Thursday, 2016-09-08

*** sabeen has joined #openstack-ironic00:03
*** snizar_ has quit IRC00:14
*** snizar has quit IRC00:14
*** sabeen has quit IRC00:21
*** sabeen has joined #openstack-ironic00:28
*** harlowja has joined #openstack-ironic00:29
*** wajdi has joined #openstack-ironic00:42
*** hoangcx has joined #openstack-ironic00:46
*** wajdi_ has joined #openstack-ironic00:47
*** wajdi has quit IRC00:50
*** jabroney1 has quit IRC00:58
*** chlong has quit IRC01:03
*** joprovost has joined #openstack-ironic01:07
*** Sukhdev_ has quit IRC01:10
*** syed_ has quit IRC01:12
*** snizar has joined #openstack-ironic01:14
*** snizar_ has joined #openstack-ironic01:14
openstackgerritJohn L. Villalovos proposed openstack/ironic: Update unit tests for neutron interface  https://review.openstack.org/36703801:19
*** jvgrant has joined #openstack-ironic01:26
*** wuhg has joined #openstack-ironic01:31
*** jvgrant has quit IRC01:31
*** sdake has joined #openstack-ironic01:32
*** rloo has joined #openstack-ironic01:39
*** rloo has left #openstack-ironic01:39
*** openstack has joined #openstack-ironic01:44
*** sabeen has quit IRC01:44
*** sabeen has joined #openstack-ironic01:46
*** rloo has joined #openstack-ironic01:46
*** rloo has left #openstack-ironic01:46
*** baoli has joined #openstack-ironic01:50
*** thrash is now known as thrash|g0ne02:01
*** baoli has quit IRC02:08
*** tuanla_ has joined #openstack-ironic02:12
*** hoangcx has quit IRC02:16
*** chlong has joined #openstack-ironic02:19
*** hoangcx has joined #openstack-ironic02:19
*** snizar_ has quit IRC02:25
*** snizar has quit IRC02:25
*** baoli has joined #openstack-ironic02:37
*** vishwana_ is now known as vishwanathj_zzz02:38
*** yuanying has quit IRC02:47
*** wajdi_ has quit IRC02:48
*** hoangcx has quit IRC02:51
*** hoangcx has joined #openstack-ironic02:54
*** joprovost has quit IRC03:00
*** rbudden has quit IRC03:02
*** vikrant has joined #openstack-ironic03:09
*** hj-hpe has quit IRC03:17
*** mkam has joined #openstack-ironic03:20
*** hoangcx has quit IRC03:27
*** praneshp has joined #openstack-ironic03:28
*** vikrant has quit IRC03:33
*** mkam has quit IRC03:34
*** vikrant has joined #openstack-ironic03:34
*** hoangcx has joined #openstack-ironic03:35
openstackgerritXavier proposed openstack/ironic: Add inspection feature for the OneView drivers  https://review.openstack.org/36706503:38
*** yuanying has joined #openstack-ironic03:51
*** vmud213 has joined #openstack-ironic03:58
*** baoli has quit IRC04:01
*** sabeen has quit IRC04:18
*** amotoki has joined #openstack-ironic04:23
*** links has joined #openstack-ironic04:28
*** harlowja_at_home has joined #openstack-ironic04:29
*** vmud213 has quit IRC04:31
openstackgerritAnkit Kumar proposed openstack/proliantutils: Fix Inspection failure in Gen7  https://review.openstack.org/35713404:32
*** vmud213 has joined #openstack-ironic04:35
*** fragatina has joined #openstack-ironic04:41
*** vmud213 has quit IRC04:44
*** fragatina has quit IRC04:45
openstackgerritJohn L. Villalovos proposed openstack/ironic: Update unit tests for neutron interface  https://review.openstack.org/36703804:47
*** vmud213 has joined #openstack-ironic04:47
openstackgerritJohn L. Villalovos proposed openstack/ironic: Update unit tests for neutron interface  https://review.openstack.org/36703804:49
*** sdake has quit IRC04:56
*** moshele has joined #openstack-ironic04:58
*** sdake has joined #openstack-ironic05:00
*** jaosorior has joined #openstack-ironic05:10
*** chlong has quit IRC05:14
*** ChubYann has quit IRC05:23
*** phuongnh has joined #openstack-ironic05:29
*** chlong has joined #openstack-ironic05:30
*** links has quit IRC05:31
*** harlowja_at_home has quit IRC05:46
*** links has joined #openstack-ironic05:46
*** fragatina has joined #openstack-ironic05:52
*** sdake_ has joined #openstack-ironic05:56
*** fragatina has quit IRC05:56
openstackgerritVasyl Saienko proposed openstack/ironic: DO NOT REVIEW  https://review.openstack.org/29643205:57
*** sdake has quit IRC05:58
*** baoli has joined #openstack-ironic06:13
*** chlong has quit IRC06:13
*** baoli has quit IRC06:17
*** e0ne has joined #openstack-ironic06:22
*** hoonetorg has quit IRC06:25
*** mjura has joined #openstack-ironic06:25
*** chlong has joined #openstack-ironic06:26
*** vmud213 has quit IRC06:40
*** vmud213 has joined #openstack-ironic06:41
*** vmud213 has quit IRC06:48
*** vmud213 has joined #openstack-ironic06:49
*** vmud213 has quit IRC06:54
*** hoangcx_ has joined #openstack-ironic06:54
*** vmud213 has joined #openstack-ironic06:55
*** hoangcx has quit IRC06:57
*** fxpester has joined #openstack-ironic06:57
*** tesseract- has joined #openstack-ironic07:04
*** e0ne has quit IRC07:22
openstackgerritPavlo Shchelokovskyy proposed openstack/ironic: Refactor _test_build_pxe_config_options tests  https://review.openstack.org/36714207:27
*** rbartal has joined #openstack-ironic07:28
*** ifarkas_afk is now known as ifarkas07:33
*** openstackgerrit has quit IRC07:33
*** openstackgerrit has joined #openstack-ironic07:34
*** praneshp has quit IRC07:36
*** dtantsur|afk is now known as dtantsur07:42
dtantsurMorning Ironic07:42
*** hoangcx_ has quit IRC07:43
dtantsuraarefiev, hey! what's the situation with grenade now? I see you're trying some reverts in tempest.07:43
aarefievdtantsur: morning, so the problem in paramiko, that patch fixes deadlocks in paramiko when data bunch transfered, but somehow causes hangs07:46
dtantsurdamn...07:46
dtantsurso reverting it is not the best option as well, right?07:46
aarefievI've ran few times w/o this patch and grenade pass all times07:47
aarefievI mean with reverted07:47
dtantsuraarefiev, could you please bring it to the ML with [qa] tag? the time is pressing, and we need assistance from qa folks, I guess...07:47
aarefievyeah, if we just revert it will fail on other jobs07:48
dtantsuraarefiev, now, the only thing I don't understand is why one we are affected by this problem07:50
aarefievdtantsur: sure, one thing I don't get it why only grenade in inspector?07:50
dtantsurlol07:50
aarefievyeah :)07:50
dtantsuraarefiev, is it possible that we have a different configuration? I think you've tried changing some timeouts..07:50
aarefievdtantsur: looks like it doesn't depend on timeouts07:51
*** ohamada has joined #openstack-ironic07:51
dtantsuraarefiev, also this patch merged successfully: https://review.openstack.org/#/c/365902/ Oo07:51
patchbotpatch 365902 - ironic-inspector - TrivialFix: Remove cfg import unused (MERGED)07:51
aarefievmerged ^07:51
aarefievahh07:51
*** jpich has joined #openstack-ironic07:52
dtantsurlemme update https://review.openstack.org/#/c/358779/ and see if it passes or fails07:52
patchbotpatch 358779 - ironic-inspector - Log hook names as we run them07:52
*** chlong has quit IRC07:52
aarefievI recheck patch with pci, lets see07:52
dtantsuraarefiev, this passed grenade too: https://review.openstack.org/#/c/365662/ please review07:54
patchbotpatch 365662 - ironic-inspector - Set node to the error if reapply fails07:54
dtantsurI wonder if there is a problem with that patch in the end...07:54
openstackgerritDmitry Tantsur proposed openstack/ironic-inspector: Log hook names as we run them  https://review.openstack.org/35877907:55
dtantsuralso updated ^^^07:55
aarefievdtantsur: looking, I wondering if it was some network issue, but why affects only inspector07:55
dtantsuraarefiev, maybe something wrong with the default introspection network?07:56
dtantsurit does not tell much about why only this patch has such severe problems...07:56
aarefievyeah, and test doesn't related to patch code07:57
*** hoangcx has joined #openstack-ironic07:57
tangchenmorning all :)07:57
dtantsurmorning tangchen07:58
*** zzzeek has quit IRC08:00
tangchendtantsur: Hi Dmitry, would you please help to confirm this trivial fix ?  https://review.openstack.org/#/c/357639/08:00
patchbotpatch 357639 - python-ironicclient - Support i18n for osc plugin help msg08:00
*** zzzeek has joined #openstack-ironic08:00
dtantsurtangchen, it's on my radar, but now I'm busy figuring out urgent things, so I might not get to it immediately08:00
*** athomas has joined #openstack-ironic08:01
tangchendtantsur: Sure, please help to review when you are free. :)  Thanks.08:01
lucasagomesmorning all08:02
tangchenmorning lucasagomes08:03
dtantsurhey lucasagomes08:05
aarefievmorning lucasagomes, tangchen08:05
tangchenHi aarefiev08:05
*** Romanenko_K has joined #openstack-ironic08:14
aarefievdtantsur: crap, and now it fails with request timeout o_0 http://logs.openstack.org/95/352295/9/check/gate-grenade-dsvm-ironic-inspector/b6e3aa7/console.html08:17
*** sdake_ is now known as sdake08:17
*** vmud213 has quit IRC08:18
aarefievsomething changed08:18
*** vmud213 has joined #openstack-ironic08:18
dtantsurwut?08:19
aarefievdtantsur: it failed with test timeout because ping hangs and now with requests timeout, whats going on?08:21
dtantsurno idea..08:29
openstackgerritDmitry Tantsur proposed openstack/python-ironic-inspector-client: DO NOT MERGE: testing gate  https://review.openstack.org/36717108:34
openstackgerritDmitry Tantsur proposed openstack/python-ironic-inspector-client: Add oslotest to test-requirements  https://review.openstack.org/36720608:35
*** vmud213 has quit IRC08:37
*** vmud213 has joined #openstack-ironic08:38
*** jaosorior is now known as jaosorior_lunch08:40
*** snecklifter has quit IRC08:45
*** sborkows has joined #openstack-ironic08:46
*** moshele has quit IRC08:47
*** moshele has joined #openstack-ironic08:48
dtantsuraarefiev, seems like grenade is completely down :( yeah, we need some external help, I guess08:54
*** derekh has joined #openstack-ironic08:55
aarefievdtantsur: I'm writing letter to openstack-dev now08:56
dtantsurthanks!08:56
sborkowsmorning all08:57
openstackgerritVasyl Saienko proposed openstack/ironic: DO NOT REVIEW  https://review.openstack.org/29643208:57
openstackgerritDmitry Tantsur proposed openstack/python-ironic-inspector-client: DO NOT MERGE: testing gate  https://review.openstack.org/36717109:03
dtantsurmorning sborkows09:09
openstackgerritImre Farkas proposed openstack/ironic: Fix DRAC failure during automated cleaning  https://review.openstack.org/36722909:10
ifarkasmorning all09:10
*** lucasagomes is now known as lucas-relocate09:21
*** mbound has joined #openstack-ironic09:30
dtantsurmorning ifarkas09:33
*** vnogin has left #openstack-ironic09:37
*** yibo has quit IRC09:37
*** sambetts|afk is now known as sambetts09:44
*** e0ne has joined #openstack-ironic09:44
*** mgould|afk is now known as mgould09:48
mgouldmorning aarefiev dtantsur jaosorior_lunch ifarkas lucas-relocate sambetts and everyone else!09:49
dtantsurhey mgould09:50
sambettsMorning all09:54
dtantsurmorning sambetts09:54
sambettso/ dtantsur :( about the grenade job09:55
dtantsuryeah... dunno what to do, let's see if somebody answers on the ML09:55
*** yibo has joined #openstack-ironic09:58
openstackgerritSam Betts proposed openstack/python-ironicclient: Add python API for node interface attach/detach  https://review.openstack.org/36442010:03
openstackgerritSam Betts proposed openstack/python-ironicclient: DO NOT MERGE Testing Client Changes with Nova  https://review.openstack.org/36674710:03
*** lucas-relocate is now known as lucasagomes10:05
*** yibo has quit IRC10:05
lucasagomesmgould, sambetts aarefiev *all* morning :-)10:05
openstackgerritMerged openstack/ironic: Log a warning for unsupported drivers and interfaces  https://review.openstack.org/36639910:07
*** hoangcx has quit IRC10:10
mgouldhey everyone, did we support whole-disk images in Kilo?10:28
mgouldalso, what's the deal with whole-disk images and EFI support? "Make sure your image has an EFIboot partition and it should Just Work"?10:29
*** phuongnh has quit IRC10:30
openstackgerritSergii Turivnyi proposed openstack/python-ironicclient: Negative tests for testing actions with Chassis.  https://review.openstack.org/34379610:32
dtantsurmgould, I guess that means that we don't need anything fancy for local boot10:39
dtantsurmgould, I'm pretty sure we supported whole-disk images, not sure if we supported them with pxe_* family though10:40
mgoulddtantsur: thanks!10:45
*** amotoki has quit IRC10:48
*** sturivnyi_ has joined #openstack-ironic10:50
*** jcoufal has quit IRC10:50
sturivnyi_Hi =) I have bunch of patches for review https://review.openstack.org/#/q/status:open+project:openstack/python-ironicclient+branch:master+topic:%22functional+CLI+tests%22. Would be grateful if you could review them all.10:55
sturivnyi_This is functional CLI tests for python-ironicclient10:57
*** baoli has joined #openstack-ironic10:57
openstackgerritDmitry Tantsur proposed openstack/ironic: [WIP] Abort cleaning and deployment if node is in maintenance  https://review.openstack.org/36682810:59
dtantsurjroll, this is pretty close now ^^^ wdyt?11:01
*** baoli has quit IRC11:05
openstackgerritAparna proposed openstack/proliantutils: Adds support in hpssa for SDD interface 'Solid State SATA'  https://review.openstack.org/36728711:09
*** baoli has joined #openstack-ironic11:13
openstackgerritSam Betts proposed openstack/ironic: Add Network Interface Attach and Detach APIs  https://review.openstack.org/32704611:19
*** baoli has quit IRC11:19
* sambetts bbiab 11:20
*** sambetts is now known as sambetts|afk11:20
*** baoli has joined #openstack-ironic11:24
TheJuliaGood morning everyone11:27
sturivnyi_Morning <TheJulia>)11:28
*** vmud213 has quit IRC11:28
TheJuliamgould: I know a number of folks that have worked with whole disk images pre-built with EFI, and as long as the image is properly setup with an EFI partition and all, it seems to be \o/ for everyone that I know that uses it.11:29
dtantsurmorning TheJulia11:30
lucasagomesTheJulia, good morning11:30
*** baoli has quit IRC11:30
mgouldTheJulia: great, thanks! Also, morning :-)11:31
TheJuliamgould: the only headache that I've seen people run into is network cards that won't run without being booted via EFI.  The deployment ended up having to use virtual media with an EFI boot config on it too.11:32
TheJuliaBut, I've only heard of that one case.11:32
mgouldTheJulia: great, thanks11:34
* mgould thinks back to the innocent happy days a year ago before he learned just how complicated booting computers could be...11:35
TheJuliaAhhh the simpler computing times...11:36
*** fragatina has joined #openstack-ironic11:36
mgould:-)11:39
mat128morning everyone11:39
*** fragatina has quit IRC11:41
TheJuliagood morning mat12811:42
openstackgerritLucas Alvares Gomes proposed openstack/ironic: Grub2 by default for PXE + UEFI  https://review.openstack.org/36730611:42
* lucasagomes cleaning some backlog11:42
*** baoli has joined #openstack-ironic11:43
*** jcoufal has joined #openstack-ironic11:46
TheJulialucasagomes: I posted feedback, super quick fix to the release notes and I think it will be good.  Also left a nit on it that I don't have any strong feelings about and can be ignored if you want.11:47
lucasagomesTheJulia, thanks a lot, will take a look11:47
lucasagomesjust gonna grab something to eat and I will fix it :-)11:48
*** lucasagomes is now known as lucas-hungry11:48
openstackgerritTang Chen proposed openstack/ironic: Prepare for functional tests  https://review.openstack.org/34942111:49
TheJuliafood sounds delicious right now11:52
*** links has quit IRC11:52
TheJuliaI just finished my first cup of coffee11:52
*** jaosorior_lunch is now known as jaosorior11:56
*** thrash|g0ne is now known as thrash11:56
*** baoli_ has joined #openstack-ironic11:56
*** baoli has quit IRC12:00
*** sambetts|afk is now known as sambetts12:00
openstackgerritPavlo Shchelokovskyy proposed openstack/ironic: iPXE allow skip caching in standalone mode  https://review.openstack.org/36732112:02
*** vikrant has quit IRC12:07
openstackgerritTuan Luong-Anh proposed openstack/ironic: Change $LOGFILE path in configuration devstack  https://review.openstack.org/36651812:07
openstackgerritTuan Luong-Anh proposed openstack/ironic: Change $LOGFILE path in configuration devstack  https://review.openstack.org/36651812:08
*** leitan has joined #openstack-ironic12:08
*** dprince has joined #openstack-ironic12:09
*** trown|outtypewww is now known as trown12:09
openstackgerritTuan Luong-Anh proposed openstack/ironic: Change $LOGFILE path in configuration devstack  https://review.openstack.org/36651812:10
*** vnogin has joined #openstack-ironic12:11
openstackgerritLucas Alvares Gomes proposed openstack/ironic: Grub2 by default for PXE + UEFI  https://review.openstack.org/36730612:11
*** vnogin has left #openstack-ironic12:11
*** bfournie has joined #openstack-ironic12:12
*** wuhg has quit IRC12:21
*** xavierr has quit IRC12:21
openstackgerritAparna proposed openstack/proliantutils: Adds support in hpssa for SDD interface 'Solid State SATA'  https://review.openstack.org/36728712:24
*** rbudden has joined #openstack-ironic12:25
openstackgerritTang Chen proposed openstack/ironic: Prepare for functional tests  https://review.openstack.org/34942112:27
TheJulialucas-hungry: wouldn't it also make sense to update the sample config? (sorry!)12:37
lucas-hungryTheJulia, ++ totally12:38
*** lucas-hungry is now known as lucasagomes12:38
lucasagomesTheJulia, lemme update it :-)12:38
* lucasagomes see a lot of unrelated changes in the sample :-O12:39
TheJulia:(12:39
openstackgerritLucas Alvares Gomes proposed openstack/ironic: Grub2 by default for PXE + UEFI  https://review.openstack.org/36730612:42
openstackgerritLucas Alvares Gomes proposed openstack/ironic: Update ironic/ironic.conf.sample  https://review.openstack.org/36734212:42
lucasagomesdone :-)12:42
TheJulialucasagomes: Thank you :)12:43
lucasagomesthank YOU for reviewing it12:43
*** amotoki has joined #openstack-ironic12:48
*** mmedvede has quit IRC12:49
*** jaosorior has quit IRC12:51
*** jaosorior has joined #openstack-ironic12:51
jrollmorning everyone12:57
lucasagomesjroll, g'morning12:57
jrollonly here for a few, have an appt to run to shortly12:57
jrolldtantsur: I guess I'm fine with that change12:58
jrollnot a huge fan but I won't block it12:58
*** skazi has quit IRC12:58
*** rloo has joined #openstack-ironic12:58
TheJuliaGood morning jroll12:59
jrollmorning lucasagomes, TheJulia :)12:59
dtantsurjroll, what are your concerns about it? also g'morning13:00
jrolldtantsur: nothing really, just changing what I'm used to. morning :)13:00
dtantsurheh, I see :)13:01
dtantsurI'll finish unit tests, do some local testing, then remove WIP13:01
dtantsurpas-ha, answering your question ^^^13:01
*** mjturek has left #openstack-ironic13:05
*** jcoufal has quit IRC13:05
*** mjturek has joined #openstack-ironic13:06
*** jcoufal has joined #openstack-ironic13:07
jrolldtantsur: patch lgtm fwiw13:07
*** tuanla_ has quit IRC13:07
*** ohamada has quit IRC13:08
*** ohamada has joined #openstack-ironic13:08
*** jheroux has joined #openstack-ironic13:09
openstackgerritKyrylo Romanenko proposed openstack/python-ironicclient: Add basic tests for OSC plugin baremetal driver commands  https://review.openstack.org/36735913:11
*** Goneri has joined #openstack-ironic13:11
*** skazi has joined #openstack-ironic13:11
*** jraju has joined #openstack-ironic13:12
*** cdearborn has joined #openstack-ironic13:12
TheJuliadtantsur: I posted some stuff for the release note entries, just suggestions really.13:14
TheJuliastuff being comments on the revision13:15
dtantsurthanks!13:15
*** mmedvede has joined #openstack-ironic13:17
*** sdake_ has joined #openstack-ironic13:20
openstackgerritAnkit Kumar proposed openstack/proliantutils: Fix Inspection failure in Gen7  https://review.openstack.org/35713413:21
*** sdake has quit IRC13:22
hwoarangHello does anyone know how to 'force' delete a broken node which is stuck in the 'inspecting' state? ironic node-delete does not seem able to do it.13:24
TheJuliahwoarang: I would think that it would eventually time out and once it has timed out I _think_ you could then delete it13:26
dtantsurhwoarang, it it in-band inspection with ironic-inspector or some OOB inspection?13:26
hwoarangin-band13:27
hwoarangbut i may be tempted to drop the entry from the DB directly if it fails to change the state soonish13:27
hwoarangTheJulia: i need you for one more thing regarding bifrost13:27
TheJuliahwoarang: sure! :)13:28
hwoarangThis change 5506f5e327c14a2008c6321a2e78a772c4b9a540 "Only set hostname on 127.0.0.1 if not present in /etc/hosts" broke some stuff for me13:28
TheJuliabroke how so?13:28
hwoarangthe reason is that in my host, the hostname in /etc/hosts is set in the local 10.X.X.X IP13:28
hwoaranghowever, epmd runs on 127.0.0.1 so when the rabbitmq-server task tries to start the service13:29
openstackgerritAnkit Kumar proposed openstack/proliantutils: Fix Inspection failure in Gen7  https://review.openstack.org/35713413:29
hwoarangit fails, because the hostname resolves to 10.X.X.X instead of 127.0.0.113:29
dtantsurhwoarang, you can abort in-band inspection via inspector command line13:29
*** baoli_ has quit IRC13:29
dtantsurhwoarang, something like 'openstack baremetal introspection abort UUID'13:29
openstackgerritSam Betts proposed openstack/ironic: Add Network Interface Attach and Detach APIs  https://review.openstack.org/32704613:29
TheJuliahwoarang: Okay, I understand :(13:29
openstackgerritDmitry Tantsur proposed openstack/ironic: Abort cleaning and deployment if node is in maintenance  https://review.openstack.org/36682813:30
dtantsurupdated ^^^13:30
hwoarangTheJulia: it all goes away when i add the hostname to 127.0.0.1 as it used to be. but clearly, according to the original gerrit patch, this break the case for others...13:30
TheJuliahwoarang: I thought the behavior had been fixed too :(13:32
hwoarangdtantsur: thanks let me try that13:33
*** joprovost has joined #openstack-ironic13:33
hwoarangTheJulia: tricky situation...13:33
*** rbartal has quit IRC13:34
hwoarangif we keep it like that then we need to be able to instruct epmd to listen on the correct address. and perhaps a few other bits and pieces which i forget about at the moment13:35
TheJuliaI seem to remember though, that epmd essentially expects the hostname to resolve to 127.0.0.1 which actually should resolve to localhost, and that is why that was there13:35
TheJuliaThis has given me too many migraines :)13:36
hwoarangright13:37
hwoarang:(13:37
hwoarangsounds like a horrible hack has to be applied somewhere :)13:39
*** mbound_ has joined #openstack-ironic13:39
*** mbound has quit IRC13:41
*** hshiina has joined #openstack-ironic13:44
hwoarangTheJulia: is it ok if I reply to the old gerrit change: https://review.openstack.org/#/c/355778/ and discuss with the author how to solve both cases?13:50
patchbotpatch 355778 - bifrost - Only set hostname on 127.0.0.1 if not present in /... (MERGED)13:50
*** Administrator_ has joined #openstack-ironic13:51
*** wajdi has joined #openstack-ironic13:52
*** zhugaoxiao has quit IRC13:54
openstackgerritKyrylo Romanenko proposed openstack/python-ironicclient: Functional tests for portgroups in ironicclient  https://review.openstack.org/29772714:00
*** anush has joined #openstack-ironic14:00
TheJuliahwoarang: absolutely!14:01
hwoarangok lets do that14:01
TheJuliahwoarang: this is something I need to ponder too :(14:01
*** amotoki has quit IRC14:01
*** rajinir has joined #openstack-ironic14:02
*** amotoki has joined #openstack-ironic14:02
*** joprovost1 has joined #openstack-ironic14:02
* jroll is back from appt14:03
*** joprovost has quit IRC14:03
*** joprovost1 is now known as joprovost14:03
mat128lucasagomes: just saw your comment on the docs patch, would you mind reviewing the initial one? https://review.openstack.org/#/c/363294/  It needs to merge before the project-config change can be merged14:06
patchbotpatch 363294 - ironic - Add a basic install guide14:06
lucasagomesmat128, hi there!14:06
mat128o/14:06
lucasagomesyes I will do, just finishing something first and then I will take a look :D14:06
mat128lucasagomes: oh sure, I didn't mean it as a "look at this now", more like "if you still want to look at install-guide stuff"14:07
*** sdake has joined #openstack-ironic14:10
*** sdake_ has quit IRC14:12
jlvillallucasagomes, On https://review.openstack.org/#/c/334967/  I'm not sure when devananda|travel will be back.14:12
patchbotpatch 334967 - ironic - Fix non-deployed nodes due to short nodes.instance...14:12
jlvillalDo you know?14:12
*** sabeen has joined #openstack-ironic14:13
openstackgerritDmitry Tantsur proposed openstack/ironic-inspector: Fix breakage with oslo.config master  https://review.openstack.org/36740714:13
*** sabeen1 has joined #openstack-ironic14:15
*** sabeen has quit IRC14:17
*** amotoki has quit IRC14:17
*** xavierr has joined #openstack-ironic14:17
openstackgerritJulia Kreger proposed openstack/ironic: Add volume_connectors and volume_targets to task  https://review.openstack.org/36619614:17
openstackgerritJulia Kreger proposed openstack/ironic: [WIP] Add storage_interface DB field and object  https://review.openstack.org/34800514:18
*** vishwanathj_zzz is now known as vishwanathj14:18
TheJuliajlvillal: I _suspect_ he will be back today14:18
jlvillalTheJulia, Cool :)14:19
*** syed_ has joined #openstack-ironic14:19
*** hj-hpe has joined #openstack-ironic14:22
*** bnemec has quit IRC14:24
*** vmud213 has joined #openstack-ironic14:25
*** bnemec has joined #openstack-ironic14:27
*** amotoki has joined #openstack-ironic14:28
*** joprovost1 has joined #openstack-ironic14:29
openstackgerritJulia Kreger proposed openstack/ironic: [WIP] Add base cinder common interface  https://review.openstack.org/35562514:29
openstackgerritJulia Kreger proposed openstack/ironic: [WIP] Add storage_interface to base driver class  https://review.openstack.org/34800614:29
*** moshele has quit IRC14:29
jrollhrm, I worry about the online-ness of ALTER TABLE there14:29
jroll(on 334967)14:29
TheJuliayeah, table would be locked. :\14:30
jrollokay, so it does nuke rolling upgrades14:30
TheJuliaWell, depends on what you term rolling upgrades :)14:30
*** joprovost has quit IRC14:30
*** joprovost1 is now known as joprovost14:30
TheJuliaAFAIK, selects and updates against the table would just hang until the alter is complete14:31
jrollonline? :)14:31
jrollyeah14:31
jrollprevious jobs with 1-hour alters have scarred me14:31
TheJuliaI can't think of another way short of having an online data migration tool built into ironic that moves data to new tables and knows where to look14:31
jrollwell, we could use a new column name14:31
TheJuliatrue14:32
jrolland translate it in the object layer14:32
TheJuliawould be the perfect way to migrate it out of the nodes table then14:32
TheJuliaout of curiosity, was that 1 hour alter with ironic?14:33
*** amotoki has quit IRC14:33
jrollTheJulia: I guess it's okay with 5.6+ http://dev.mysql.com/doc/refman/5.6/en/innodb-create-index-overview.html14:34
TheJuliaactually, don't answer that :)14:34
jrollTheJulia: no, previous jobs, janky django app with x million row tables14:34
jrollwith many columns14:34
TheJuliaoh neat, so queries should continue to work according to that14:35
jrollassuming folks are running 5.614:36
*** vmud213 has quit IRC14:36
TheJuliayeah14:36
jrolltrusty is 5.514:36
jlvillaljroll, I did ask him about rolling upgrades and he responded near the bottom14:36
jrolljlvillal: yes, I read that, and the answer was not sufficient for me14:36
jrolljlvillal: my question is, how much does it hurt to read-lock the nodes table in a busy ironic environment14:36
jlvillalAh, okay.14:36
* jlvillal doesn't know14:37
TheJuliaI can't imagine it being all that bad with the existing structure of the table, which is the only reason I feel moderately okay with the change14:37
jrollgiven we don't officially support rolling upgrades, it's probably fine14:37
*** scarab_ has joined #openstack-ironic14:38
TheJuliaonce that change is in place though, table schema changes could get very bad14:39
* TheJulia is liking the alternate table idea more and more14:39
jrollI'd love to have a key/value table for each of our json blobs14:39
jrollor use postgres json support :P14:39
TheJuliaheh14:40
openstackgerritLucas Alvares Gomes proposed openstack/ironic: Grub2 by default for PXE + UEFI  https://review.openstack.org/36730614:40
jrollJayF: I might like your deployer hat's opinion on this, please see the comments around upgrades and table locks and such https://review.openstack.org/#/c/334967/814:41
patchbotpatch 334967 - ironic - Fix non-deployed nodes due to short nodes.instance...14:41
*** mkam has joined #openstack-ironic14:41
TheJuliajroll: or just move to a database that can be distributed super easily, and start building the concept of geographic locations in *ducks*14:42
jrollTheJulia: brb switching everything to riak14:42
*** jvgrant has joined #openstack-ironic14:43
dtantsurwe need something that is Web Scale, MySQL is not Web Scale, hence all the problems14:43
*** baoli has joined #openstack-ironic14:43
JayFjroll: that's a huge ass field14:43
* TheJulia awaits a link to the webscale video14:43
*** michauds has joined #openstack-ironic14:43
jrollJayF: yeeeeees?14:44
jrollJayF: more curious about upgrade implications, like, do you think that will cause a significantly longer deploy than usual, etc14:44
JayFI don't know enough about dbs to really make a good statement on that14:45
JayFbut I will say if someone puts anything near a gig of data in that field14:45
JayFperformance will tank14:45
JayFit seems very unwise to leave that field essentially unbounded14:45
jrolldtantsur: obligatory https://github.com/dcramer/mangodb/14:45
jrollJayF: well, it holds a configdrive, so 32mb is possible without trying hard14:46
jroller, not quite because compression, but ya know14:46
TheJuliajroll: writes would also get blocked, no deployments, no heartbeats, until the table copy is finished.  I just changed my vote to -1 on it14:46
JayFfrom TheJulia "My only fear is this is something that can never really be un-done, and could result in some very large database entries."14:46
JayFand I agree with that, a lot14:46
JayFsure, but there's a lot of room between 32mb and 4gb14:46
jrollTheJulia: well, we don't currently support running database migrations while services are up, is the argument for it14:46
jrollcan longtext specify a size?14:46
*** causten has quit IRC14:47
jrolllike, clearly someone has hit more than 64k14:47
JayFidk, dbs not my strongsuit14:47
jrollif they took the time to do this14:47
JayFbut 64k is not that much, 4gb is way too much14:47
JayFthere's gotta be a better middle ground14:47
*** rpioso has joined #openstack-ironic14:47
jlvillalWe have a customer that has hit greater than 64KB. And they don't use swift...14:47
jrollyeah, I'm not sure if there is a MEDIUMTEXT14:47
jroll:P14:47
TheJuliajroll: I think so, but only artificially I think if I'm remembering what I read last month correctly.  Realistically we break after 64MB, but even that is rather big and would be a very long database migration14:48
jrollTheJulia: 64KB14:48
*** rbrud has joined #openstack-ironic14:48
dtantsur64KB is enough for everyone14:48
TheJuliaI mean, if I supply a URL for the config drive instead of the config drive contents, it can't be larger than 64MB14:48
* dtantsur ducks14:49
jrollheh14:49
*** jvgrant has quit IRC14:49
jrollTIL there *is* a MEDIUMTEXT, which is max 16 MiB14:49
dtantsurbut yeah, tripleo uses config drive pretty extensively, it's a miracle we never maxed 64 KB14:49
JayFFor something like this, I'm also tempted to say like14:49
dtantsurhmm, 16 MiB is a huge improvement14:50
TheJuliawell, they are compressed14:50
JayFWONTFIX14:50
jlvillalThe configdrive I believe is sent via the REST API to the conductor. And then stored. If setup to use Swift, it will save the configdrive to Swift and instead store the URL14:50
JayFbecause it's just not an awesome pattern14:50
JayFI'd like to see someone run this patch14:50
JayFand successfully deploy a configdrive of triple-digit-mb size14:50
*** mkam has quit IRC14:50
JayFI think we're pretending making a a db field larger is all we'd have to do14:50
JayFMadasi: you around yet? I bet you'd have an interesting insight into this14:50
dtantsurwell, I'd prefer to not store config drive in this way at all14:51
jrollwell, we max at 32MB14:51
lucasagomesthe only reason we keep the config drive in the node is because of rebuild14:51
lucasagomesbut we shouldn't, cause nova can regeneate the configdrive at rebuild time14:51
dtantsurI'm fine with keeping it, I'm not fine with keeping it this way :)14:51
TheJulia++14:51
JayFlucasagomes: ah yeah, we were talking about that a few weeks back too14:51
lucasagomesso we should change our API/CLI to accept passing it14:51
JayFjroll: ^ I think we should WONTFIX, and the real fix is to stop storing configdrive altogether14:51
TheJulialucasagomes: I think those changes got proposed14:51
JayFI think there's at least a bug for it14:52
* lucasagomes looks14:52
JayFhttps://bugs.launchpad.net/ironic/+bug/157593514:52
openstackLaunchpad bug 1575935 in Ironic "[RFE] Rebuild should also accept a configdrive" [Wishlist,Confirmed] - Assigned to Stephane Miller (stephaneeee)14:52
jrollJayF: if this human wants to solve the problem quickly, wdyt about node.configdrive field (or in another table), make that larger to fix their bug, then when we real-fix it we can drop that column14:52
lucasagomesJayF, nice14:53
JayFjroll: how about that human installs swift and uses our reasonable option for using superlarge configdrives? or makes a smaller script to pull down his bigger one?14:53
JayFjroll: there's a million workarounds that don't involve us ballooning our db14:53
*** baoli has quit IRC14:53
dtantsurdo we always default to swift if it's present?14:53
jrollJayF: temporarily ballooning*14:53
JayFdtantsur: no14:53
jrollwe don't, because we've no way to tell if swift is there14:53
*** baoli has joined #openstack-ironic14:54
dtantsuraha, so this needs configuration as well?14:54
jrollright14:54
jrollalso 'just install swift' isn't a very friendly answer :/14:54
lucasagomesJayF, edited that bug to -rfe, it's just a bug14:54
dtantsurI'm asking because I remember config drives on nodes in tripleo, where we definitely have swift14:54
lucasagomesdtantsur, no, we default to saving it to the node14:55
lucasagomeswe can ask ironic to save it in swift by changing a config14:55
JayFjroll: "make a db field 2^16 larger" is not a very friendly patch14:55
dtantsurif it's the default, I'm afraid we have to fix it14:55
jrollso what's the path to fixing this for real? make nova re-send the configdrive, and change all the code to pass it through? probably not much work yeah? (keep in mind we can't drop it from ironic for a cycle after the change lands in nova)14:56
lucasagomessomeone knows stephane miller ?14:56
jrolllucasagomes: that's cinerama14:56
jrollI believe14:56
lucasagomesa-ha, right :-)14:56
TheJuliathat is her14:56
JayFjroll: yep, exactly, and then in a later release ironic doesn't store configdrive at all14:56
lucasagomescinerama, are you working on bug/1575935 ?14:56
JayFjroll: except during deployment, I think, we'd have to keep it somewhere14:56
jrollJayF: that's why I'm saying for now, add a table/column14:56
jrolloh you're right14:56
jrollblah14:56
TheJuliaI don't think she is right now.  She also likely will not be on for another hour14:57
JayFI mean, a different field, with a max length <4gb and I'm a little OK with it14:57
*** dprince has quit IRC14:57
lucasagomesTheJulia, thanks, yeah she might check the ping later. Left a comment on that bug as well14:57
jrollright, MEDIUMTEXT somewhere outside of instance_info14:57
JayFI'm very nonplussed about the idea os making instance_info that large14:57
dtantsurjroll++14:57
*** dprince has joined #openstack-ironic14:57
jrollJayF: I agree, I never said such a thing :)14:57
jrollor I haven't in some time14:58
JayFjroll: a patch that has lots of +2 votes says that, lol14:58
TheJuliajroll: ++, also since the drive is compressed, one should be able to fix the maximum size into that field while deploying14:58
* jroll puts up a vote14:58
dtantsurTheJulia, unless they put something that does not compress well there :)14:58
TheJuliadtantsur: true...  :(14:58
dtantsuranyway, whatever limit we put, we should fail early, on the API stage14:58
TheJuliabut if it is tons of text! :)14:58
dtantsuryeah, most often it is tons of text, maybe some binaries14:59
JayFdtantsur: I guess another possible fix is: API rejects configdrives that are too large if not stored in swift, fails cleanly, documentation + configs say "if you need >Xkb use swift"14:59
*** jraju has quit IRC14:59
jrollI need a browser with readline support, too many ^W closing windows while I'm typing :(14:59
TheJuliaHow have we not had someone hitting this before now though... have we really managed to convince people not to put binaries in their config drives all along?14:59
dtantsurJayF, my problem with it is that you don't know until you hit it15:00
JayFTheJulia: Because having huge config scripts is a major antipattern?15:00
dtantsurJayF, so we're kind of moving the problem to users15:00
JayFTheJulia: the proper thing is to bootstrap a *real* config management system using cloud-init, not to have it be huge from all your configs15:00
JayFdtantsur: ^ IMO the user is moving the problem to us, trying to use a big configdrive15:00
dtantsurtripleo config drives are JSONs, they compress well15:00
TheJuliaJayF: touché15:00
dtantsurJayF, well, at the very least we should document it and reject too big config drives on API stage15:01
JayFTheJulia: I'm pretty sure that's actually in cloud-init documentation, too15:01
JayFdtantsur: I agree with that, and I agree we shouldn't store configdrives long term (the bug posted above)15:01
JayFI'm unhappy but can deal with a separate field being made that's bigger, but I feel like we should be a little opinionated here15:02
*** causten has joined #openstack-ironic15:02
JayFand say this use case for configdrive is not awesome, and if you want it, you have to be configured differently than default15:02
dtantsurside question: which option do I set to enable config drives in Swift?15:02
* jroll lets this conversation finish before he posts an opinion15:02
jrolldtantsur: configdrive_use_swift15:02
JayFdtantsur: https://github.com/openstack/ironic/blob/master/etc/ironic/ironic.conf.sample#L68915:02
jrolland a couple related ones15:02
*** mkam has joined #openstack-ironic15:03
dtantsurmeh, it's not in puppet-ironic. will have to add then. thanks15:03
lucasagomesdtantsur, ++ since we always have swift that would be a good thing to have15:04
dtantsurwill hack on it tomorrow, yeah. though I hope we'll never hit the limit there.15:04
lucasagomesdtantsur, right, I don't think we inject anything in the config drive15:05
lucasagomesright? so it's unlikely15:05
dtantsurlucasagomes, for tripleo? os-net-config settings, but that's a relatively small JSON15:05
openstackgerritSam Betts proposed openstack/ironic: Add Network Interface Attach and Detach APIs  https://review.openstack.org/32704615:06
lucasagomesah right, yeah those are pretty small15:06
*** scarab_ has quit IRC15:10
*** jjohnson2 has joined #openstack-ironic15:13
*** fxpester has quit IRC15:14
openstackgerritImre Farkas proposed openstack/ironic: DRAC: list unfinished jobs  https://review.openstack.org/36745115:14
jlvillalOur customer ran into too large of a configdrive due to the size of their team and all the SSH public keys they tried to put into the config drive15:16
jrollO_O15:16
jlvillalThat's what they told me :)15:17
* TheJulia seconds jroll's O_O15:17
sambettsthats a lot of SSH keys...15:18
*** mbound_ has quit IRC15:18
TheJuliathey really shouldn't compress well, but that is still a lot15:19
sambettswhere they using the OpenStack keypairs thing? or some other way to get the SSH keys into the instances?15:19
sambettss/where/were15:19
TheJuliaI'm curious why they are setting that up front instead of using a disposable key pair that gets ripped out after the fact from the instance once it is up and running15:19
*** praneshp has joined #openstack-ironic15:20
rloohi and morning everyone, sambetts, TheJulia, jlvillal, jroll, lucasagomes, dtantsur, JayF15:20
JayFo/15:20
TheJuliagood morning rloo15:20
lucasagomesrloo, morning15:20
jrollmorning rloo :)15:20
sambettshey rloo15:20
*** sdake_ has joined #openstack-ironic15:21
*** sabeen1 has quit IRC15:21
*** sdake has quit IRC15:22
*** mjura has quit IRC15:24
dtantsurmorning rloo15:26
*** sabeen has joined #openstack-ironic15:27
*** sborkows has quit IRC15:27
*** amotoki has joined #openstack-ironic15:29
*** dtantsur is now known as dtantsur|afk15:29
*** dprince has quit IRC15:30
*** dprince has joined #openstack-ironic15:31
*** hshiina has quit IRC15:31
*** amotoki has quit IRC15:34
openstackgerritMerged openstack/ironic: Update ironic/ironic.conf.sample  https://review.openstack.org/36734215:36
*** dprince has quit IRC15:37
*** mgoddard_ has joined #openstack-ironic15:37
*** jvgrant has joined #openstack-ironic15:40
*** mgoddard has quit IRC15:40
*** mgoddard_ has quit IRC15:44
*** MattMan has quit IRC15:46
*** MattMan has joined #openstack-ironic15:47
* jlvillal wonders if we could have some sort of proposal bot job for the ironic.conf.sample ?15:50
*** vmud213 has joined #openstack-ironic15:56
openstackgerritJohn L. Villalovos proposed openstack/ironic: Update unit tests for neutron interface  https://review.openstack.org/36703815:56
openstackgerritSergii Turivnyi proposed openstack/python-ironicclient: Negative tests for testing actions with Chassis.  https://review.openstack.org/34379615:57
*** mgoddard has joined #openstack-ironic15:58
openstackgerritMatthew Treinish proposed openstack/ironic-inspector: DNM: Test a tempest ssh patch  https://review.openstack.org/36748016:01
mtreinishdtantsur|afk, aarefiev: ^^^16:02
jlvillalmtreinish: DNM: Desperately Needs Merging?  ;)16:05
mtreinishhaha16:05
mtreinishjlvillal: so the patch that's testing is doing terrible things in unit tests on my local machine (I should have run those before pushing it)16:07
mtreinishit's sitting at ~15GB of ram consumed and growing16:08
jlvillalmtreinish: That has never ever ever happened to me. Nope never ;)16:08
* jlvillal too often runs the tox battery of tests after he pushes ...16:09
*** praneshp has quit IRC16:09
mtreinishwell it has stopped growing at 16.3G of ram used16:09
*** rbrud has quit IRC16:10
JayFI think there are better ways for inspector to determine how much ram a node has than to use it all up /s16:10
sambettsmtreinish: woah?!16:10
TheJuliaJayF: lol16:11
*** mbound has joined #openstack-ironic16:19
*** vsaienk0 has quit IRC16:19
*** tesseract- has quit IRC16:21
sambettsjroll: What do you think about the possiblity of splitting the Ironic tempest API tests away from the end to end deployment test? I find the junk requests in the Ironic conductor/api logs generated by the API tests makes it hard to debug issues16:22
*** trown is now known as trown|outtypewww16:23
jrollsambetts: +1, I've been wanting to do that for a while now16:24
*** mbound has quit IRC16:24
jrollsambetts: also means we should be able to make tempest run parallel for scenario jobs16:24
sambettsboth cases should be parallisable if they are split would be awesome16:25
sambettswhich would be awesome*16:25
sambettsjroll: do you think infra will mind if we add another job per patch ?16:26
jrollsambetts: should be fine16:27
sambettscool :) I'll add that to my todo list then16:27
jroll<316:27
*** rama_y has joined #openstack-ironic16:29
*** dprince has joined #openstack-ironic16:31
*** amotoki has joined #openstack-ironic16:31
*** amotoki has quit IRC16:31
*** amotoki has joined #openstack-ironic16:31
*** fxpester has joined #openstack-ironic16:31
*** ametts has joined #openstack-ironic16:32
*** fragatina has joined #openstack-ironic16:37
*** fragatina has quit IRC16:41
*** rbrud has joined #openstack-ironic16:41
*** clenimar has joined #openstack-ironic16:43
mat128JayF: there are, but not for dtantsur: https://bugzilla.redhat.com/show_bug.cgi?id=136630916:44
openstackbugzilla.redhat.com bug 1366309 in openstack-ironic-inspector "ironic introspection fails on GCE node" [Unspecified,Closed: wontfix] - Assigned to dtantsur16:44
lucasagomesfolks I'm calling it a day16:47
lucasagomeshave a great evening all16:47
mat128o/ Lucas16:48
lucasagomessee ye later :-)16:48
*** lucasagomes is now known as lucas-dinner16:48
*** derekh has quit IRC16:49
*** devananda|travel is now known as devananda16:56
*** jheroux has quit IRC16:57
*** anush has quit IRC17:00
*** jaosorior has quit IRC17:01
*** ElCoyote_ has quit IRC17:02
*** ElCoyote_ has joined #openstack-ironic17:04
*** ohamada has quit IRC17:05
*** leitan has quit IRC17:06
*** vmud213 has quit IRC17:13
NobodyCamNight lucas-dinner17:17
NobodyCamand Good Morning Ironic'ers :)17:17
mat128morning NobodyCam17:18
*** leitan has joined #openstack-ironic17:19
NobodyCammorning mat128 :)17:19
jlvillalGood morning NobodyCam mat12817:19
NobodyCammorning jlvillal :)17:20
jlvillalAnd everyone else!17:20
*** michauds has quit IRC17:30
openstackgerritSam Betts proposed openstack/ironic: Add Network Interface Attach and Detach APIs  https://review.openstack.org/32704617:32
* sambetts -> home 17:33
sambettsNight everyone!17:33
*** sambetts is now known as sambetts|afk17:34
NobodyCamnight sambetts|afk17:34
sambetts|afko/ NobodyCam17:34
NobodyCam:)17:34
openstackgerritMario Villaplana proposed openstack/ironic: Add power state change notifications  https://review.openstack.org/32186517:34
rloojlvillal: last chance, i'm going to forget about it after this. what did you want to know about https://review.openstack.org/#/c/364440/17:35
patchbotpatch 364440 - python-ironicclient - osc node power on & off commands17:35
jlvillalrloo: Sorry, my boss wants to meet now for a 1:117:36
rloojlvillal: lucky you ;)17:36
jlvillalWoo hoo! :)17:36
*** rama_y has quit IRC17:44
*** krtaylor has quit IRC17:49
*** jpich has quit IRC17:50
*** sabeen has quit IRC17:53
*** e0ne has quit IRC17:58
*** fragatina has joined #openstack-ironic18:01
*** ifarkas is now known as ifarkas_afk18:04
*** vishwanathj is now known as vishwanathj_18:06
*** Romanenko_K has quit IRC18:07
*** ChubYann has joined #openstack-ironic18:14
*** rama_y has joined #openstack-ironic18:17
openstackgerritMario Villaplana proposed openstack/ironic: Add power state change notifications  https://review.openstack.org/32186518:20
*** athomas has quit IRC18:21
rloojroll: do we care if http://docs.openstack.org/releasenotes/python-ironicclient/current-series.html won't exist after this patch: https://review.openstack.org/#/c/366980/?18:22
patchbotpatch 366980 - python-ironicclient - standardize release note page names and ordering18:22
rloojroll: I mean, I don't particularly care but I can fix if TheRoyalWe care :)18:23
*** bfournie has quit IRC18:23
jrollrloo: in theory I care, but not really18:24
JayFheh18:24
JayFthe unrendered versino of that page on github18:24
JayFis higher in google than the rendered version when searching for "ironicclient release notes"18:24
*** krtaylor has joined #openstack-ironic18:24
JayFso I'd say it's likely OK :P not killing any major google karma or anything18:24
rloojroll: ha ha. ok, then forget it. i also wonder if that page will actually disappear (there were issues before with pages not being removed, so will see)18:25
jrollyeah, doubt it will18:25
jrollthey just write whatever the job writes out18:25
jrolls/write/publish18:26
rloojroll: in which case, someone can submit another patch to link from the old to the new page. i'm going to +2 doug's patch.18:26
jrollrloo: sounds good18:26
*** sdake_ is now known as sdake18:27
mat128JayF: not here https://www.dropbox.com/s/eijela6fmippv5v/Screen%20Shot%202016-09-08%20at%202.29.15%20PM.png?dl=018:29
mat128weird18:29
*** michauds has joined #openstack-ironic18:33
JayFmat128: look for current-series18:33
JayFmat128: the current-series file is only in the end of your thing18:33
mat128Ahhh18:34
mat128right18:34
mat128my bad :)18:34
*** aNupoisc has joined #openstack-ironic18:34
*** mgoddard has quit IRC18:43
*** Sukhdev has joined #openstack-ironic18:51
mat128anyone here who wants to look at https://review.openstack.org/#/c/363294/ ? :)18:52
patchbotpatch 363294 - ironic - Add a basic install guide18:52
JayFhttps://review.openstack.org/#/c/356811/ more docs that could use review love18:53
patchbotpatch 356811 - ironic - Add metrics to administrator guide18:53
*** david-lyle has quit IRC18:53
*** david-lyle has joined #openstack-ironic18:53
*** sabeen has joined #openstack-ironic18:54
openstackgerritDevananda van der Veen proposed openstack/ironic-specs: [WIP] API Evolution  https://review.openstack.org/36758318:58
devanandajroll: even though we aren't having an api meeting for a couple weeks, I've got a draft up ^ .19:00
*** jralph_ has joined #openstack-ironic19:00
devanandamore substance still to write in these (and some more specs to add) but I'd like to know what folks think of the format19:00
patchbotError: I haven't seen substance.19:00
*** vishwanathj_ is now known as vishwanathj19:00
*** mbound has joined #openstack-ironic19:00
devanandabreaking things down into separate specs like this, with one overall spec19:01
jrolldevananda: awesome, thanks, will review later19:03
* jroll probably done for the day19:03
*** dprince has quit IRC19:16
*** e0ne has joined #openstack-ironic19:21
*** moshele has joined #openstack-ironic19:25
*** mgoddard has joined #openstack-ironic19:36
*** mgould is now known as mgould|afk19:37
*** fxpester has quit IRC19:49
*** sdake_ has joined #openstack-ironic19:50
*** sdake has quit IRC19:52
openstackgerritStephanie Miller proposed openstack/ironic: Add volume_connector table to save connector information  https://review.openstack.org/20098320:03
*** moshele has quit IRC20:03
openstackgerritVasyl Saienko proposed openstack/ironic: DO NOT REVIEW  https://review.openstack.org/29643220:05
openstackgerritFatih Degirmenci proposed openstack/bifrost: Test OPNFV CI verify job. Please ignore.  https://review.openstack.org/36760820:11
*** mgoddard has quit IRC20:16
*** e0ne has quit IRC20:17
*** e0ne has joined #openstack-ironic20:18
openstackgerritStephanie Miller proposed openstack/ironic: Introduce new RPCs to support volume connector operation  https://review.openstack.org/21458520:28
openstackgerritStephanie Miller proposed openstack/ironic: Add volume_targets table to save target information  https://review.openstack.org/28521820:29
openstackgerritStephanie Miller proposed openstack/ironic: Introduce VolumeTarget object  https://review.openstack.org/28521920:30
*** Goneri has quit IRC20:31
openstackgerritStephanie Miller proposed openstack/ironic: Introduce new RPCs to support volume target operations  https://review.openstack.org/28522020:31
openstackgerritJulia Kreger proposed openstack/ironic: Add volume_connectors and volume_targets to task  https://review.openstack.org/36619620:35
openstackgerritJulia Kreger proposed openstack/ironic: [WIP] Add storage_interface DB field and object  https://review.openstack.org/34800520:36
openstackgerritJulia Kreger proposed openstack/ironic: [WIP] Add storage_interface to base driver class  https://review.openstack.org/34800620:36
openstackgerritJulia Kreger proposed openstack/ironic: [WIP] Add base cinder common interface  https://review.openstack.org/35562520:36
*** hamzy has quit IRC20:38
*** hogepodge has joined #openstack-ironic20:39
hogepodgehi, I've been playing around with ironic20:40
hogepodgedoes anyone have a sample inventory file that they've used for provisioning NUCs using AMT?20:40
openstackgerritStephanie Miller proposed openstack/ironic: Introduce new RPCs to support volume target operations  https://review.openstack.org/28522020:41
* hogepodge glances sideways at devananda20:41
TheJuliahogepodge: using bifrost?20:41
hogepodgeTheJulia: yes20:42
TheJuliahogepodge: You know, I finally got an AMT device that is supported by the AMT driver, and never tried it out :(20:42
hogepodgeTheJulia: so I have auth-less ironic set up, and my management interfaces are active, I just don't have any solid docs on how to configure the amt driver20:42
*** bfournie has joined #openstack-ironic20:43
david-lylehogepodge: I tried that out for a while and could never get it to work20:43
hogepodgedavid-lyle: well that makes me sad20:44
david-lylebut I'm no expert20:44
TheJuliathe amt driver, and interface... can be problematic.20:44
hogepodgeTheJulia: yes?20:44
hogepodgeTheJulia: there are two drivers, right?20:44
TheJuliahogepodge: the driver_info parameters can be found https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/amt/common.py#L3820:44
TheJuliahogepodge: so, only one in tree20:44
hogepodgeTheJulia: bifrost installed pxe_amt automatically20:45
TheJuliathe other ?two? are out of tree.  The main question is what version is the AMT controller in your NUC?20:45
hogepodgebut it looked like agent_amt was also an option?20:45
TheJuliahogepodge: so agent_amt does appear to be an option if enabled in the config file.  Mainly it has to do with method of deployment differences.  There is no difference as far as I'm aware with regards to the hardware supported20:46
openstackgerritMerged openstack/python-ironicclient: standardize release note page names and ordering  https://review.openstack.org/36698020:47
*** absubram has joined #openstack-ironic20:48
hogepodgeTheJulia: AMT version 8.1.4020:49
TheJuliahogepodge: so, according to http://docs.openstack.org/developer/ironic/drivers/amt.html that Might just work20:50
TheJuliahogepodge: so kind of like https://github.com/openstack/bifrost/blob/master/playbooks/inventory/baremetal.json.example#L31 lines 30-32, just amt_username, amt_password, amt_address20:51
hogepodgeTheJulia: Ah, I see that. Thanks! I'll give it a try20:52
rloohogepodge: i'd be interested if it works for you. I've been thinking about using a NUC, but in the past, I heard that the AMT part was flaky. Don't know if that has been improved or not.20:52
hogepodgeTheJulia: hopefully I'll have 8 machines paved by end of day20:52
hogepodgerloo: I'm running on an ubuntu orange box that broke horribly after a recent update, and I wanted to give up on maas anyway, so this gave me the perfect opportunity20:53
rloohogepodge: crossing my fingers :)20:53
JayFrloo: intel wants to give all ironic cores NUCs for testing? That sounds great! :P20:55
* TheJulia fully supports this idea :)20:58
*** harlowja_ has joined #openstack-ironic20:59
*** causten has quit IRC20:59
rlooJayF: Did you hear that? I didn't know. I have been thinking of getting something besides just my laptop... how ironic is it that i have no hardware?21:00
JayFI was mainly joking :P21:00
rlooJayF: :)21:00
JayFI don't have a hardware test lab for Ironic21:00
*** harlowja has quit IRC21:00
JayFI mean, I guess OnMetal has a hardware test lab21:00
rlooJayF: how do you try things out?21:00
JayFbut we don't deploy devstack there generally or do end-to-end testing21:00
rlooJayF: yeah, you can use OnMetal, right?21:00
JayFrloo: devstack, mostly21:01
JayFrloo: unless it's a feature we already put in onmetal21:01
rlooJayF: but what do you run devstack on?21:01
* jlvillal still trying to get a few old servers to use. Still has not had any luck.21:01
JayFeither a VM or an OnMetal server21:01
JayFI have a plan to buy a shiny new desktop computer and turn my old one into a devstack box :)21:01
rlooJayF: oh, that's a good plan.21:02
rlooJayF: I was only thinking of a nuc, in case I could figure out how to get Intel to give me one. or two :)21:02
*** e0ne has quit IRC21:02
JayFif you wanna contribute to project-jay-new-desktop, I could use an i7...21:03
JayFhehe21:03
rlooJayF: haha. I wish.21:03
*** causten has joined #openstack-ironic21:05
*** jcoufal has quit IRC21:06
rlooJayF, TheJulia: did either of you want to +A, or did you want more eyes on https://review.openstack.org/#/c/367306/?21:06
patchbotpatch 367306 - ironic - Grub2 by default for PXE + UEFI21:06
devanandaif folks are interested, I can write up a thing on running ironic in a VM on my laptop, to deploy images to a NUC on my LAN21:09
devanandabut not right now - I need to run out to an appointment. bbiah21:09
*** mmedvede has quit IRC21:09
*** devananda is now known as devananda|bbl21:09
openstackgerritJay Faulkner proposed openstack/ironic: Add metrics to administrator guide  https://review.openstack.org/35681121:09
JayFrloo: wanna take another gander at that; I just responded to your comments ^21:09
jlvillalNot sure if this is out of date or not. From mrda: http://lifelog.michaeldavies.org/2015/09/ironic-on-nuc-part-1-setting-things-up.html21:09
rloodevananda: ++ (leave it to him to entice us and then run away!)21:09
rlooJayF: sure, is jenkins fast today? I also saw another doc patch from you on the install guide. maybe i will get to it too. today.21:10
JayFrloo: I always find it fastish on docs-only changes, but mainly wanted you to look at the edits so maybe we could do a quick irc turnaround if you're still unhappy with it21:11
JayFrloo: re: install-guide, mat128 and I are working on migrating the install/admin guide from developer docs into the proper location. We might even get it done by newton release \o/21:11
rlooJayF: OH. ok, in a sec. was looking up 'built-in function'. if we add a function to our code, it cannot be a built-in function, since we're adding it.21:12
rlooJayF: related to a patch I was jsut reviewing :)21:12
openstackgerritJay Faulkner proposed openstack/ironic: Add metrics to administrator guide  https://review.openstack.org/35681121:13
JayFI reviewed myself and already fixed a nit :P21:13
rlooJayF: even better!21:13
*** causten has quit IRC21:14
jlvillalrloo: https://review.openstack.org/#/c/365491/21:14
patchbotpatch 365491 - ironic - Add __ne__ built-in function for API Version object21:14
jlvillalI just commented on that with a -1.21:15
rloojlvillal: i just saw your comment21:15
rloojlvillal: you sure about that?21:15
openstackgerritStephanie Miller proposed openstack/bifrost: Allow user to get iPXE files  https://review.openstack.org/36763521:15
rloojlvillal: i don't want to spend more time on it if i don't have to :)21:15
jlvillalrloo: That's what the docs say.21:15
*** e0ne has joined #openstack-ironic21:15
openstackgerritJay Faulkner proposed openstack/ironic: Add metrics to administrator guide  https://review.openstack.org/35681121:15
jlvillalrloo: I would like more documentation in the commit message to sway my opinion21:15
*** cdearborn has quit IRC21:15
hogepodgedevananda|bbl: I am interested, hurry back! ;-)21:16
jlvillalrloo: I'll dig a bit more21:19
*** ametts has quit IRC21:19
mrdajlvillal: yes, that's out of date.  I should update that for the OOT status of AMT now. (And I'd love to get an agent_ version working too, if I can find the time...)21:20
openstackgerritStephanie Miller proposed openstack/bifrost: [WIP] openSUSE support  https://review.openstack.org/36763721:20
mrdaAnd good morning :)21:20
jlvillalHi mrda21:20
rloothx jlvillal21:20
rloomorning mrda21:20
mrdao/21:20
openstackgerritVasyl Saienko proposed openstack/ironic: DO NOT REVIEW  https://review.openstack.org/29643221:23
*** e0ne has quit IRC21:25
*** e0ne has joined #openstack-ironic21:26
*** e0ne has quit IRC21:26
*** wajdi_ has joined #openstack-ironic21:28
*** causten has joined #openstack-ironic21:29
jlvillalrloo: Some code testing and looks the patch is correct. Cool :)  My theory is that the default __ne__ compares the hash value.  Learned something new. And think the functools docs should be updated.21:31
*** wajdi has quit IRC21:31
rloojlvillal: thx for looking into it!21:31
rloojlvillal: i was going to edit the commit and then +A. unless you want to.21:31
*** ametts has joined #openstack-ironic21:31
jlvillalrloo: I can do that. I want to put a link to the docs that they quoted21:32
rloojlvillal: thx!21:32
*** wajdi_ has quit IRC21:32
openstackgerritFatih Degirmenci proposed openstack/bifrost: Test OPNFV CI verify job. Please ignore.  https://review.openstack.org/36760821:35
cineramahi mrda21:37
openstackgerritJohn L. Villalovos proposed openstack/ironic: Add __ne__() function for API Version object  https://review.openstack.org/36549121:38
jlvillalrloo: ^^21:38
rloojlvillal: you have my permission to +2 and +A it all by yourself if you want21:40
jlvillalrloo: Done!21:40
jlvillalrloo: I hope you don't then complain about the commits message ;)21:40
jlvillals/commits/commit/21:41
rloojlvillal: how could an edit'd commit by a core, be worse than what was there already? i think we should just do it to quickly improve things.21:41
jlvillalrloo: True21:42
rloojlvillal: i'm trying to figure out how we can be more efficient yet not let quality suffer :)21:42
jlvillalUnless I haven't had my morning coffee.  Then...21:42
rloojlvillal: doesn't make sense to comment with the grammatical changes, and wait for the person to update it, blah blah.21:42
jlvillal+121:43
rloojlvillal: having said that, i don't recall if that is a sanctioned 'thing' or if i had discussed with someone or if i had seen others doing it.21:43
jlvillalI think I've seen others do it.21:43
rloojlvillal: i'll make a note to update some doc somewhere about it...21:44
openstackgerritMerged openstack/ironic: Update unit tests for neutron interface  https://review.openstack.org/36703821:44
openstackgerritMerged openstack/ironic: [api-ref] Remove temporary block in conf.py  https://review.openstack.org/36647321:45
openstackgerritJohn L. Villalovos proposed openstack/ironic: Update special methods to use (self, other)  https://review.openstack.org/36764321:46
mrdahey cinerama21:46
openstackgerritFatih Degirmenci proposed openstack/bifrost: Test OPNFV CI verify job. Please ignore.  https://review.openstack.org/36760821:47
rlooJayF: so I still don't understand what you mean by 'data points'. Is each timing metric that is emitted, considered a data point? If so, I don't see how you can say 100-500. It depends on the amount of activity going on.21:48
JayFrloo: so with a little bit of background: most systems that would back statsd, like graphite21:48
rlooJayF: if you do eg 'ironic node-show' for 1000 nodes, you'll get 1000 data points?21:49
JayFeach distinctly named metric requires a certain amount of space on disk and processing time21:49
JayFthe number of times that named metric is emitted doesn't matter21:49
JayFthe number of distinctly named metrics do21:49
rlooJayF: ok, that's what I wanted to know.21:49
JayFso I'm trying to say, without getting into the implementation details of statsd+graphite, that ironic will create 100-500 whisper dbs on disk on your graphite server :)21:50
JayFmaybe I should say "distinctly named" data points?21:50
rlooJayF: basically, there is a timing metric for each method, and there are 100-500 different methods that are timed?21:51
mariojvJayF: what patch set is this in reference to?21:51
rloomariojv: https://review.openstack.org/#/c/356811/21:51
patchbotpatch 356811 - ironic - Add metrics to administrator guide21:51
mariojvthanks21:51
JayFmariojv: https://review.openstack.org/#/c/356811/21:51
patchbotpatch 356811 - ironic - Add metrics to administrator guide21:51
JayFrloo: pretty much, yeah21:52
JayFrloo: but it's also, if you exercise that code21:52
jlvillalJayF: mariojv: Would https://github.com/openstack/ironic/blob/master/ironic/conductor/manager.py#L2329-L2363  be a good candidate for a metric?21:52
jlvillalJayF: mariojv I was thinking of the case where it ends up using Swift.21:52
rlooJayF: do we really have > 100 timing metrics already? wow.21:52
mariojvjlvillal: i agree, think i commented there on some other patch set21:52
JayFjlvillal: probably so21:53
JayFrloo: yep, it doesn't take long21:53
JayFrloo: and in the world of metrics, that's not a crazy large number21:53
rlooJayF: cool.21:53
*** mmedvede has joined #openstack-ironic21:53
jlvillalmariojv: JayF Thanks21:54
JayFrloo: so wdyt? Should I change it to 'distinctly named' metrics, rewrite it differently? exclude it? leave it be?21:54
mariojvsort of tangentially related, i found this useful for estimating size of a whisper db (graphite storage backend) a while ago: https://gist.github.com/jjmaestro/577406321:55
patchbot(graphite a a ago: backend) db estimating for found https://gist.github.com/jjmaestro/5774063 i of of related, size storage tangentially this useful while whisper21:55
rlooJayF: should be changed. i'm thinking about it.21:55
rlooJayF: ok, i looked at the spec to refresh myself on terminology. so the 'metric name' is what distinguishes what you are talking about, right?21:56
*** ametts has quit IRC21:57
JayFrloo: well (prefix + name), but yeah21:57
mariojvi like "distinctly named", i think that makes it clear enough if other backends were added in the future too21:58
mariojvit seems generalizable enough to encompass the different types of metrics we may collect as well (time series, counter, etc)22:00
rlooJayF: ok, i suggested something but i'm not sure it is that great. But I think you know what I mean now :)22:01
JayFI might just put in the long form explanation22:02
JayFuse distinctly named metrics, which in some backends may map to required disk capacity to store them22:02
JayFor something like that22:02
rlooJayF: I am mostly gone now. will pop in a bit, off/on this evening.22:02
rlooJayF: that seems good to me22:03
JayFIt's fine, I'll be gone soon too and have tomorrow off :)22:03
JayFI just want to get these docs merged up, since all that's left is a little polish22:03
*** hj-hpe has quit IRC22:04
openstackgerritJay Faulkner proposed openstack/ironic: Add metrics to administrator guide  https://review.openstack.org/35681122:05
*** mbound has quit IRC22:08
openstackgerritDavid Edery proposed openstack/ironic: Fix non-deployed nodes due to short nodes.instance_info  https://review.openstack.org/33496722:09
*** dtantsur|afk has quit IRC22:12
*** baoli has quit IRC22:13
*** baoli has joined #openstack-ironic22:13
*** michauds has quit IRC22:14
*** dtantsur has joined #openstack-ironic22:14
*** absubram has quit IRC22:15
*** harlowja has joined #openstack-ironic22:18
openstackgerritMerged openstack/bifrost: Add possibility to set source branch for ironic-inspector  https://review.openstack.org/36330022:19
*** mkam has quit IRC22:20
*** harlowja_ has quit IRC22:22
openstackgerritVasyl Saienko proposed openstack/ironic: DO NOT REVIEW  https://review.openstack.org/29643222:24
*** sabeen has quit IRC22:31
*** sdake has joined #openstack-ironic22:34
*** sdake_ has quit IRC22:36
*** sai has quit IRC22:49
*** krtaylor has quit IRC22:51
*** sai has joined #openstack-ironic22:54
*** sdake has quit IRC22:59
*** Sukhdev has quit IRC23:00
*** rbrud has quit IRC23:02
*** leitan has quit IRC23:05
*** baoli has quit IRC23:06
hogepodgeTheJulia: ok, I can control the power state of the nucs using ironic, which is good. However, I'm stuck on the nucs finding the dhcp server to pxe boot from23:11
*** sai has quit IRC23:11
*** rook has quit IRC23:11
*** lucas-dinner has quit IRC23:12
*** watanabe_isao has joined #openstack-ironic23:22
*** rook has joined #openstack-ironic23:26
*** sai has joined #openstack-ironic23:26
*** rook is now known as Guest9098323:26
*** lucasagomes has joined #openstack-ironic23:26
TheJuliahogepodge: using what bifrost installed? installed your own?23:45
*** krtaylor has joined #openstack-ironic23:45
hogepodgeTheJulia: I installed bifrost from the provided instructions23:46
hogepodgeTheJulia: I'm seeing the bootp/dhcp request from the node, but dnsmasq isn't responding23:47
devananda|bblhogepodge: check for local firewall23:48
*** devananda|bbl is now known as devananda23:48
TheJulia^^^ that and looking up the other likely possibility23:48
hogepodgedevananda: yup, disabled. Got rid of it! :-D23:49
TheJuliahogepodge: check your /etc/dnsmasq.conf, make sure interface= is set to the right interface23:49
hogepodgeTheJulia: yeah, br0, which is the local private network23:50
devanandaTheJulia: is bifrost by default configuring dnsmasq with a whitelist, or a default response?23:50
devanandahogepodge: how is the NUC connected to your machine?23:50
TheJuliahogepodge: and you see the requests on br0?23:50
devanandaalso, are you tcpdump'ing on the host or the VM where bifrost is installed?23:50
hogepodgedevananda: through a switch23:50
devanandahogepodge: any other DHCP servers on that network?23:51
hogepodgeTheJulia: yeah, I'm seeing the packets arrive with tcpdump -i br023:51
TheJuliadevananda: binds to specific interface, can operate via config if the macs are known via static mappings, otherwise does default range on the configured interface23:51
hogepodgedevananda: I installed bifrost directly on a node. living large!23:52
devanandahogepodge: ah! I see23:52
TheJuliahogepodge: is br0 what has the ip address for that network bound to it?23:52
devanandaTheJulia: that's what I thought, but I didn't recall the config name that changes it between static mappings / default range.23:52
hogepodgeTheJulia: that could be it, where is that configured?23:52
hogepodgeTheJulia: yeah23:52
TheJuliaswapping that into brain now23:53
TheJuliainventory_dhcp and inventory_dhcp_static_ip23:53
devanandahogepodge: if bifrost is configured to do static dhcp mappings and it doesn't know the MAC of your NUC, that would explain the lack of response. check your dnsmasq config files.23:53
TheJuliaas boolean settings upon install23:53
TheJuliaThe other possibility, is if it is not configured at all to do that, or was not installed that way... you dnsmasq could be doing something funky because it is a bridge interface, but that seems kind of remote in my mind23:55
hogepodgeTheJulia: that was it, dnsmasq was serving on the wrong range23:55
TheJuliaWell, that will do it :)23:56
*** watanabe_isao has quit IRC23:56
openstackgerritMerged openstack/ironic: Add __ne__() function for API Version object  https://review.openstack.org/36549123:59

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