Friday, 2017-09-01

« Thursday, 2017-08-31 Index Saturday, 2017-09-02 »
openstackgerritAnup Navare proposed openstack/ironic-inspector master: Improves documentation for discovery  https://review.openstack.org/47764000:26
*** dyasny_ has quit IRC00:26
*** ijw has joined #openstack-ironic00:26
*** ijw has quit IRC00:31
*** ijw has joined #openstack-ironic00:32
*** zshi_ has joined #openstack-ironic00:34
*** karimb has quit IRC00:35
*** baoli has joined #openstack-ironic00:53
*** baoli has quit IRC00:54
*** baoli has joined #openstack-ironic00:55
*** hw_wutianwei has joined #openstack-ironic00:55
*** wong has joined #openstack-ironic00:56
*** stendulker has joined #openstack-ironic00:59
*** phuongnh has joined #openstack-ironic01:01
*** sasha2 has quit IRC01:04
*** mat128 has joined #openstack-ironic01:04
*** sasha2 has joined #openstack-ironic01:04
wongaha!morning01:04
*** litao__ has joined #openstack-ironic01:05
*** ijw has quit IRC01:08
*** rbudden has quit IRC01:08
*** yangyapeng has joined #openstack-ironic01:10
*** zhangfei has joined #openstack-ironic01:27
*** trungnv has quit IRC01:40
*** rama_y has joined #openstack-ironic01:46
*** rama_y has quit IRC01:47
*** rama_y has joined #openstack-ironic01:47
*** zshi_ has quit IRC01:47
*** priteau has joined #openstack-ironic01:50
*** mguz__ has quit IRC01:53
*** priteau has quit IRC01:55
*** mguz__ has joined #openstack-ironic01:56
TheJuliacrushil: no worries, I was just waving through the internet01:57
TheJuliawong: good morning!01:57
*** stendulker has quit IRC02:01
*** rbudden has joined #openstack-ironic02:02
*** rbudden has quit IRC02:02
openstackgerritShangXiao proposed openstack/ironic-inspector master: i[Trivialfix]Fix typos in ironic-inspector  https://review.openstack.org/49986602:05
*** pmannidi has joined #openstack-ironic02:08
openstackgerritShangXiao proposed openstack/ironic-inspector master: Replace http with https for doc links in ironic-inspector  https://review.openstack.org/49986802:14
crushilTheJulia, Haha. \o/02:15
*** rama_y has quit IRC02:20
wong:)02:21
*** TuanLA has joined #openstack-ironic02:21
wonghi, how to monitor the node?02:23
*** thegreenhundred has joined #openstack-ironic02:31
*** jlabarre has quit IRC02:31
*** mguz__ has quit IRC02:49
*** mguz__ has joined #openstack-ironic02:52
*** zshi_ has joined #openstack-ironic03:22
*** stendulker has joined #openstack-ironic03:26
*** baoli has quit IRC03:26
*** baoli has joined #openstack-ironic03:27
*** baoli has quit IRC03:31
*** links has joined #openstack-ironic03:32
*** udesale has joined #openstack-ironic03:37
*** mguz__ has quit IRC03:40
*** Nisha_Agarwal has joined #openstack-ironic03:43
Nisha_Agarwalmorning Ironic!!!03:46
wongmorning03:47
Nisha_Agarwalwong, o/03:47
*** sasha2 has quit IRC03:47
*** sasha2 has joined #openstack-ironic03:48
*** sasha2 has joined #openstack-ironic03:48
wongare there any projects to monitor  nodes?03:48
*** baoli has joined #openstack-ironic03:51
*** priteau has joined #openstack-ironic03:51
*** crushil has quit IRC03:52
*** priteau has quit IRC03:56
*** mguz has joined #openstack-ironic03:57
*** sasha2 has quit IRC04:09
*** baoli has quit IRC04:34
*** mguz has quit IRC04:34
*** yangyapeng has quit IRC04:41
*** MattMan_1 has quit IRC04:50
*** MattMan_1 has joined #openstack-ironic04:50
*** jaianshu has joined #openstack-ironic04:55
*** yangyapeng has joined #openstack-ironic04:57
*** pmannidi has quit IRC04:59
*** mguz has joined #openstack-ironic05:03
*** hoonetorg has quit IRC05:05
*** mat128 has quit IRC05:07
*** hoonetorg has joined #openstack-ironic05:22
*** marios has joined #openstack-ironic05:44
*** marios_ has joined #openstack-ironic05:45
*** rivers has quit IRC05:45
*** rivers has joined #openstack-ironic05:47
*** mjura has joined #openstack-ironic05:48
*** mjura has quit IRC05:49
*** mjura has joined #openstack-ironic05:49
*** links has quit IRC05:50
*** priteau has joined #openstack-ironic05:52
*** priteau has quit IRC05:57
*** yangyapeng has quit IRC05:57
*** mguz has quit IRC05:59
*** links has joined #openstack-ironic06:02
*** yangyapeng has joined #openstack-ironic06:04
*** parlos has joined #openstack-ironic06:15
*** pcaruana has joined #openstack-ironic06:17
openstackgerritDao Cong Tien proposed openstack/ironic master: WIP: Implement console log support for Socat console  https://review.openstack.org/49990906:20
*** jtomasek has joined #openstack-ironic06:20
*** parlos has quit IRC06:22
*** parlos has joined #openstack-ironic06:22
*** parlos has quit IRC06:23
*** parlos has joined #openstack-ironic06:23
*** parlos has quit IRC06:24
*** parlos has joined #openstack-ironic06:24
*** parlos has quit IRC06:25
*** parlos has joined #openstack-ironic06:25
*** parlos has quit IRC06:26
*** parlos has joined #openstack-ironic06:26
*** parlos has joined #openstack-ironic06:27
*** parlos has quit IRC06:28
*** mguz has joined #openstack-ironic06:29
*** jaganathan has joined #openstack-ironic06:32
*** phuongnh has quit IRC06:32
*** rcernin has joined #openstack-ironic06:33
*** phuongnh has joined #openstack-ironic06:33
*** Nisha_Agarwal has quit IRC06:44
openstackgerritShivanand Tendulker proposed openstack/ironic stable/pike: Boot from volume fails with 'iscsi' deploy interface  https://review.openstack.org/49991106:44
*** jaganathan has quit IRC06:46
*** charcol has quit IRC06:47
*** jaganathan has joined #openstack-ironic06:49
openstackgerritMerged openstack/bifrost master: Fix the dhcp checks on centos  https://review.openstack.org/49949807:03
*** hjensas has joined #openstack-ironic07:07
pas-hawong, AFAIK ceilometer can be configured to collect some info from nodes over SNMP07:08
pas-haand good morning ironic :)07:08
wongbut if we put it in tenant network, we can't connect it07:11
*** tesseract has joined #openstack-ironic07:12
*** mgoddard has joined #openstack-ironic07:17
*** d0ugal has quit IRC07:20
*** karimb has joined #openstack-ironic07:33
karimbdtantsur|afk anything more needed for https://review.openstack.org/#/c/301226/07:38
patchbotpatch 301226 - ironic-staging-drivers - Ovirt support07:38
karimb?07:38
openstackgerritMerged openstack/ironic-inspector master: i[Trivialfix]Fix typos in ironic-inspector  https://review.openstack.org/49986607:38
*** racedo has joined #openstack-ironic07:38
*** Nisha_Agarwal has joined #openstack-ironic07:39
*** lifeless_ is now known as lifeless07:39
*** ralonsoh has joined #openstack-ironic07:49
*** livelace has joined #openstack-ironic07:49
*** TuanLA has quit IRC07:50
*** TuanLA has joined #openstack-ironic07:51
*** priteau has joined #openstack-ironic07:53
*** priteau has quit IRC07:58
*** yangyapeng has quit IRC08:01
*** lucas-hungry is now known as lucasagomes08:04
*** yangyapeng has joined #openstack-ironic08:05
vdrokgood morning ironic!08:05
openstackgerritparesh sao proposed openstack/ironic master: Add change in 'iscsi' deploy interface to call boot interface  https://review.openstack.org/49992608:12
wongmorning/afternoon :)08:14
*** jaianshu_ has joined #openstack-ironic08:14
*** jaianshu has quit IRC08:18
vdrokhi wong :)08:18
*** yangyapeng has quit IRC08:31
*** yangyapeng has joined #openstack-ironic08:32
*** yangyapeng has quit IRC08:36
*** adisky__ has quit IRC08:49
*** syqian__ has quit IRC08:50
*** syqian__ has joined #openstack-ironic08:50
*** yangyapeng has joined #openstack-ironic08:50
*** yangyapeng has quit IRC08:51
*** syqian__ has quit IRC08:53
stendulkermorning vdrok, wong08:55
wongmorning :)08:56
vdrokgood morning stendulker08:59
*** yangyapeng has joined #openstack-ironic09:00
*** priteau has joined #openstack-ironic09:02
*** e0ne has joined #openstack-ironic09:09
*** milan has joined #openstack-ironic09:11
milanmorning Ironic! :)09:11
*** sambetts|afk is now known as sambetts09:11
milan#pixiesay -mflexing rainy weather --- sleepy Friday09:12
PixieBootsᕙʕ⇀ᴥ⇀ʔᕗ: rainy weather09:12
milanmorning aarefiev sambetts! :)09:12
aarefievmorning!09:12
sambettsMorning milan o/09:12
milanaarefiev, I guess you're right about the lock release09:13
milanaarefiev, and I've just realised the self.app isn't shut-down09:13
aarefievmilan: hmm, we don't shutdown app?09:14
milanaarefiev, I don't think so09:15
milanaarefiev, this is how it's supposed to be done for the simple server (werkzeug): http://flask.pocoo.org/snippets/67/09:15
*** Nisha_Agarwal has quit IRC09:15
milanmodulo the api part09:15
* milan double checks09:16
milanaarefiev, no git grep match09:17
*** etingof has quit IRC09:18
milanaarefiev, sambetts guys, how do you feel about it missing? should I put it in https://review.openstack.org/#/c/487909 or rather a separate patch?09:18
patchbotpatch 487909 - ironic-inspector - Allow periodics to terminate inspector09:18
*** hshiina has joined #openstack-ironic09:19
aarefievI'd do it in separate one09:19
milanack, if sambetts acks ;)09:21
aarefievmilan: I'm also wondering if sys.exit may trigger it09:24
milanhmmm maybe09:25
* milan looks09:26
*** udesale__ has joined #openstack-ironic09:27
*** udesale has quit IRC09:27
sambettshmm, it does make me wonder because we have multiple threads running, if one thread calls sys.exit does it kill the other threads09:27
vdrokmorning aarefiev milan and sambetts09:28
*** udesale has joined #openstack-ironic09:29
milanmorning vdrok! :)09:29
aarefievvdrok o/ :)09:29
*** etingof has joined #openstack-ironic09:30
*** udesale__ has quit IRC09:31
milansambetts, https://stackoverflow.com/questions/905189/why-does-sys-exit-not-exit-when-called-inside-a-thread-in-python09:31
milanbut we're joining all, save from any threads that the simple server might have spun09:32
sambettsmilan: what I mean is that periodic worker that is running the perodic tasks is one thread, and the run() function that is running self.app.run is another thread, so does the sys.exit in shutdown run in the periodic thread or in the main thread running self.app.run>?>#09:35
sambetts?09:35
milansambetts, ah, I see... /me checks09:37
*** amarao has joined #openstack-ironic09:38
*** e0ne has quit IRC09:38
*** amarao1 has joined #openstack-ironic09:38
*** amarao has quit IRC09:38
*** e0ne has joined #openstack-ironic09:38
milansambetts, it may be any thread because of the watchdog method09:38
*** amarao1 has quit IRC09:39
*** amarao has joined #openstack-ironic09:39
sambettsmilan: yeah... self.shutdown is called in a new thread, which means that sys.exit is called inside a thread which won't shutdown the whole app correctly09:40
milanyup but the code performs the shutting down of the periodics and executor, waiting for it to finish09:40
*** udesale has quit IRC09:40
sambettsmilan: I think we need to call self.app.shutdown to something too right??09:41
*** udesale has joined #openstack-ironic09:41
sambettsor something*09:41
milansambetts, yeah seems like that09:41
milanI guess I better put it in this patch so we don't introduce a hanging inspector?09:41
*** e0ne has quit IRC09:42
*** e0ne has joined #openstack-ironic09:42
sambettsyeah thats what I'm thinking, because before this patch shutdown() can only be called when self.app stops for whatever reason09:43
sambettsso there is no reason to worry about it09:43
sambettsbut with your patch self.app will be left running09:43
sambettsif shutdown is called by the periodic09:43
milanright09:43
*** udesale has quit IRC09:44
* milan amends, shouldn't be a big addition09:45
sambettsthanks :)09:45
milanaarefiev, sambetts thanks for the discussion! :)09:46
aarefievhmm, does sys.exit shutdowns only green thread?09:46
sambettsaarefiev: if sys.exit is called inside a thread it acts just like thread.exit()09:48
sambettsalthough green threads might change the behaviour09:49
sambettsunsure on that09:49
aarefievI'm wondering how it works now09:49
aarefievwith milan' patch09:50
sambetts:/ we don't have a CI test that tests shutting down the service do we? I guess if we added grenade it would test that09:51
milanaarefiev, so my take on it is any periodic can trigger now the wsgi.shutdown09:51
milanso we'd better explicitly stop the app in case it has spun any threads09:52
*** yangyapeng has quit IRC09:52
milanaarefiev, sambetts, devstack now isntead of sigabrt sends sigterm and we not handling it, the default is process terminates09:53
milanso this shutdown is never called in CI09:53
milansince devstack switched away from screen to systemd09:53
milanis still devstack branching a thing? ;)09:54
milanso that's going to be another patch: allow handling of the sigterm -> shutdown09:54
aarefievmilan, yeah, but did you try your patch in negative scenario09:54
sambetts+1 to another patch09:54
aarefievdid it shutdown service?09:54
milanaarefiev, yeah I was able to test it manually, couple of revisions ago09:55
*** yangyapeng has joined #openstack-ironic09:55
milanhave to do it again though09:55
milanto be on the safe side09:55
aarefievso I'm wondering if eventlet patches exit method09:55
* milan tried custom exceptions on custom places 09:55
aarefievor exception goes up to native thread09:56
milandunno, it's all monkey patched, but the exceptions propagate to anyone that waits09:56
milaniirc it's implemented through event.Event09:57
milanso what ever exception happens in a greenlet, you won't see unless you wait09:57
* milan checks if it applies to SystemExit too09:58
*** karthiks is now known as karthiks_afk09:58
sambettsblah /me hates eventlet, green threads and all the monkey patching09:58
aarefieveventlet thread may catch SystemExit exception and shutdown or it may raise it up to native thread09:58
aarefievI'm not sure now :)09:58
milanaarefiev, https://github.com/eventlet/eventlet/blob/4d2cdca17a44b19405d4cdeafc898a500fe2485c/eventlet/greenthread.py#L17510:09
milanthat's the waiting10:09
milanit polls the event object that holds either exception or return value10:10
aarefievdoes it reraise exception10:12
aarefievafter sending the event10:12
milanyup10:14
milanonce you wait on it I mean10:14
milanbut what really stops the whole thing is the hub10:14
milanhub.abort()10:14
* milan can't find how the SystemExit is handled right now10:17
milanbut https://github.com/eventlet/eventlet/blob/4d2cdca17a44b19405d4cdeafc898a500fe2485c/eventlet/greenthread.py doesn't seem to have special handling for SystemExit10:17
milanso I guess it's the same as any other exc10:18
*** d0ugal has joined #openstack-ironic10:18
* milan -> lunch bbl, but going to add:10:19
milan* the semaphore.release before sys.exit()10:19
milan* the app shut down call before periodics call10:19
*** yangyapeng has quit IRC10:21
*** hw_wutianwei has quit IRC10:21
*** TuanLA has quit IRC10:30
*** wong has quit IRC10:31
*** amarao has quit IRC10:33
*** hw_wutianwei has joined #openstack-ironic10:34
*** jaianshu_ has quit IRC10:37
*** ricardoas has quit IRC10:43
*** mgoddard has quit IRC10:44
*** parlos has joined #openstack-ironic10:59
*** mrch has joined #openstack-ironic11:00
*** lucasagomes is now known as lucas-afk11:06
*** fxpester has quit IRC11:10
*** stendulker_ has joined #openstack-ironic11:11
*** fxpester has joined #openstack-ironic11:12
*** mgoddard has joined #openstack-ironic11:13
*** stendulker has quit IRC11:14
*** stendulker_ has quit IRC11:15
*** fellypefca has joined #openstack-ironic11:26
fellypefcaGood Morning, Ironic :D TGIF11:29
fellypefcamilan hey, yesterday I revived an all-green-beuty too 8) heh my 468129 has passed in all CI's11:31
*** mat128 has joined #openstack-ironic11:35
*** hw_wutianwei has quit IRC11:39
*** zhangfei has quit IRC11:51
*** dprince has joined #openstack-ironic11:53
*** srobert has joined #openstack-ironic11:56
*** srobert has quit IRC11:56
*** srobert has joined #openstack-ironic11:57
*** rbudden has joined #openstack-ironic11:57
milanmorning fellypefca! Oh yeah, happens once in fool moon doesn't it ;)12:03
milanTGIF indeed! :D12:03
fellypefcamilan yeah haha :)12:04
*** litao__ has quit IRC12:07
*** trown|outtypewww is now known as trown12:08
openstackgerritparesh sao proposed openstack/ironic master: Boot from volume fails with 'iscsi' deploy interface  https://review.openstack.org/49992612:20
milanaarefiev, sambetts, so turns out app runs in the "main" thread, you can try yourself: app = flask.Flask('Foo'). Moreover, http://flask.pocoo.org/snippets/67/ works only in a request context so not going to work (actually raises a RuntimeError)12:21
sambettsmilan: right, but does the sys.exit in the eventlet thread actually successfully kill the main thread and any child threads it might have started?12:22
milansambetts, exactly what I'm trying to figure out right now12:23
openstackgerritparesh sao proposed openstack/ironic master: Boot from volume fails with 'iscsi' deploy interface  https://review.openstack.org/49992612:23
sambettsmilan: hmmm we could add the /shutdown endpoint that they are going on about in that snippet, and lock it down so its not accessible as a useable endpoint, and then make a request from def shutdown() to that endpoint12:26
milanyeah as a separate patch :D12:27
sambettsI can't believe there is no self.app.stop... thats so dumb...12:27
milansambetts, cause flask is expected to be run from a real server12:27
sambettsunicorn or something?12:27
milanthe app is the debugging werkzeug thing but yeah, it is stupid indeed12:27
milansambetts, or uwsgi or whatever12:28
*** jlabarre has joined #openstack-ironic12:29
*** udesale has joined #openstack-ironic12:29
*** phuongnh has quit IRC12:33
*** krtaylor has joined #openstack-ironic12:40
fellypefcasambetts Could you take a look when you have some free time?:) https://review.openstack.org/#/c/46812912:41
patchbotpatch 468129 - ironic - Introduce hpOneView and ilorest to OneView12:41
milansambetts, http://paste.openstack.org/show/620207/12:49
*** jcoufal has joined #openstack-ironic12:51
milanand aarefiev too :P ^12:51
aarefievmilan: and what is the result :)12:52
* milan guess should save just to release the semaphore and don't care about the app12:52
*** zhangfei has joined #openstack-ironic12:53
milanthe SystemExit kills the app12:53
milansomehow12:53
milan:D12:53
sambettsmilan: are you running that in python3? I'm getting syntax errors running that paste12:53
milanyup python312:53
* aarefiev too lazy to copy-paste script12:53
aarefievThat's what I expect12:54
sambettsI wonder if its because its a greenthread12:55
sambetts:/12:55
milancan well be12:55
milanlet's blame the monkey patch for it :D12:55
*** edleafe is now known as figleaf12:57
milanor let's merge it and I'll follow-up with an explicit semaphore.release() ;)12:57
milanaarefiev, sambetts ^ :P12:57
*** sasha2 has joined #openstack-ironic12:59
sambettsmilan: just tested with a real thread instead of eventlet and it didn't exit the main thread12:59
sambettsI didn't remove the monkey patch though12:59
milanwow12:59
sambettsso there must be some wizardry in eventlet itself that handles it12:59
*** thegreenhundred has quit IRC13:02
*** SerenaFeng has joined #openstack-ironic13:04
sambettsi wonder if its because eventlet isn't using real threads, so everything is running in the native thread therefore sys.exit is in the native thread13:08
*** exodusftw has quit IRC13:11
milansambetts, can be13:12
milanI've found just this https://github.com/eventlet/eventlet/blob/4d2cdca17a44b19405d4cdeafc898a500fe2485c/eventlet/hubs/pyevent.py13:12
milanthe pyevent hub seems to care about the SystemExit13:12
*** crushil has joined #openstack-ironic13:13
aarefievIt's seems reasonable to me13:13
*** baoli has joined #openstack-ironic13:14
sambettsso as long as we don't move away from eventlet then it should be ok I guess13:14
milanack13:17
*** exodusftw has joined #openstack-ironic13:17
aarefievI guess python coroutines work in same way13:18
milanor someone will file a very urgent bug O:-)13:20
milan* one day13:20
*** efried is now known as fried_rice13:20
*** Goneri has joined #openstack-ironic13:24
*** rivers has quit IRC13:26
*** sasha2 has quit IRC13:26
*** hshiina has quit IRC13:26
*** rivers has joined #openstack-ironic13:26
*** sasha2 has joined #openstack-ironic13:32
TheJuliagood morning everyone13:35
fellypefcaGood morning TheJulia :)13:35
*** zshi_ has quit IRC13:36
*** dyasny has joined #openstack-ironic13:36
milanmorning TheJulia! :)13:38
*** e0ne has quit IRC13:40
*** e0ne has joined #openstack-ironic13:41
*** udesale has quit IRC13:42
openstackgerritMilan Kováčik proposed openstack/ironic-inspector master: Refactoring the firewall  https://review.openstack.org/47183113:42
openstackgerritMilan Kováčik proposed openstack/ironic-inspector master: Allow periodics to terminate inspector  https://review.openstack.org/48790913:42
openstackgerritMilan Kováčik proposed openstack/ironic-inspector master: Introducing a dnsmasq PXE filter driver  https://review.openstack.org/46644813:42
openstackgerritMilan Kováčik proposed openstack/ironic-inspector master: TEST check dnsmasq PXE filter in devstack  https://review.openstack.org/46871213:42
milanaarefiev, sambetts, thanks for the reviews&discussion13:43
milanbtw let's make Friday the Inspector day? ;)13:43
* milan -> more coffee :)13:44
*** links has quit IRC13:44
*** jaganathan has quit IRC13:49
*** baoli has quit IRC13:59
*** rivers has quit IRC14:03
*** SerenaFe_ has joined #openstack-ironic14:04
*** mrch has quit IRC14:04
*** rivers has joined #openstack-ironic14:04
*** SerenaFeng has quit IRC14:05
*** sasha2 has quit IRC14:06
openstackgerritBob Fournier proposed openstack/ironic-inspector master: Change example dnsmasq.conf in install doc for uefi  https://review.openstack.org/50006314:08
*** mjura has quit IRC14:11
*** zhangfei has quit IRC14:14
riversHi, if Im not using "whole disk image" to boot baremetal, should i set "local boot" option for ironic node & flavor ? Or is that setting just for Whole disk image ? Thanks14:18
*** srobert has quit IRC14:18
*** sasha2 has joined #openstack-ironic14:22
*** fellypefca has quit IRC14:26
*** SerenaFe_ has quit IRC14:29
*** tcrowe has quit IRC14:34
vdrokrivers: if the image is whole disk or not determined by it having kernel and ramdisk properties in glance. if your image does not have those, you don't need to set the local boot14:35
*** baoli has joined #openstack-ironic14:39
openstackgerritBob Fournier proposed openstack/ironic-inspector master: Change example dnsmasq.conf in install doc for uefi  https://review.openstack.org/50006314:39
openstackgerritmelissaml proposed openstack/python-ironicclient master: Fix to use "." to source script files  https://review.openstack.org/50007914:43
riversvdrok: Ok, cause i set local boot, and the image is not whole disk... What im seeing is that upon deploy, Im taken to the login screen for CoreOs deploy image... I thought the baremetal node is supposed to reboot itself, aftern loading the user image ? Is that correct ?14:44
vdrokrivers: so is the image_source whole disk or not? first you do deployment with coreos deploy ramdisk, then the node should reboot into the image_source image14:45
*** skazi_ has joined #openstack-ironic14:45
riversrivers: No the image is not a whole disk .... I see the ironic node provision state go from "deploying" to "wait-call-back" , and on the bare-metal host, Im taken to the login screen....I dont see it rebooting itself,14:47
riversvdlok: No the image is not a whole disk .... I see the ironic node provision state go from "deploying" to "wait-call-back" , and on the bare-metal host, Im taken to the login screen....I dont see it rebooting itself,14:47
sambettsrivers: wait call back means the node is deploying, it is not complete until the node is active14:48
sambettsthe deploy ramdisk you using is coreos based, which is why you see coreos booted on the server during deployment14:49
riversYeah but it never reboots, and never goes to "active" state ... after a while i just see the nova instance fail with the generic message "No Valid host' found, even though a host was found and deployed14:50
sambettsrivers: if the deployment fails you need to look at ironic and see what error its seen14:50
*** skazi_ has quit IRC14:50
sambettsrivers: or if it is stuck in wait call back then it probably means the ramdisk was never able to call back to the ironic API service14:50
sambettsand will timeout14:51
sambettseventually14:51
riversThe only error i see or warning in ironic-conductor.log, is "Timed out waiting for RPC response: Timeout while waiting on RPC response - topic: "<unknown>", RPC method: "<unknown>" info: "<unknown>" _raise_timeout"14:51
sambettsrivers: my guess is that the baremetal node is never able to call back to ironic and heartbeat to continue deployment14:52
sambettsrivers: is the baremetal node able to reach the ironic API IP address from its IP address?14:53
riverssambetts: how do i verify that ? Will that info be in the logs somewhere /14:54
sambettsrivers: if the ironic node was able to heartbeat successfully you should see agent_url populated in the driver_internal_info field on the node that is deploying14:55
sambettsif you do a ironic node show <node>14:55
*** crushil has quit IRC14:56
sambettsas a operator you should know if there is a route configured between your provisioning network and your ironic API IP addres14:56
riverssambetts: No, so the only info i see in driver_internal_info is "driver_internal_info   | {u'is_whole_disk_image': False}" , thats all14:56
sambettsthen I expect this is a networking issue14:57
riversOk, yeah i think i see the issue, the "ironic user " is configured as the external OAM network, while i launched the baremetal host on a private internal network15:01
riversThanks, I'll dig deeper15:02
sambettsthe provisioning network should be a network that is routable from the control plane (where ironic is running) and can route to the control plane, this allows the baremetal nodes to talk to ironic, and ironic to talk to the baremetal nodes15:03
sambettsthe normal way to do this is to make the provisioning network (flat network) a provider network, with a real world gateway router and subnet15:04
sambettsthat is routable from the outside world15:04
sambettsand can route to the outside world15:04
sambettssimilar to how the public network (floating ip) network in neutron is traditionally configured15:05
*** crushil has joined #openstack-ironic15:05
*** EmilienM has quit IRC15:12
*** EmilienM has joined #openstack-ironic15:12
*** fxpester has quit IRC15:16
*** rpioso has joined #openstack-ironic15:18
*** links has joined #openstack-ironic15:19
*** parlos has quit IRC15:19
NobodyCamGood Morning Ironic'ers... and...15:20
NobodyCamTGIF15:20
NobodyCam:)15:20
*** rama_y has joined #openstack-ironic15:21
*** thegreenhundred has joined #openstack-ironic15:22
*** yangyapeng has joined #openstack-ironic15:22
*** yangyapeng has quit IRC15:26
pas-haNobodyCam: yes indeed it is :) and good morning15:27
NobodyCamhey hey Good morning pas-ha :)15:27
*** rcernin has quit IRC15:29
*** pcaruana has quit IRC15:30
*** rama_y has quit IRC15:42
vdrokgood morning NobodyCam15:56
NobodyCamHey hey Good Morning vdrok :)15:57
NobodyCamit's Friday!15:58
*** livelace has quit IRC15:58
*** ralonsoh has quit IRC16:02
*** baoli_ has joined #openstack-ironic16:06
*** baoli has quit IRC16:08
*** marios has quit IRC16:08
*** rama_y has joined #openstack-ironic16:14
rpiosoAre DeployInterface.prepare() and BootInterface.prepare_instance() guaranteed to be called before ManagementInterface.set_boot_device()?16:20
sambettsrivers: bootinterface.prepare_instance should be the one calling set_boot_device16:20
*** parlos has joined #openstack-ironic16:21
rpiososambetts: So the answer is an emphatic "Yes!" :)16:22
sambettsrpioso: in some cases, (like in the agent deploy interface with whole disk images) it might not be the case because of legacy reasons, but there is a patch up to fix that16:23
rpiososambetts: I see what you described in PXEBoot.prepare_instance().16:25
rpiososambetts: Cool!  I'll take a look at the agent deploy interface.16:25
rpiososambetts: May BootInterface.prepare_instance() take an extended amount of time to return to the caller, say a few minutes?16:27
rpiososambetts: That would allow the BMC time to affect the change to the boot mode and boot device.16:27
*** srobert has joined #openstack-ironic16:29
*** aarefiev is now known as aarefiev_afk16:29
sambettsrpioso: that should be fine, there isn't a timeout applied to specific actions like that16:29
*** e0ne has quit IRC16:29
rpiososambetts: And it may be done synchronously?16:30
*** links has quit IRC16:31
sambettsif thats what your hardware requires16:33
*** dyasny_ has joined #openstack-ironic16:35
rpiososambetts: What about the set boot device API?  What are the expectations regarding its interactions with or dependency on a node's boot mode setting?16:36
rpiososambetts: Is it expected to set the boot device for the boot mode that the hardware is already configured to use?16:37
rpiososambetts: Even if the h/w's boot mode does not match the ironic node config?16:38
*** dyasny has quit IRC16:38
sambettsrpioso: the boot device API will directly call ManagementInterface.set_boot_device regardless of the state the node is in16:39
sambettsand ironic doesn't keep track of the expected boot device state, and even if it did that ^ would override it, exactly like setting the power state16:39
rpiososambetts: Yep, I see that.  I'm not clear on what's expected of ManagementInterface.set_boot_device().16:40
sambettsset_boot_device is expected to set the boot device on the node, that is all16:40
sambettsnothing smart about it, just an abstraction layer for different hardware16:41
rpiososambetts: Should it set the boot device for the boot mode that's configured in ironic, or just set it for whatever boot mode the server is set to use?16:41
rpiososambetts: The server's boot mode config may not match ironic's.16:42
sambettsrpioso: i'm not sure I understand, set_boot_device takes in a parameter which is the boot device to set, and thats the device it should set16:42
rpiososambetts: I'm hoping it should simply set the boot device for whatever boot mode the hardware is set to use.16:43
sambettsit sets the boot device to what ever its told to set the boot device too16:43
sambettsits just a wrapper around the hardware specific method for setting the boot device16:43
sambettsit shouldn't be aware about any ironic node state16:44
sambettsset_boot_device("pxe") sets the primary boot device on the baremetal server to PXE boot16:44
sambettsset_boot_device("cdrom") sets the primary boot device on the baremetal server to the CDROM drive16:45
*** milan has quit IRC16:45
rpiosoSome BMC's have separate boot device lists for each boot mode.  E.g., legacy BIOS (IPL)'s list can be different from UEFI's.16:46
rpiosoAnd to change the specific list, sometimes the associated mode must be active.16:47
anupnGood morning Ironic'ers16:47
rpiosoGood morning, anupn16:48
anupnrpioso, Good morning :)16:48
rpioso:)16:48
openstackgerritVladyslav Drok proposed openstack/ironic-specs master: Owner-aware baremetal nodes  https://review.openstack.org/41551216:49
sambettsrpioso: if capabilties=boot_mode:<boot_mode> is set then it is assumed that the operator wants the node to always boot in that mode, therefore I would expect that you only need to set the boot device for that boot mode16:50
*** dprince has quit IRC16:50
*** etingof has quit IRC16:50
yusefHello, is anyone available from the core-ironic team to review my changes for the Ironic-Sushy project? The change is in regards to sessions and is available for review at https://review.openstack.org/#/c/471942/. Thank you.16:55
patchbotpatch 471942 - sushy - Implement Redfish Sessions16:55
*** Goneri has quit IRC16:55
*** marios_ has quit IRC16:56
rpiososambetts: That sounds reasonable.  However, looking at the ilo example, I don't see that being done.16:56
sambettsrpioso: likely that was implemented before boot_mode was introduce16:56
rpiososambetts: I may be missing something.16:57
*** penick has joined #openstack-ironic16:57
rpiososambetts: Is the expectation you described documented somewhere?16:58
sambettsrpioso: https://docs.openstack.org/ironic/latest/install/include/boot-mode.html16:58
sambettsrpioso: 3rd bullet point down16:58
rpiososambetts: Hrm.  That's discussing boot mode, instead of boot device.17:00
sambettsrpioso: right, but that says that boot mode is static, so then you only need to care about setting the boot device for the active boot mode17:01
*** penick has quit IRC17:02
rpiososambetts: Understand.17:03
rpiosoI believe the boot mode biz was added as part of the UEFI effort, so yes, it came after boot device.17:04
*** Goneri has joined #openstack-ironic17:05
rpiososambetts: I have another question, if you have a few more minutes.  It's about deprecating a hardware type's supported interface implementation.17:05
rpiososambetts: If a new interface implementation is a superset of the old one, should the old one be deprecated?  And, if so, how is that done?17:06
*** mgoddard has quit IRC17:08
sambettsif its a superset of the old one, and there is no reason to use the old one, then it can/should be deprecated, the interface will need to be deprecated for one cycle, and publish a deprecation warning when it is loaded, and then can be removed in the next cycle17:08
rpiososambetts: Thank you!17:11
rpiososambetts: And a belated thank you for your assistance a couple of weeks ago.  I had momentarily walked afk, and when I returned, you were afk :)17:12
sambettsno problem :)17:12
anupnHi sambetts, figured that inside tinycore it should look for /usr/local/lib/python3.317:13
rpiososambetts: Hope you have a great weekend.17:13
sambettsrpioso: you too17:13
* rpioso is hungry17:13
*** rpioso is now known as rpioso|eat17:13
*** e0ne has joined #openstack-ironic17:15
*** Goneri has quit IRC17:17
* sambetts --> home 17:19
sambettsnight all17:19
*** sambetts is now known as sambetts|afk17:19
*** thegreenhundred has quit IRC17:24
*** trown is now known as trown|lunch17:27
*** harlowja has quit IRC17:30
*** yangyapeng has joined #openstack-ironic17:37
*** yangyapeng has quit IRC17:42
*** mgoddard has joined #openstack-ironic17:49
*** parlos has quit IRC17:50
*** racedo has quit IRC17:51
*** bnemec is now known as beekneemech17:56
*** rpioso|eat is now known as rpioso18:11
*** harlowja has joined #openstack-ironic18:12
NobodyCamNight sambetts|afk18:20
NobodyCamhave a good weekend18:20
*** trown|lunch is now known as trown18:29
*** mgoddard has quit IRC18:38
*** e0ne has quit IRC18:38
*** Goneri has joined #openstack-ironic18:50
*** ijw has joined #openstack-ironic19:03
*** penick has joined #openstack-ironic19:23
*** mgoddard has joined #openstack-ironic19:24
*** etingof has joined #openstack-ironic19:25
rama_yHi TheJulia, question regarding boot-from-volume19:26
rama_yI am using Ubuntu to boot from volume and it is unable to find root device19:27
rama_yTheJulia, any ideas?19:27
TheJuliarama_y: none right now19:30
rama_yTheJulia, here is the paste of the output: http://paste.openstack.org/show/620231/19:31
*** rpioso has quit IRC19:34
rama_yTheJulia: did you try booting from Ubuntu volume? I am able to boot from the same image (used for create the volume).19:35
TheJuliano, I personally have not. It is likely looking for something specific in the data structure19:36
TheJuliayou can see, it things its root is different, but it doesn't find it.19:36
rama_yYes19:37
openstackgerritPeter Piela proposed openstack/ironic-ui master: Cleanup ironic-ui REST API  https://review.openstack.org/50018119:44
*** yangyapeng has joined #openstack-ironic19:54
*** ppiela_ has joined #openstack-ironic19:56
*** yangyapeng has quit IRC19:58
*** ppiela has quit IRC19:58
*** baoli_ has quit IRC20:04
*** mat128 has quit IRC20:06
*** mat128 has joined #openstack-ironic20:06
*** yangyapeng has joined #openstack-ironic20:07
*** mat128 has quit IRC20:10
*** yangyapeng has quit IRC20:11
*** penick has quit IRC20:13
*** rpioso has joined #openstack-ironic20:16
*** rama_y has quit IRC20:28
*** rama_y_ has joined #openstack-ironic20:28
*** rcernin has joined #openstack-ironic20:31
*** Goneri has quit IRC20:34
*** mgoddard has quit IRC20:38
*** ijw has quit IRC20:50
*** ijw has joined #openstack-ironic20:52
*** jcoufal has quit IRC20:53
*** rcernin has quit IRC20:55
*** ijw has quit IRC20:56
*** ijw has joined #openstack-ironic20:56
*** srobert has quit IRC21:01
*** ijw has quit IRC21:05
riverssambetts: A quick question, does the deploy image start IPA process on the baremetal node, or is it the user image that does so ?21:11
riversdoes the deploy image start IPA process on the baremetal node, or is it the user image that does so ?21:11
*** jtomasek has quit IRC21:32
*** ijw has joined #openstack-ironic21:36
*** rivers has quit IRC21:37
*** rivers has joined #openstack-ironic21:38
*** priteau has quit IRC21:39
*** rivers has quit IRC21:58
*** rivers has joined #openstack-ironic22:02
*** tesseract has quit IRC22:09
jlvillalrivers: The deploy image is what is running IPA.22:11
jlvillalrivers: IPA is used to 'deploy' the user image onto the node.22:11
*** baoli has joined #openstack-ironic22:22
*** baoli has quit IRC22:26
*** figleaf is now known as edleafe22:31
*** ijw has quit IRC22:33
*** ijw has joined #openstack-ironic22:36
*** karimb has quit IRC22:41
*** karimb has joined #openstack-ironic22:43
*** karimb has quit IRC22:44
JayFrivers: IPA is a rest API for *imaging servers*. Never ever run or install it on user-facing operating systems -- only the deploy image (ramdisk).22:50
*** rama_y_ has quit IRC23:02
*** thegreenhundred has joined #openstack-ironic23:16
*** rwsu has quit IRC23:20
*** thegreenhundred has quit IRC23:24
jlvillalJayF: Hey! Good to see you here :)23:24
*** rama_y has joined #openstack-ironic23:25
*** jlabarre has quit IRC23:38
*** jlabarre has joined #openstack-ironic23:41
NobodyCamhey hey JayF :) here for the long weekend :p23:57
NobodyCamspeaking of I think its time for me it head out, so I can start enjoying said long weekend!23:58
« Thursday, 2017-08-31 Index Saturday, 2017-09-02 »

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!