds proposed openstack/nova: libvirt: Make the guest console configurable  https://review.openstack.org/12252317:51
*** kiranr has joined #openstack-nova17:51
openstackgerritNicolas Simonds proposed openstack/nova: imagebackend: Split out resize_image logic from create_image  https://review.openstack.org/18785717:51
openstackgerritNicolas Simonds proposed openstack/nova: Use utf8_bin collation on the flavor extra-specs table in MySQL  https://review.openstack.org/19030017:51
openstackgerritNicolas Simonds proposed openstack/nova: Fix/implement revert-resize for RBD-backed images  https://review.openstack.org/18739517:51
mgagnemriedem: it looks like it's already handled in the manager17:52
mgagnemriedem: https://git.openstack.org/cgit/openstack/nova/commit/?id=78f7c744f675d618048968850259422338dd495517:52
*** gabor-t has quit IRC17:53
*** achanda has joined #openstack-nova17:54
*** ijw_ has quit IRC17:54
*** electroc_ has quit IRC17:55
*** kiran-r has quit IRC17:55
openstackgerritDavanum Srinivas (dims) proposed openstack/nova-specs: Allow list all instances for non-admin based on policy  https://review.openstack.org/19115117:55
mriedemmgagne: oh yeah that was added relatively recently17:55
mgagneand it's part of kilo so I'm happy17:56
mriedemif you're happy then i'm happy17:56
*** tonytan4ever has joined #openstack-nova17:56
*** sambol has joined #openstack-nova17:56
*** electrocucaracha has joined #openstack-nova17:57
*** kiranr has quit IRC17:58
*** sdake has joined #openstack-nova17:58
*** Sukhdev has quit IRC17:58
*** Nikolay_St has joined #openstack-nova17:59
*** SourabhP has joined #openstack-nova18:00
*** browne has quit IRC18:01
*** Sukhdev has joined #openstack-nova18:01
*** sdake_ has quit IRC18:01
dimsso we begin the debate again - bug or feature - whether we need micro version :( https://review.openstack.org/#/c/191151/18:02
dansmithvilobhmm: okay18:02
*** Marga__ has quit IRC18:03
*** browne has joined #openstack-nova18:03
vilobhmmdansmith : thanks :)18:03
*** ociuhandu has quit IRC18:06
openstackgerritDavanum Srinivas (dims) proposed openstack/nova-specs: Allow list all instances for non-admin based on policy  https://review.openstack.org/19115118:06
*** annashen has joined #openstack-nova18:07
openstackgerritAnkit Agrawal proposed openstack/nova: Test patch race condition in delete (do not merge)  https://review.openstack.org/19173218:08
*** VW has joined #openstack-nova18:10
mriedemdims: i don't see that anyone asked for a microveresion18:10
mriedemdims: seems like a bug fix18:10
*** sdake_ has joined #openstack-nova18:10
*** IanGovett has joined #openstack-nova18:11
mriedemin general the code should work based on what's configured in the policy18:11
*** vladikr_ has quit IRC18:11
*** isd has joined #openstack-nova18:11
mriedemdims: i'd probably ask alex_xu and sdague18:12
dimsmriedem: didn't we say we need microversion in the ipv6 bug?18:12
*** vladikr_ has joined #openstack-nova18:12
mriedemyeah...18:12
mriedemso that's probably precedent i guess18:12
dimsright18:12
*** annashen has quit IRC18:13
mriedemi look forward to the day when i have a patchy gray beard and there is nova v2.43918:13
dimslol18:13
sdagueso a microversion would mean we would keep the old behavior regardless of policy18:13
*** sdake has quit IRC18:14
sdaguethis seems weird in that regard18:14
*** VW has quit IRC18:14
mriedemyeah, it seems as long as the policy is being enforced in the code we should be honoring it18:15
mriedembut i'm getting tripped up with the old @require_admin_context stuff that we kept broken in v2 api18:15
*** annashen has joined #openstack-nova18:15
mriedemand we didn't have microversions for those changes in the v2.1 api18:15
mriedemunless that's just, v2.1 itself18:16
*** agireud has joined #openstack-nova18:16
mgagneweighters or weights? :D18:17
*** ijw_ has joined #openstack-nova18:17
*** achanda has quit IRC18:17
mgagneI see both in the code18:18
*** ijw_ has quit IRC18:18
edmondswdims, I really think that should get fixed for all versions.18:18
*** ijw_ has joined #openstack-nova18:19
edmondswnot clear why anyone would need to be able to discover that they are talking to a fixed nova instance18:19
dimsedmondsw: happy to do the work. need consensus from the cores18:19
edmondswand if there is a good reason, then there should be some other way...18:19
edmondswsure, I get that... not a one man show here :)18:20
dimsedmondsw: one sec let me grab you the other review where the precedent was set18:20
dimsedmondsw: https://review.openstack.org/#/c/187835/18:20
edmondswrelated question... we're not going to spin off multiple microversions per release, are well? Is there not already a new microversion for Liberty?18:21
*** sambol has quit IRC18:21
edmondsws/well/we/18:21
mriedemedmondsw: yes there will be multiple microversions per release18:21
edmondswwhat's the point of that?18:21
mriedemthere were 3 in kilo, and already at least one new one in liberty18:21
*** cdent has quit IRC18:21
mriedemhttp://docs.openstack.org/developer/nova/api_microversion_history.html18:22
edmondswso kilo shipped with 2.1, 2.2, and 2.3?18:23
sdagueedmondsw: people deploy nova from source all the time18:23
sdaguefrom master18:23
sdaguethey don't wait for releases18:23
edmondswsdague, ugh!!!18:23
sdagueedmondsw: not ugh, that's been part of the model since day one18:23
edmondswthen they should pay the price in losing version discovery18:23
sdagueedmondsw: no, not at all, continuous delivery is a core value in Nova18:24
sdagueand most of the openstack projects18:24
*** markvoelker has quit IRC18:25
*** markvoelker has joined #openstack-nova18:25
mriedemi'm still not sure how this particular change relates to microversions since if the admin doesn't change the policy it's still not going to work for non-admins18:27
*** isq_ has quit IRC18:27
mriedemso you don't know if it works (discoverable) because the admin has not configured it or because the nova you're talking to doesn't have that microversion18:27
*** VW has joined #openstack-nova18:27
*** baoli has quit IRC18:28
*** baoli has joined #openstack-nova18:28
*** Marga_ has joined #openstack-nova18:29
dansmithedmondsw: no really, CDers should be the primary, and package-based people can deal with more funkiness :)18:30
sdaguemriedem: yeh, the fact that it's non default policy makes me feel like we probably don't need a bump for it18:31
edmondswsdague, ok... that's interesting... I'm all for continuous delivery, but I didn't think we were claiming that18:31
sdagueedmondsw: yes, since forever18:31
sdaguepublic clouds like rax are CD18:31
edmondswso why the convos about releasing more frequently?18:31
mriedemedmondsw: rax used to be like every 2 weeks or something18:31
mriedem2 weeks from trunk18:31
mriedemnow that's like 4 drops a year or something right?18:32
mriedemas was HP cloud18:32
*** VW has quit IRC18:32
* mriedem reboots18:32
*** mriedem has quit IRC18:32
edmondswI agree with what mriedem said about about microversions not applying in this case because of the need for a policy change as well18:33
*** VW has joined #openstack-nova18:33
sdagueif policy was discoverable, we'd need that, but right now that's all black box and not, so it's a shortcoming of our discovery18:34
*** penick has quit IRC18:34
*** agireud has quit IRC18:35
*** baoli has quit IRC18:35
melwittI think the only possible concern would be if someone has left the policy as default, which is allow all_tenants for everyone, and doesn't know it. current behavior only admin can do all_tenants, then if it changes, suddenly everyone can do it18:36
*** penick has joined #openstack-nova18:37
openstackgerritDan Smith proposed openstack/nova: Remove compat_instance()  https://review.openstack.org/19270818:37
edmondswmelwitt, that would be a deployer's mistake... I wouldn't expect our code to have to account for that18:37
*** mingy has joined #openstack-nova18:37
sdaguemelwitt: the default policy blocks this as well though, right?18:40
*** Marga_ has quit IRC18:40
openstackgerritNicolas Simonds proposed openstack/nova: libvirt: Don't fetch kernel/ramdisk files if you already have them.  https://review.openstack.org/16366118:41
melwittedmondsw: yeah. I admit I get a little confused about when it's okay to fix wrong behavior without a microversion18:41
*** afazekas has joined #openstack-nova18:41
*** ZZelle_ has joined #openstack-nova18:41
melwittsdague: I guess I'm not sure what default policy means. our policy.json in tree has "compute:get_all_tenants": "", which means all can do18:42
edmondswdefault policy has all kinds of security issues... hopefully nobody who cares about security actually uses those18:43
openstackgerritHans Lindgren proposed openstack/nova: Drop scheduler RPC 3.x support  https://review.openstack.org/19283618:43
melwittedmondsw: fair point18:44
edmondswsadly...18:44
sdaguemelwitt: so... that's a good point. dims can you explain the change a bit more? I'm confused now.18:45
sdagueedmondsw: lots of people use default policy, if there are issues we should tighten up, we should do that.18:46
sdaguemorganfainberg told me they had an ops session on policy, and in the room the number of folks that had successfully modified their policy without breaking the world was a small handful. People assume we're sending them sane defaults.18:47
edmondswyes, we should clean that up18:47
*** markvoelker_ has joined #openstack-nova18:47
*** markvoelker has quit IRC18:47
edmondswI'm trying to find time to work on that18:47
morganfainbergsdague: 2 or 3 people raised their hands.18:48
morganfainbergNot even a handful.18:48
dimssdague: see old and new behavior lines 68-77 - https://review.openstack.org/#/c/191151/6/specs/liberty/approved/allow-list-all-instances-for-non-admin.rst,cm18:48
edmondswI'd have raised mine if I'd been there18:48
sdagueok, less than a handful. And one of them was Nate from the NSA.18:48
morganfainbergYep.18:48
*** nihilifer has quit IRC18:48
*** Longgeek has joined #openstack-nova18:48
dansmithsdague: important qualifier :D18:48
morganfainbergAnd just 403ing a single api call isn't what we were asking. But I'm sure the hands would have been the same.18:49
morganfainbergEven just looking at blocking a single api call.18:49
edmondswthat is extremely sad18:50
*** mtecer has joined #openstack-nova18:50
dimssdague: i'll be happy not to do a microversion for this :)18:50
morganfainbergIt's a sign we need to make default policy a rich / usable set of definitions.18:50
morganfainbergAnd not assume everyone will customize it.18:51
*** ildikov has joined #openstack-nova18:51
edmondswmorganfainberg, the starting point should have been that policy doesn't allow much, and you have to open things up if you need them18:52
samueldmqmorganfainberg, ++ with different meanings to different admins (cloud, domain and project) :)18:52
edmondswinstead we've left so much open, that people would need to customize to close things off18:52
edmondswbut in either event, people will need to customize.. we can't have defaults that cover every environment18:53
*** Longgeek has quit IRC18:53
morganfainbergedmondsw: I am not making that judgement call. I'm just saying we need a better set of defaults because clearly most deployments use the defaults. And updates from defaults need to be better.18:53
sdaguedims: ok, so I -1ed it because I think we need to talk about the security impact, because there is one for existing default install envs18:53
sdaguebut once you get that updated, I don't think it needs a microversion18:53
edmondswif we don't need a microversion, we don't even need a spec, do we?18:54
*** mriedem has joined #openstack-nova18:54
*** ociuhandu has joined #openstack-nova18:54
dimssdague: default policy allows these calls, but does not return the right result...18:55
sdaguedims: right18:55
*** oro_ has quit IRC18:55
sdaguebut that combo of behavior is what people expect18:55
sdaguewe're blocking all_tenants in code18:55
dimsshould i be fixing the policy too?18:55
sdaguedims: absolutely18:55
sdagueotherwise you just openned a new security hole for people18:56
dimsso "is_admin:True"18:56
sdagueyep18:56
*** jog0 has joined #openstack-nova18:56
*** saza has joined #openstack-nova18:56
sazahello18:56
sdagueand because policy isn't code, I think UpgradeImpact SecurityImpact should be used on the patch, so people know they should make a change at the same time to policy18:56
sazanova network18:56
dimssdague: still need a spec and blueprint?18:57
*** sudipto has quit IRC18:57
*** sahid has joined #openstack-nova18:57
sdaguedims: you've already got one, I'd say yes. You can ask johnthetubaguy in the morning what he thinks.18:57
*** krtaylor has quit IRC18:57
sazai have question nova-network18:57
dimssdague: ack thanks18:58
sazacan i ask querstion on this channel ?18:58
mriedemsaza: just ask18:58
samueldmqsdague, does not restrict operations on a project to tokens scoped to that specific project ? for any calls18:58
samueldmqsdague, https://github.com/openstack/nova/blob/master/nova/api/openstack/wsgi.py#L804-L81218:58
samueldmqs/not/nova18:58
mriedemsaza: if it's a support question, see topic18:58
saza br100:  172.16.0.254/2418:58
sazanova flat network18:58
sazathis gateway for subnet.18:59
sazaon compute note.18:59
sazaon vm can ping this ip <172.16.0.254>18:59
sdaguesamueldmq: so that's a little different, that's asking specifically for a different project than yours18:59
*** Marga_ has joined #openstack-nova18:59
sdaguebut all_tenants wouldn't hit that check18:59
sazabut extenal  network can't ping this ip .19:00
sdaguebecause you didn't set project_id to anything19:00
mriedemsaza: you're asking a support question, that's not for this channel, which is a dev channel19:00
samueldmqsdague, k .. I was wondering in the update quota case19:00
mriedemsaza: you want the #openstack channel or ask.openstack.org19:00
mriedemsaza: or the general openstack mailing list19:00
sazabut external  network ( ex. 172.16.0.100  can ping  vm  (19:00
samueldmqsdague, where someone updating a givne project's quota would be a domain admin (or someone in a higher project in the hierarhcy)19:00
sazathx19:00
sdaguesamueldmq: yeh, I don't know how all this works in a domains context19:01
*** VW has quit IRC19:01
*** ctrath has quit IRC19:01
edmondswsamueldmq, I believe the domain admin would need to get a project-scoped token to update that project's quotas19:02
*** mtecer has quit IRC19:02
*** saza has left #openstack-nova19:02
*** otter768 has joined #openstack-nova19:02
edmondswdomain-scoped tokens aren't good for much19:02
samueldmqedmondsw, the domain admin may even not have a role assigned on that project19:02
edmondswsamueldmq, then they're out of luck19:03
samueldmqedmondsw, domains are about identity, managing users and groups separately, and contain projects as well19:03
samueldmqedmondsw, I mean, he could be able to delegate quotas, but couldn't list servers, for ex19:03
edmondswsamueldmq, yes, I'm familiar... more familiar with keystone than with nova19:03
samueldmqedmondsw, nice19:03
edmondswI didn't say I like the way domain tokens work.. ;)19:04
samueldmqedmondsw, tokens behavior is controlled depending on the scope and roles19:05
samueldmqedmondsw, but all depends on your policy19:05
samueldmqedmondsw, RBAC :)19:05
edmondswsamueldmq, well... yes and no. I wish that were really true19:05
samueldmqedmondsw, why aren't they?19:06
*** tojuvone has quit IRC19:06
edmondswbecause there is too much that is coded in a way that doesn't allow the policy to really control it to that extent19:06
*** otter768 has quit IRC19:07
samueldmqedmondsw, we're improving that19:09
edmondswgood!19:09
samueldmqedmondsw, dynamic policies is coming ot start improving our access control mechanism19:09
samueldmqedmondsw, https://wiki.openstack.org/wiki/DynamicPolicies19:09
samueldmqsdague, cc ^19:09
*** Sukhdev_ has joined #openstack-nova19:10
edmondswsamueldmq, I'm familiar with the dynamic policy work... I think it's actually putting the cart before the horse...19:10
samueldmqedmondsw, sdague that's still wip, we're still working with people to get the right solution19:10
edmondswwe need to fix all these policy issues throughout our code and then worry about how to configure policy dynamically19:10
samueldmqsdague, I am considerign nova requirements in my first try in the 'Overview Solution - Liberty Scope' section19:10
samueldmqsdague, where warns are logged in the case someone change policy from defaults19:11
*** cfriesen has quit IRC19:11
samueldmqedmondsw, what are the issues throughout our code? give me an example ?19:11
samueldmqnot sure I follow19:12
*** fifieldt_ has quit IRC19:12
*** Sukhdev has quit IRC19:13
edmondswe.g., nova and cinder have a single policy for a class of things in some places, rather than for the individual CRUD actions19:13
*** IanGovett has quit IRC19:13
edmondswceilometer has policy hardcoded for probably half of their APIs19:13
edmondswetc.19:13
*** ijw_ has quit IRC19:14
edmondswthere are a lot of issues around ownership... where policy really needs to allow you to specify that only the owner of the element can do something, but the service isn't recording ownership for that element19:14
edmondswetc.19:14
*** rook-spec is now known as rook19:15
*** openstackgerrit has quit IRC19:16
samueldmqedmondsw, we're improving the mechanisms around our curent policy.json, if people don't respect that policy that's a separate issue19:16
samueldmqedmondsw, we will educate people to get better default policies19:17
samueldmqedmondsw, which includes going away from hard-coded19:17
*** openstackgerrit has joined #openstack-nova19:17
edmondswdon't get me wrong, though... I will be glad to see dynamic policy19:17
edmondswand maybe working on it will bring more attention to some of these other things19:17
samueldmqedmondsw, sure, no worries .. you're more than welcome to come and help, everyone is19:18
*** matrohon has joined #openstack-nova19:19
samueldmqedmondsw, I am working now on having some agreements internally on keystone, will be sendng a message in the ML this week19:19
edmondswsamueldmq, yep, I believe I added some comments to that spec the other day19:19
samueldmqedmondsw, we need to involve other projects on this (nova started) .. we need a weekly meeting for this subject19:19
samueldmqedmondsw, I am trying to run with those things .. the clock is ticking, and we plan to get something valuable in Liberty19:20
*** afazekas has quit IRC19:21
*** ctrath has joined #openstack-nova19:22
*** nihilifer has joined #openstack-nova19:23
*** mtecer has joined #openstack-nova19:27
mriedemdhellmann: what are your thoughts on all of the deprecation warning code in nova that is referring to 2015.2 and 2016.1 for when things are removed, as ericwb points out in https://review.openstack.org/#/c/192406/ ?19:27
mriedemsince i have one of those proposed here https://review.openstack.org/#/c/192753/1/nova/virt/vmwareapi/driver.py19:27
dhellmannmriedem: I suppose those should be updated?19:28
mriedemdhellmann: yeah, but in your change?19:28
mriedemi'd think yes19:28
*** krtaylor has joined #openstack-nova19:28
mriedemand 2016.1 M will be 13.0.0 M right?19:28
mriedemmaybe not even called M anymore? i don't know the latest on wtf is going on.19:29
*** tjones1 has quit IRC19:29
dhellmannmriedem: I will happily rebase my change on top of that, but I really can't do all of that for you guys.19:29
*** agireud has joined #openstack-nova19:29
mriedemdhellmann: yeah, understand, since you have 50 of these cagts19:29
mriedem*cats19:29
dhellmannright19:29
ijuwangDo we open a bug for implemented spec that contains api that's not implemented? For example, http://specs.openstack.org/openstack/nova-specs/specs/juno/implemented/user-defined-shutdown.html19:29
*** fifieldt_ has joined #openstack-nova19:30
dhellmannmriedem: we have some similar issues in heat, fwiw19:30
mriedemijuwang: if anything it'd just be a spec update probably19:30
mriedemdhellmann: yeah, and novaclient19:30
mriedemhttps://github.com/openstack/python-novaclient/blob/master/novaclient/v2/shell.py#L6619:31
*** Marga_ has quit IRC19:31
mriedemok....19:31
mriedemso maybe i'll just take one for the team here and update that19:31
*** Marga_ has joined #openstack-nova19:31
dhellmannmriedem: that's very sporting of you19:32
*** Mike_D_laptop has quit IRC19:32
mriedemctrath: if you could legally push code this would be a good one for you :)19:32
*** Nikolay_St has quit IRC19:32
ijuwangmriedem: ok, thanks19:34
sahidmriedem: hello, we have a backport ready for kilo, if you are happy with it https://review.openstack.org/#/c/189350/19:37
*** electrocucaracha has quit IRC19:37
mriedemsahid: i saw that before, i wanted to dig into the original change on master to make sure i understood it first19:37
mriedemsince that seems too easy19:37
sahidmriedem: ah ok19:38
sahidmriedem: please let me know of the review if you need any information19:38
*** arnaud___ has joined #openstack-nova19:39
openstackgerritMatt Riedemann proposed openstack/nova: cells: remove deprecated mute_weight_value option  https://review.openstack.org/19285319:39
sahidO_o - what i wanted to say...19:39
mriedemi understood :)19:39
mriedemlet you know if i have questoins19:39
*** electrocucaracha has joined #openstack-nova19:40
*** Marga_ has quit IRC19:40
mriedemalaski: dansmith: sdague: ^ removing that option will help with the semver change patch19:40
sahidahah cool thank you19:40
*** thedodd has joined #openstack-nova19:40
dansmithbauzas: jaypipes: https://review.openstack.org/#/c/19270819:41
*** isd has quit IRC19:42
dansmithI think I'm clear to do that ^^ based on the justification in the commit message19:42
*** VW has joined #openstack-nova19:42
dansmithbut would like you guys to check my work :)19:42
*** IanGovett has joined #openstack-nova19:42
bauzasdansmith: looking19:43
*** arnaud___ has quit IRC19:43
*** electrocucaracha has quit IRC19:44
*** beagles is now known as beagles|afk19:45
*** agireud has quit IRC19:45
*** nkrinner has joined #openstack-nova19:46
*** agireud has joined #openstack-nova19:48
*** Longgeek has joined #openstack-nova19:49
dansmithbauzas: no rush, it's at the end of a set19:51
edleafedansmith: in https://review.openstack.org/#/c/192708/3/nova/compute/resource_tracker.py,cm L913, should that parameter still be 'object_or_dict'?19:51
edleafedansmith: or should it always be an object at this point?19:51
*** penick has quit IRC19:51
dansmithedleafe: not changed because of this patch19:52
dansmithand I think it's still valid anyway19:52
edleafeso you could still pass an instance dict?19:52
dansmithedleafe: one could, yes19:53
dansmithwhether that can still happen I'm not sure, but I haven't seen anything lately that would change that19:53
dansmithI think this has to do with the fact that we take in an object, and convert it to a dict, and some things will re-pass that dict to us19:53
edleafedansmith: ok, thanks. I wasn't sure that obj_to_primitive() would accept a dict19:54
*** nihilifer_ has joined #openstack-nova19:54
dansmithedleafe: we won't call obj_to_primitive() with a dict in this code anyway19:54
dansmithbecause of the isinstance()19:54
edleafedansmith: yep, got it19:55
bauzasedleafe: the main problem is that Instance or Flavor can be either a dict or not, depending on the caller19:55
bauzasedleafe: just because sometimes the instance is primitived before that19:55
*** isd has joined #openstack-nova19:56
bauzasedleafe: so there are many possibilities where it can be different, hence nikola trying to provide that helper method for managing all the cases19:56
edleafebauzas: yeah, like dansmith pointed out, I missed the isinstance() qualifier19:56
*** nihilifer has quit IRC19:56
bauzasedleafe: yeah, I just tried to explain why we need this helper19:57
*** exploreshaifali has quit IRC19:57
bauzasjust because we need to fix all the callers before removing that19:57
*** baoli has joined #openstack-nova19:57
*** penick has joined #openstack-nova19:58
*** penick has quit IRC19:58
*** Nikolay_St has joined #openstack-nova19:58
bauzasdansmith: so, I maybe have a question https://review.openstack.org/#/c/192708/3/nova/scheduler/utils.py,cm19:58
dansmithbauzas: that's the bit I figured would be concerning, yes :D19:58
bauzasdansmith: build_request_spec is for calling the Scheduler RPC API, but we didn't removed the Scheduler RPC v3 yet19:59
dansmithum19:59
dansmithhmm19:59
dansmithI was thinking this went to compute for build_and_run_instances(), and it doesn't use anything there19:59
bauzasdansmith: so, I don't think it's really a problem, because the instance is mainly consumed at the right named consume_from_instance() method19:59
bauzasdansmith: nope, it goes thru select_dests() too20:00
*** electrocucaracha has joined #openstack-nova20:00
dansmithdoes anything actually use the flavor stuff in sysmeta there?20:00
*** tonytan4ever has quit IRC20:00
bauzasdansmith: that's my point20:00
bauzasdansmith: I don't think so20:00
dansmithhmm, okay20:01
dansmithif it doesn't we could either (a) just do this or (b) put it back in just in this location I guess20:01
bauzasdansmith: I mean, I'm 99% sure about that since I'm currently working on the HostManager atm :D20:01
dansmithokay20:01
bauzasdansmith: but I'm not sure the instance could be checked for the flavor stuff in a filter, I need to check that20:01
*** penick has joined #openstack-nova20:02
bauzasdansmith: anyway, all of that build_req_spec() will be doomed once someone does his job20:02
*** tjones1 has joined #openstack-nova20:02
dansmithbauzas: okay20:03
bauzasdansmith: but we still need to care about that given we would still leave the old select_dest() method for the compat20:03
bauzasso, checking now who calls instance_props in the scheduler, and if the flavor stuff is checked then20:03
dansmithbauzas: so what we want to avoid is these things being in the instance_properties and then ending up creating instances with that in sysmeta20:04
*** neophy has quit IRC20:04
bauzasdansmith: as the scheduler doesn't save the instances, we're fine IMHO20:04
dansmithif we can reliably add it into scheduler/utils and then make sure it is removed before we create the instance in the db, then we can do that  until your object work is done20:04
dansmithbauzas: well, it passes those bits along to someone that does (conductor) right?20:04
bauzasdansmith: the conductor calls build_req_spec() and then select_dest()20:05
bauzasdansmith: but the spec is not then used20:05
bauzasI mean, after the scheduler call20:05
*** Marga_ has joined #openstack-nova20:06
dansmithwell, whatever creates the instance does it from instance_properties, as a template, right?20:06
bauzasdansmith: which is different from filter_properties, which is passed to the compute IIRC20:06
*** Longgeek has quit IRC20:06
bauzasdansmith: the instance is created before the sched call, right?20:06
alaskithe instance is created in the compute api20:06
bauzasalaski: yeah, my point20:07
*** tonytan4ever has joined #openstack-nova20:07
alaskifor now...20:07
dansmithbauzas: ah, okay20:07
bauzasalaski: don't freak me out :)20:07
dansmithhere is where I insert a hilarious "gosh I should know this" comment where we all laugh about how it's nearly impossible to follow all this madness20:07
*** nic has quit IRC20:08
bauzasdansmith: and wait for cells v220:09
bauzas... :)20:09
alaskiit's pretty crazy20:09
bauzasthe only reasonable reason I'm part of cells v2 is to lurk at the design discussions20:09
*** baoli has quit IRC20:10
bauzasdansmith: so, given that, I'm pretty okay with your change20:10
bauzasdansmith: there is no flavor-ish thing needed20:11
*** baoli has joined #openstack-nova20:11
bauzasI mean by the scheudler20:11
*** thiagop has joined #openstack-nova20:11
*** tdurakov_ has quit IRC20:12
dansmithalaski: bauzas: okay, so I'll slide the crap back into that one call, and make sure that it doesn't end up with stuff back in sysmeta in the actual cleaned instances20:12
bauzasdansmith: that's pretty isolated on the scheduler side20:13
*** thiagop has left #openstack-nova20:13
bauzasdansmith: I mean, what's given by that utils.build_req_spec() method goes to the scheduler as a dead-end, just for consuming fields20:14
*** jog0 has quit IRC20:14
dansmithyeah, but if filters could be looking at it...20:14
bauzasdansmith: and if you see the RequestSpec object that I modeled, I only provided some instance fields20:14
bauzasdansmith: not the whole Instance object20:15
dansmithright20:15
*** artom has quit IRC20:15
bauzasdansmith: I checked that, and no filters are looking at it20:15
bauzasdansmith: and if I had found one, I would have slapfaced me20:15
*** artom has joined #openstack-nova20:15
bauzasdansmith: just because I made that review earlier on the spec writing when I exploded the Instance stuff20:16
*** dane-fichter has quit IRC20:16
dansmithbauzas: well, in-tree filters anyway20:17
bauzasdansmith: right20:17
dansmithif it's easy for me to add it in without any other work, I will20:17
*** artom has quit IRC20:17
dansmithif it starts getting difficult, I'll recant :)20:17
*** artom has joined #openstack-nova20:17
bauzasdansmith: but since we're working on providing a clear API for what's going to the scheduler, out-of-tree filters are also impacted anyway20:18
dansmithyeah20:18
bauzasjust because it won't be that easier to pass something unversioned to the sched now20:18
bauzaseasy*20:18
*** achanda has joined #openstack-nova20:19
*** artom has quit IRC20:19
*** artom has joined #openstack-nova20:20
*** jog0 has joined #openstack-nova20:20
*** dpaterson has quit IRC20:20
mriedemctrath: this is that bug i was talking about https://bugs.launchpad.net/nova/+bug/1186354 - i think we agreed that if you're hitting those quota APIs and you're using neutron for networking, just fail since it's not going to be correct and we don't want to add the proxy code, even though we could20:22
openstackLaunchpad bug 1186354 in OpenStack Compute (nova) "Limits API doesn't work with Neutron" [Medium,Confirmed]20:22
mriedemshould maybe bring that one up in the ML first20:22
*** achanda_ has joined #openstack-nova20:23
*** mleroy_ has joined #openstack-nova20:23
mriedemthis was the original change way back in havana i think to proxy the calls to neutron, so some form of that could be resurrected if we wanted to proxy https://review.openstack.org/#/c/43822/20:23
*** thangp_ has quit IRC20:24
mriedemalthough http://docs.openstack.org/developer/nova/project_scope.html?highlight=proxy#no-more-api-proxies20:24
openstackgerritAndrew Laski proposed openstack/nova: Cells: cell scheduler anti-affinity filter  https://review.openstack.org/19287520:24
*** whenry has joined #openstack-nova20:26
*** achanda has quit IRC20:26
*** ivasev has quit IRC20:26
*** thangp has joined #openstack-nova20:27
*** VW has quit IRC20:28
*** lpetrut has quit IRC20:29
*** ctrath has quit IRC20:29
*** ctrath has joined #openstack-nova20:32
*** mtecer has quit IRC20:33
*** cfriesen has joined #openstack-nova20:34
*** mwagner_afk has quit IRC20:34
*** sahid has quit IRC20:35
openstackgerritMing Yang proposed openstack/nova: Refactor get_filtered_hosts in filter host manager  https://review.openstack.org/19287820:37
mingy(I'm just doing this while I'm reading the code... it helps me understand what's going on. I hope no one minds.)20:37
*** comstud has joined #openstack-nova20:37
*** achanda_ has quit IRC20:38
*** baoli has quit IRC20:39
*** whenry has quit IRC20:39
*** baoli has joined #openstack-nova20:40
*** vladikr_ has quit IRC20:41
*** tlbr has quit IRC20:47
*** tlbr has joined #openstack-nova20:47
openstackgerritWalter A. Boring IV (hemna) proposed openstack/nova: Switch to using os-brick  https://review.openstack.org/17556920:47
*** READ10 has quit IRC20:47
*** lan has quit IRC20:47
openstackgerritQin Zhao proposed openstack/nova: Support flavor object in migrate_disk_and_power_off  https://review.openstack.org/19278620:48
openstackgerritMatt Riedemann proposed openstack/nova: Update version for Liberty  https://review.openstack.org/19240620:50
openstackgerritMatt Riedemann proposed openstack/nova: Remove hv_type translation shim for powervm  https://review.openstack.org/19288720:50
openstackgerritMatt Riedemann proposed openstack/nova: neutron: remove deprecated allow_duplicate_networks config option  https://review.openstack.org/19288820:50
*** lan has joined #openstack-nova20:52
*** josecastroleon1 has quit IRC20:53
*** yamahata has joined #openstack-nova20:54
SourabhPbeagles|afk: ping20:54
*** josecastroleon has joined #openstack-nova20:54
*** jskinner has quit IRC20:54
*** haypo has joined #openstack-nova20:54
bauzasdansmith: quick question, what would be better than a StringField for a JSON property, saying that I can't make sure that it would be a DictOfStrings or rather a DictofDicts etc.20:58
bauzas?20:58
*** artom has quit IRC20:58
*** nkrinner has quit IRC20:58
*** penick has quit IRC20:59
openstackgerritClaudiu Belu proposed openstack/nova-specs: Adds Hyper-V Shielded VMs spec  https://review.openstack.org/19289220:59
bauzasdansmith: see https://github.com/openstack/nova/blob/master/nova/scheduler/filters/json_filter.py#L143 for the question20:59
*** dulek_home has joined #openstack-nova21:00
*** dulek_home has quit IRC21:01
*** dprince has quit IRC21:01
*** dulek has joined #openstack-nova21:02
mriedemdhellmann: i updated https://review.openstack.org/#/c/192406/ but does that not work until we actually tag the 12.0.0 release for liberty?21:02
*** agireud has quit IRC21:02
*** otter768 has joined #openstack-nova21:03
*** annashen has quit IRC21:05
*** edmondsw has quit IRC21:06
*** electrocucaracha has quit IRC21:07
openstackgerritPhilip Marc Schwartz proposed openstack/nova: Add flag to force experimental run of db contract  https://review.openstack.org/19289421:07
*** otter768 has quit IRC21:08
*** electrocucaracha has joined #openstack-nova21:09
*** mleroy_ has quit IRC21:09
*** Sukhdev_ has quit IRC21:10
*** penick has joined #openstack-nova21:10
*** haomaiwang has joined #openstack-nova21:13
*** READ10 has joined #openstack-nova21:14
*** haomaiw__ has quit IRC21:15
*** dulek has quit IRC21:18
*** atuvenie has quit IRC21:20
openstackgerritDavanum Srinivas (dims) proposed openstack/nova: [WIP] Kicking the oslo.service tires  https://review.openstack.org/19290021:20
*** isd has quit IRC21:20
*** haomaiw__ has joined #openstack-nova21:21
dansmithbauzas: well, anything other than a JSONfield would be better21:21
dansmithbauzas: like: GiraffeField21:21
*** haomaiwang has quit IRC21:21
*** mwagner_afk has joined #openstack-nova21:22
mriedemGarField?21:23
*** Marga_ has quit IRC21:24
dansmithbah dum pish21:24
*** openstackgerrit has quit IRC21:24
mriedemi'm actually here all week21:24
dansmithheh21:24
mriedemi guess openstackgerrit couldn't handle that level of pun21:24
*** Marga_ has joined #openstack-nova21:24
*** openstackgerrit has joined #openstack-nova21:24
dimsLol21:25
*** dulek_home has joined #openstack-nova21:25
*** thangp has quit IRC21:25
*** sdake_ is now known as sdake21:25
*** dulek_home has quit IRC21:25
*** dulek has joined #openstack-nova21:26
*** nihilifer_ is now known as nihilifer21:26
openstackgerritMatt Riedemann proposed openstack/python-novaclient: Change future nova version number references based on new values  https://review.openstack.org/19290221:27
*** erlon has quit IRC21:30
openstackgerritMatt Riedemann proposed openstack/python-novaclient: Change future nova version number references based on new values  https://review.openstack.org/19290321:30
*** VW has joined #openstack-nova21:32
*** angdraug has joined #openstack-nova21:32
*** penick has quit IRC21:33
*** Marga_ has quit IRC21:33
*** Marga_ has joined #openstack-nova21:34
*** achanda has joined #openstack-nova21:36
*** tonytan4ever has quit IRC21:36
openstackgerritMatt Riedemann proposed openstack/nova: vmware: update help text for vmware.cluster_name option  https://review.openstack.org/19275321:37
*** baoli_ has joined #openstack-nova21:37
*** yamahata has quit IRC21:37
*** radek_ has quit IRC21:38
*** baoli has quit IRC21:40
mriedemSourabhP: i think you're looking for this https://review.openstack.org/#/c/162468/21:40
*** jlk has joined #openstack-nova21:40
jlkQuick question, can somebody help me find in the nova code where a hypervisor's hostname is determined for the sake of creating the service entry, and whether or not that can be configured in nova's config file?21:41
openstackgerritPhilip Marc Schwartz proposed openstack/nova: Add flag to force experimental run of db contract  https://review.openstack.org/19289421:41
mriedemjlk: isn't that my_host in nova.conf?21:41
mriedemsorry my_ip21:42
jlkah, I knew about my_ip, didn't about my_host21:42
jlkno, my_ip is a bit different.21:42
mriedemb/c i made it up21:42
SourabhPmriedem: thanks for the pointer!21:42
jlkthat's about how to /reach/ the host, for the sake of things like migration21:42
*** achanda has quit IRC21:42
jlkbut I'm more concerned about what hostname gets written into the service table when nova-compute comes online21:42
*** annashen has joined #openstack-nova21:42
*** alex_klimov has joined #openstack-nova21:43
jlkor if it's hard set to whatever it can resolve from the system hostname21:43
*** matrohon has quit IRC21:44
mriedemjlk: should be able to trace that to whatever is using the resource tracker in the compute manager21:44
jlkah, maybe it's just host=21:44
jlknicely undocumented :)21:46
mriedempatches welcome21:46
mriedem        if not host:21:47
mriedem            host = CONF.host21:47
mriedemthere you go21:47
dansmithtweets unwelcome21:47
jlkn/m, i tis documented21:47
mriedemnova.service.Service.create21:47
mriedemthe nova.cmd.compute entry point creates the compute service and doesn't pass a host, so that comes from CONF.host21:48
jlkI think I failed ot find it because I was looking for "hostname"21:48
mriedemhost = localhost(StrOpt) Name of this node. This can be an opaque identifier. It is not necessarily a hostname, FQDN, or IP address. However, the node name must be valid within an AMQP key, and if using ZeroMQ, a valid hostname, FQDN, or IP address21:48
jlkyup, thanks, I've found the bits I was looking for. Always a pleasure!21:48
mriedemhttp://docs.openstack.org/kilo/config-reference/content/list-of-compute-config-options.html21:48
mriedem\o/21:48
mriedemon that i can leave for the day21:48
*** tjones1 has quit IRC21:51
mriedemctrath: you are now the proud owner of https://bugs.launchpad.net/nova/+bug/118635421:52
openstackLaunchpad bug 1186354 in OpenStack Compute (nova) "Limits API doesn't work with Neutron" [Medium,Confirmed] - Assigned to Cale Rath (ctrath)21:52
mriedemctrath: you should probably also join the nova bug team if you want to triage stuff https://launchpad.net/~nova-bugs21:53
*** agireud has joined #openstack-nova21:53
mriedems/want/need/21:53
*** ianbrown has joined #openstack-nova21:53
mriedemjust dont sign up to get bug email for all new bugs21:53
*** xyang2 has quit IRC21:57
*** ZZelle_ has quit IRC21:57
*** mriedem has quit IRC21:58
*** jog0 is now known as jogo21:58
*** burt has quit IRC22:00
*** jlk has left #openstack-nova22:00
*** thedodd has quit IRC22:01
*** Longgeek has joined #openstack-nova22:03
*** jogo has quit IRC22:03
*** wverdugo5001 has quit IRC22:06
*** jgrimm is now known as zz_jgrimm22:07
*** mdrabe has quit IRC22:11
*** VW has quit IRC22:11
*** alex_klimov has quit IRC22:12
*** VW has joined #openstack-nova22:13
*** agireud has quit IRC22:14
*** baoli_ has quit IRC22:22
rookwrt Kilo & SRIOV - should the databse be populated with the pci information? and if so, what table?22:23
rookdatabase*22:23
*** haypo has quit IRC22:24
*** penick has joined #openstack-nova22:24
*** dims has quit IRC22:26
*** jecarey has quit IRC22:30
*** baoli has joined #openstack-nova22:31
*** Longgeek has quit IRC22:36
*** ijuwang has quit IRC22:36
*** baoli has quit IRC22:38
*** kylek3h has quit IRC22:42
*** aysyanne has quit IRC22:42
*** Nikolay_St has quit IRC22:42
*** krtaylor has quit IRC22:44
*** zzzeek has quit IRC22:44
*** jasondotstar has joined #openstack-nova22:45
*** IanGovett has quit IRC22:45
*** josecastroleon has quit IRC22:47
*** josecastroleon has joined #openstack-nova22:48
*** ctrath has quit IRC22:49
*** ociuhandu has quit IRC22:49
*** Marga_ has quit IRC22:49
*** ctrath has joined #openstack-nova22:50
*** Marga_ has joined #openstack-nova22:50
*** oro_ has joined #openstack-nova22:51
*** asselin has joined #openstack-nova22:55
*** dboik_ has quit IRC22:56
*** IanGovett has joined #openstack-nova22:56
*** dboik has joined #openstack-nova22:56
*** ctrath has quit IRC22:56
*** _diana_ has joined #openstack-nova22:59
*** tpeoples has quit IRC22:59
*** nihilifer has quit IRC23:01
*** oro_ has quit IRC23:02
*** otter768 has joined #openstack-nova23:04
*** IanGovett has quit IRC23:06
*** Marga_ has quit IRC23:07
*** otter768 has quit IRC23:08
_diana_Hi folks, I have what is hopefully an easy PCI virtual function vs physical function question. Does anyone have a couple of minutes to explain why is_physical_function returns true when it matches a virtual function? https://github.com/openstack/nova/blob/master/nova/pci/utils.py#L97 I would have assumed that a virtual function match would mean that is_physical_function would return false, but I've obviously m23:09
*** tpeoples has joined #openstack-nova23:09
*** obondarev has quit IRC23:11
*** obondarev has joined #openstack-nova23:12
*** tjones1 has joined #openstack-nova23:13
*** yamahata has joined #openstack-nova23:15
*** puniuee has joined #openstack-nova23:15
puniuee@all I need help in figuring out something23:16
puniueetwo of my nova.tests.unit.compute.test_compute_cells.CellsComputeAPITestCase are failing23:16
puniueei am stuck on this from 2 days and I don't know, how the mock calls and stub calls actually work23:17
*** park_ has joined #openstack-nova23:22
*** dims has joined #openstack-nova23:24
*** zzzeek has joined #openstack-nova23:25
*** Marga_ has joined #openstack-nova23:27
*** VW has quit IRC23:28
*** park_ has quit IRC23:28
*** dims has quit IRC23:28
melwittpuniuee: I recommend looking through the docs to learn how they work https://code.google.com/p/pymox/wiki/MoxDocumentation or https://mock.readthedocs.org/en/latest/index.html depending on whether the tests in question are using mox or mock23:28
*** dims has joined #openstack-nova23:28
*** park_ has joined #openstack-nova23:29
*** krtaylor has joined #openstack-nova23:32
*** dims has quit IRC23:33
*** claudiub has quit IRC23:34
puniueethanks @melwitt23:35
*** yamahata has quit IRC23:40
*** mriedem has joined #openstack-nova23:42
*** diogogmt has quit IRC23:45
*** mriedem1 has joined #openstack-nova23:45
*** _diana_ has quit IRC23:45
*** jskinner has joined #openstack-nova23:46
*** mriedem has quit IRC23:47
*** lucas-dinner has quit IRC23:51
puniueeplease someone also tell me an example of usage of NoopQuotaDriver23:51
puniueeit says there" with nova compute  cells, the parent cell should do quota checking, but the child cell should not"23:51
*** park_ has quit IRC23:52
*** otter768 has joined #openstack-nova23:53
puniueei mean a flow where parent cell checks the quota and child does not23:57
*** pixelbeat_ has quit IRC23:59
*** sdake_ has joined #openstack-nova23:59
*** oomichi has joined #openstack-nova23:59

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