Thursday, 2016-05-19

*** zenoway has quit IRC00:00
*** pushkaru has quit IRC00:01
*** shaohe_feng has quit IRC00:03
*** jamielennox is now known as jamielennox|away00:04
*** shaohe_feng has joined #openstack-nova00:04
*** jamielennox|away is now known as jamielennox00:04
*** hieulq_ has joined #openstack-nova00:06
*** winston-d_ has joined #openstack-nova00:06
*** AmyLee7 has quit IRC00:07
*** harlowja_ has joined #openstack-nova00:11
*** sdake has quit IRC00:13
*** shaohe_feng has quit IRC00:13
*** markvoelker has quit IRC00:13
*** shaohe_feng has joined #openstack-nova00:14
*** harlowja has quit IRC00:15
*** diogogmt has quit IRC00:23
*** shaohe_feng has quit IRC00:23
*** shaohe_feng has joined #openstack-nova00:24
*** dimtruck is now known as zz_dimtruck00:26
*** diogogmt has joined #openstack-nova00:27
*** thorst_ has joined #openstack-nova00:27
*** rama_y has quit IRC00:29
*** rama_y has joined #openstack-nova00:29
*** shaohe_feng has quit IRC00:33
*** shaohe_feng has joined #openstack-nova00:34
*** rama_y has quit IRC00:35
*** mikal_ is now known as mikal00:35
*** diogogmt has quit IRC00:36
*** rfolco has quit IRC00:37
*** mtanino has quit IRC00:43
*** shaohe_feng has quit IRC00:44
*** ijw_ has joined #openstack-nova00:44
*** ijw has quit IRC00:46
*** hoangcx has joined #openstack-nova00:46
*** shaohe_feng has joined #openstack-nova00:47
*** stewie925 has quit IRC00:48
*** Kennan has quit IRC00:50
*** fragatina has quit IRC00:52
*** Apoorva has quit IRC00:54
*** shaohe_feng has quit IRC00:54
*** raunak has quit IRC00:54
*** shaohe_feng has joined #openstack-nova00:55
*** cheneydc has joined #openstack-nova00:59
*** nic has quit IRC01:00
*** Kennan has joined #openstack-nova01:01
lifelessalaski: around ?01:02
lifelessalaski: there's something odd with test_patch_unboundmethod_with_staticmethod01:03
*** Kevin_Zheng has joined #openstack-nova01:03
*** gjayavelu has quit IRC01:03
lifelessalaski: if I use INST_C in that test, it returns 1, not tgtslf, 1 - e.g. the patch isn't applied ? :/01:03
*** shaohe_feng has quit IRC01:04
*** shaohe_feng has joined #openstack-nova01:05
*** iceyao has joined #openstack-nova01:05
*** moshele has joined #openstack-nova01:09
*** jinxing has joined #openstack-nova01:10
*** diogogmt has joined #openstack-nova01:14
*** shaohe_feng has quit IRC01:14
*** shaohe_feng has joined #openstack-nova01:17
*** lappy-rook has quit IRC01:19
*** mriedem has joined #openstack-nova01:19
*** ducttape_ has joined #openstack-nova01:20
*** Sukhdev has quit IRC01:22
*** doug-fish has joined #openstack-nova01:24
*** shaohe_feng has quit IRC01:25
*** shaohe_feng has joined #openstack-nova01:25
*** EinstCrazy has joined #openstack-nova01:26
*** raunak has joined #openstack-nova01:29
*** doug-fish has quit IRC01:29
*** Jeffrey4l has joined #openstack-nova01:30
*** raunak has quit IRC01:34
*** baoli has joined #openstack-nova01:35
*** shaohe_feng has quit IRC01:35
*** whenry has joined #openstack-nova01:36
*** skoude has quit IRC01:36
*** shaohe_feng has joined #openstack-nova01:36
*** moshele has quit IRC01:37
*** skoude has joined #openstack-nova01:37
*** Sukhdev has joined #openstack-nova01:39
openstackgerritkevin shen proposed openstack/nova: fix the limit of injected file content length is incorrect problem  https://review.openstack.org/31786501:40
*** thorst_ has quit IRC01:42
*** Fdaisuke has joined #openstack-nova01:42
*** al_indigo has quit IRC01:42
*** thorst_ has joined #openstack-nova01:42
*** Sukhdev has quit IRC01:45
*** shaohe_feng has quit IRC01:45
*** shaohe_feng has joined #openstack-nova01:46
*** gjayavelu has joined #openstack-nova01:46
*** iceyao has quit IRC01:48
*** smatzek_ has joined #openstack-nova01:49
*** baoli has quit IRC01:51
*** baoli has joined #openstack-nova01:51
*** thorst_ has quit IRC01:51
*** ljxiash has joined #openstack-nova01:54
openstackgerritSiri Kim proposed openstack/nova: Title: bugfix  https://review.openstack.org/31840501:54
*** yamahata has joined #openstack-nova01:55
*** shaohe_feng has quit IRC01:55
*** shaohe_feng has joined #openstack-nova01:56
*** ducttape_ has quit IRC01:56
*** baoli has quit IRC01:57
*** pradk has joined #openstack-nova01:59
*** whenry has quit IRC02:00
*** EinstCra_ has joined #openstack-nova02:00
*** EinstCrazy has quit IRC02:03
*** lixiaoy1 has joined #openstack-nova02:04
*** iceyao has joined #openstack-nova02:05
*** iceyao has quit IRC02:06
*** shaohe_feng has quit IRC02:06
*** iceyao_ has joined #openstack-nova02:08
*** vishwanathj has quit IRC02:09
*** shaohe_feng has joined #openstack-nova02:09
*** hieulq_ has quit IRC02:09
*** pradk has quit IRC02:10
*** shaohe_feng has quit IRC02:16
*** shaohe_feng has joined #openstack-nova02:17
*** yonglihe has joined #openstack-nova02:17
*** unicell has quit IRC02:20
*** unicell has joined #openstack-nova02:22
*** unicell has quit IRC02:22
*** sacharya_ has joined #openstack-nova02:26
*** shaohe_feng has quit IRC02:26
*** shaohe_feng has joined #openstack-nova02:27
*** ducttape_ has joined #openstack-nova02:28
*** sacharya has quit IRC02:29
*** jinxing has quit IRC02:30
*** sree has joined #openstack-nova02:30
*** houming has joined #openstack-nova02:30
*** BjoernT has joined #openstack-nova02:32
mriedemdid we ever store the image href (url) in the instances table?02:33
mriedemi hope not02:33
mriedemjohnthetubaguy: i'm +2 on the glance v2 spec now https://review.openstack.org/#/c/301741/02:33
*** sree has quit IRC02:33
*** mriedem has quit IRC02:33
*** ducttape_ has quit IRC02:34
*** yamahata has quit IRC02:34
*** links has joined #openstack-nova02:35
*** baoli has joined #openstack-nova02:35
*** shaohe_feng has quit IRC02:36
*** jinxing has joined #openstack-nova02:37
*** shaohe_feng has joined #openstack-nova02:37
*** baoli has quit IRC02:39
*** BjoernT has quit IRC02:39
*** baoli has joined #openstack-nova02:39
*** smatzek_ has quit IRC02:40
*** sree has joined #openstack-nova02:40
*** mdrabe has joined #openstack-nova02:41
*** baoli has quit IRC02:42
*** baoli has joined #openstack-nova02:43
*** mdrabe has quit IRC02:43
openstackgerritZhenyu Zheng proposed openstack/python-novaclient: Clean up deprecated CLI options  https://review.openstack.org/31779502:43
*** shaohe_feng has quit IRC02:47
*** shaohe_feng has joined #openstack-nova02:47
*** woodster_ has quit IRC02:48
*** thorst_ has joined #openstack-nova02:48
*** lakshmiS has quit IRC02:52
*** mdrabe has joined #openstack-nova02:54
*** thorst_ has quit IRC02:57
openstackgerritZhenyu Zheng proposed openstack/python-novaclient: Clean up deprecated CLI options  https://review.openstack.org/31779502:57
*** shaohe_feng has quit IRC02:57
*** jinxing has quit IRC02:58
*** EinstCra_ has quit IRC02:58
*** iceyao_ has quit IRC02:58
*** shaohe_feng has joined #openstack-nova02:58
*** tangchen_ has joined #openstack-nova02:58
*** iceyao has joined #openstack-nova02:59
*** EinstCrazy has joined #openstack-nova02:59
*** jinxing has joined #openstack-nova02:59
*** ducttape_ has joined #openstack-nova03:01
*** raunak has joined #openstack-nova03:03
*** shaohe_feng has quit IRC03:07
*** shaohe_feng has joined #openstack-nova03:08
*** baoli has quit IRC03:09
*** zhangyufei has joined #openstack-nova03:10
*** ducttape_ has quit IRC03:12
*** ijw_ has quit IRC03:14
*** gjayavel_ has joined #openstack-nova03:15
*** rama_y has joined #openstack-nova03:16
*** AmyLee7 has joined #openstack-nova03:17
*** shaohe_feng has quit IRC03:17
*** ozialien10 has quit IRC03:18
*** ozialien10 has joined #openstack-nova03:19
*** gjayavelu has quit IRC03:19
*** shaohe_feng has joined #openstack-nova03:21
*** mdrabe has quit IRC03:21
*** gmann_ has joined #openstack-nova03:27
*** shaohe_feng has quit IRC03:28
*** shaohe_feng has joined #openstack-nova03:28
*** zz_dimtruck is now known as dimtruck03:31
*** chip__ has joined #openstack-nova03:38
*** shaohe_feng has quit IRC03:38
*** shaohe_feng has joined #openstack-nova03:38
*** ljxiash has quit IRC03:41
*** garyk1 has joined #openstack-nova03:47
*** shaohe_feng has quit IRC03:48
*** shaohe_feng has joined #openstack-nova03:49
*** ijw has joined #openstack-nova03:53
*** thorst_ has joined #openstack-nova03:54
*** ljxiash has joined #openstack-nova03:56
*** shaohe_feng has quit IRC03:58
*** shaohe_feng has joined #openstack-nova03:59
*** ljxiash has quit IRC04:00
*** links has quit IRC04:00
*** thorst_ has quit IRC04:01
*** prateek has joined #openstack-nova04:02
*** gmann_ has quit IRC04:04
*** flwang has quit IRC04:04
*** amotoki has joined #openstack-nova04:06
*** gmann_ has joined #openstack-nova04:07
*** shaohe_feng has quit IRC04:09
*** shaohe_feng has joined #openstack-nova04:10
*** rama_y has quit IRC04:10
*** prateek has quit IRC04:11
*** flwang has joined #openstack-nova04:11
*** gmann_ has quit IRC04:11
*** links has joined #openstack-nova04:16
*** Jeffrey4l has quit IRC04:16
*** zhangyufei has quit IRC04:18
*** dimtruck is now known as zz_dimtruck04:18
*** jinxing has quit IRC04:19
*** ducttape_ has joined #openstack-nova04:19
*** shaohe_feng has quit IRC04:19
*** shaohe_feng has joined #openstack-nova04:20
*** Sukhdev has joined #openstack-nova04:20
*** ducttape_ has quit IRC04:21
*** prateek has joined #openstack-nova04:21
*** can8dnSix has joined #openstack-nova04:21
*** Sukhdev has quit IRC04:24
*** Sukhdev has joined #openstack-nova04:24
*** shaohe_feng has quit IRC04:29
*** shaohe_feng has joined #openstack-nova04:30
*** AmyLee7 has quit IRC04:33
*** shaohe_feng has quit IRC04:39
*** shaohe_feng has joined #openstack-nova04:40
*** armax has quit IRC04:40
*** kengo_sakai has joined #openstack-nova04:41
*** zz_dimtruck is now known as dimtruck04:43
*** kengo_sakai has quit IRC04:46
*** chip__ has quit IRC04:47
*** chip__ has joined #openstack-nova04:48
*** shaohe_feng has quit IRC04:50
*** hemna has joined #openstack-nova04:52
*** shaohe_feng has joined #openstack-nova04:53
*** gongysh has joined #openstack-nova04:53
*** hemna has quit IRC04:53
*** ljxiash has joined #openstack-nova04:53
*** hemna has joined #openstack-nova04:54
*** armax has joined #openstack-nova04:54
*** mnestratov has quit IRC04:56
*** hemna has quit IRC04:56
*** ijw has quit IRC04:57
*** thorst_ has joined #openstack-nova04:59
*** shaohe_feng has quit IRC05:00
*** shaohe_feng has joined #openstack-nova05:01
*** armax has quit IRC05:01
*** kengo_sakai has joined #openstack-nova05:02
*** armax has joined #openstack-nova05:03
*** gongysh has quit IRC05:04
*** thorst_ has quit IRC05:06
*** sudipto has joined #openstack-nova05:06
*** cheneydc has quit IRC05:07
*** syfy323 has quit IRC05:09
*** cheneydc has joined #openstack-nova05:10
*** shaohe_feng has quit IRC05:10
*** shaohe_feng has joined #openstack-nova05:11
*** doug-fish has joined #openstack-nova05:11
*** can8dnSix has quit IRC05:12
*** klkumar_ has joined #openstack-nova05:13
*** doug-fish has quit IRC05:15
*** syfy323 has joined #openstack-nova05:17
*** shaohe_feng has quit IRC05:20
*** shaohe_feng has joined #openstack-nova05:21
*** chip__ has quit IRC05:22
*** chip__ has joined #openstack-nova05:24
*** irenab has joined #openstack-nova05:25
*** kengo_sakai has quit IRC05:27
*** markvoelker has joined #openstack-nova05:27
openstackgerritAlex Xu proposed openstack/nova: Enable 'null' value for user_data in V2.1 API  https://review.openstack.org/31844705:29
*** shaohe_feng has quit IRC05:31
*** shaohe_feng has joined #openstack-nova05:31
*** markvoelker has quit IRC05:34
*** syfy323 has quit IRC05:35
*** dimtruck is now known as zz_dimtruck05:35
*** syfy323 has joined #openstack-nova05:36
*** zhangyufei has joined #openstack-nova05:36
*** unicell has joined #openstack-nova05:36
*** jinxing has joined #openstack-nova05:37
*** shaohe_feng has quit IRC05:41
*** shaohe_feng has joined #openstack-nova05:42
*** yamahata has joined #openstack-nova05:44
*** zhangyufei has quit IRC05:51
*** shaohe_feng has quit IRC05:51
*** salv-orlando has joined #openstack-nova05:51
*** shaohe_feng has joined #openstack-nova05:52
*** zhangyufei has joined #openstack-nova05:55
*** fawadkhaliq has joined #openstack-nova05:57
*** Sukhdev has quit IRC05:57
*** apoorv has joined #openstack-nova05:57
*** doug-fish has joined #openstack-nova05:57
*** Jeffrey4l has joined #openstack-nova05:58
*** mjura has joined #openstack-nova06:00
*** armax has quit IRC06:00
*** doug-fish has quit IRC06:01
*** shaohe_feng has quit IRC06:01
*** malos has quit IRC06:02
*** shaohe_feng has joined #openstack-nova06:02
*** raunak has quit IRC06:02
*** thorst_ has joined #openstack-nova06:04
*** mnestratov has joined #openstack-nova06:06
*** mjura has quit IRC06:07
*** mjura has joined #openstack-nova06:07
*** moshele has joined #openstack-nova06:08
*** Oku_OS-away is now known as Oku_OS06:10
*** thorst_ has quit IRC06:10
*** shaohe_feng has quit IRC06:12
*** shaohe_feng has joined #openstack-nova06:12
*** rcernin has joined #openstack-nova06:13
*** Jeffrey4l has quit IRC06:18
*** jgriffith has quit IRC06:19
*** edand has joined #openstack-nova06:19
*** unicell has quit IRC06:20
*** unicell has joined #openstack-nova06:20
*** lpetrut has joined #openstack-nova06:22
*** shaohe_feng has quit IRC06:22
*** gmann_ has joined #openstack-nova06:23
*** sfinucan has joined #openstack-nova06:24
*** shaohe_feng has joined #openstack-nova06:25
*** jgriffith has joined #openstack-nova06:27
*** jgriffith is now known as Guest6086406:27
openstackgerritMoshe Levi proposed openstack/nova: Remove unused itype parameter from get migration context  https://review.openstack.org/31797706:27
*** tbachman has quit IRC06:28
*** tbachman has joined #openstack-nova06:31
*** salv-orlando has quit IRC06:32
*** shaohe_feng has quit IRC06:32
openstackgerritAlex Xu proposed openstack/nova: Fix the versions API for api-ref  https://review.openstack.org/30872006:33
*** sudipto has quit IRC06:34
*** shaohe_feng has joined #openstack-nova06:35
*** sudipto has joined #openstack-nova06:36
*** shaohe_feng has quit IRC06:42
*** nkrinner_afk is now known as nkrinner06:43
*** belmoreira has joined #openstack-nova06:45
*** shaohe_feng has joined #openstack-nova06:45
*** malos has joined #openstack-nova06:46
*** gjayavel_ has quit IRC06:47
*** gjayavelu has joined #openstack-nova06:50
*** laszloh has joined #openstack-nova06:51
*** jinxing has quit IRC06:51
*** zhangyufei has quit IRC06:51
*** EinstCrazy has quit IRC06:51
*** iceyao has quit IRC06:51
*** liverpooler has joined #openstack-nova06:51
*** zhangyufei has joined #openstack-nova06:52
*** jinxing has joined #openstack-nova06:52
*** EinstCrazy has joined #openstack-nova06:52
*** iceyao has joined #openstack-nova06:52
*** shaohe_feng has quit IRC06:53
*** shaohe_feng has joined #openstack-nova06:53
*** jinxing has quit IRC06:54
*** EinstCrazy has quit IRC06:54
*** zhangyufei has quit IRC06:54
*** iceyao has quit IRC06:54
*** iceyao has joined #openstack-nova06:55
*** EinstCrazy has joined #openstack-nova06:56
*** zhangyufei has joined #openstack-nova06:56
*** jinxing has joined #openstack-nova06:57
*** iceyao has quit IRC06:57
*** zhangyufei has quit IRC06:57
*** EinstCrazy has quit IRC06:57
*** sacharya_ has quit IRC06:57
*** andreas_s has joined #openstack-nova06:58
*** EinstCrazy has joined #openstack-nova06:58
*** iceyao has joined #openstack-nova06:58
*** zhangyufei has joined #openstack-nova06:58
*** sridharg has joined #openstack-nova06:59
*** gjayavelu has quit IRC06:59
*** hoangcx has quit IRC06:59
*** jaosorior has quit IRC06:59
*** jaosorior has joined #openstack-nova07:00
*** EinstCrazy has quit IRC07:00
*** jinxing has quit IRC07:00
*** iceyao has quit IRC07:00
*** zhangyufei has quit IRC07:00
*** iceyao has joined #openstack-nova07:01
*** zhangyufei has joined #openstack-nova07:01
*** EinstCrazy has joined #openstack-nova07:01
*** jinxing has joined #openstack-nova07:02
*** shaohe_feng has quit IRC07:03
*** shaohe_feng has joined #openstack-nova07:03
*** cfriesen has quit IRC07:04
openstackgerritZhenyu Zheng proposed openstack/python-novaclient: Clean up deprecated CLI options  https://review.openstack.org/31779507:05
*** gjayavelu has joined #openstack-nova07:06
*** roeyc has joined #openstack-nova07:07
*** roeyc has quit IRC07:08
*** thorst_ has joined #openstack-nova07:08
*** roeyc has joined #openstack-nova07:08
*** realitix has joined #openstack-nova07:09
*** jichen has joined #openstack-nova07:09
realitixHello everyone. Do you know why we can't rename a flavor ? Is there a good reason or is it just unimplemented ?07:11
*** shaohe_feng has quit IRC07:13
*** shaohe_feng has joined #openstack-nova07:13
openstackgerritSudipta Biswas proposed openstack/nova-specs: Complete Nova support for Glance's v2  https://review.openstack.org/30174107:15
*** thorst_ has quit IRC07:15
openstackgerritZhenyu Zheng proposed openstack/python-novaclient: Clean up deprecated CLI options  https://review.openstack.org/31779507:17
*** gmann_ has quit IRC07:19
*** sfinucan has quit IRC07:22
*** shaohe_feng has quit IRC07:23
*** jlanoux has joined #openstack-nova07:24
*** mpavone has joined #openstack-nova07:24
*** shaohe_feng has joined #openstack-nova07:27
*** gjayavelu has quit IRC07:29
*** shaohe_feng has quit IRC07:34
*** lpetrut has quit IRC07:34
*** shaohe_feng has joined #openstack-nova07:35
*** dedery has joined #openstack-nova07:35
*** irenab_ has joined #openstack-nova07:36
*** ihrachys has joined #openstack-nova07:37
*** irenab has quit IRC07:38
*** irenab_ is now known as irenab07:38
*** fawadkhaliq has quit IRC07:39
*** fawadkhaliq has joined #openstack-nova07:39
*** matrohon has joined #openstack-nova07:40
*** ihrachys has quit IRC07:41
*** lpetrut has joined #openstack-nova07:43
*** shaohe_feng has quit IRC07:44
*** shaohe_feng has joined #openstack-nova07:45
*** salv-orl_ has joined #openstack-nova07:47
*** salv-orl_ has quit IRC07:47
*** danpawlik has joined #openstack-nova07:47
*** markmc``` is now known as markmc07:48
*** salv-orlando has joined #openstack-nova07:48
*** zhangyufei has quit IRC07:49
*** zenoway has joined #openstack-nova07:50
*** dmk0202 has joined #openstack-nova07:50
*** zhangyufei has joined #openstack-nova07:51
*** gmann_ has joined #openstack-nova07:52
openstackgerritZhongcheng Lao proposed openstack/nova: VMware: Omit exceptions during snapshot cleanup  https://review.openstack.org/21608607:53
*** haukebruno has joined #openstack-nova07:53
*** markus_z has joined #openstack-nova07:54
*** shaohe_feng has quit IRC07:54
*** shaohe_feng has joined #openstack-nova07:55
bauzasgood morning Nova07:56
*** EinstCra_ has joined #openstack-nova07:57
*** sacharya has joined #openstack-nova07:58
*** zzzeek has quit IRC08:00
*** EinstCrazy has quit IRC08:00
*** iceyao_ has joined #openstack-nova08:02
*** zzzeek has joined #openstack-nova08:02
*** iceyao has quit IRC08:03
gibigood morning bauzas08:03
*** sacharya has quit IRC08:03
*** shaohe_feng has quit IRC08:04
*** jamielennox is now known as jamielennox|away08:05
*** shaohe_feng has joined #openstack-nova08:08
*** cheran has quit IRC08:09
openstackgerritDavanum Srinivas (dims) proposed openstack/nova: [WIP] Testing latest u-c  https://review.openstack.org/30334508:10
*** vnogin has joined #openstack-nova08:10
*** hoangcx has joined #openstack-nova08:10
*** lpetrut has quit IRC08:11
*** iceyao has joined #openstack-nova08:12
*** iceyao has quit IRC08:13
*** thorst_ has joined #openstack-nova08:14
*** iceyao has joined #openstack-nova08:14
*** danpb has joined #openstack-nova08:14
*** shaohe_feng has quit IRC08:15
*** markus_z has quit IRC08:15
*** shaohe_feng has joined #openstack-nova08:16
*** iceyao_ has quit IRC08:16
*** markus_z has joined #openstack-nova08:16
*** yamahata has quit IRC08:17
*** mpavone has quit IRC08:18
openstackgerritHiroyuki Eguchi proposed openstack/nova-specs: Versioned notification for keypairs  https://review.openstack.org/30094208:19
*** ihrachys has joined #openstack-nova08:20
*** daemontool has joined #openstack-nova08:21
*** tangchen_ has quit IRC08:21
*** thorst_ has quit IRC08:21
*** ljjjustin has quit IRC08:21
*** ljjjustin has joined #openstack-nova08:23
*** mpavone has joined #openstack-nova08:23
*** shaohe_feng has quit IRC08:25
*** lucas-dinner is now known as lucasagomes08:26
*** ozialien10 has quit IRC08:27
*** ozialien10 has joined #openstack-nova08:27
*** shaohe_feng has joined #openstack-nova08:28
*** aloga has quit IRC08:28
*** lpetrut has joined #openstack-nova08:29
*** hoangcx has quit IRC08:29
*** hoangcx has joined #openstack-nova08:29
*** aloga has joined #openstack-nova08:29
*** jistr has joined #openstack-nova08:30
*** liusheng has quit IRC08:34
*** liusheng has joined #openstack-nova08:35
*** shaohe_feng has quit IRC08:35
*** shaohe_feng has joined #openstack-nova08:36
lyarwood_mdbooth: morning, could I ask for a review of https://review.openstack.org/318266 if you have time?08:36
*** EinstCrazy has joined #openstack-nova08:38
*** yassine has joined #openstack-nova08:40
*** yassine is now known as Guest8606108:40
*** EinstCra_ has quit IRC08:40
*** lpetrut has quit IRC08:41
*** ljxiash has quit IRC08:42
lifelessalaski: ahha, found it.08:42
*** ljxiash has joined #openstack-nova08:42
openstackgerritZhenyu Zheng proposed openstack/python-novaclient: Clean up deprecated CLI options  https://review.openstack.org/31779508:42
lifelessalaski: we setattr to unpatch things, even if what we patched didn't actually exist on the object previously08:43
lifelessalaski: I'm going to have a stab at fixing that08:43
*** iceyao has quit IRC08:45
*** shaohe_feng has quit IRC08:45
*** shaohe_feng has joined #openstack-nova08:46
*** ljxiash has quit IRC08:47
*** cdent has joined #openstack-nova08:48
*** iceyao has joined #openstack-nova08:51
*** PaulMurray has joined #openstack-nova08:51
*** al_indigo has joined #openstack-nova08:53
*** shaohe_feng has quit IRC08:56
*** shaohe_feng has joined #openstack-nova08:57
*** paul-carlton2 has quit IRC08:58
*** gmann_ has quit IRC08:58
*** derekh has joined #openstack-nova09:00
openstackgerritHieu LE proposed openstack/nova: Config options: remove import_opts from cloudpipe section  https://review.openstack.org/31851709:02
*** shaohe_feng has quit IRC09:06
*** al_indigo has quit IRC09:07
*** shaohe_feng has joined #openstack-nova09:09
mdboothlyarwood_: Sure09:11
lyarwood_mdbooth: thanks09:12
*** shaohe_feng has quit IRC09:16
*** shaohe_feng has joined #openstack-nova09:17
*** sambetts|afk is now known as sambetts09:17
*** lpetrut has joined #openstack-nova09:18
*** PaulMurray has quit IRC09:18
*** ptm_away has joined #openstack-nova09:19
*** thorst_ has joined #openstack-nova09:19
*** chenfan has joined #openstack-nova09:21
*** eliqiao has quit IRC09:21
*** chenfan has quit IRC09:23
*** ihrachys has quit IRC09:23
*** chenfan has joined #openstack-nova09:24
*** houming has quit IRC09:24
*** fawadkhaliq has quit IRC09:24
*** fawadkhaliq has joined #openstack-nova09:24
*** chenfan has quit IRC09:25
*** thorst_ has quit IRC09:26
*** shaohe_feng has quit IRC09:26
openstackgerritBalazs Gibizer proposed openstack/nova: Transform generic instance action notifications  https://review.openstack.org/31365409:27
*** chenfan has joined #openstack-nova09:28
*** shaohe_feng has joined #openstack-nova09:28
*** chenfan has quit IRC09:29
*** ccard has joined #openstack-nova09:29
*** chenfan has joined #openstack-nova09:31
*** ccard_ has quit IRC09:31
*** EinstCrazy has quit IRC09:36
*** EinstCrazy has joined #openstack-nova09:36
*** shaohe_feng has quit IRC09:37
*** shaohe_feng has joined #openstack-nova09:38
*** iceyao has quit IRC09:40
*** iceyao has joined #openstack-nova09:41
*** iceyao_ has joined #openstack-nova09:42
*** iceyao__ has joined #openstack-nova09:43
*** iceyao has quit IRC09:46
*** ericksonsantos has quit IRC09:46
*** laszloh has quit IRC09:46
*** shaohe_feng has quit IRC09:47
*** iceyao_ has quit IRC09:47
*** shaohe_feng has joined #openstack-nova09:48
*** ericksonsantos has joined #openstack-nova09:49
*** eslamelhusseiny has joined #openstack-nova09:49
*** EinstCrazy has quit IRC09:51
openstackgerritSharat Sharma proposed openstack/nova: Adding "security_group_default_rules" parameter  https://review.openstack.org/31555709:54
*** rk4n has joined #openstack-nova09:55
*** hoangcx has quit IRC09:56
*** shaohe_feng has quit IRC09:57
*** shaohe_feng has joined #openstack-nova09:58
*** eslamelhusseiny has quit IRC09:59
*** cheneydc has quit IRC10:00
*** al_indigo has joined #openstack-nova10:01
*** porunov has joined #openstack-nova10:01
*** porunov has left #openstack-nova10:02
*** al_indigo has quit IRC10:03
openstackgerritDao Cong Tien proposed openstack/nova: Enables Py34 tests for unit.api.openstack.compute.test_server_tags  https://review.openstack.org/31854310:03
mdboothlyarwood_: That detach_volume patch requires a lot of context!10:03
mdboothSo, there's a comment in there just above the lines you patches (line 1286)10:04
mdboothWhich says that this is called post live migration10:04
mdboothAnd in that case we *would* want to disconnect the volume, despite not being on the instance's host10:04
*** matrohon has quit IRC10:04
mdboothHowever, afaict it is not actually called post live migration10:04
*** hoangcx has joined #openstack-nova10:05
*** edand has quit IRC10:07
*** shaohe_feng has quit IRC10:07
*** zhangyufei has quit IRC10:08
*** shaohe_feng has joined #openstack-nova10:08
*** zhangyufei has joined #openstack-nova10:11
*** gjayavelu has joined #openstack-nova10:13
*** salv-orl_ has joined #openstack-nova10:14
*** rmart04 has joined #openstack-nova10:14
*** mvk_ has quit IRC10:15
lyarwood_mdbooth: hey sorry hilights aren't working in here for some reason10:17
*** salv-orlando has quit IRC10:17
mdboothlyarwood_: Just thinking myself into another knot, btw.10:17
*** gjayavelu has quit IRC10:17
lyarwood_mdbooth: and yeah sorry for the lack of context in the commit msg10:18
mdboothSo your test does: if connector.get('host') == instance.host10:18
*** shaohe_feng has quit IRC10:18
*** claudiub has joined #openstack-nova10:18
mdboothHowever, aren't we interested in if it matches the current host?10:18
lyarwood_mdbooth: right, and that connector is pulled from the current host10:18
lyarwood_mdbooth: not from the connection_info10:18
*** shaohe_feng has joined #openstack-nova10:18
mdboothWhoa10:19
mdboothCan't we get that information more directly?10:19
mdbooth            host=CONF.host)10:20
mdbooth(from get_volume_connector)10:20
lyarwood_mdbooth: yeah we can just do that instead10:20
mdboothWould if instance.host == CONF.host work there?10:20
lyarwood_mdbooth: but your point re post live migration is a deal breaker for this10:21
mdboothlyarwood_: I don't think it is, actually10:21
mdboothI think that comment is a bug10:21
mdboothI've commented on that in the review10:21
lyarwood_ah and we don't call this?10:21
*** cdent has quit IRC10:21
mdbooth...sec, let me post the interim10:21
lyarwood_thanks10:21
mdbooththar you go10:21
mdboothlyarwood_: However, *please* double check that10:22
mdboothBecause it's not exactly straightforward10:22
mdboothAlso might be worth specifically testing it if you have the cycles10:22
lyarwood_yeah I have a multi host devstack env ready to go to test this10:22
lyarwood_I'll get on with that now, thanks again10:23
mdboothlyarwood_: Can we not add that guard the level up?10:25
mdboothIt seems more intuitive to me to not call detach_volume() at all if you know it isn't required10:25
mdboothAnd the code above doesn't look relevant in the remote rebuild case, either10:26
mdboothI suspect it's just logging an error, in fact10:26
lyarwood_mdbooth: in the compute layer?10:26
mdboothlyarwood_: Just reminding myself where it's called from, but I guess so, yes10:26
lyarwood_mdbooth: I really wanted to avoid breaking other virt drivers tbh10:27
*** zhangyufei has quit IRC10:27
*** sahid has joined #openstack-nova10:27
mdboothlyarwood_: But would it?10:27
mdboothIf detach_volume has the same semantics across all drivers10:27
mdboothi.e. cleanup the host10:27
*** zhangyufei has joined #openstack-nova10:27
lyarwood_mdbooth: hmm I was worried about the single compute host drivers tbh10:27
mdboothThen that should be ok, right?10:27
sahiddanpb: can you re+W this serie when you have a moment https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/libvirt-clean-driver i had to rebase it10:28
*** shaohe_feng has quit IRC10:28
*** shaohe_feng has joined #openstack-nova10:29
*** zhangyufei has quit IRC10:29
lyarwood_mdbooth: I'll move it up and run it past CI, hopefully it's valid.10:30
*** zhangyufei has joined #openstack-nova10:30
*** zhangyufei has quit IRC10:30
danpbsahid: when i get done wading through email backlog...10:32
*** sfinucan has joined #openstack-nova10:35
*** rtheis has joined #openstack-nova10:35
*** sdague has joined #openstack-nova10:37
*** shaohe_feng has quit IRC10:38
*** shaohe_feng has joined #openstack-nova10:39
*** ansiwen_ is now known as ansiwen10:41
*** smatzek_ has joined #openstack-nova10:46
*** apoorv has quit IRC10:47
*** shaohe_feng has quit IRC10:48
*** yonglihe has quit IRC10:48
*** hieulq has quit IRC10:49
*** shaohe_feng has joined #openstack-nova10:49
*** mvk_ has joined #openstack-nova10:50
*** smatzek_ is now known as smatzek10:50
*** apoorv has joined #openstack-nova10:53
*** flwang1 has joined #openstack-nova10:54
openstackgerritsahid proposed openstack/nova: Introduce VNC Security Proxy Framework  https://review.openstack.org/11548310:55
openstackgerritsahid proposed openstack/nova: Add VeNCrypt (TLS/x509) Security Proxy Driver  https://review.openstack.org/11548410:55
*** cdent has joined #openstack-nova10:55
flwang1hi guys, can i put a customized filter to another place instead of nova/scheduler/filter? thanks10:55
bauzasflwang1: sure, easy thing10:55
flwang1bauzas: how? ;)10:55
bauzasflwang1: http://docs.openstack.org/developer/nova/filter_scheduler.html#filtering10:56
flwang1bauzas: just wondering how to make the puppet code easy10:56
bauzasflwang1: at the end of that section10:56
bauzasflwang1: you don't even need to have the same namespace for your own custom filter10:56
flwang1oh, really? that's sound super cool10:57
sfinucandanpb: This has already been merged, but I would appreciate an eye over the NUMA/CPU pinning docs. Would be good to know I haven't said anything too outrageous http://docs.openstack.org/admin-guide/compute-numa-cpu-pinning.html10:57
*** shaohe_feng has quit IRC10:59
bauzasflwang1: tl;dr: what you need is to implement a new class inheriting from the BaseHostFilter and either return True or False10:59
*** shaohe_feng has joined #openstack-nova10:59
bauzasflwang1: then, it's just a configuration thing, by adding your class to the scheduler_available_filters opt like scheduler_available_filters += mynamespace.MyFilter11:00
flwang1bauzas: but the place must be a package which can be accessible in python, right?11:00
bauzasflwang1: and then, use it by adding it to the scheduler_default_filters11:00
bauzasflwang1: of course, yes11:00
bauzasflwang1: you can also put your filters in the same namespace than the others, ie. nova.scheduler.filters11:01
sdaguealex_xu: some comments back on the spec to make sure you are good with it11:01
sdaguehttps://review.openstack.org/#/c/312209/311:01
openstackgerritSean Dague proposed openstack/nova-specs: Deprecate API Proxies  https://review.openstack.org/31220911:01
bauzasflwang1: in that case, you don't need to have your own namespace, nor modifying the scheduler_available_filters, because that module and class will be found by the all_filters module call11:01
flwang1bauzas: i'm just thinking how to do that with lowest cost in puppet11:02
bauzasflwang1: well, it's just a File variable, right?11:02
bauzasthen, you just need to depend your conf modification call from your File var11:02
flwang1bauzas: so basically, just put the file under nova/scheduler/filter is the easiest way, right?11:02
flwang1i'm not a puppet guy actually :)11:03
bauzasflwang1: the easiest but most confusing one, yes :)11:03
*** jamielennox|away is now known as jamielennox11:03
bauzasbecause operators could think your custom filter is in the nova tree11:04
flwang1bauzas: from that view, yes, i can see your point11:04
flwang1but i'm also wondering where to put the new filter them11:04
bauzasplus the fact that if we modify the scheduler filters location, then you're screwed11:04
flwang1since we do need a namespace(python pkg) for it11:04
bauzaswell, you just need a tree that can be found from the python path...11:05
flwang1is there any best practice for this?11:06
bauzasit's a python thing11:06
bauzasnot really needing Nova experts :)11:06
*** raj_singh_ has quit IRC11:06
*** raj_singh_ has joined #openstack-nova11:07
flwang1ok, i see. thanks a lot11:07
openstackgerritJames Page proposed openstack/nova: Fix use of invalid assert calls  https://review.openstack.org/31856811:07
*** ptm_away is now known as PaulMurray11:08
*** shaohe_feng has quit IRC11:09
*** shaohe_feng has joined #openstack-nova11:10
openstackgerritBalazs Gibizer proposed openstack/nova: Transform generic instance action notifications  https://review.openstack.org/31365411:15
*** palexster has quit IRC11:17
*** palexster has joined #openstack-nova11:17
*** danpb has quit IRC11:18
*** rk4n has quit IRC11:18
*** jinxing has quit IRC11:19
*** shaohe_feng has quit IRC11:19
openstackgerritBalazs Gibizer proposed openstack/nova: Transform generic instance action notifications  https://review.openstack.org/31365411:19
*** jinxing has joined #openstack-nova11:20
*** shaohe_feng has joined #openstack-nova11:20
*** rodrigods has quit IRC11:20
*** rodrigods has joined #openstack-nova11:20
*** jinxing has quit IRC11:21
*** sree has quit IRC11:23
Kevin_Zhengjohnthetubaguy: ping11:24
*** thorst_ has joined #openstack-nova11:27
*** doug-fish has joined #openstack-nova11:28
*** doug-fish has quit IRC11:29
*** shaohe_feng has quit IRC11:29
*** shaohe_feng has joined #openstack-nova11:30
*** lucasagomes is now known as lucas-hungry11:30
*** fawadkhaliq has quit IRC11:30
*** mtanino has joined #openstack-nova11:31
*** iceyao__ has quit IRC11:32
openstackgerritZhenyu Zheng proposed openstack/python-novaclient: Use tempes.lib instead of tempes_lib  https://review.openstack.org/31858511:32
*** iceyao has joined #openstack-nova11:34
*** cdent has quit IRC11:38
*** shaohe_feng has quit IRC11:40
*** danpb has joined #openstack-nova11:40
*** shaohe_feng has joined #openstack-nova11:40
*** hoangcx has quit IRC11:41
*** shaohe_feng has quit IRC11:50
*** shaohe_feng has joined #openstack-nova11:51
openstackgerritAtsushi SAKAI proposed openstack/nova: api-ref: Add fault parameter details  https://review.openstack.org/31795511:52
lbeliveaumoshele: hi11:56
moshelelbeliveau: hi :) I saw your comment11:57
moshelelbeliveau: I in a meeting will talk to you later11:57
lbeliveaumoshele: ok11:57
*** lucas-hungry is now known as lucasagomes11:57
lbeliveaumoshele: not sure about why there is a check there to make sure the flavor hasn't changeg11:57
moshelelbeliveau: but I think we hit different bug  :)11:58
moshelelbeliveau: it is not always happens at least in my set11:58
moshelelbeliveau: my setup11:58
lbeliveaumoshele: ok, let me know11:58
*** ducttape_ has joined #openstack-nova11:59
*** sacharya has joined #openstack-nova12:00
*** shaohe_feng has quit IRC12:00
*** shaohe_feng has joined #openstack-nova12:01
openstackgerritBalazs Gibizer proposed openstack/nova: Transform generic instance action notifications  https://review.openstack.org/31365412:02
*** rk4n has joined #openstack-nova12:05
*** cdent has joined #openstack-nova12:05
*** sacharya has quit IRC12:06
*** penick has quit IRC12:06
*** jistr has quit IRC12:08
openstackgerritjichenjc proposed openstack/nova: verify api-ref metadata.inc  https://review.openstack.org/31860012:09
*** shaohe_feng has quit IRC12:10
openstackgerritJuan Antonio Osorio Robles proposed openstack/nova: Add http_proxy_to_wsgi to api-paste  https://review.openstack.org/31860312:12
openstackgerritSean Dague proposed openstack/nova: api-ref: verify flavors.inc  https://review.openstack.org/31413312:13
*** shaohe_feng has joined #openstack-nova12:13
*** al_indigo has joined #openstack-nova12:14
*** ljxiash has joined #openstack-nova12:15
*** raildo-afk is now known as raildo12:15
flip214I'd like to know how to proceed with https://review.openstack.org/#/c/256292/... spec was originally proposed as https://review.openstack.org/#/c/134153/12:16
flip214Thanks for any advice.12:16
al_indigoHello guys. I have periodic random problems with nova-conductor and I don't know what to do with it. I get several times a day "ERROR oslo_service.periodic_task MessagingTimeout: Timed out waiting for a reply to message ID" errors and they make our cloud think that some compute nodes are lost or makes instances start in ERROR state.12:16
*** markvoelker has joined #openstack-nova12:16
*** rk4n has quit IRC12:17
*** rk4n has joined #openstack-nova12:18
*** gszasz has joined #openstack-nova12:18
al_indigoat the same time there is no any high load and I don't have ideas what causes such behaviour. The nova-conductor is resided at the same node with nova-api and neutron; controller is 32CPU/192GB RAM, network bandwidth is 2x10Gbit/s each to each. Database is postgres and there is no any high load on it. There are only 16 compute nodes in the cluster, nova version is stable/liberty 142323a73e379ab657af4dab6fe8b36c6710f1db12:19
openstackgerritJames Page proposed openstack/nova: Fix use of invalid assert calls  https://review.openstack.org/31856812:19
*** jtp92 has joined #openstack-nova12:19
*** ljxiash has quit IRC12:19
jtp92Good morning everyone!  Does anyone have experience with the VMware driver for Nova?12:20
*** eantyshev has quit IRC12:20
openstackgerritLee Yarwood proposed openstack/nova: compute: Avoid remote volume detach calls  https://review.openstack.org/31826612:20
*** shaohe_feng has quit IRC12:21
al_indigoconfig is the following: https://gist.github.com/al-indigo/fad6c84961ab51e00ebe803771c86af712:21
*** shaohe_feng has joined #openstack-nova12:21
*** markvoelker has quit IRC12:21
al_indigoMaybe I have something wrong in it? Could anyone look through it12:21
al_indigo*please12:21
jtp92I have a controller node, a compute node (not sure if necessary), and a cluster of nodes available through vCenter Server.  I found some documentation on what to add to the Nova config file, but no detailed write-up on how to actually deploy this type of environment.12:22
jtp92Currently I configured the Controller as normal, and I also configured the compute node as normal but added the vmware config lines to the nova.conf file.  Nova fails to start (can't find detailed reasons why)12:23
alex_xusdague: i agree with attach vol/interface should be kept. but checking the implementation of addfixedip, it just call the neutron api to update the port to have all the ips from the subnets. if the thing which was done by that API, it is nothing about nova, then it should be pure proxy api.12:23
*** syjulian has joined #openstack-nova12:23
*** markvoelker has joined #openstack-nova12:24
sdaguealex_xu: except it also triggers the nw_info_cache update12:25
sdagueright?12:25
*** apoorv has quit IRC12:25
*** gcb has quit IRC12:25
alex_xusdague: yeah, but if user do something at neutron side without proxy api, the nova should update nw_info_cache also12:25
sdaguealex_xu: what is the call path to attach a fixed_ip if this is removed12:26
*** penick has joined #openstack-nova12:27
*** jistr has joined #openstack-nova12:28
*** ducttape_ has quit IRC12:28
*** edmondsw has joined #openstack-nova12:28
alex_xusdague: just call the neutron api directly, and check the subnets of network which port belong to. Then update the port with neutron api. just as the code doing https://github.com/openstack/nova/blob/master/nova/network/neutronv2/api.py#L97512:28
*** jistr is now known as jistr|bbl12:28
*** shaohe_feng has quit IRC12:31
*** shaohe_feng has joined #openstack-nova12:31
openstackgerritZhenyu Zheng proposed openstack/python-novaclient: Use tempes.lib instead of tempes_lib  https://review.openstack.org/31858512:32
*** sdague has quit IRC12:32
*** prateek has quit IRC12:33
*** mdrabe has joined #openstack-nova12:35
openstackgerritAnkit Agrawal proposed openstack/nova: Move is_volume_backed_instance to compute.utils  https://review.openstack.org/27048212:35
openstackgerritAnkit Agrawal proposed openstack/nova: Fix evaluation of host disk usage by volume-backed instances  https://review.openstack.org/20087012:35
*** links has quit IRC12:36
*** gcb has joined #openstack-nova12:38
*** _prad_ has joined #openstack-nova12:40
*** kolesovdv has joined #openstack-nova12:40
*** shaohe_feng has quit IRC12:41
*** markvoelker has quit IRC12:41
openstackgerritHans Lindgren proposed openstack/nova: Save all instance extras in a single db call  https://review.openstack.org/30543712:42
*** markvoelker has joined #openstack-nova12:42
*** shaohe_feng has joined #openstack-nova12:42
*** sdague has joined #openstack-nova12:43
*** kylek3h has joined #openstack-nova12:44
*** eharney has joined #openstack-nova12:45
*** efried has joined #openstack-nova12:45
*** markvoelker_ has joined #openstack-nova12:46
openstackgerritHans Lindgren proposed openstack/nova: Make flavor-manage api call destroy with Flavor object  https://review.openstack.org/24429112:46
*** prateek has joined #openstack-nova12:47
alex_xusdague: emm...looks like the virt drivers other than libvirt needs inject_network/reset_network, but i'm not very sure. At least inject_network/reset_network are admin api on nova side, if user call neutron api directly, and thing go wrong without inject_network/reset_network, that sounds bug also.12:48
*** baoli has joined #openstack-nova12:48
*** markvoelker_ has quit IRC12:48
*** baoli_ has joined #openstack-nova12:49
*** penick has quit IRC12:49
*** markvoelker_ has joined #openstack-nova12:49
*** ivase has joined #openstack-nova12:50
*** markvoelker has quit IRC12:50
*** salv-orl_ has quit IRC12:50
*** shaohe_feng has quit IRC12:51
*** shaohe_feng has joined #openstack-nova12:52
*** baoli has quit IRC12:52
*** cleong has joined #openstack-nova12:53
*** markvoelker_ has quit IRC12:55
*** markvoelker has joined #openstack-nova12:55
*** winston-d_ has quit IRC12:55
*** akshai has joined #openstack-nova12:56
sdaguealex_xu: sorry, lost internet12:58
sdagueanyway, I think we should leave the actions12:58
alex_xusdague: np12:58
openstackgerritLee Yarwood proposed openstack/nova: libvirt: Limit live migration progress warnings  https://review.openstack.org/31864412:58
*** akshai has quit IRC12:58
alex_xusdague: ok, i guess there should be another evolution of api related to neutron interactive after nova-net removed...so we can leave that to that point12:59
*** palexster has quit IRC13:00
*** AmyLee7 has joined #openstack-nova13:00
*** whenry has joined #openstack-nova13:00
alex_xusdague: if we keep it, should we stop the nova-net functional for that api, maybe some problem for securitygroup related api13:00
*** roeyc has quit IRC13:01
*** garyk1 has quit IRC13:01
*** palexster has joined #openstack-nova13:01
*** garyk has joined #openstack-nova13:02
*** shaohe_feng has quit IRC13:02
*** roeyc has joined #openstack-nova13:02
*** shaohe_feng has joined #openstack-nova13:02
*** malos has quit IRC13:03
*** akshai has joined #openstack-nova13:03
*** palexster_ has joined #openstack-nova13:05
*** palexster_ has quit IRC13:05
*** ljxiash has joined #openstack-nova13:06
*** sdague has quit IRC13:07
*** palexster has quit IRC13:08
*** palexster has joined #openstack-nova13:09
*** palexster has quit IRC13:09
*** palexster has joined #openstack-nova13:09
*** sdague has joined #openstack-nova13:10
*** palexster has quit IRC13:10
alaskilifeless: around now, though you're probably not :) I added one small comment on the test you tweaked, but otherwise the changes lgtm13:10
*** MikeG451 has joined #openstack-nova13:11
*** sridharg has quit IRC13:11
* cdent writes a mock and shakes his tiny fist13:11
*** markvoelker_ has joined #openstack-nova13:12
*** markvoelker_ has quit IRC13:12
*** takedakn has joined #openstack-nova13:12
*** shaohe_feng has quit IRC13:12
*** rlrossit has joined #openstack-nova13:12
*** dave-mccowan has joined #openstack-nova13:12
*** markvoelker_ has joined #openstack-nova13:12
*** shaohe_feng has joined #openstack-nova13:13
gibisdague: hi!, It is just a friendly reminder to look at https://review.openstack.org/#/c/286675/ (versioned notification transformation spec) if you have time13:13
realitixHello everyone. Do you know why we can't rename a flavor ? Is there a good reason or is it just unimplemented ?13:14
*** markvoelker has quit IRC13:15
sdaguealex_xu: honestly, with the other interfaces gone, I think we can be sloppy about preventing it in the nova-net case13:16
alex_xusdague: ok13:17
*** palexster has joined #openstack-nova13:17
alex_xusdague: fyi, i found some comments about addfixedip/removefixedip at https://bugs.launchpad.net/nova/+bug/1430512/comments/1 and https://bugs.launchpad.net/nova/+bug/1430512/comments/313:17
openstackLaunchpad bug 1430512 in OpenStack Compute (nova) "Adding a fixed_ip to an instance with neutron only updates the first port using the first subnet" [Low,In progress] - Assigned to Gary Kotton (garyk)13:17
*** diogogmt has quit IRC13:17
*** jinxing has joined #openstack-nova13:17
*** lindycoder has joined #openstack-nova13:19
*** ctrath has joined #openstack-nova13:19
sdaguealex_xu: ok, cool, that's something we'll need to look into13:20
openstackgerritLee Yarwood proposed openstack/nova: compute: Avoid remote volume detach calls  https://review.openstack.org/31826613:20
*** lindycod_ has joined #openstack-nova13:21
*** shaohe_feng has quit IRC13:22
*** sridharg has joined #openstack-nova13:23
*** lindycoder has quit IRC13:24
*** shaohe_feng has joined #openstack-nova13:25
*** sridharg has quit IRC13:25
*** sridharg has joined #openstack-nova13:25
*** rk4n_ has joined #openstack-nova13:25
*** rk4n has quit IRC13:26
*** irenab has quit IRC13:26
openstackgerritAlex Xu proposed openstack/nova: Enable 'null' value for user_data in V2.1 API  https://review.openstack.org/31844713:28
openstackgerritHironori Shiina proposed openstack/nova: Ironic: change flat network provider to 'flat'  https://review.openstack.org/29789513:29
*** liverpooler has quit IRC13:30
*** AmyLee7 has quit IRC13:32
*** mriedem has joined #openstack-nova13:32
*** shaohe_feng has quit IRC13:32
*** mtanino has quit IRC13:33
*** shaohe_feng has joined #openstack-nova13:33
*** jichen has quit IRC13:33
*** penick has joined #openstack-nova13:35
openstackgerritOpenStack Proposal Bot proposed openstack/nova: Updated from global requirements  https://review.openstack.org/31695113:35
*** AmyLee7 has joined #openstack-nova13:36
*** smatzek has quit IRC13:38
sdaguejohnthetubaguy / mriedem you want to take a gander at the proxy deprecation spec again - https://review.openstack.org/#/c/312209/ ?13:39
mriedemsdague: maybe after the meeting13:40
*** READ10 has joined #openstack-nova13:42
*** shaohe_feng has quit IRC13:43
*** garyk has quit IRC13:44
sdaguemriedem: cool13:44
*** markvoelker_ has quit IRC13:44
sdaguemriedem: so, you head on vacation tomorrow, what's the run down I need for the n1 tagging. Is there anything specific we have to land by then?13:44
diana_clarkemriedem: Morning! Now that stable/liberty is security bugs only, I wondered if you have plans to cut another stable/liberty release soon-ish?13:45
diana_clarkemriedem: Asking b/c I'd like to package the last few bugs that made it in before the cutoff. A stable/mitaka release would be great too ;)13:45
mriedemdiana_clarke: i'd have to look at the change log, i've been meaning to do that13:45
*** shaohe_feng has joined #openstack-nova13:45
*** mgoddard_ has joined #openstack-nova13:45
mriedemdiana_clarke: we'll probably do stable/mitaka release around n-113:45
diana_clarkemriedem: Thanks!13:46
mriedemsdague: there isn't anything specific that i know of that has to be merged by n-113:46
alex_xusdague: johnthetubaguy mriedem, fyi, for api high priority bug, https://review.openstack.org/318447, hope you can take a look at when you have time.13:46
*** BobBall_AWOL is now known as BobBall13:47
*** takedakn has quit IRC13:47
*** mrkz has joined #openstack-nova13:48
*** mgoddard has quit IRC13:48
mriedemalex_xu: thanks, comments inline. i was trying to get a functional regression test working for that bug last night but missed the extra validation stuff you had for that TypeError13:49
mriedemi couldn't figure out why i kept getting a 40013:49
mriedemsdague: i woke up at 4am thinking we had something when getting an instance that put baremetal node properties in it and we proxied it to ironic, had a hard time getting back to sleep13:50
mriedemwas a nightmare13:50
*** akshai has quit IRC13:52
*** ametts has joined #openstack-nova13:52
*** penick has quit IRC13:52
*** gagehugo has joined #openstack-nova13:53
sdaguedude, you need this vacation13:53
mriedemi just need to review fewer api-ref patches13:53
*** shaohe_feng has quit IRC13:53
sdagueon the upside - https://web.archive.org/web/20051016203842/http://diveintomark.org/archives/2004/08/16/specs is pretty funny, and relevant to spec review13:53
sdaguemriedem: well, only 60% left... ;)13:53
*** shaohe_feng has joined #openstack-nova13:54
*** belmoreira has quit IRC13:55
*** belmoreira has joined #openstack-nova13:55
mriedem+2 on the api deprecation spec13:57
mriedemlooks like i need to go back over the glance v2 spec after the meeting then https://review.openstack.org/#/c/301741/13:58
mriedemwas +2 on it before bed13:58
mdboothdiana_clarke: It's just Lvm and Flat which don't use the image cache for generated disks, btw.13:58
mdboothThe other 3 do.13:58
mriedemsdague: does this need to be re-kicked? https://review.openstack.org/#/c/316298/13:59
mriedemdependencies are all merged13:59
diana_clarkemdbooth: thanks for checking13:59
sdaguemriedem: looking13:59
mriedemnova meeting in #openstack-meeting14:00
PaulMurraymriedem, sdague I was looking at that just now - I think so14:00
sdaguemriedem: yeh, project-config with depends on needs a A state flip to trigger14:00
sdaguebecause ... details14:00
*** jistr|bbl is now known as jistr14:00
*** akshai has joined #openstack-nova14:00
*** smatzek has joined #openstack-nova14:01
*** woodster_ has joined #openstack-nova14:02
openstackgerritZhenyu Zheng proposed openstack/python-novaclient: Use tempes.lib instead of tempes_lib  https://review.openstack.org/31858514:03
*** shaohe_feng has quit IRC14:03
*** tbachman has quit IRC14:04
*** shaohe_feng has joined #openstack-nova14:04
*** ducttape_ has joined #openstack-nova14:05
*** ducttape_ has quit IRC14:07
*** salv-orlando has joined #openstack-nova14:07
*** ducttape_ has joined #openstack-nova14:07
openstackgerritHans Lindgren proposed openstack/nova: Make flavor-manage api call destroy with Flavor object  https://review.openstack.org/24429114:07
*** ducttape_ has quit IRC14:08
*** ducttape_ has joined #openstack-nova14:08
*** raj_singh_ has quit IRC14:08
openstackgerritHans Lindgren proposed openstack/nova: Replace flavors.get_all_flavors_sorted_list() with object call  https://review.openstack.org/26880114:10
openstackgerritHans Lindgren proposed openstack/nova: Refactor and objectify flavor fakes used in api tests  https://review.openstack.org/26880014:10
*** mgoddard_ has quit IRC14:11
*** mgoddard has joined #openstack-nova14:11
*** salv-orlando has quit IRC14:11
*** shaohe_feng has quit IRC14:13
*** shaohe_feng has joined #openstack-nova14:14
*** marksei has joined #openstack-nova14:17
markus_zsdague: Should the functional tests in Nova hit through to the underlying libvirt/kvm? Is there a general rule about the borders of these tests?14:18
*** cfriesen has joined #openstack-nova14:18
mriedemmarkus_z: the functional tests don't hit a real libvirt14:18
markus_zsdague: The background is, that we executed them on s390x and some failed14:18
sdaguemarkus_z: they don't touch real libvirt14:19
*** marksei has quit IRC14:20
*** vladikr has quit IRC14:20
*** marksei has joined #openstack-nova14:20
openstackgerritDiana Clarke proposed openstack/nova: Ploop: add create_from_image & create_from_func  https://review.openstack.org/31768914:20
markus_zmriedem: sdague: OK, then I assume that a mock of the arch is missing. Some code of s390x was executed.14:20
*** marksei has quit IRC14:20
*** unicell1 has joined #openstack-nova14:20
*** unicell has quit IRC14:21
*** wangfeng has joined #openstack-nova14:21
*** xenogear has joined #openstack-nova14:21
openstackgerritDiana Clarke proposed openstack/nova: Rbd: add create_from_image & create_from_func  https://review.openstack.org/31769414:21
openstackgerritDiana Clarke proposed openstack/nova: Flat: add create_from_image & create_from_func  https://review.openstack.org/31770114:22
markus_zI dig deeper into it. The unit test "LibvirtConnTestCase.test_init_host" also executed s390x code paths.14:22
openstackgerritAnkit Agrawal proposed openstack/nova: Move is_volume_backed_instance to compute.utils  https://review.openstack.org/27048214:23
openstackgerritAnkit Agrawal proposed openstack/nova: Fix evaluation of host disk usage by volume-backed instances  https://review.openstack.org/20087014:23
*** shaohe_feng has quit IRC14:24
*** AmyLee7 has quit IRC14:24
*** syjulian has quit IRC14:24
*** shaohe_feng has joined #openstack-nova14:25
*** jinxing has quit IRC14:25
mriedemmoshele: you're not in the nova meeting14:26
moshelemriedem: is it now?14:26
*** edand has joined #openstack-nova14:27
mriedemyup14:27
*** belmoreira has quit IRC14:27
alaskisdague: the oslo.policy series is at https://review.openstack.org/#/c/313141/, seems to be in shape and is just waiting on more reviewers14:27
*** sajeesh has joined #openstack-nova14:28
*** realitix has quit IRC14:28
*** ctrath has quit IRC14:28
*** jinxing has joined #openstack-nova14:29
*** fawadkhaliq has joined #openstack-nova14:29
*** edand_ has joined #openstack-nova14:30
*** mtanino has joined #openstack-nova14:31
*** edand has quit IRC14:31
openstackgerritBob Ball proposed openstack/nova-specs: XenAPI: Support Nova services independently from the hypervisor  https://review.openstack.org/27745214:31
*** diogogmt has joined #openstack-nova14:33
*** fawadkhaliq has quit IRC14:33
*** iceyao has joined #openstack-nova14:33
*** pushkaru has joined #openstack-nova14:34
*** shaohe_feng has quit IRC14:34
*** shaohe_feng has joined #openstack-nova14:34
*** penick has joined #openstack-nova14:37
*** raunak has joined #openstack-nova14:37
*** belmoreira has joined #openstack-nova14:38
*** raunak has quit IRC14:39
*** prateek has quit IRC14:39
*** penick has quit IRC14:39
openstackgerritClaudiu Belu proposed openstack/nova: Hyper-V: Implement nova rescue  https://review.openstack.org/12715914:40
*** AmyLee7 has joined #openstack-nova14:40
*** AmyLee7 has quit IRC14:40
*** penick has joined #openstack-nova14:41
*** irenab has joined #openstack-nova14:41
*** yamahata has joined #openstack-nova14:43
*** ivase has quit IRC14:43
*** roeyc has quit IRC14:43
*** sdake_ has joined #openstack-nova14:44
*** shaohe_feng has quit IRC14:44
*** shaohe_feng has joined #openstack-nova14:45
*** unicell1 has quit IRC14:46
*** sridharg has quit IRC14:46
cdentdansmith: I need your help with something when you have a moment. I'm trying to figure out how to deal with lazy relationships as in: https://review.openstack.org/#/c/275432/14:46
moshelemriedem: hi14:46
dansmithcdent: okay, let me get this change pushed up first14:47
cdentdansmith: thanks14:47
mriedemmoshele: hey14:47
openstackgerritDiana Clarke proposed openstack/nova: Ploop: add create_from_image & create_from_func  https://review.openstack.org/31768914:47
moshelemriedem: I found out the the resize is race it you do it fast enough it doesn't release the old resource14:47
flip214mriedem: ping, about https://blueprints.launchpad.net/cinder/+spec/drbd-transport14:48
moshelemriedem:  by fast I mean small image and resize-confirm it very fast14:48
*** jlanoux has quit IRC14:48
*** ivase has joined #openstack-nova14:49
mriedemmoshele: this is resize in general you said?14:49
moshelemriedem: yes14:49
*** salv-orlando has joined #openstack-nova14:49
*** tonytan4ever has joined #openstack-nova14:49
mriedemand doesn't release which old resource?14:49
mriedempci?14:49
mriedemor doesn't cleanup from the old host?14:49
*** whenry has quit IRC14:50
mriedemflip214: looks like the drbd connector went into os-brick version 0.7.014:51
mriedemflip214: given we require >=1.3.0 http://git.openstack.org/cgit/openstack/requirements/tree/global-requirements.txt#n118 we should be ok there14:51
moshelemriedem: the following section will not execute https://github.com/openstack/nova/blob/master/nova/compute/resource_tracker.py#L315-L32614:51
*** penick has quit IRC14:51
flip214mriedem: yes, thanks.14:52
*** dedery has quit IRC14:52
*** rk4n_ has quit IRC14:52
mriedemflip214: want to create a nova bluerpint for the nova changes? https://blueprints.launchpad.net/nova/+spec/drbd-transport14:52
flip214but the nova changes from https://review.openstack.org/#/c/256292/ were rejected..14:52
moshelemriedem: so it won't remove the usage of the old guest14:52
*** ivase has quit IRC14:53
*** fesp has joined #openstack-nova14:53
flip214mriedem: I had one: https://blueprints.launchpad.net/nova/+spec/drbd-driver but it never got looked at, AFAICT14:54
*** jdurgin1 has joined #openstack-nova14:54
mriedemflip214: ok, please update that with the details and link to the nova path14:54
mriedem*patch14:54
flip214the spec that john asked about was https://review.openstack.org/#/c/134153/1414:54
mriedemflip214: also, i left comments in your nova change14:54
flip214thanks14:54
mriedemflip214: we don't need the spec14:54
*** shaohe_feng has quit IRC14:54
cdentdansmith: gonna switch locations, back in 15 min, thanks14:55
*** cdent has quit IRC14:55
mriedemflip214: the biggest issue you're going to have is that you can't add a new compute rootwrap filter to nova for this, you need to use oslo.privsep14:55
*** edand_ has quit IRC14:55
*** shaohe_feng has joined #openstack-nova14:55
mriedemi left a comment in the nova change about this14:55
mriedemos-brick is using privsep now14:55
*** unicell has joined #openstack-nova14:55
mriedemand at the summit we said nova wouldn't take any new compute rootwrap filters, anything new needs to be using privsep14:55
*** ivase_ has joined #openstack-nova14:56
mdboothdansmith: So, interestingly I completely misread the locking code in imagebackend. It actually never locks an instance disk at all, only the image cache file.14:56
danpbmriedem:  os-vif is using privsetp too !14:56
dansmithmdbooth: okay14:56
*** whenry has joined #openstack-nova14:56
mdboothdansmith: I'm still processing that :)14:56
mriedemdanpb: yeah14:57
dansmithmdbooth: the terminology in there is definitely confusing as it's hard to tell whether we're talking about images, disks, etc at any point in there14:57
mdboothSo, compatibility-wise, I need to ensure that the locking in ImageCacheLocalDir is compatible with the locking in the various backends.14:57
flip214mriedem: https://blueprints.launchpad.net/nova/+spec/drbd-driver updated.14:57
dansmithyou know this of course14:57
mriedemflip214: https://etherpad.openstack.org/p/newton-privsep was from the austin summit, has some details on privsep14:57
mdboothdansmith: Right. It also creates the same lock in 2 different ways.14:57
flip214and yes, I'll switch to privsep, now that this has stabilized... is wasn't when I submitted this14:57
flip214thanks for the pointer14:58
mdboothI'd misread that to be disk and image cache.14:58
sdaguealaski: who on the oslo side is needed to review this?14:58
mriedemdanpb: if kashyap has already mentioned it, the live migration job on xenial nodes is failing to use the gate64 cpu model that gets setup by devstack14:58
mdboothdansmith: I'm still trying to decide whether we *should* be locking the instance disk.14:58
kashyapmriedem: I didn't.  Today by EOD I wanted to write current status to -devel list14:58
mriedemdanpb: would be good to get that sorted out so we can use the live migration job again. we moved it to xenial nodes to use newer libvirt/qemu.14:59
mdboothi.e. what concurrent operations might reasonably happen to it?14:59
mriedemkashyap: ok, thanks14:59
mriedemmoshele: i have to run to a meeting, will read traceback when i get back14:59
mdboothWell, reasonably or unreasonably, tbh.14:59
*** mriedem is now known as mriedem_meeting14:59
kashyapmriedem: Today was a yak-shave day upgrading systems to F24, and trying to fix some build deps for various components14:59
*** tyrola has joined #openstack-nova14:59
*** jaosorior has quit IRC14:59
alaskisdague: I don't actually know, because the oslo review team is one large group I don't know who looks at policy changes14:59
sdaguehttps://review.openstack.org/#/admin/groups/556,members15:00
moshelemriedem: I will send a mail about in the ML15:00
sdaguealaski: it's actually a lot of keystone core15:00
danpbmriedem_meeting: no, first i'm hearing of it15:00
*** danpb has left #openstack-nova15:00
*** jaosorior has joined #openstack-nova15:00
*** danpb has joined #openstack-nova15:00
alaskisdague: ahh, I didn't realize they split out like that as well15:00
*** mgoddard_ has joined #openstack-nova15:00
kashyapdanpb: I'll write to upstream -devel list with current notes.15:00
sdagueit looks like keystone core + oslo core is the review team15:00
danpbkashyap: what's the deal with cpu model ?15:00
kashyapdanpb: Despite having 'gate64' CPU model in cpu_map.xml15:01
openstackgerritPavel Kholkin proposed openstack/nova: [WIP]: nova-api-hypervisor-cpu-info change  https://review.openstack.org/30418015:01
kashyapdanpb: libvirt fails to detect it:15:01
*** ctrath has joined #openstack-nova15:01
flip214mriedem_meeting: JFI, we edited https://blueprints.launchpad.net/nova/+spec/drbd-driver in parallel, my changes are lost ;[15:01
kashyap[...]15:02
kashyap2016-05-13 21:33:08.194 15003 ERROR nova.compute.manager [instance: 9bc566d2-c856-4f36-bc7c-48d82870aa30] libvirtError: internal error: early end of file from monitor, possible problem: 2016-05-13T21:33:07.143245Z qemu-system-x86_64: Unable to find CPU definition: gate6415:02
kashyap2016-05-13 21:33:43.625+0000: 3010: warning : x86Decode:1583 : Preferred CPU model gate64 not allowed by hypervisor; closest supported model will be used15:02
kashyapdanpb: And the associated libvirt debug log (upcoming) doesn't record anything from the CPU driver.15:02
kashyaplibvirt debug log: logs.openstack.org/03/315703/1/experimental/gate-tempest-dsvm-multinode-live-migration/1288a98/logs/libvirt/libvirtd.txt.gz15:03
danpbwhere does this 'gate64' cpu model come from ?15:03
*** jaosorior has quit IRC15:03
kashyapdanpb: From here: https://review.openstack.org/#/c/168407/9/tools/cpu_map_update.py15:03
*** dstepanenko__ has quit IRC15:03
openstackgerritEd Leafe proposed openstack/nova: Correct the network config option help text  https://review.openstack.org/31872615:03
edleafedansmith: ^^15:03
dansmithedleafe: thanks15:03
*** syjulian has joined #openstack-nova15:03
kashyapdanpb: It was added (by Infra folks) a year ago to make live migration work15:03
*** jaosorior has joined #openstack-nova15:03
danpbkashyap: soo must be a case that the xenial nodes are not exposing one of the cpu flags list there15:04
*** mgoddard has quit IRC15:04
kashyapdanpb: Yeah, yesterday I went looking into Xenial's libvirt source to see if they're carrying any non-upstream patches:15:04
kashyapThese are their patches list http://paste.openstack.org/show/497514/15:05
*** shaohe_feng has quit IRC15:05
kashyapAnd I don't see anything CPU models related stuff.  (That "dont-include-non-migrateable-features-in-host-model" patch is already upstream.)15:05
danpbwell patches aren't neccessarily the problem - you've no way of knowing what cpu model they've configured their guest to expose15:05
*** shaohe_feng has joined #openstack-nova15:05
gibirlrossit: regarding the ovo registration in https://review.openstack.org/#/c/309454/15:05
danpbyou'll need to look at the virsh capabilities output to see what nova is seeing15:06
danpband identify the missing cpu flag(s)15:06
kashyapdanpb: So, should we request -infra folks the missing CPU flags?15:06
kashyaps/the/to add the/15:06
gibirlrossit: my test result disagres with your comment that ovo base classes  without register_if(False) doesn't work15:06
danpbno the CI needs to be configured to cope with whatever the nodes expose15:06
*** jamesdenton has joined #openstack-nova15:07
gibirlrossit: can it be that it is enough to call register_if only on the concrete classes and that will take care of the initialization of the base classes too?15:07
kashyapdanpb: Okay, I'll make it a point to check what `virsh capabilities` expose on the 'Xenial' release.15:08
kashyapThanks!15:08
dansmithmriedem_meeting: when I add an if clause for an index because of db2, does that mean I have to leave it out of the model? otherwise the test complains about them not being in sync15:09
*** jaosorior has quit IRC15:10
*** doug-fish has joined #openstack-nova15:10
*** claudiub has quit IRC15:10
*** adisky has quit IRC15:10
*** wangfeng has quit IRC15:10
*** links has joined #openstack-nova15:11
*** zz_dimtruck is now known as dimtruck15:12
*** doug-fish has quit IRC15:12
sdaguealaski: can you poke into #openstack-dev, talking with ayoung on oslo.policy reviews there15:12
ayoungsdague, thanks15:12
openstackgerritDaniel Berrange proposed openstack/nova: os-vif: convert libvirt driver to use os-vif library  https://review.openstack.org/26967215:13
openstackgerritalex proposed openstack/nova: Remove destroy instance which states is "accepted" after evacuate.  https://review.openstack.org/31873115:13
*** heyongli has quit IRC15:13
*** alex_xu has quit IRC15:13
*** cdent has joined #openstack-nova15:15
*** shaohe_feng has quit IRC15:15
siva_krishnandanpb I have few questions regarding libvirt. Can you please help me with that ?15:18
*** doug-fish has joined #openstack-nova15:20
*** jlanoux has joined #openstack-nova15:20
*** yamahata has quit IRC15:20
*** DaiDV has joined #openstack-nova15:21
danpbsiva_krishnan: possibly what's the q15:21
*** ijw has joined #openstack-nova15:22
siva_krishnandanpb can you share me your mail id, so that I can forward the mail that I sent to openstack-dev mailing list15:23
danpbi'm on openstack-dev already15:23
*** fawadkhaliq has joined #openstack-nova15:24
siva_krishnandanpb can you pull the mail with following line as subject [openstack-dev][nova]Libvirt error code for failing volume in nova  ?15:25
*** whenry has quit IRC15:25
dansmithmriedem_meeting: nevermind, I'm stupid15:25
*** whenry has joined #openstack-nova15:25
dansmithcdent: okay shoot15:26
openstackgerritDan Smith proposed openstack/nova: Add resource provider tables to the api database  https://review.openstack.org/31568115:26
openstackgerritDan Smith proposed openstack/nova: Make compute nodes update their own inventory records  https://review.openstack.org/31704915:26
openstackgerritDan Smith proposed openstack/nova: Make Inventory and ResourceProvider objects use the API DB instead  https://review.openstack.org/31568215:26
*** links has quit IRC15:26
*** akshai has quit IRC15:27
cdentdansmith: in https://review.openstack.org/#/c/275432/ the Inventory table as an orm.relationship for a resource_provider15:27
*** kolesovdv has quit IRC15:27
cdentthe way it is currently written that will be lazily loaded when asked15:27
cdentit is easy, however, for that lazy loading to happen after the session has gone out of scope15:27
cdentthrowing an error15:28
dansmithcdent: yeah, you should either load it explicitly on the query, or load it directly (not via sqla lazy load) later15:28
*** belmoreira has quit IRC15:28
cdentwhich is the fix for that: a) make it not be lazy b) engage the session when ever we want to pull on that lazy loaded attribute c) something else15:28
*** claudiub has joined #openstack-nova15:28
*** andreas_s has quit IRC15:29
dansmithcdent: your access to the model should be purely within the scope of an object operation, so you should know whether you're going to lazy-load it or not, right?15:29
cdentwhere this seems to be coming up with when use the results of getting an InventoryList15:29
*** hemnafk is now known as hemna15:30
cdentdansmith: I'm somewhat lost15:31
dansmithcdent: all who wander are not lost15:32
dansmithno, screwed that up15:32
dansmithnot all who wander.. meh whatever :)15:32
artom_dansmith, question about storing metadata in the database15:32
cdentthis has come up because I was moving around some tests for Allocation and AllocationList15:32
*** artom_ is now known as artom15:32
*** AmyLee7 has joined #openstack-nova15:32
dansmithcdent: can you point me at a paste or something?15:32
dansmithartom: okay15:33
cdentyeah, let me see if I can construct something15:33
*** dmk0202 has quit IRC15:33
artomdansmith, if we put the JSON blob in instance_extra, this means the API will pull something from the DB that's been put there by the virt driver. I recall you not being keen on the idea of the API barfing up JSON directly from the virt driver15:33
*** claudiub has quit IRC15:33
*** sudipto has quit IRC15:33
siva_krishnandanpb were you able to get hold off my mail ?15:33
dansmithartom: if the virt driver is putting an abstract model that the api barfs up that's cool15:33
*** akshai has joined #openstack-nova15:34
artomdansmith, but can an abstract model be stored in a column of instance_extra?15:34
artomWouldn't it need its own table?15:34
dansmithartom: what I mean by abstract model is something not driver-specific15:34
dansmithartom: of course, all of instance_extra is that :)15:34
*** ljxiash has quit IRC15:35
artomdansmith, right, but it's still a JSON blob that's served up to the user (with potentially some schema validation)15:35
artomdansmith, and thus something that the virt layer has generated becomes part of the API15:35
dansmithartom: no, when I say store a json blob I mean a serialized obejct15:35
dansmithartom: the object provides the generic model which would be used by all the virt drivers15:36
dansmithartom: that evokes another question.. surely we need to implement this interface for the other drivers so that they're not exposing "no devices" to the guest right?15:36
*** kfarr has joined #openstack-nova15:37
artomdansmith, aha, I see - so it would go virt layer -> metadata objects -> serialize -> DB -> unserialize -> metadata API -> construct user-facing jSON -> chuck it at user15:37
dansmithartom: i.e. it's not okay to provide a tag for a nic and have vmware or xenapi guests say "you have no tagged devices" to the guest, IMHO15:37
artomdansmith, yeah, we need a way of indicating that it's not implemented15:37
dansmithartom: yes15:37
artomdansmith, for that wouldn't it be better to make the device tags its own resource? We would then be able to use return codes15:38
dansmithartom: I don't think you get to make it not implemented by some drivers.. it's metadata and generic, so it should be implemented by them as well15:38
artom500 not implemented, 404 if there are no tagged devices for example15:38
artomdansmith, well, should, yes, but until it is...15:38
dansmithdoes any of the rest of the metadata api behave that way?15:38
*** vladikr has joined #openstack-nova15:38
artomdansmith, utterly not :/15:39
dansmithartom: well, the other option is to not expose it in the metadata api until you have it implemented everywhere15:39
artomdansmith, that could take while... where 'a while' may end up being infinity ;)15:40
*** Swami has joined #openstack-nova15:40
danpbi think blocking implementation in any driver until all implement it is rather overly extreme15:40
dansmithright but if you don't do it as part of this, it'll never happen and we have guest-facing api that differs (a lot) between drivers15:40
danpbif the virt driver could report whether they support this or not15:40
dansmithdanpb: if this wasn't guest-facing api I'd agree15:40
danpbthen we can reject it if the user specifies it when booting a guest15:40
danpbeg, nova/virt/driver.py could have a capability flag to indicate if it supports device tagging15:41
dansmithdanpb: and for multi-driver clouds?15:41
danpbthat way when we get a public API request to boot a guest, or hotplug a disk / nic we can report whether its supportable or not15:41
dansmithI think that's problematic because of the guest-facing aspect of this15:41
*** mjura has quit IRC15:42
dansmithI don't think it's okay to not have xenapi and vmware and hyperv implement this and thus have a hole in metadata15:42
danpbi don't see that as being a real problem15:42
*** fungi has joined #openstack-nova15:42
*** clarkb has joined #openstack-nova15:42
clarkbkashyap: hello15:42
kashyapclarkb: Hi15:42
danpbi mean it would be no different to the same guest image being booted on Newton vs Mitaka15:43
kashyapdanpb: I see you're in a different conversatoin here, when you get a sec:  clarkb from -infra tells us these are the CPU flags from `Xenial` -- http://paste.openstack.org/show/497747/15:43
dansmithdanpb: but once Ocata comes along, and you've got clearly-ocata things in metadata, but are missing things that were added in newton, then it becomes different15:43
danpbif the guest image needs this metadata to be provided, then perhaps we should tag the guest image itself as requiring it15:43
danpbeg   img_dev_metadata_version=1.015:44
dansmithmordred: I bet you have a reality-based opinion here15:44
danpbto indicate that this image is only bootable if it can be given device metdata version 1.015:44
mordreduhoh15:44
dansmithmordred: let me catch you up:15:44
danpbif the hypervisor couldn't honour that then it can raise an error15:44
kashyapclarkb: On -infra, which file location has changed you say?15:44
kashyap(Or likely have changed.)15:45
dansmithmordred: we are working on a metadata chunk that will let you tag devices like nics and disks when you create them, and then the guest will be able to see what devices the hypevisor hooked them up to, by tag, in metadata15:45
clarkbkashyap: I dont know that it has changed, I just think that is more likely. The one which libvirt reads the cpu model defs from15:45
dansmithmordred: so you can see that the second nic is the one you tagged as "private"15:45
dansmithmordred: or that sdb is "fordatabasestuff"15:45
kashyapYeah, I corrected myself with the hedge 'likely'.15:45
dansmithmordred: so I think it's very important that we keep that view consistent to the guests, and not have it be missing because nobody wanted to go implement this for, say, the xenapi driver15:46
mordreddansmith: yes please. this sounds fantastic15:46
mordreddansmith: agree15:46
dansmithmordred: I feel like our metadata api is our guest-facing ABI and should be _extremely_ stable and consistent15:46
clarkbkashyap: another possibility is that we are getting lobvirt to reread the config properly. On fedora service reload was sufficient, on truaty we had to stop start, maybe now we need something different15:46
mordreddansmith: I should not know you're running xen ever15:46
mordreddansmith: +1000015:46
kashyapclarkb: It's cpu_maps.xml. You can see 'gate64' here: http://logs.openstack.org/03/315703/1/experimental/gate-tempest-dsvm-multinode-live-migration/1288a98/logs/libvirt/cpu_map.xml15:46
danpbdansmith: we've definitely making them consistent as we've defined a clear general object model for the format15:46
danpbi don't see why we need a big bang to implement it in every single driver at the same time15:46
mordredwell....15:47
dansmithdanpb: right but just not having that present because we're on xen and nobody wanted to write that code is not okay to me.. either we support that driver or we don't, and this is not a hypervisor-doesn't-support-this-thing kind of deal15:47
mordredso as an end user - I'm still going to have to deal with clouds that have it and clouds that don't15:47
dansmithdanpb: it won't happen if we land the metadata api and move on15:47
dansmithmordred: right, but after some point not having it but having newer things, because the provider is running xen is inconsistent15:48
kashyapclarkb: Let me write a quick email to the list, so people can chime in there -- it's a bit more cleaner15:48
*** klkumar_ has quit IRC15:48
mordredwhich is probably handled by the metadata api date version things already?15:48
clarkbkashyap: ok... I dont think we need any of this15:48
clarkbsomeone just needs to boot a xenial host, fix it, then patch devstack15:48
dansmithmordred: we don't bump that date for our own non-EC2 additions to that, AFAIK :(15:48
mordredoh15:48
mordredwell nevermind then15:48
clarkbIdeally someone involved with live migrate as this is the only reason we have the workaround15:48
dansmithmordred: so all you have to go on is "if this thing is there, then I assume you're newer" and thus making that inconsistent across drivers is... uncool15:49
danpbthe device role tag metadata document is itself versioned independantly of the metadata api though15:49
kashyapclarkb: I can try it on a 'Xenial' VM.15:49
danpbas it has an associated oslo.vo  version15:49
mordredis the new thing a whole new document?15:49
dansmithdanpb: viewable from the metadata api? and what version would you show for "not present" ?15:49
mordredlike network_info.json ?15:49
dansmithmordred: does it matter if it's a separate document?15:50
dansmithmordred: meaning, aren't people going to look for something super new and assume that things older than that are present?15:50
sdaguedansmith: isn't there already a mechanism for that in metadata service, because you get a new directory per version15:50
cdentdansmith: I seem to have more variables present than I thought, and I've got a meeting in a few minutes, so I'll cry in your general direction later if I must.15:51
*** sacharya has joined #openstack-nova15:51
*** whenry has quit IRC15:51
dansmithsdague: mechanism for what? versioning?15:51
sdagueyeh15:51
danpbdansmith: why does a version matter if it isn't present15:51
artomdansmith, well, if it's a separate document return codes could be used to indicate status15:51
dansmithsdague: I think we don't do that for our own stuff do we? I've never known us to change it15:51
danpbits no different to just running on any existing nova version if the driver doesn't export it15:51
mordreddanpb: nah - I think I'd look for the thing and only assume from having found it that that thing is present15:51
mordredgah15:51
mordreddansmith: ^^15:51
*** ctrath has quit IRC15:51
dansmithdanpb: I'm not saying it does15:52
mordreddansmith: like, I would not assume network_info.json is going to exist if I find the new devive mapping info in metadata.json15:52
mordredI'll be testing each thing independently15:52
dansmithmordred: okay15:52
*** jaypipes has joined #openstack-nova15:52
danpbBTW, I really *do* totally want the other virt drivers to implement this15:52
dansmithmordred: so because this hasn't existed since the beginning, it's okay to never add support for the other hypervisors for this? since you'll always have to deal with it being missing?15:52
*** ctrath has joined #openstack-nova15:53
mordredya15:53
dansmithdanpb: I know, but it won't happen if we don't arrange for it15:53
mordredsince I have to deal with that anyway15:53
mordredbecuase of old clouds15:53
dansmithmordred: well, hang on15:53
danpbi just think that if we block on having that, then we're going to end up with zero virt drivers having it for several more releases15:53
mordreddansmith: I think it's a bad idea to not implement it everywhere15:53
dansmithmordred: you have asked for it via the api and we said okay, but we put you on a node that won't ever generate it15:53
mordredoh... so this is now interesting15:53
mordredI may change my mind here15:53
dansmithdanpb: yeah, I know, but we do things that take time to be consistent15:54
artomdansmith, well, hence danpb's idea of not booting it on a node with a virt driver that doesn't support it15:54
mordredthis is a new user-face rest api thing I just opted in to isn't it?15:54
dansmithartom: yeah, adding yet another thing you have to specify manually to get correct behavior15:54
mordreddansmith: in that case, then no, I do not think it's ok for some drivers to not implement it - that woudl be massively suck from an end user pov15:54
*** zenoway has quit IRC15:54
danpbdansmith: yeah, that's why i think we could do with an image property to indicate that this disk image needs this feature15:54
sdaguedansmith: I guess you are right, we only add things - https://github.com/openstack/nova/blob/cdfbb9a668fdcf289ffdfa5252d102e9d3e2ec35/nova/api/metadata/base.py#L325-L33015:54
dansmithmordred: right15:54
danpbso we can intelligently pick the right node to put it on, if there's a choice of some nodes with & without the feature15:54
dansmithdanpb: and I want fewer of those things15:54
mordredalso - image properties are exceptionally opaque interface for end users, fwiw15:54
dansmithmordred: right, that15:55
artomdansmith, well, it would only devolve to that case if the cloud has 0 nodes that support tagging15:55
sdaguemordred: that's mostly a documentation problem, right?15:55
artomOtherwise the scheduler should find a right one15:55
mordredsdague: api discoverability too15:55
danpbwe could avoid the image property if we can look at the actuall VM boot request to see if the user has actually requested tags or not15:55
dansmithartom: no, you have to specify it in order to make it work15:55
artomdansmith, specify what?15:55
mordredsdague: "what image properties does this cloud think are meaningful"15:55
sdaguemordred: ok, sure15:55
dansmithdanpb: and fail after we make it to the compute node because you asked for a thing you were told was available but will never be available on this cloud? that sucks :)15:56
dansmithI want to avoid someone booting three hundred attempts, all failed, before calling support and them saying "oh we don't do that thing. we run xenapi"15:56
sdaguedansmith: ++15:56
mordredsdague: did you know that you need to put "vm_mode: hvm" on images in rackspace to have them work with performance flavors? as a for instance15:56
*** zenoway has joined #openstack-nova15:56
dansmithespecially since implementing it in each driver is something we know we need to do, but nobody did it because we merged the api and moved on :)15:56
sdaguemordred: I did not know that. Sounds awesome.15:57
mordredor "xenapi_use_agent: false" if you don't want the boot to wait for the xenapi agent to phone home15:57
danpbthat's not a new problem though and we've not blocked stuff because of that15:57
dansmiththat, to me, is why the project owning a driver in tree is expensive.. because we should maintain it and not just  let it rot15:57
danpbwe've got countless features which are only implemneted in one driver15:57
dansmiththis matters more to me, because of the guest-facing api15:57
dansmithwhich we've not had before in terms of a virt-specific feature that affects metadata, that I know of15:58
*** iceyao has quit IRC15:58
danpbevery aspect of virt machine configuration is guest facing ABI15:58
danpbwhether its via metadata service or via the hardware platform, its all guest facing15:58
dansmithbut we control the metadata api15:58
dansmithand it's the thing we're trying to abstract, not the hypervisor stuff we don't control15:58
artomDon't we already have the situation where you fail to boot an instance if you ask for a NUMA topology or some PCI device that doesn't exist in that cloud?15:58
dansmithanyway, that's MHO15:58
*** armax has joined #openstack-nova15:59
artomWhat's different if we refuse to boot an instance because there are no hosts with drivers that support tagging?15:59
dansmithartom: capacity aside, you should fail fast in the scheduler for that15:59
artomdansmith, and it's impossible to fail fast in the scheduler with tagging?15:59
dansmithartom: no, we could write a filter probably, but the filter would be "some drivers haven't been maintained" and not "this device is not available"16:00
dansmithwhich is kinda embarrassing no? :)16:00
mriedem_meetingflip214: ok, want to update the nova bp again?16:00
*** mriedem_meeting is now known as mriedem16:00
*** Oku_OS is now known as Oku_OS-away16:00
auggygood morning!16:00
danpbdansmith: i don't think it is saying that they've not been maintained16:01
artomdansmith, well, we could disguise it as "that driver can't do it"16:01
*** sdake_ is now known as sdake16:01
artomOr "strongly prefers not to do it, on principle"16:01
artom;)16:01
danpbdansmith: it is just reflecting the reality that we don't have unlimited resources to implement features across every driver at once16:01
sfinucandansmith: Aye - more like "this isn't a priority for anyone using this driver"16:01
*** lakshmiS has joined #openstack-nova16:01
*** klkumar has joined #openstack-nova16:01
artomOr a 202 accepted that becomes a 200 when the implementation is finished :D16:01
danpband don't want to delay getting useful features to one driver that have 90% of openstack user survey share16:01
*** stewie925 has joined #openstack-nova16:01
*** mpavone has quit IRC16:01
dansmithsfinucan: not using, maintaining :)16:01
*** ctrath has quit IRC16:02
danpbwhile we implement it for another driver which has 5% user share and no immediate demand for the feature16:02
*** mgoddard_ has quit IRC16:02
*** mgoddard_ has joined #openstack-nova16:02
dansmithdanpb: it's the users that have the demand, and they don't control what hypervisor was chosen16:02
dansmiththat sets this apart from something like NUMA to me16:02
*** penick has joined #openstack-nova16:02
sfinucandansmith: Sounds like what was discussed at the operator meetup16:03
artomWait, who's demanding NUMA that are not users?16:03
danpbwell the immediate near term demand for the feature is from NFV use case16:03
sfinucanoperator: we need this, devs: provide contributions?16:03
danpband those users certainly do control what hypervisor their imags are running on16:03
*** marcusvrn_ has quit IRC16:03
danpbsure, other non-NFV users will benefit from this feature too16:03
dansmithdanpb: that is the worst argument ever :)16:03
mriedemflip214: the drbd bp is approved now, i've removed my -2 from the nova change, but still -1 for the comments inline16:04
dansmitheveryone who has ever booted with more than one disk or network will use this :)16:04
artomdansmith, not if you tell them the cloud's chosen virt driver doesn't have it implemented yet ;)16:04
*** rcernin has quit IRC16:04
vladikrdansmith, not necessarily, most probably they won't care about the order16:04
danpbdansmith: it really isn't  - we can give the feature to the people who really need it right now, and every one else gets it in 1 year time16:04
danpbdansmith: or we can give it to no one and make everyone wait a year16:05
dansmithvladikr: I completely disagree :)16:05
danpbits madness to artifically make everyone wait for it to be imeplemented across every driver16:05
sdaguehow much effort is it to get it added to hyperv?16:05
sdaguethat's sort of the #2 driver16:05
dansmithdanpb: are you and artom going to work on the other drivers after you land the api?16:06
artomThere *is* a chance people will just tag stuff "butts" because they can ;)16:06
dansmithsdague: my other concern is that these things get designed with no input from the others, and then there is a problem16:06
*** jdurgin1 has quit IRC16:06
danpbdansmith: no, because we're not domain experts in hyperv apis16:06
*** igordcard has quit IRC16:06
dansmithdanpb: right, so it'll never get done for those other people16:06
*** mriedem is now known as mriedem_afk16:06
artomdansmith, I personally wouldn't mind, but like danbp has said, I have 0 clue about other drivers16:06
danpbdansmith: this has been throw spec review where other driver maintainers commented that the design was acceptrable16:06
artomSo it would be a long and error fraught process16:06
sdaguedanpb: sure, but you could coordinate with the experts and try to get folks signed up for doing it this cycle as well16:06
danpband even indicated they'd be willing to support it16:07
sdaguethis seems universal enough that having a consistent story would be really good16:07
*** unicell has quit IRC16:07
dansmithdanpb: yeah and we've deviated from the spec's original design quite a bit remember, especially around when in the process we need to generate this data16:07
danpbsdague: different driver teams have different priorities for developing features16:07
danpbdansmith: my latest comments on the reviews were to actually go back to what the spec design said16:08
*** unicell has joined #openstack-nova16:08
danpbas the current patches up there are flawed16:08
sdaguedanpb: sure, but we should try to coordinate16:08
dansmithdanpb: of making an rpc call from the metadata api?16:08
danpbdansmith: no, in terms of the way we're representing the data inside nova16:08
*** ccard_ has joined #openstack-nova16:08
artomsdague, I don't think anyone's disagreeing that we shouldn't coordinate, but the reality on the ground means that coordinating would probably delay the feature landing by a significant amount16:09
danpball the versioned object model we defined in the spec & originally implemented got dropped in latest patches16:09
dansmithdanpb: heh, okay16:09
stewie925good morning all16:09
artomSo, I think we have a plan to bo back to the original metadata objects representation while still using the database16:10
danpbanyway I think its a great disservice to our users if we artificially delay this until all drivers implement it16:10
danpbi don't really have much more to say16:10
artomThe question we're all yelling about now is what to do with unequal driver support16:10
*** ctrath has joined #openstack-nova16:10
artomEither indicate it in the API somehow16:10
artomOr wait until all drivers have it16:10
artomPersonally I prefer indicating16:11
artomI think danpb agrees?16:11
*** sahid has quit IRC16:11
*** fawadkhaliq has quit IRC16:11
danpbthe approved spec already says that we'll do it in the libvirt driver for QEMU only16:11
dansmithartom: if we had any indication that there were implementations for the other drivers coming, then I'd be less concerned16:11
danpbunless someone wants to submit a patch to unapprove the spec16:11
stewie925I just attempted to create an instance in cell environment and am getting error message ""No valid host was found. There are not enough hosts available."16:11
danpb"The initial implementation will be done for the Libvirt driver, however, other virt driver maintainers are encouraged to provide the same functionality."16:12
stewie925but I have captured the n-api and n-sch logs and theyre in http://paste.openstack.org/show/497753/16:12
danpbhttp://specs.openstack.org/openstack/nova-specs/specs/newton/approved/virt-device-role-tagging.html16:12
*** ccard has quit IRC16:12
*** nkrinner has quit IRC16:12
artomdansmith, so how would be get said indications (or obtain evidence of lack thereof?)16:12
danpbso its not like this is a changein plan in any way16:12
*** nkrinner has joined #openstack-nova16:12
*** nkrinner is now known as nkrinner_afk16:12
artomMailing list?16:12
rlrossitgibi: sorry, I was in meetings16:13
dansmithdanpb: surely you appreciate that problems become clearer once you start implementing a thing, and that if you land a spec to do something crazy, it's okay to reconsider when other concerns are surfaced, right?16:13
rlrossitgibi: but it looks like https://github.com/openstack/oslo.versionedobjects/blob/master/oslo_versionedobjects/base.py#L47-L51 says it'll take fields from all supers and apply them as setters/getters on itself, but it won't do it on the super classes16:13
rlrossitmy guess is you aren't using the super classes at all, so you aren't affected by it16:13
dansmithdanpb: so pointing to the letter of what was approved doesn't really add much to the argument, IMHO16:13
dansmithdanpb: I'm just saying that this concerns me and that I think the reality will be that it never gets implemented elsewhere and that I think that's a problem..16:14
danpbas i mentioned earlier though, this was discussed during the review of the spec and feedback from other virt driver maintainers explicitly solicited16:14
openstackgerritDiana Clarke proposed openstack/nova: Rbd: add create_from_image & create_from_func  https://review.openstack.org/31769416:15
openstackgerritDiana Clarke proposed openstack/nova: Flat: add create_from_image & create_from_func  https://review.openstack.org/31770116:15
sdaguedanpb: I think what's being asked is a bunch of the other virt driver authors said they liked it back in Nov 201516:15
danpbdansmith: if it never gets implemented in a particular driver imho that's more a reflection on the state of that driver's supportability than anything else16:15
sdaguedanpb: and that the ask is please *also* spend some time rounding up those folks and trying to get them to prioritize it16:16
dansmithdanpb: I don't disagree16:16
sdagueto make for a good community land of the feature in multiple drivers during this release16:16
*** claudiub has joined #openstack-nova16:16
sdagueit's an organizing overhead for sure, but it makes all of nova better16:16
dansmithclarkb: ohai16:16
artomdansmith, there was some indications in the original Mitaka spec review from other drivers IIRC16:16
dansmithoops16:16
dansmithclaudiub: ohai16:17
danpbi don't think we should be blocking  features on other well maintained drivers because of possibility that other virt driver maintainers may drop the ball16:17
artomdanpb, example from BobBall here: https://review.openstack.org/#/c/195662/8/specs/mitaka/approved/virt-device-role-tagging.rst@32416:17
sdagueartom: right, so please talk with those folks and try to get them on board. Is that an unreasonable ask?16:17
claudiubdansmith: hellou16:17
dansmithdanpb: and I think that we're not just writing a proxy to libvirt here16:17
* danpb leaves to get some useful work done now16:17
*** danpb has left #openstack-nova16:17
claudiubam I in trouble? :)16:17
dansmithclaudiub: wondering what your plans are for implementing the device tagging spec, if any?16:17
dansmithclaudiub: I know you were interested in it during review, but I also kinda hope we can land more than just one driver's support this cycle16:18
artomsdague, totally reasonable :)16:18
claudiubdansmith: it's actually in my plans for this cycle.16:18
dansmithclaudiub: sweet,t hat makes me feel better16:18
artomsdague, I would just argue that an all or nothing approach to landing this is too extreme, and if we adopt it we can essentially forget the whole thing entirely16:18
sdagueartom: I think that just doing some leg work in talking to those folks and presenting what peoples plans are16:18
sdaguewould be great16:18
sdagueartom: I honestly don't think all or nothing is what we're talking about. It's > 1 would be really good16:19
claudiubdansmith: just one question though, can the hyper-v implementation piggyback on another spec, or do I have to submit another one?16:19
dansmithsdague: artom: right..agree..16:19
sdagueand make everyone happier in getting the feature prioritized to land16:19
dansmithclaudiub: I think the spec is landed and covers you16:19
openstackgerritKaren Bradshaw proposed openstack/nova: api-ref, os-fping.inc  https://review.openstack.org/31876416:19
claudiubyeay, music to my ears then. :)16:19
artomsdague, dansmith, ah, so if we have at least one other driver in process of implementing this...16:20
dansmithclaudiub: we'd love to see even some prototype code up in parallel with the libvirt bits16:20
BobBallWe'll implement in this cycle too, of course :)16:20
BobBallartom: ^^16:20
artomYeah, that's a reasonnable compromise16:20
sdagueBobBall: great16:20
dansmithartom: all we're trying to do is get that ball rolling... and as you see, it's not that hard :)16:20
artomdansmith, oh for sure, and thanks for that :)16:20
*** vishwanathj has joined #openstack-nova16:20
artomI was afraid the argument was all or nothing16:21
sdagueartom: yeh, I think the libvirt team is the team most ahead on this one, and if they took the extra responsibility of just generally touching base with other driver authors and keeping an eye on what's moving forward, that would be great16:21
dansmithBobBall: any idea when? it'd be cool to have some code up in parallel to the libvirt bits16:21
artomSo, if we go back to the technical side of things, this means not all drivers will land this at the same time, yes? So we need a way to indicate/control tags and booting with tags16:22
artomCan we get the API side of that sorted out?16:22
*** yamahata has joined #openstack-nova16:22
dansmithartom: I don't want to have to add an image flag for it, and I think we're probably on a path to not need to16:23
BobBalldansmith: I've not been tracking this discussion very closely... When will the libvirt bits be up? :) Very happy to work for getting them up in parallel16:23
dansmithartom: but yeah, at least we should fail the boot if we have tags and no support16:23
dansmithBobBall: the bits are up, but have been redone multiple times16:23
vladikrhttps://review.openstack.org/#/q/topic:bp/virt-device-role-tagging16:24
vladikrBobBall, ^^16:24
dansmithBobBall: I'd like the guys working on the libvirt side to coordinate with the other driver owners to synchronize exactly that.. when the model is up that you can implement16:24
artomBobBall, err, yeah, we need to re-jigger the bits, so wait for that please :)16:24
BobBallI was looking at the wrong tag :)16:24
BobBallYeah - we're up for that dansmith.  Want someone to talk to now artom, so they can get the lay of the land, etc?16:25
dansmithBobBall: I think the aforementioned rejiggering might should happen first to avoid confusion, but maybe try to do that sync next week if you'll be around?16:25
BobBallI will indeed16:26
artomBobBall, sure - I won't engage just yet, as we're in a re-write ourselves, but I definitely need a list of folks form various drivers to harass ;)16:26
dansmithI actually don't know who is working on vmware these days16:26
dansmithwhich is probably not good16:26
artomclaudiub is on the hyperv side of things, yar?16:27
BobBallartom: My team is in PRC, so I'm an easy list of 1 for this sort of timezone (I'm in UK though - so won't be around much later than this)16:27
dansmithartom: yes16:27
artomBobBall, PRC as in... Taiwan?16:27
claudiubjaypipes: hellou, you around?16:27
dansmithartom: if xen and hyperv are up at least in prototype form I think we're in much less-concerned mode16:27
sdaguewell, honestly, if vmware is back to rotting in tree, that's what it is. Having hyperv, xen, libvirt would be pretty solid16:27
dansmithyeah16:28
BobBallartom: Peoples Republic of China :)16:28
artomNo wait, the other "proper" China is PRC16:28
dansmithsdague: and if so, we probably need to be poking people16:28
*** tonytan4ever has quit IRC16:28
artomDamn communists ;)16:28
stewie925hehehe - taiwan is ROC (Republic of China)16:29
*** fesp has quit IRC16:29
artom(Says the Soviet-born dude)16:29
*** Apoorva has joined #openstack-nova16:29
clarkbkashyap: fwiw sudo virsh cpu-models x86 does not work locally for me. 'error: failed to get CPU model names' this is whtout any modifications or devstack or anything16:30
stewie925hello alaski and melwitt - only when you have a chance...  my boot instance in cell environment failed with "no valid hosts" error and I posted the logs in http://paste.openstack.org/show/497753/ - much appreciated!16:32
*** tonytan4ever has joined #openstack-nova16:33
clarkbhuh works for ppc64 so clearly pebkac16:34
*** jinxing has quit IRC16:34
openstackgerritSam Betts proposed openstack/nova: Eliminate resource tracker to host manager races  https://review.openstack.org/30834816:34
*** raunak has joined #openstack-nova16:35
*** klkumar has quit IRC16:36
*** paul-carlton2 has joined #openstack-nova16:37
melwittstewie925: next you want to check n-cpu.log for errors. in n-sch it selected a host and after that something failed16:38
stewie925oh thank you melwitt - checking on n-cpu logs now16:39
melwittalso check n-cond, n-cell-child, n-cell-region for errors. and maybe n-net too16:39
*** akshai has quit IRC16:40
stewie925oh thanks - will make a note to check all n- logs as well16:40
mgagnemelwitt: tyvm for your help with scheduler last week. it's now freaking fast :D The only other config we tweaked is max_concurrent_builds to allow more concurrent builds16:40
kashyapclarkb: Sorry, was AFK.  Yeah, I could reproduce on x86 too -- "error: this function is not supported by the connection driver: virConnectGetCPUModelNames"16:41
clarkbkashyap: doesn't make it easy to see what is valid or not :)16:42
melwittmgagne: awesome! so all you did was reorder the filters and increase max_concurrent_builds? or was there more you had to do?16:42
*** PaulMurray has quit IRC16:42
mgagneusing caching_scheduler too to bypass the overhead of initializing all those host state objects16:42
kashyapclarkb: Let me check w/ virt folks on OFTC16:42
melwittmgagne: ah, cool. that's right I forgot you had mentioned that last time16:43
mgagnebaremetal takes a long time before cleaning anyway so we more or less care about them no being available for another minute. we also increased scheduler_max_attempts16:43
mgagnejust to cover the case where scheduler allocates you a node that was claimed in the past minute.16:44
sdagueclarkb: x86_6416:44
*** tyrola has quit IRC16:44
sdaguesudo virsh cpu-models x86_64 works fine on my xenial box16:44
clarkbsdague: aha16:44
clarkbsdague: so that doesn't match what is in the xml16:44
clarkbxml says x8616:45
*** nic has joined #openstack-nova16:45
melwittmgagne: makes sense16:45
*** markus_z has quit IRC16:45
*** fawadkhaliq has joined #openstack-nova16:46
sdagueclarkb: huh, yep, guess it doesn't16:46
*** zengine has joined #openstack-nova16:46
*** jamesdenton has quit IRC16:47
*** zengine has quit IRC16:47
clarkbkashyap: ok I have locally confirmed the file path on xenial matches and that a sudo systemctl restart libvirt-bin is sufficient to make that a valid cpu16:47
*** zengine has joined #openstack-nova16:47
kashyapclarkb: You mean, the file path has not changed16:48
clarkbkashyap: correct16:48
*** edand_ has joined #openstack-nova16:48
kashyapclarkb: What test did you do locally w/ DevStack?16:48
*** moshele has quit IRC16:48
*** earlephilhower has joined #openstack-nova16:48
clarkbkashyap: I just ran sudo devstack/tools/cpu_map_update.py /usr/share/libvirt/cpu_map.xml && sudo systemctl restart libvirt-bin && virsh cpu-modles x86_6416:49
*** Jabadia_ has joined #openstack-nova16:50
*** sfinucan has quit IRC16:50
kashyapclarkb: Does your `virsh uri` show "qemu:///system"?  If not -- we see that error when you run `virsh cpu-models x86` (For me it was showing lxc:///)16:50
*** doug-fish has quit IRC16:50
clarkbkashyap: sdague pointed out I had to use x86_6416:50
*** doug-fish has joined #openstack-nova16:51
*** mvk_ has quit IRC16:51
clarkbwhich is totally confusing because it doesn't match the config file but whatever it works16:51
*** sambetts is now known as sambetts|afk16:52
kashyapclarkb: Hmm.  So, the issue is probably why isn't libvirt converting 'gate64' model into something it can understand...16:53
*** mriedem has joined #openstack-nova16:53
*** penick has quit IRC16:54
*** mriedem_afk has quit IRC16:54
*** penick has joined #openstack-nova16:54
*** doug-fis_ has joined #openstack-nova16:54
clarkbthe next thing devstack does is a sudo service libvirt-bin stop && sudo service libvirt-bin start16:55
*** doug-fish has quit IRC16:55
clarkbwhen I do this locally I don't get a running libvirt-bin service16:55
*** penick has quit IRC16:55
*** zenoway has quit IRC16:55
*** mriedem1 has joined #openstack-nova16:55
*** derekh has quit IRC16:55
clarkbif I use systemctl I get a thing that works16:56
*** zenoway has joined #openstack-nova16:56
clarkbI wonder if this is the problem16:56
clarkbwe might want to update devstack's service start and stop to detect if systemctl is present and use that if so else if ubuntu use service16:57
kashyapclarkb: Oh, it's the difference between upstart & systemd16:57
sdagueclarkb: yeh, that's probably the crux of it16:57
clarkbkashyap: yes, I am beginning to suspect libvirt on xenial does not work properly if using hte old service management16:57
sdagueclarkb: well, the issue is if sudo service libvirt-bin stop && sudo service libvirt-bin sta16:57
sdagueand the first one is async submit to systemd16:58
sdagueit's not really done when you do start16:58
clarkbah16:58
*** yamahata has quit IRC16:58
sdaguethat is my guess16:58
*** r-daneel has joined #openstack-nova16:58
dansmithclaudiub: hah, did jaypipes catch you with asking to see the rescue test run? https://review.openstack.org/#/c/127159/16:58
clarkbsdague: would a systemctl stop && systemctl start be synchrnous?16:58
*** yamahata has joined #openstack-nova16:58
sdagueclarkb: I don't know, but a restart is best16:58
*** mtanino has quit IRC16:58
clarkbsdague: restart does not work on trusty16:58
sdaguebecause you've told systemd the intent16:58
*** whenry has joined #openstack-nova16:59
*** doug-fis_ has quit IRC16:59
sdaguewe should just make it a conditional for the backend, otherwise we're going to be chasing bugs one way or another16:59
*** mriedem has quit IRC16:59
*** doug-fish has joined #openstack-nova16:59
sdagueclarkb: are we actually sure restart doesn't work on trusty?17:00
sdagueor did it not work on precise17:00
sdagueand it's a carry forward17:00
dansmithclaudiub: meaning it looks like it failed on the first run :D17:00
clarkbsdague: I am 100% sure it did not work at the time I got this going a year or so ago17:00
*** igordcard has joined #openstack-nova17:00
clarkbsdague: it is possible that trusty has since fixed their broken libvirt service management17:00
*** zenoway has quit IRC17:01
sdagueok, well, systemd does work17:01
clarkbI spent about a day debugging why restart didi work17:01
sdagueso lets do the working thing with systemd, and leave the trusty work around, conditional on trusty17:01
*** penick has joined #openstack-nova17:01
kashyapclarkb: Updating 'functions-common' to use systemctl work -- Maybe we can post a change to see what breaks?17:01
claudiubdansmith: well, the commit sat for so long, it was no longer valid. :)17:01
clarkbit was a huge waste of time17:01
dansmithclaudiub: ah, okay, fair then :D17:01
sdaguefor all the griping people had about systemd, a ton of stuff is so much more reliable with it :)17:01
claudiubdansmith: image_meta changed since this code was implemented. it was dict-like before, now it isn't. I've updated that, ran tempest for the rescue tests in my env, worked.17:02
kashyapsdague: Fully agreed.  It's just those who don't invest some time learning it complain17:02
dansmithclaudiub: yep, well, I'll have a look when the results come back17:02
claudiubcool. :)17:03
clarkbsdague: kashyap if only unit file names didn't change the behavior of srevices... oh well (this is the root of at least one bug in centos7)17:03
dansmithclaudiub: looks like also a unit test fail from jenkins?17:03
claudiubdansmith: indeed. something changed in imagecache since last jenkins ran on this patch.17:03
sdagueok, I'm going to get outside for a bit, clarkb / kashyap if you push up a patch for review, I'll hit it when I get back so we can unstick this17:03
kashyapclarkb: Oh, what way a file name change triggers _behavior_ of the service?17:03
dansmithclaudiub: ah, yep17:03
kashyapsdague: No worries.  Get some air!17:03
clarkbkashyap: @s mean run many of these, useful to inetd type replacement17:04
*** doug-fish has quit IRC17:04
clarkbkashyap: and people write their unit files assuming that behavior then don't get the filename right then centos7 git daemon doesn't work17:04
clarkbkashyap: are you going to write the devstack change? you can also depends-on that change on the live migrate change side to see if that makes it all work17:05
kashyapclarkb: I can.  Got a URL to the live-migrate change?17:05
clarkbkashyap: I don't17:05
kashyapclarkb: No worries, I'll try find out, once I make some quick dinner17:06
*** lucasagomes is now known as lucas-afk17:06
kashyapclarkb: Where is systemctl located in Ubuntu? /bin I suppose?17:07
clarkbyup17:07
openstackgerritClaudiu Belu proposed openstack/nova: Hyper-V: Implement nova rescue  https://review.openstack.org/12715917:09
mriedem1sdague: is this your way of calling me an asshole for pointing out typos in your specs? :) https://web.archive.org/web/20051016203842/http://diveintomark.org/archives/2004/08/16/specs17:09
*** mriedem1 is now known as mriedem17:09
*** doug-fish has joined #openstack-nova17:11
*** doug-fish has quit IRC17:11
*** doug-fish has joined #openstack-nova17:12
*** akshai has joined #openstack-nova17:13
*** paul-carlton2 has quit IRC17:13
*** paul-carlton2 has joined #openstack-nova17:14
*** fragatina has joined #openstack-nova17:15
*** jlanoux has quit IRC17:15
*** fragatina has quit IRC17:16
*** fragatina has joined #openstack-nova17:17
syjulianjaypipes: hi jay, would you be able to review this patch fothe nova.virt.firewall.IpTablesFirewallDriver bug? https://review.openstack.org/#/c/311758/17:18
*** ivase_ has quit IRC17:18
syjulianjaypipes: this is the bug https://bugs.launchpad.net/nova/+bug/129588917:18
openstackLaunchpad bug 1295889 in OpenStack Compute (nova) "No unit tests for nova.virt.firewall.IptablesFirewallDriver" [Low,In progress] - Assigned to Julian Sy (syjulian)17:18
mriedemedleafe: is there a template somewhere for the config option help text?17:22
mriedemor just the spec?17:22
*** BobBall is now known as BobBall_AWOL17:22
edleafemriedem: just the spec, but to be honest, I just copy/pasted markus_z's and edited from there17:24
*** zzelle_ has joined #openstack-nova17:24
dansmithartom: is it worth me commenting on this: https://review.openstack.org/#/c/264016/36/nova/virt/libvirt/driver.py ?17:24
dansmithartom: how much is going to change?17:25
*** ivase has joined #openstack-nova17:25
artomdansmith, that specific diff will change entirely...17:26
dansmithokay17:26
dansmithgood, because it's fatally flawed :D17:26
dansmithoh wait, actually it's not17:26
dansmithglad I didn't comment17:26
vladikrdansmith, we will post a new draft in couple of mins17:26
artomdansmith, I'm giving you the Fry eyes right now17:27
dansmithvladikr: okie17:27
artomdansmith, can you quickly comment here? In case we're about to repeat the same mistake?17:27
dansmithartom: you didn't make a mistake, so no :)17:27
dansmithwell, not the mistake I thought17:27
dansmithartom: vladikr: that hunk was nice and small for pointing other drivers at17:28
*** zengine has quit IRC17:28
dansmitha nice "do this like this" link17:28
artomExcept... don't do it like this just yet :)17:28
dansmithwell, right, I said "was"17:28
dansmith"could have been"17:28
artomSo much lost potential17:28
artomThanks for calling it a hunk though, it feels better about its body now17:29
*** ivase has quit IRC17:29
flip214mriedem: thank you very much, I'll redo the rootwrap => privsep parts.17:30
*** paul-carlton2 has quit IRC17:30
mriedemdansmith: what hunks are nice for you now?17:32
dansmithmriedem: small hunks of concise code.. why? what hunks were _you_ thinking of?17:32
mriedemhttps://cgrzeszczak.files.wordpress.com/2012/11/untitled.png?w=34117:33
mriedemthat one17:33
*** jwcroppe_ has quit IRC17:33
*** fawadkhaliq has quit IRC17:33
openstackgerritDiana Clarke proposed openstack/nova: Rbd: add create_from_image & create_from_func  https://review.openstack.org/31769417:34
openstackgerritDiana Clarke proposed openstack/nova: Flat: add create_from_image & create_from_func  https://review.openstack.org/31770117:34
openstackgerritDiana Clarke proposed openstack/nova: Ploop: add create_from_image & create_from_func  https://review.openstack.org/31768917:34
*** ivase has joined #openstack-nova17:35
*** moshele has joined #openstack-nova17:35
flip214mriedem: grepping the nova code for privsep turns up nothing. would the DRBD transport be the first thing to use privsep in nova? https://review.openstack.org/#/q/privsep doesn't really help either...17:41
*** lpetrut has quit IRC17:42
*** ctrath has quit IRC17:43
flip214hmmm, this part just calls os-brick anyway, and that's already converted... so there's nothing rootwrap or privsep specific left, is there?17:44
*** r-daneel has quit IRC17:44
openstackgerritDiana Clarke proposed openstack/nova: Rbd: add create_from_image & create_from_func  https://review.openstack.org/31769417:45
openstackgerritDiana Clarke proposed openstack/nova: Flat: add create_from_image & create_from_func  https://review.openstack.org/31770117:45
openstackgerritDiana Clarke proposed openstack/nova: Ploop: add create_from_image & create_from_func  https://review.openstack.org/31768917:45
*** tonytan4ever has quit IRC17:45
openstackgerritPhilipp Marek proposed openstack/nova: Connecting Nova to DRBD storage nodes directly  https://review.openstack.org/25629217:45
mriedemflip214: there isn't anything in nova using privsep yet, no17:45
flip214mriedem: thanks.17:45
flip214new patch just pushed.17:45
flip214thanks a lot for the *very* quick feedback!17:46
mriedemsince the nova volume driver doesn't need to call drbdadm then we don't need the rootwrap filter, and os-brick is using privsep already, yeah, so we should be ok there17:46
openstackgerritMike Fedosin proposed openstack/nova: Introduce a CONF flag to determine glance client version  https://review.openstack.org/31596317:47
openstackgerritMike Fedosin proposed openstack/nova: split the glance API path based on config  https://review.openstack.org/31600717:47
mriedemflip214: is there a third party ci for drbd in cinder/os-brick where i can see this passing?17:47
*** akshai has quit IRC17:47
*** pushkaru has quit IRC17:47
*** daemontool has quit IRC17:47
flip214mriedem: uh, I don't know.17:47
*** Guest86061 has quit IRC17:47
flip214the DRBD driver in Cinder is in openstack-infra, as it's pure Open Source... don't know about the os-brick and/or Nova parts, though.17:48
flip214For single-machine CI it wouldn't be triggered anyway.17:48
mriedemflip214: ok, so is there any drbd ci for openstack, anywhere?17:48
*** rlrossit has quit IRC17:48
flip214Multi-machine CI was an ongoing topic in -infra last time I asked, perhaps there's some example now that can be adapted.17:49
mriedemwe have multinode jobs17:49
*** haukebruno has quit IRC17:49
mriedembut cinder isn't using any yet that i know of17:49
mriedemhemna: do you know anything about ci for drbd in cinder?17:49
*** rama_y has joined #openstack-nova17:49
*** jwcroppe has joined #openstack-nova17:51
flip214mriedem: the jenkins job is called gate-tempest-dsvm-full-drbd-devstack-nv17:51
flip214http://ci-watch.tintri.com/project?project=cinder&time=7+days has it in the overview, for example17:52
hemnahey17:52
mriedemflip214: ok, that's helpful17:52
mriedemi'll see what that runs on in project-config17:52
hemnaoh looks like flip214 has it covered :)17:52
flip214or https://review.openstack.org/#/c/318749/ has a recent run17:52
flip214hemna: yeah, thanks.17:52
flip214mriedem: right below the sheepdog gate check.17:53
*** rlrossit_ has joined #openstack-nova17:53
flip214thanks for the attention, I'm /off now.17:54
flip214have a nice vacation!17:54
mriedemthanks17:54
*** akshai has joined #openstack-nova17:54
*** rmart04 has quit IRC17:55
*** tonytan4ever has joined #openstack-nova17:57
mriedemhrm, tempest.api.compute.volumes.test_attach_volume.AttachVolumeTestJSON.test_attach_detach_volume is skipped in the drbd job17:59
*** jichen has joined #openstack-nova17:59
openstackgerritDan Smith proposed openstack/nova: Completed migrations are not "in progress"  https://review.openstack.org/31883218:00
*** Apoorva_ has joined #openstack-nova18:00
*** ducttape_ has quit IRC18:01
*** Jabadia_ has quit IRC18:02
mriedemoh b/c it uses ssh18:03
*** Apoorva has quit IRC18:04
*** cheran has joined #openstack-nova18:04
openstackgerritKaren Bradshaw proposed openstack/nova: api-ref: os-consoles.inc  https://review.openstack.org/31753618:06
*** notmorgan has quit IRC18:11
*** ducttape_ has joined #openstack-nova18:11
*** edand_ has quit IRC18:12
*** moshele has quit IRC18:13
*** doug-fis_ has joined #openstack-nova18:16
*** eharney has quit IRC18:18
dansmithalaski: so, sdague always comments that checking that columns were added during a migration isn't really the goal of the migration testing framework18:18
dansmithand I agree, but I never know what to put there18:18
*** doug-fish has quit IRC18:19
alaskidansmith: I'm not really concerned with the columns, just that each table exists18:19
dansmithalaski: I guess the conditional tests are worth it, although they're kinda silly because they're never going to pass18:19
dansmithalaski: well, the column check was pretty much for that, unless there is a table helper, but I can add one per table I gess18:19
dansmither, guess18:19
*** ctrath has joined #openstack-nova18:19
dansmithalaski: so you're thinking one per table and the conditional things like the indexes?18:20
alaskiyep18:20
*** ctrath has quit IRC18:20
alaskithere doesn't seem to be a table helper18:20
alaskibut db_utils.get_table not blowing up would be fine18:21
alaskiand at some point we should nail down what the point of migration testing is, since I can never remember18:21
dansmithalaski: well, one column assert per table would be easy right?18:21
dansmithalaski: well, when we did data migrations, that was the point :)18:22
dansmithdata migrations in schema migrations I mean18:22
alaskidansmith: yeah, a column test would be just as fine18:22
dansmithokay18:22
alaskioh right, data migrations18:22
dansmithI kinda wrote it off as "this doesn't matter I just need something"18:22
dansmithbut I was also rather annoyed at the time18:22
*** eharney has joined #openstack-nova18:23
*** ctrath has joined #openstack-nova18:24
alaskiif we're going to keep these tests around some cursory testing would be nice. but I really like it for the indexes because the way they're defined always concerns me since there's no .create() call or anything18:24
dansmithso the reason why I don't think it's worth it,18:24
dansmithis that we have the model sync test,18:25
*** Vek has quit IRC18:25
dansmithwhich will fail if we didn't actually create things exactly like they're specified in the model18:25
dansmithso if we failed to hit the "if not db2" thing, we'd fail that other test18:25
*** whenry has quit IRC18:25
alaskitrue18:25
dansmithand that's a lot stronger than hand-poking at things18:25
alaskiso maybe we should scrap these tests18:25
alaskiat least for the api db18:25
*** READ10 has quit IRC18:26
*** notmorgan has joined #openstack-nova18:26
dansmithmaybe18:26
*** cfriesen has quit IRC18:26
*** dillaman has quit IRC18:26
*** lennyb has quit IRC18:26
*** russellb has quit IRC18:26
*** dosaboy has quit IRC18:26
*** anthonyper has quit IRC18:26
*** wanghao has quit IRC18:26
*** dmellado has quit IRC18:26
*** bradm has quit IRC18:26
*** j_king has quit IRC18:26
*** crinkle has quit IRC18:26
*** lifeless has quit IRC18:26
*** mfisch has quit IRC18:26
*** klindgren has quit IRC18:26
*** edleafe has quit IRC18:26
*** rmk has quit IRC18:26
*** j_king has joined #openstack-nova18:26
*** bradm has joined #openstack-nova18:26
*** cfriesen has joined #openstack-nova18:26
*** dosaboy has joined #openstack-nova18:26
*** lifeless has joined #openstack-nova18:26
*** klindgren has joined #openstack-nova18:26
*** wanghao has joined #openstack-nova18:26
*** dillaman has joined #openstack-nova18:26
*** russellb has joined #openstack-nova18:26
*** lennyb has joined #openstack-nova18:27
*** rmk has joined #openstack-nova18:27
*** mfisch has joined #openstack-nova18:27
*** mfisch is now known as Guest7626818:27
*** dmellado has joined #openstack-nova18:27
*** gagehugo has quit IRC18:27
*** crinkle has joined #openstack-nova18:27
*** anthonyper has joined #openstack-nova18:27
alaskilooking at the api db tests I don't see anything that wouldn't be caught by the model sync test18:28
dansmithyeah because these are all new18:28
*** zhenguo_ has quit IRC18:30
*** edleafe has joined #openstack-nova18:31
openstackgerritDan Smith proposed openstack/nova: Add resource provider tables to the api database  https://review.openstack.org/31568118:31
openstackgerritDan Smith proposed openstack/nova: Make compute nodes update their own inventory records  https://review.openstack.org/31704918:31
openstackgerritDan Smith proposed openstack/nova: Make Inventory and ResourceProvider objects use the API DB instead  https://review.openstack.org/31568218:31
alaskiI'd be happy to scrap them or use them18:31
*** esnyder has quit IRC18:31
*** kgalanov has quit IRC18:31
*** andrewbogott has quit IRC18:31
*** zhiyan has quit IRC18:31
*** vdrok has quit IRC18:31
dansmithalaski: we still want the base test because it walks them I think, which the rest don't do18:32
dansmithalaski: but just the required check phase is not so useful.. we could maybe just allow people putting pass in there in case we come up with a thing we need to test that way18:32
*** kgalanov has joined #openstack-nova18:33
*** esnyder has joined #openstack-nova18:33
*** zhenguo_ has joined #openstack-nova18:34
*** andrewbogott has joined #openstack-nova18:34
*** whenry has joined #openstack-nova18:34
*** vdrok has joined #openstack-nova18:34
*** zhiyan has joined #openstack-nova18:34
mriedemdansmith: alaski: on the last keypair migration patch, i'm basically in the same boat with alaski, looks like just wanting an abort on the main db delete if the api db create fails https://review.openstack.org/#/c/313664/18:35
mriedemi think once that patch is done we can mark the bp complete right?18:35
alaskidansmith: sure. if it's pass or a good test I'm happy, just don't want a middle ground18:35
dansmithmriedem: so the problem with that is, I think because of enginefacade, I can't open a transaction against both at the same time18:36
dansmithmriedem: maybe if I create another context to use to fake it or something...18:36
dansmithalaski: yeah, the fake test busywork is the worst case18:36
mriedemcan you just recreate in the main db from the same values that were in the deleted record?18:36
mriedemhokey but it's an option18:36
*** moshele has joined #openstack-nova18:37
dansmithmriedem: that's really not a solution right?18:37
dansmithI mean, if the power goes out, you've deleted one and can't recreate it18:37
mriedemyeah18:37
*** efoley has joined #openstack-nova18:38
*** MikeG451 has quit IRC18:38
dansmithit solves half of the problem I guess, but doesn't seem worth it18:38
mriedemwait, what constraint is keeping you from creating in the api db first?18:38
dansmithI can try another fake context to fool it18:38
mriedemthe check in the keypair objecT?18:38
dansmithmriedem: the object constraint of checking that the thing doesn't exist in the old db18:38
dansmithmriedem: so I can do a different type of create if you'd prefer that, I've just been trying to keep these to using the mainline interfaces for such things18:39
alaskiyou could bypass create and use _create_in_db, but that's hacky too18:39
mriedemuse a fake context if tha tworks18:39
*** tri2sing has joined #openstack-nova18:39
dansmithalaski: yeah, don't want to do that if we can help it18:39
mriedemlet me just say,18:39
mriedementerprise NFV disaster recovery18:39
dansmithlet me try a fake context18:40
dansmithI bet this isn't going to work18:42
dansmithbecause the transaction won't be committed, so when create goes to check it'll still look present, right?18:42
*** baoli_ has quit IRC18:42
dansmithalaski: mriedem ^18:42
*** sean-k-m1 is now known as sean-k-mooney18:42
*** jistr has quit IRC18:43
alaskiI'm not sure, but I think so18:43
*** ihrachys has joined #openstack-nova18:43
dansmithI'll just add a non-remotable wrapper for the meat18:43
mriedemidk, why isn't it possible to stack up transaction writers? aren't they just decorators?18:43
dansmithmriedem: your session goes in your context and you can only have one18:44
dansmithmriedem: so you need another context for the second db, which will be a different session18:44
alaskiI feel like the best thing here would be to have a mechanism for creating in the api even if it exists in the cell db, then catch and ignore a dupe write in the migration method18:44
dansmithalaski: yeah, I'm just going to do that, it's not that hard..18:45
openstackgerritOpenStack Proposal Bot proposed openstack/nova: Updated from global requirements  https://review.openstack.org/31886518:46
*** rcernin has joined #openstack-nova18:46
*** ctrath has quit IRC18:48
*** lpetrut has joined #openstack-nova18:48
*** tri2sing has quit IRC18:49
*** smatzek has quit IRC18:49
*** ctrath has joined #openstack-nova18:50
*** andrewbogott has quit IRC18:51
*** andrewbogott has joined #openstack-nova18:51
jaypipessyjulian: will do my best to get to that today, yes.18:51
*** dimtruck is now known as zz_dimtruck18:51
*** jtp92 has left #openstack-nova18:52
dansmithalaski: yeah, that was easy, not sure why I talked myself out of doing that in the first place18:52
openstackgerritDan Smith proposed openstack/nova: Add online migration to move keypairs from main to API database  https://review.openstack.org/31366418:52
alaskibecause you hadn't been sufficiently annoyed by reviewers yet18:53
*** whenry has quit IRC18:53
*** gszasz has quit IRC18:53
*** renatoarmani has joined #openstack-nova18:53
*** renatoarmani has quit IRC18:54
*** whenry has joined #openstack-nova18:54
dansmithcdent: see my pointer response to your comment and tell me if I'm blind or confused18:54
*** Sukhdev has joined #openstack-nova18:55
*** diogogmt has quit IRC18:55
cdentdansmith: my understanding, which could well be wrong, was the the special casing only mattered for single column uniq constraint/index pairs18:55
cdentthat with multi column constraints the situation was different18:56
cdentbut I dunno, thus invoking mriedem18:56
dansmithcdent: hmm, okay18:56
cdentI hope he will enlighten us18:57
openstackgerritDan Smith proposed openstack/nova: Add online migration to move keypairs from main to API database  https://review.openstack.org/31366418:57
dansmithI'm a bit annoyed that we have this in there anyway, since apparently 0% of people run DB218:57
*** doug-fis_ has quit IRC18:58
cdentyeah, it all seems a bit bogus18:58
mriedemcan someone point me to the thing i need to look at?18:58
cdentmriedem: https://review.openstack.org/#/c/315681/9/nova/db/sqlalchemy/api_migrations/migrate_repo/versions/016_resource_providers.py@6918:59
*** dillaman has quit IRC19:00
*** salv-orlando has quit IRC19:01
*** claudiub has quit IRC19:04
*** pushkaru has joined #openstack-nova19:04
*** MikeG451 has joined #openstack-nova19:04
syjulianjaypipes: thanks jay19:04
*** moshele has quit IRC19:04
mriedemjaypipes: doesn't mysql also create an index over the columns in a unique constraint, but mysql allows duplicate indexes?19:05
openstackgerritVladik Romanovsky proposed openstack/nova: libvirt: New configuration classes to parse device address element  https://review.openstack.org/30414719:05
openstackgerritVladik Romanovsky proposed openstack/nova: libvirt: attach configdrive after instance XML  https://review.openstack.org/30333519:05
openstackgerritVladik Romanovsky proposed openstack/nova: objects: adding an update method to virtual_interface  https://review.openstack.org/31061819:05
openstackgerritVladik Romanovsky proposed openstack/nova: Device tagging metadata API support  https://review.openstack.org/30451119:05
openstackgerritVladik Romanovsky proposed openstack/nova: Device tagging API support  https://review.openstack.org/30451019:05
openstackgerritVladik Romanovsky proposed openstack/nova: Compute manager device tagging support  https://review.openstack.org/26401719:06
openstackgerritVladik Romanovsky proposed openstack/nova: Libvirt driver implementation of device tagging  https://review.openstack.org/26401619:06
openstackgerritVladik Romanovsky proposed openstack/nova: db: add a device_metadata column to instance_extra  https://review.openstack.org/31888319:06
openstackgerritVladik Romanovsky proposed openstack/nova: objects: new InstanceDevicesMetadata object  https://review.openstack.org/31888419:06
*** zengine has joined #openstack-nova19:07
*** sdague has quit IRC19:08
*** zenoway has joined #openstack-nova19:09
*** liverpooler has joined #openstack-nova19:09
openstackgerritChris Dent proposed openstack/nova: Add Allocation and AllocationList objects  https://review.openstack.org/28244219:09
openstackgerritChris Dent proposed openstack/nova: Extend ResourceProvider to have resources and aggregates  https://review.openstack.org/28496319:09
openstackgerritChris Dent proposed openstack/nova: Add a get_by_uuid for aggregates  https://review.openstack.org/28489519:09
openstackgerritChris Dent proposed openstack/nova: Correct the Inventory object  https://review.openstack.org/31528819:09
openstackgerritChris Dent proposed openstack/nova: Add name to ResourceProvider object  https://review.openstack.org/28194519:09
openstackgerritChris Dent proposed openstack/nova: WIP resource-pools/placement api  https://review.openstack.org/29310419:09
*** doug-fish has joined #openstack-nova19:09
*** sdague has joined #openstack-nova19:09
*** raunak has quit IRC19:10
mriedemdansmith: cdent: comments inline19:10
sdaguemriedem: :)19:10
*** salv-orlando has joined #openstack-nova19:10
*** diogogmt has joined #openstack-nova19:12
*** zenoway has quit IRC19:13
dansmithmriedem: so I get that I should do that thing cdent said, but I'm not sure if the earlier thing you quoted implies a change19:14
*** diogogmt has quit IRC19:15
mriedemdansmith: on Index('inventories_resource_provider_resource_class_idx', ?19:15
cdentthe thing described (either no name, or a unique name) is the desired behavior, as I understand it.19:15
mriedemoh19:16
mriedemah the first thing19:16
*** ijw has quit IRC19:16
dansmithcdent: no, not that19:16
cdentmriedem: dunno, I'm talking about the first thing, not sure if dansmith is, as there are two things19:16
dansmithmriedem: the "this is the problem index" one19:16
mriedemyeah, so the only reason i bring that up is, when you create a UC in db2, if any of the columns are nullable, then it doesn't create a UC, it creates a unique index19:16
mriedemwhich is really only a problem later if you try to have a fkey on that column19:16
mriedemwhich for name we won't19:16
dansmithsigh19:17
mriedemit's a problem for fkeys because fkey requires the UC or primary key, which we wouldn't have on name19:17
dansmithI just want that thing we created in the other db to be moved to this db19:17
mriedemclarified the comment inline19:18
dansmithmriedem: you don't care about db2 anymore?19:18
mriedemdansmith: i don't19:18
dansmithwhy (the f) are we still dancing around this?19:18
mriedemlike i said, feel free to remove any if ibm_db_sa checks in there19:18
mriedemand let it break19:18
* dansmith 's head explodes19:18
sdaguemriedem: so we can remove db2 with fire?19:19
mriedemsure19:19
* cdent fetches the champagne19:19
sdaguedansmith: do you want to write the patch or approve it?19:20
dansmithsdague: lol.. tough choice.19:20
dansmithsdague: I feel like someone somewhere would use me writing the patch as a reason not to vote me onto the lollipop committee or something19:20
sdaguedansmith: you never run for the lollipop committee19:20
dansmithhah19:21
dansmithfunny you knew what I was saying there19:21
sdague:)19:21
openstackgerritDan Smith proposed openstack/nova: Add resource provider tables to the api database  https://review.openstack.org/31568119:22
openstackgerritDan Smith proposed openstack/nova: Make compute nodes update their own inventory records  https://review.openstack.org/31704919:22
openstackgerritDan Smith proposed openstack/nova: Make Inventory and ResourceProvider objects use the API DB instead  https://review.openstack.org/31568219:22
openstackgerritSarafraj Singh proposed openstack/nova: Deprecate barbican options  https://review.openstack.org/31768419:22
*** cdent has left #openstack-nova19:22
*** cdent has joined #openstack-nova19:23
*** vladikr has quit IRC19:24
sdaguedansmith: I'm happy to write the patch19:24
dansmithsdague: I'm happy to +2 the ever-loving snot out of it19:24
*** jwcroppe has quit IRC19:28
*** cdent has quit IRC19:28
*** ducttape_ has quit IRC19:30
*** jwcroppe has joined #openstack-nova19:30
*** Guest76268 is now known as mfisch19:30
*** baoli has joined #openstack-nova19:30
*** mfisch has quit IRC19:30
*** mfisch has joined #openstack-nova19:30
*** ducttape_ has joined #openstack-nova19:33
*** ducttape_ has joined #openstack-nova19:33
*** rama_y has quit IRC19:36
*** diogogmt has joined #openstack-nova19:37
*** BjoernT has joined #openstack-nova19:40
openstackgerritSean Dague proposed openstack/nova: remove db2 support from tree  https://review.openstack.org/31890619:40
sdaguethat rewrites migrations... so it should be looked at carefully to make sure I didn't screw up an if block19:40
*** Apoorva_ has quit IRC19:43
stewie925would it be normal for an instance (flavor = micro) to spawn for > 5 minutes on a cell environment?19:43
dansmithstewie925: yes, wait at least 60 minutes19:44
stewie925hey dan thanks!19:44
stewie925whoops 60 minutes19:44
dansmithstewie925: and in case we haven't been introduced, I'm the one that says all the overly sarcastic things19:44
stewie925lol I'll make note of that19:45
stewie925hope it doesnt take that long lol19:46
*** jamesdenton has joined #openstack-nova19:46
stewie925crap - it errored out ugh19:47
stewie925"failed to allocate network, not rescheduling "   :(19:48
auggysdague: i think i'm a little confused about https://review.openstack.org/#/c/315216/19:48
auggyi saw a code path for the additional limits parameters i documented, but maybe that's only for an older version?19:48
sdagueauggy: got the reference?19:50
auggysdague: yeah let me find it19:50
auggysdague: also i had found tests that expected those params in the response19:51
*** nic has quit IRC19:51
*** zenoway has joined #openstack-nova19:52
*** rama_y has joined #openstack-nova19:52
mriedemdansmith: why do we have SERVICE_VERSION = 11 but SERVICE_VERSION_HISTORY is at 4.12?19:53
mriedemor that's unrelated?19:53
mriedemi guess 4.12 is compute rpc api version19:53
dansmithright19:53
*** pushkaru has quit IRC19:53
mriedemok, so me checking in the api, i just need to update SERVICE_VERSION19:53
mriedemand don't care about SERVICE_VERSION19:53
*** efoley has quit IRC19:53
mriedem*SERVICE_VERSION_HISTORY19:53
mriedemfor get me a network i mean19:54
auggysdague: ok so i'll double double check that this is getting called - https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/views/limits.py19:54
*** ducttape_ has quit IRC19:54
dansmithyou need another entry in there, but it will be the same as the current verson19:54
sdagueauggy: that doesn't have the rate limit keys in it19:54
mriedemdansmith: oh i see19:55
mriedemlist of dicts, with duplicates19:55
*** diogogmt has quit IRC19:55
sdagueauggy: oh, sorry, I see19:55
auggysdague: https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/views/limits.py#L6819:55
auggythat might not actually be getting called by anything tho, so i should double check that19:56
auggybut i traced that and saw the unit test coverage and assumed it was accurate19:56
auggy"traced" =  just followed a trail manually19:56
*** ducttape_ has joined #openstack-nova19:57
sdagueauggy: https://github.com/openstack/nova/blob/d75fb4b0e457317054e20e1f493c2730aa172a2d/nova/api/openstack/compute/limits.py#L3919:57
auggyif it's not actually getting called or being used then i can submit a change to remove it19:57
auggysdague: right, that confused me a little19:57
sdagueso, I think it's vestigial19:57
auggygah19:57
auggyi thought maybe whatever created that nova.limits data was using this view19:58
sdaguehttps://github.com/openstack/nova/blob/d75fb4b0e457317054e20e1f493c2730aa172a2d/nova/api/openstack/compute/legacy_v2/limits.py#L20419:59
sdagueit's only on the legacy_v2 stack19:59
bauzasmriedem: you'll be in vacation for the Newton-1 milestone, right?19:59
sdaguewhich is getting deleted... soon19:59
mriedembauzas: i'm back to work on 5/3119:59
mriedembauzas: n-1 is 6/219:59
bauzasmriedem: okay, cool then19:59
mriedemso i'll be back right before it19:59
auggysdague: yarrgh ok yeah19:59
auggyman i can't believe i missed that19:59
*** nic has joined #openstack-nova19:59
sdagueauggy: so, just say rate is an array, which will be empty19:59
auggysdague: ok, i will also file doc bugs for places where i found other documentation20:00
sdagueand the rest of those params after it we'll never speak of :)20:00
bauzasmriedem: no problem, was asking because of the release change for the milestone20:00
auggysdague: haha wfm20:00
mriedembauzas: sdague is covering it20:00
auggysdague: should i file a bug or just start a change to remove that crud20:00
sdaguewell, the code is getting removed20:01
auggysdague: or is that something that should just wait until the v2 effort is done? i just don't want to forget about it20:01
auggyah ok20:01
sdagueyeh, it should just go with the v2 effort20:01
auggysdague: ok i'll see what is on the list and maybe see if i can help with that20:01
auggysdague: with removing that bit since i was just in it20:01
sdagueI guess we could patch the v2 stack to never even look at the env variable, let me do that quick20:01
bauzasmriedem: okay, I asked because I did the patch for Mitaka, but I'm fine with sdague doing it20:01
bauzasI'll review the reno notes btw.20:01
*** gyee has joined #openstack-nova20:02
mriedembauzas: thanks20:03
auggysdague: thanks, that definitely is what sent me down this path ;)20:03
*** diogogmt has joined #openstack-nova20:04
*** Apoorva has joined #openstack-nova20:05
*** gjayavelu has joined #openstack-nova20:05
*** Sukhdev has quit IRC20:06
openstackgerritSean Dague proposed openstack/nova: be more explicit that rate limits are gone in v2.1  https://review.openstack.org/31892320:07
sdagueauggy: how does that look?20:07
*** sacharya_ has joined #openstack-nova20:08
auggysdague: a++ would +1 again20:08
*** diogogmt has quit IRC20:08
auggyalthough i'm just waiting for jenkins to do it's thang20:08
*** Apoorva has quit IRC20:09
*** diogogmt has joined #openstack-nova20:09
auggysdague: i'm working on an update to my limits params patch and should have examples submitted shortly as well20:10
openstackgerritjichenjc proposed openstack/nova: verify api-ref metadata.inc  https://review.openstack.org/31860020:10
*** sachary__ has joined #openstack-nova20:11
*** sacharya has quit IRC20:11
*** pushkaru has joined #openstack-nova20:11
*** jamesdenton has quit IRC20:12
*** ducttape_ has quit IRC20:13
*** matrohon has joined #openstack-nova20:13
*** sacharya_ has quit IRC20:15
*** doug-fish has quit IRC20:15
*** markvoelker has joined #openstack-nova20:15
openstackgerritSylvain Bauza proposed openstack/nova: API change for verifying the scheduler when live migrating  https://review.openstack.org/31833420:16
openstackgerritSylvain Bauza proposed openstack/nova: API change for verifying the scheduler when evacuating  https://review.openstack.org/31557220:16
bauzas*cough cough* ladies and gentlemen, here comes the last changes for my non-priority blueprint20:16
*** ducttape_ has joined #openstack-nova20:16
bauzasjaypipes: dansmith: alaski: fasten your seatbelts, I'll now review your changes and nitpick them :p20:16
dansmithbauzas: your change yesterday put all mine in merge conflict, so don't expect kindness :P20:17
bauzashehe20:17
* alaski keeps his hands and feet inside the vehicle while it is in motion20:17
*** ijw has joined #openstack-nova20:18
bauzasgiven now my non-prio BP is implemented before N-1, I can now hit the other changes with merge conflicts by implementing some priority specs :D (devil voice)20:18
openstackgerritAugustina Ragwitz proposed openstack/nova: api-ref: limits.inc validate parameters  https://review.openstack.org/31521620:18
bauzasdansmith: jaypipes: alaski: so, what should I review first ? the inventories changes, the API DB migrations or the cellsv2 scheduler changes ?20:19
bauzasI mean, for the sched prio20:20
*** markvoelker has quit IRC20:20
alaskifor scheduler priority stuff don't worry about cells changes yet20:20
openstackgerritTom Patzig proposed openstack/nova-specs: Add "disable local disks" option to flavors  https://review.openstack.org/26767320:20
bauzasokay, I'll begin with biting dansmith's https://review.openstack.org/#/c/315681/20:21
*** raunak has joined #openstack-nova20:21
*** jichen has quit IRC20:21
dansmithbauzas: I broke it just now, putting a fix up in a sec20:21
bauzasokay20:21
bauzasdansmith: which one did I wedged it ?20:22
dansmith_I_ broke it20:22
bauzasI mean, the merge conflict I made20:22
dansmiththe db migration20:22
bauzask20:22
*** akshai has quit IRC20:22
*** Apoorva has joined #openstack-nova20:23
*** ayoung has quit IRC20:24
openstackgerritPushkar Umaranikar proposed openstack/nova: Config Options: Improve help text for Ipv6 options  https://review.openstack.org/31829120:24
*** diogogmt has quit IRC20:25
openstackgerritHans Lindgren proposed openstack/nova: Move some flavor fakes closer to where they are being used  https://review.openstack.org/31893420:25
openstackgerritDan Smith proposed openstack/nova: Add resource provider tables to the api database  https://review.openstack.org/31568120:25
openstackgerritDan Smith proposed openstack/nova: Make compute nodes update their own inventory records  https://review.openstack.org/31704920:25
openstackgerritDan Smith proposed openstack/nova: Make Inventory and ResourceProvider objects use the API DB instead  https://review.openstack.org/31568220:26
*** flwang1 has quit IRC20:26
openstackgerritAugustina Ragwitz proposed openstack/nova: api-ref: limits.inc validate parameters  https://review.openstack.org/31521620:26
openstackgerritPushkar Umaranikar proposed openstack/nova: Config Options: Improve help text for debugger  https://review.openstack.org/31182220:27
*** clenimar has joined #openstack-nova20:28
*** diogogmt has joined #openstack-nova20:29
mriedemthe keystone project id validation spec should be good to go if another spec reviewer wants to take a look https://review.openstack.org/#/c/294337/20:31
*** Sukhdev has joined #openstack-nova20:32
*** daemontool has joined #openstack-nova20:32
dansmithI shall cast my eyes across it20:33
dansmithmriedem: there are two hyper-v patches that have +2 on them that have been waiting since before image meta was an object20:33
*** cleong has quit IRC20:33
auggythanks mriedem20:33
dansmithmriedem: https://review.openstack.org/#/c/184039/2820:33
auggymriedem: i did see the formatting change but i was too much of a lazy ass to fix it20:34
*** moshele has joined #openstack-nova20:34
dansmithmriedem: the third needs a tweak since the imagemeta serial attribute was codified, but the first two can belanded20:34
auggybut if others feel it's a big enough deal i can fix it ;)20:34
*** Vek has joined #openstack-nova20:34
mriedemdansmith: there is probably no ci testing on the serial console one right?20:34
mriedemi don't think tempest does serial console testing20:34
dansmithmriedem: we don't have it on libvirt either.. it'd be a little rough expecting them to connect to a tcp socket and validate something I think20:35
dansmithcertainly doable, but not something we do elsewhere that I know of20:35
clarkbmriedem: tempest does dump the nova console-log output on failures not sure if it does so as a proper test20:35
dansmithclarkb: we're talking about bidirectional serial console20:35
clarkbah20:35
dansmithmriedem: either way, the two that are ready are more hyperv plumbing for the third one that needs a tweak, which actually exposes it20:36
*** moshele has quit IRC20:36
*** mdrabe has quit IRC20:37
*** DaiDV has quit IRC20:37
*** doug-fish has joined #openstack-nova20:38
*** dmk0202 has joined #openstack-nova20:39
*** hemanthm is now known as hemanthm|afk20:40
*** rdopiera has quit IRC20:40
*** whenry has quit IRC20:41
*** rdopiera has joined #openstack-nova20:42
*** doug-fish has quit IRC20:43
*** fragatina has quit IRC20:43
*** doug-fish has joined #openstack-nova20:43
dansmithwaaaaaiit a sec20:43
dansmithcan we really allow that spec to go in as authored by someone @cpan.org?20:44
dansmithjust seems like bad form20:44
auggydansmith: ha20:44
* auggy is a little bit of a troll20:44
*** nic has quit IRC20:47
*** salv-orl_ has joined #openstack-nova20:48
*** zzelle_ has quit IRC20:49
*** baoli has quit IRC20:50
*** rdopiera has quit IRC20:50
*** rdopiera has joined #openstack-nova20:51
*** tonytan4ever has quit IRC20:51
*** salv-orlando has quit IRC20:51
openstackgerritMatt Riedemann proposed openstack/nova: Skip network allocation if 'none' is requested  https://review.openstack.org/30963820:52
openstackgerritMatt Riedemann proposed openstack/nova: nova-net: handle 'auto' network request in allocate_for_instance  https://review.openstack.org/29790620:52
openstackgerritMatt Riedemann proposed openstack/nova: Skip network validation if explicitly requesting no networks  https://review.openstack.org/30960220:52
openstackgerritMatt Riedemann proposed openstack/nova: neutron: handle 'auto' network request in allocate_for_instance  https://review.openstack.org/31627520:52
*** ivase has quit IRC20:52
openstackgerritMatt Riedemann proposed openstack/nova: Add helpers to NetworkRequest(List) objects for auto/none cases  https://review.openstack.org/29787120:52
openstackgerritMatt Riedemann proposed openstack/nova: WIP: Add REST API support for get me a network  https://review.openstack.org/31639820:52
openstackgerritMatt Riedemann proposed openstack/nova: neutron: validate auto-allocate is available  https://review.openstack.org/30957420:52
*** zz_dimtruck is now known as dimtruck20:57
*** ivase has joined #openstack-nova20:58
*** ivase_ has joined #openstack-nova20:59
*** AmyLee71 has joined #openstack-nova21:00
*** mmandel has joined #openstack-nova21:01
*** ivase__ has joined #openstack-nova21:01
*** AmyLee7 has quit IRC21:01
*** zengine has quit IRC21:02
*** ivase has quit IRC21:02
mriedemdansmith: opinion on this - if you request network='none' (i don't want networking on my instance) and the computes aren't new enough, fail the request? or change to None and let the mitaka computes handle it as they do today, which is you get networking if you don't specify anything and a network is available to use, else you don't get networking.21:02
mriedemi'm thinking the former since you specifically request 'none' and we can't honor it21:03
mriedemat least with 'auto' we can try21:03
bauzasjaypipes: dansmith: I'm not an i18n expert, but is it cool to see an Unicode field in a latin-1 charset table?21:03
dansmithmriedem: I thought yesterday you said keep the default in both cases, which might be different for a while during a migration21:03
*** ivase_ has quit IRC21:03
*** smatzek has joined #openstack-nova21:03
*** fragatina has joined #openstack-nova21:03
mriedemkeeping the default in both cases is easier to code, sure21:03
mriedemone less condition21:03
openstackgerritPushkar Umaranikar proposed openstack/nova: Stop using mox in test_firewall  https://review.openstack.org/30653621:04
mriedembut we can just say, nope, sorry, can't honor this21:04
bauzasdansmith: jaypipes: that looks possible (both are 8 bits) that said21:04
dansmithmriedem: I dunno, every time I try to think pedantically about this, someone reminds me this is a pretty silly defaults case, so I dunno21:04
*** lindycod_ has quit IRC21:04
dansmithbauzas: unicode is not 8 bits21:04
mriedemyeah, you'd have to request a lower microversion to get around it21:04
dansmithmriedem: so I dunno, I'm inclined to say as long as neither will be too surprising or problematic for the user, do the simpler/easier thing21:05
*** ivase__ has quit IRC21:05
*** rtheis has quit IRC21:06
bauzasdansmith: well, unicode is not an encoding, so yeah it depends21:06
bauzasbut I'm probably nitpicking21:06
*** ivase has joined #openstack-nova21:07
bauzasbecause that's already in the cell DB21:07
dansmithbauzas: is this for my api patch?21:07
bauzasyup21:07
dansmithwhat table are you concerned about?21:07
bauzasResourceProviders21:07
dansmithI know that cdent intentionally made some of those fields latin121:07
bauzasbut like I said, that's already something done in the cell db21:07
dansmithwell, don't avoid being picky for something like this, but I think he probably made certain decisions as that was all done recently, and I remember a charset thing21:08
*** tpepper1 has joined #openstack-nova21:08
bauzasjust looking at the 318 DB migration21:09
*** raildo is now known as raildo-afk21:10
dansmithmight look at the review that added that migration21:10
openstackgerritWalter A. Boring IV (hemna) proposed openstack/nova: WIP Don't check cinder volume states during attach  https://review.openstack.org/31578921:10
bauzasright21:10
openstackgerritAugustina Ragwitz proposed openstack/nova: WIP: limits.inc example verification  https://review.openstack.org/31571421:11
*** ivase has quit IRC21:11
*** yamahata has quit IRC21:11
jaypipesbauzas: latin-1 is a subset of utf8, no?21:11
clarkbjaypipes: nope21:11
dansmitho.O21:11
openstackgerritMatt Riedemann proposed openstack/nova-specs: Get Me a Network  https://review.openstack.org/28320621:11
clarkbascii is though, but the non ascii bits of latin-1 don't overlap with utf8 iirc21:11
*** smatzek has quit IRC21:11
clarkbbecause they do differen things with that high bit21:11
bauzasutf8 is an encoding thing, AFAIK21:11
bauzaslatin1 is a charset21:11
jaypipesclarkb: ah, duh, yeah, was mixing ascii and latin-1. thx21:11
bauzaswhich are 2 different things21:12
clarkbbauzas: latin-1 is an encoding21:12
bauzasAFAIR, utf8 is how we encode the list of chars21:12
bauzasgraaah21:12
jaypipesbauzas: anyway, what is this in regards to? :)21:12
dansmithI have to run off for a bit, so if I have to change latin-1 to latin-2 or something, just leave a comment on the review21:13
* dansmith hopes there is no latin-2, making that joke hilarious21:13
bauzasjaypipes: I was a bit surprised to see an unicode mixed with latin-1, but like I said, it wasn't worth mentioning it yet in the review21:13
bauzasI need to load in mind a few bits before commenting21:14
clarkb0xFF in latin-1 is ÿ but is invalid in utf8 iirc21:14
bauzasanyway, wikipedia seems perfectly fine with that https://en.wikipedia.org/wiki/Latin-1_Supplement_%28Unicode_block%2921:15
*** ihrachys has quit IRC21:16
openstackgerritDiana Clarke proposed openstack/nova: Image: add create_from_image & create_from_func  https://review.openstack.org/31894821:16
clarkbya unicode and latin-1 do overlap I think. Just pointing out that latin-1 and utf8 do not21:16
bauzasI see21:17
*** ihrachys has joined #openstack-nova21:17
bauzasbecause unicode is a charset for noting all the chars we have, right?21:18
bauzasand then using an encoding that can be either latin-1 or utf-8, right?21:18
clarkbright21:19
bauzasokay21:19
openstackgerritDiana Clarke proposed openstack/nova: Image: add create_from_image & create_from_func  https://review.openstack.org/31894821:19
* bauzas loves OpenStack for that precise reason that I can learn things every day21:19
*** tpepper1 has quit IRC21:19
*** nic has joined #openstack-nova21:20
*** dmk0202 has quit IRC21:22
*** chip__ has quit IRC21:23
*** thorst_ has quit IRC21:23
jaypipesbauzas: I offer you a 💩21:23
openstackgerritClenimar Filemon proposed openstack/nova: Use Identity v3 params when connecting to Ironic  https://review.openstack.org/30015421:23
bauzasheh21:24
melwittaccording to stackoverflow latin-1 is single-byte encoding that can represent the first 256 unicode characters http://stackoverflow.com/questions/7048745/what-is-the-difference-between-utf-8-and-iso-8859-121:24
*** chip__ has joined #openstack-nova21:24
*** flwang1 has joined #openstack-nova21:24
*** matrohon has quit IRC21:25
stewie925hello melwitt - I have posted the error logs here - http://paste.openstack.org/show/497808/21:25
*** thorst_ has joined #openstack-nova21:25
stewie925had to re-create instance using public network nic21:26
*** pushkaru has quit IRC21:27
stewie925but failing with "Failed to allocate the network(s), not rescheduling"21:27
*** syjulian has quit IRC21:27
stewie925sorry I mean "had to create instance with PRIVATE nic"21:28
*** thorst_ has quit IRC21:29
*** ctrath has quit IRC21:29
melwittstewie925: you have more than one network in 'nova network-list'?21:31
stewie925yes I do21:31
stewie925a public and a private - I think thats how it was created when I ran stack.sh21:31
auggydansmith: there is latin-2 - http://www.ascii-codes.com/cp852.html21:32
openstackgerritTom Patzig proposed openstack/nova: Add ironic 'api_insecure' option to nova.conf  https://review.openstack.org/31116321:33
*** eharney has quit IRC21:33
openstackgerritSarafraj Singh proposed openstack/nova: Deprecate barbican options  https://review.openstack.org/31768421:34
lifelessalaski: cool21:34
stewie925melwitt: I did a grep on the instance-id and there are only entries in n-api, n-cell-region, n-cell-child, and n-cpu (nothing in n-sch or n-cond)21:34
*** imcsk8 has quit IRC21:34
*** ducttape_ has quit IRC21:35
melwittstewie925: hm. devstack only creates one network AFAIK, as seen in this log from a cells tempest job run http://logs.openstack.org/85/317785/1/check/gate-tempest-dsvm-cells/f8302ed/logs/devstacklog.txt.gz21:35
*** baoli has joined #openstack-nova21:36
stewie925melwitt, maybe I can just delete one of the networks, if that may be the cause of the issue?21:36
alaskilifeless: if anything else comes up let me know21:36
melwittstewie925: so it works when you pass the 'private' network with --nic? usually I don't specify the --nic option and it just picks by itself21:36
stewie925melwitt: I initially did not specify the --nic, but nova errored and asked me to specify it - maybe because it finds there are two networks21:37
*** imcsk8 has joined #openstack-nova21:37
openstackgerritMatt Riedemann proposed openstack/nova: WIP: Add REST API support for get me a network  https://review.openstack.org/31639821:37
openstackgerritMatt Riedemann proposed openstack/nova: Bump the service version for get-me-a-network support  https://review.openstack.org/31895321:37
*** ctrath has joined #openstack-nova21:37
melwittstewie925: oh, okay. does instance boot work when you use the 'private' network?21:37
*** doug-fish has quit IRC21:38
stewie925the error I have now is when I use the private ntetwork, actually21:38
melwittstewie925: also in the error case, I would expect to find errors in n-net21:38
*** ctrath has quit IRC21:38
stewie925ohhh - well i did not see a n-net in the list of services when I brought up devstacks screen21:39
*** sdake has quit IRC21:39
lifelessalaski: will do, I've poked another fixtures core for a review21:40
* bauzas head explodes with charsets vs. encodings21:40
melwittstewie925: that's probably why it's failing then. what's the status of 'nova-network' when you do 'nova service-list'21:40
stewie925checking - brb21:40
melwittbauzas: it's pretty new to me. learned from reading up based on your question21:40
stewie925melwitt: argh I see nove-compute is down21:41
bauzasmelwitt: I'm coming from the Web world, where you usually do unicode with utf-8 and basta.21:41
stewie925and nova-network is not listed21:41
*** efried has quit IRC21:41
*** efried has joined #openstack-nova21:42
jaypipesbauzas, melwitt, dansmith: https://review.openstack.org/#/c/316031/ should be ready to rock and roll. fixes an issue found in Ironic with new resource providers queries...21:42
stewie925maybe I should re-stack, then after successful stack I will check 'nova service-list'  ?21:42
melwittstewie925: nova-network should be down too if it's not running in your screen session. oh, that's a bit unhelpful, not running but no entry to say it's 'down'21:42
stewie925let me see the local.conf to see if enable_service n-net is present21:42
bauzasjaypipes: well, the patch uses the DB cell models, right?21:42
melwittstewie925: you shouldn't have to name it explicitly, it should do it by default21:43
bauzasjaypipes: which is a bit orthogonal to dansmith's change21:43
bauzasjaypipes: I mean https://review.openstack.org/#/c/316031/5/nova/objects/resource_provider.py@20521:43
mriedemalaski: dansmith: sdague: thread in the ops list about hitting the default 1000 limit for listing instances at scale, kind of interesting some of the views on that21:43
mriedeme.g. the one guy that just runs a direct query against the db21:44
stewie925uh oh21:44
*** doug-fish has joined #openstack-nova21:44
stewie925guess what melwitt - n-net was disabled21:44
melwittbauzas: yeah. since utf-8 can represent the complete unicode set you need to use it if you want the thing to be multilingual. I'm not knowledgeable about when to choose latin-121:44
stewie925I think that was how the local.conf was originally set up21:44
stewie925let me remove and rerun stack21:44
melwittstewie925: oh geez, that's not good21:45
*** thorst_ has joined #openstack-nova21:45
stewie925yeah ... strange21:45
bauzasmelwitt: well, I know that answer21:45
stewie925right now I have two VMs - one with the regular devstack,   and 2nd is this VM with devstack 'cell'21:45
bauzasmelwitt: utf8 can be very greedy21:45
*** lakshmiS has quit IRC21:45
bauzasmelwitt: in terms of bytes per char21:45
stewie925brb, checking the original devstack (no cell) VM - I know I was able to create instance successfully there21:46
melwittbauzas: yeah, so I guess if you know you can live with only the first 256, use the latin-1 to be efficient21:46
bauzasmelwitt: that's why you usually prefer a certain other encoding if you don't care about supporting all languages21:46
bauzasmelwitt: the problem is the last 12721:46
mriedemalright, i'm out of here. dan and sean and probably others have my personal email if people need to get a hold of me while on vacation. otherwise see you all on the 31st.21:46
bauzasmriedem: \o21:46
bauzasmriedem: Europe, right?21:47
mriedemyes21:47
melwittmriedem: o/ bon voyage21:47
bauzask,enjoy your trip21:47
bauzassafe travels too21:47
*** mdrabe has joined #openstack-nova21:47
mriedemDěkuji!21:47
*** mriedem has quit IRC21:47
bauzasa-ha, Praha... one of the locations I'd love to visit21:48
bauzaswell, Prague21:48
*** doug-fish has quit IRC21:49
*** thorst_ has quit IRC21:49
*** diogogmt has quit IRC21:50
*** doug-fish has joined #openstack-nova21:50
*** kfarr has quit IRC21:52
*** Sukhdev has quit IRC21:53
*** MikeG451 has quit IRC21:54
*** AmyLee71 has quit IRC21:54
*** doug-fish has quit IRC21:54
*** Sukhdev has joined #openstack-nova21:56
*** rlrossit_ has quit IRC21:56
*** daemontool has quit IRC21:57
dansmithauggy: blast!21:58
*** Swami has quit IRC22:01
openstackgerritRobert Ellis proposed openstack/nova: Changed 202 for lock/unlock return code to 204.  https://review.openstack.org/31426722:01
*** ducttape_ has joined #openstack-nova22:02
openstackgerritCarl Baldwin proposed openstack/nova-specs: Neutron Routed Networks  https://review.openstack.org/26389822:03
stewie925hey melwitt - before i re-stack with local.conf's   'disable n-net' removed22:03
stewie925should I just leave the nova-cells.conf and nova.conf alone22:04
melwittstewie925: yeah22:04
stewie925thank you22:05
*** edmondsw has quit IRC22:06
*** AmyLee7 has joined #openstack-nova22:06
*** ayoung has joined #openstack-nova22:08
openstackgerritOpenStack Proposal Bot proposed openstack/nova: Updated from global requirements  https://review.openstack.org/31886522:13
*** ducttape_ has quit IRC22:16
*** gjayavelu has quit IRC22:16
*** diogogmt has joined #openstack-nova22:19
*** nic has quit IRC22:21
*** rcernin has quit IRC22:23
*** tbachman has joined #openstack-nova22:24
*** mdrabe has quit IRC22:24
*** jaypipes has quit IRC22:25
*** zenoway has quit IRC22:25
*** zenoway has joined #openstack-nova22:26
*** markvoelker has joined #openstack-nova22:27
*** markvoelker has quit IRC22:29
*** markvoelker has joined #openstack-nova22:29
*** ducttape_ has joined #openstack-nova22:29
*** sdake has joined #openstack-nova22:30
*** zenoway has quit IRC22:30
*** kylek3h has quit IRC22:32
dansmithbauzas: I'm not sure what you're asking me to do22:35
dansmithbauzas: are you saying I have to create the table and then apply that special ddl execution to change the definition?22:35
*** mmandel has quit IRC22:35
*** lpetrut has quit IRC22:37
dansmithsurely there is some way to just specify the charset on the column in the table itself?22:38
*** sdague has quit IRC22:39
*** ducttape_ has quit IRC22:40
openstackgerritbailin.zhang proposed openstack/nova: Create VM failed, unplug vifs  https://review.openstack.org/30621922:40
*** tbachman_ has joined #openstack-nova22:41
*** ducttape_ has joined #openstack-nova22:42
*** tbachman has quit IRC22:42
*** tbachman_ is now known as tbachman22:42
*** tbachman has quit IRC22:44
*** doug-fish has joined #openstack-nova22:44
*** doug-fish has quit IRC22:44
*** sdake has quit IRC22:45
stewie925melwitt: very strange - a few nova services are down22:47
*** tbachman has joined #openstack-nova22:47
stewie925nova-conductor, nova-network, and nova-compute22:47
stewie925when I reran stack and it was successful (15 min ago), I checked and all nova services were up22:48
*** fragatin_ has joined #openstack-nova22:49
stewie925ok.. all services are back up now except nova-compute22:50
openstackgerritDan Smith proposed openstack/nova: Add resource provider tables to the api database  https://review.openstack.org/31568122:51
openstackgerritDan Smith proposed openstack/nova: Make compute nodes update their own inventory records  https://review.openstack.org/31704922:51
openstackgerritDan Smith proposed openstack/nova: Make Inventory and ResourceProvider objects use the API DB instead  https://review.openstack.org/31568222:51
melwittstewie925: when you're going to restack you should unstack.sh first and also make sure no screen session is still running, kill service processes if any are still running, etc before stack.sh again22:51
melwittstewie925: there should be an error in n-cpu.log if it's not up22:51
*** fragatina has quit IRC22:52
stewie925melwitt: I actually ran from a snapshot earlier - then I had to redo everything22:52
melwittstewie925: or if it failed to start altogether, stack.sh should have shown an error and exited non-zero22:52
stewie925I see that all nova services are now up...22:52
melwittokay22:53
stewie925thank you for these notes, melwitt22:53
*** josecastroleon has quit IRC22:54
stewie925oh btw melwitt - right after I ran stack.sh I still seee a private and public network, is that okay?22:55
*** diogogmt has quit IRC22:56
melwittstewie925: probably, I don't know why it's creating two networks for you, might be something in your local.conf. I don't have a current devstack handy to check22:57
*** ducttape_ has quit IRC22:57
stewie925thank you melwitt - no problem - hope this env works now with n-net already up :)22:57
*** markvoelker has quit IRC22:58
*** mrkz has quit IRC22:58
melwittthe local.conf I use is very minimal for cells http://paste.openstack.org/show/497816/22:59
*** zenoway has joined #openstack-nova22:59
*** dimtruck is now known as zz_dimtruck23:01
*** ametts has quit IRC23:04
*** ihrachys has quit IRC23:04
*** zenoway has quit IRC23:04
*** sdake has joined #openstack-nova23:04
*** Sukhdev has quit IRC23:05
*** Sukhdev has joined #openstack-nova23:08
*** stewie925 has quit IRC23:08
*** salv-orl_ has quit IRC23:12
*** sdake has quit IRC23:14
*** sdake has joined #openstack-nova23:15
*** ducttape_ has joined #openstack-nova23:15
artomdansmith, thanks for the reviews :) Don't spend *too* much time on it, as it's entirely possible tests are failing all over the place, because this is just the first patchset reorg to use the new "blob in database" approach23:16
dansmithartom: I know, I just want to make sure you have early feedback on structural things23:16
openstackgerritSivasathurappan Radhakrishnan proposed openstack/nova: Improve the help text for the rdp options  https://review.openstack.org/31897423:16
artomdansmith, http://www.goodpaper.com/images/21-010.jpg23:17
dansmithartom: feel free to ignore it until it's convenient, but if it saves a round-trip23:17
artom(Yes, I know this isn't reddit and we don't converse in image macros)23:17
dansmiththen it makes sense23:17
dansmithheh23:17
*** amotoki has quit IRC23:21
*** sdake has quit IRC23:21
*** amotoki has joined #openstack-nova23:21
*** ayoung has quit IRC23:22
*** Swami has joined #openstack-nova23:23
*** zenoway has joined #openstack-nova23:23
*** ducttape_ has quit IRC23:26
*** amotoki has quit IRC23:26
*** Sukhdev has quit IRC23:27
*** sorrison_laptop has joined #openstack-nova23:27
*** imcsk8 has quit IRC23:27
*** imcsk8 has joined #openstack-nova23:27
*** zenoway has quit IRC23:28
*** zz_dimtruck is now known as dimtruck23:31
*** stewie925 has joined #openstack-nova23:32
stewie925hi thank you Melwitt for sharing your local.conf - it sure is very minimal23:32
stewie925mine has a lot of junk in it23:32
*** BjoernT has quit IRC23:37
*** dimtruck is now known as zz_dimtruck23:41
*** baoli has quit IRC23:44
*** zenoway has joined #openstack-nova23:48
*** ayoung has joined #openstack-nova23:49
*** sorrison_laptop has quit IRC23:49
*** jamielennox is now known as jamielennox|away23:49
*** sorrison_laptop has joined #openstack-nova23:49
*** fragatin_ has quit IRC23:50
*** fragatina has joined #openstack-nova23:51
*** zz_dimtruck is now known as dimtruck23:51
*** sorrison_laptop has quit IRC23:52
*** sorriso__ has joined #openstack-nova23:52
*** sdake has joined #openstack-nova23:52
*** zenoway has quit IRC23:52
*** ducttape_ has joined #openstack-nova23:54
*** kylek3h has joined #openstack-nova23:56
*** sorrison_laptop has joined #openstack-nova23:56
*** sorriso__ has quit IRC23:56

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