Thursday, 2015-03-05

*** morazi has quit IRC00:06
*** mandre has joined #tripleo00:08
openstackgerritMerged openstack/diskimage-builder: Fix race in svc-map
*** prad has quit IRC00:15
*** penick has quit IRC00:16
alopgreghaynes: I'll give you a +1 on 155207 in exchange for
*** Marga_ has quit IRC00:24
*** Marga_ has joined #tripleo00:25
*** penick has joined #tripleo00:42
*** jerryz_ has joined #tripleo00:52
*** chuckC_ has quit IRC00:58
openstackgerritMerged openstack/diskimage-builder: Use package-installs on dpkg-based elements
*** penick has quit IRC01:15
*** sdake_ has joined #tripleo01:28
*** sdake__ has joined #tripleo01:30
*** sdake_ has quit IRC01:33
*** Marga_ has quit IRC01:37
jpeelersdake__: you around?02:04
sdake__snakes on a plane jpeeler whats up02:04
jpeeleri've been looking at the testing stuff, and i have too many questions to really get going02:05
jpeelerfor example, do you have any idea how service discovery should work?02:05
jpeeleri guess the docker daemon could be queried for running containers... is that adequate/acceptable?02:05
ccrouchcan't we just use keystone?02:06
jpeelerthings like rabbit aren't registered there though02:06
jpeelerworking on a full functional testing solution02:07
ccrouchtrue, but most all the stuff with a proper api should be02:07
ccrouchexcept for rabbit and mariadb02:07
jpeelerhrm, are there no other services that openstack depend on?02:09
*** Marga_ has joined #tripleo02:09
jpeeleri'm also a top down type of person. hard to figure out how to test apis and such when we have one working container (that being rabbit)02:09
*** Marga_ has quit IRC02:09
sdake__for the first test ccase02:09
sdake__just create a rabbitmq container02:09
sdake__and use rabbitmctl to see if it is operating02:09
*** Marga_ has joined #tripleo02:10
sdake__stop the container02:10
sdake__that should be the contents of the test case ;-)02:10
jpeeleroh i forgot we can just go into the container. see conceptually still working it out02:10
jpeelermaybe it's too fancy to discover running containers. but eventually the containers won't all be running on the same host02:11
ccrouch(08:09:01 PM) jpeeler: hrm, are there no other services that openstack depend on?02:15
ccrouchnot that i'm aware of. outside of all the craziness required for HA, but lets walk before we can run :-)02:15
*** penick has joined #tripleo02:16
*** jpeeler1 has joined #tripleo02:17
jpeeler1i got disconnected... if anybody said anything in the past 6 minutes02:18
ccrouch(08:15:24 PM) ccrouch: (08:09:01 PM) jpeeler: hrm, are there no other services that openstack depend on?02:18
ccrouch(08:15:24 PM) ccrouch: not that i'm aware of. outside of all the craziness required for HA, but lets walk before we can run :-)02:18
jpeeler1ah ok02:18
jpeeler1yeah, well as sdake__ pointed out, can take advantage of going into the containers as a first round02:19
jpeeler1eventually checking network communication with the proper apis would be good02:19
*** regebro has quit IRC02:19
*** andreaf_ has joined #tripleo02:20
*** cinerama` has joined #tripleo02:20
ccrouchjpeeler: yep02:20
jpeeler1although... entering the container forces the tests to run on the host. i kind of like the idea of all the tests running from a test container02:20
ccrouch+1 on test container02:20
*** rwsu_ has joined #tripleo02:20
ccrouchi bet you could use keystone for service discovery trivially02:20
jpeeler1one day!02:21
*** andreaf has quit IRC02:21
*** jpeeler has quit IRC02:21
*** jang1 has quit IRC02:21
*** cinerama has quit IRC02:21
*** mandre has quit IRC02:21
*** pcrews has quit IRC02:21
*** rwsu has quit IRC02:21
*** chmouel has quit IRC02:21
*** penick has quit IRC02:21
*** penick has joined #tripleo02:21
*** andreaf_ is now known as andreaf02:21
*** jang1 has joined #tripleo02:21
jpeeler1right now we have a rabbit container running post-kubernetes02:21
*** chmouel has joined #tripleo02:22
ccrouchi've got rabbit, mariadb, keystone, ceilometer and glance: going via fig02:24
ccrouchjust have had time to submit the changes.02:24
ccrouchmost stuff, outside of ceilometer just worked02:25
jpeeler1i did sort of assume that would not be the case, but awesome02:25
sdake__ccrouch its great that you are so rockin you can type "git review" on those containers ;-)02:25
sdake__i think daneyon rewrite keystone02:26
sdake__rewrote rather02:26
ccrouchsdake__: I've been behind on my review too :-(02:26
*** ChuckC has joined #tripleo02:26
sdake__would be nice to get er all workin again post -k8s :)02:27
ccrouchpretty much nothing needed rewriting02:27
*** pcrews has joined #tripleo02:27
sdake__some cat i met wants to rewrite crudini to jinja02:27
sdake__ccrouch isn't that was tripleo-image-elements does?02:27
* ccrouch racks brain02:27
sdake__his model is write /opt/service/ninja.wtf02:27
sdake__and /opt/service/jina.keys02:28
sdake__and bind miount /opt/service02:28
sdake__and have run which jinja applies the key value apirs to the service02:28 is a jinja application program - probably like 20 lines of python or something02:28
rhalliseyI've got mariadb, rabbit, and keystone stood up no kube02:29
jpeeler1ccrouch: that's a lot of commits!02:29
rhalliseyall env vars :P02:29
ccrouchjpeeler1: i like small commits :-)02:29
sdake__ya, what eh is proposing is using a bindmount rather then the environment to configure the container02:30
sdake__the environment could be used to apply a ninja template too02:30
ccrouchsdake__: this is what i was talking about when I said people are going to come along an rewrite the containers :-)02:30
ccroucheveryone is going to have their own may to configure them :-)02:31
ccrouchkolla will be the reference impl02:31
sdake__it would be nice if kolla did deployment too so we weren't a manually run container list02:32
ccrouchthere will be a tonne of value in the container orchestration i.e. fig, and also in the testsuite02:32
ccrouchsdake__: totally agreed02:32
ccrouchjpeeler1: ^ no pressure on the testsuite :-)02:32
sdake__jpeeler1 just do something verybasic02:34
sdake__to get it going02:34
sdake__gotta land airplane bye02:34
jpeeler1sdake__: see ya02:34
ccrouchjpeeler1: +1 on starting off basic02:34
rhalliseyjpeeler1, how were you trying to setup your containers?02:34
ccrouchbut i see a tonne of value in the testsuite over time02:34
jpeeler1i do plan on doing something basic, but i want it to conceptually be sound02:34
*** rwsu_ is now known as rwsu-afk02:35
jpeeler1rhallisey: i haven't really done anything beyond look at how the rabbit container is currently launched02:35
jpeeler1rhallisey: did you launch your containers with compose?02:36
ccrouchjpeeler1: I'd be fine with something conceptually sound for v1.1 :-)02:36
rhalliseyjpeeler1, I just wrote a script that initialized variables and does a bunch of 'docker run -d <my container>'02:37
ccrouchi gotta head out out but this works with docker-compose rc102:37
*** sdake__ has quit IRC02:38
jpeeler1ooh thanks ccrouch02:38
rhalliseyoh cool02:38
ccrouchrhallisey: fig.yml is *so* easy to write, its probably quicker than bash02:38
rhalliseyjust 3 docker runs :P02:38
rhalliseybut ya I'd prefer that I guess02:39
rhalliseymy way is probably not the way to do it02:39
ccrouchrhallisey: it's almost identical to the fig.yml to be fair02:39
rhalliseythat's true02:40
rhalliseylooking at it now :)02:40
jpeeler1ccrouch: do you know anything about dox?02:40
ccrouchfig/docker-compose just takes some of the edges off02:40
rhalliseyccrouch, ya cool02:40
ccrouchyou still need to specifiy all the same env vars and ports for instance02:40
ccrouchjpeeler1: not really02:40
rhalliseyccrouch, thank for the reference02:42
rhalliseywhen I move over to fig eventually.. this will be useful02:42
*** cinerama` is now known as cinerama02:42
jpeeler1rhallisey: heh, how are you not?02:42
rhalliseyjpeeler1, I'm just using bash02:43
ccrouchjpeeler1: rhallisey: one other general thing i had to do was to deal with the lack of dependency mgmt in fig and docker more generally02:43
ccrouchwith kubernetes we got it for free because k8s would just keep restarting containers until they worked :-)02:44
ccrouchbut fig doesnt have that i dont think02:44
ccrouchso I stole this from tripleo02:44
jpeeler1I guess that's okay though02:44
ccrouchrhallisey: that sleep in the .sh script reminded me of it02:45
rhalliseyhaha that way cool than my 'sleep 10'02:45
jpeeler1how's that differ from check_for_db?02:46
ccrouchcheck_for_db exits on failure immediately02:46
ccrouchthus killing the container02:46
jpeeler1ah that's right02:46
ccrouchwhich is not a problem in K8s02:46
ccrouchbut is a downer in fig02:46
jpeeler1should probably just replace check_for_db with the blocking version perhaps?02:47
ccrouchi haven't nicely scripted it up, the raw wait for statement is in lots of the scripts :-/02:47
ccrouchyep, i think we just need to decide on the right "query" to wait on02:48
ccrouchsometimes i think you need different ones for keystone depending on how you are using it02:48
ccrouchi found a nice race condition on keystone tenant and role creation too :-)02:50
ccrouchanyways there are a few wrinkles like that, but nothing dramatic02:50
rhalliseyccrouch, add a sleep 1002:51
jpeeler1kube didn't have dependency management through right?02:51
ccrouchjpeeler1: nope02:51
jpeeler1how was that not experienced previously?02:51
ccrouchits philosophy was to keep restarting stuff until all the dependencies had resolved :-)02:51
ccrouchah the race condition?02:52
ccrouchi don;t know, it didn't happen to me every time02:52
ccrouchjust sometimes02:52
ccrouchit required the glance-api and keystone containers to be going through their "crux user-create" parts at exactly the same time02:53
jpeeler1it sounds like fig either needs restart support or dependency support02:53
ccrouchdependency support is a bit of a sacred cow i think02:54
jpeeler1what do you mean?02:54
jpeeler1see ya rhallisey02:54
*** pradeep has joined #tripleo02:54
lifelessyou can't kill it02:54
lifelessbut its not helpful ?02:54
*** rhallisey has quit IRC02:54
ccrouchmaybe that wasn't the right term :-)02:55
* ccrouch thinks02:55
*** yog_ has joined #tripleo02:56
ccrouchno a more accurate term escapes. Its a topic that i think has generated a lot of discussion, but at least on the k8s side was deemed "not our problem"02:57
ccrouchbut! iirc fig does have restart support02:57
jpeeler1well now i'm confused02:58
ccrouchrestart policies is what i was thinking of02:59
ccrouchin plain docker and now exposed through fig to i believe03:00
ccrouchjpeeler1: whats the confusion?03:00
jpeeler1yes it does appear to be available03:00
jpeeler1how does that not mitigate the need for the wait_for function?03:00
*** yamahata has quit IRC03:00
*** ChuckC has quit IRC03:01
ccrouchit may very well do, I just didn't try it out.03:01
ccrouchi guess i just didnt think it was a very elegant solution :-/03:01
jpeeler1well it's either restart, dependencies, or wait_for03:02
*** pradeep1 has joined #tripleo03:03
jpeeler1i would prefer dependencies to be listed, but maybe that actually IS too much work03:03
*** pradeep has quit IRC03:03
ccrouchwell restart and wait_for are implemented already so I would go with one of those two :-)03:03
jpeeler1heh yeah03:04
ccrouchwell they are listed sort of, in the links section03:04
ccrouchthere is just no concept of those links being "ready"03:04
ccrouchbefore your container is started03:04
jpeeler1yeah guess that's true03:04
ccrouchthats really the dependency part03:05
ccrouchand its not really openstack services fault03:06
ccrouchits all the little scripting we do in start.sh03:06
jpeeler1oh neat, i didn't realise the hosts file was modified by the links03:06
ccrouchoh yeah, its pretty slick03:06
jpeeler1so no service discovery necessary03:06
*** ChuckC has joined #tripleo03:06
ccrouchnot for linked containers no03:07
jpeeler1would it be crazy to link the test container to all of them?03:07
ccrouchwell you would link all of them to the test container i think03:07
ccrouchbut i dont think thats insane03:08
*** eghobo has quit IRC03:08
ccrouchit certainly sounds good enough for a v1.0 :-)03:08
jpeeler1yeah i think so03:08
ccrouchsounds like you have a plan :-)03:09
ccrouchok i'm off03:09
jpeeler1more so than i did, thanks for the chat ccrouch03:09
jpeeler1see ya03:09
ccrouchttyl, cheers!03:10
*** ccrouch is now known as ccrouch-afk03:10
*** jpeeler1 is now known as jpeeler03:10
*** jpeeler has quit IRC03:10
*** jpeeler has joined #tripleo03:10
*** alop has quit IRC03:17
*** pradeep has joined #tripleo03:21
*** pradeep1 has quit IRC03:24
*** pradeep1 has joined #tripleo03:35
*** pradeep has quit IRC03:35
*** penick has quit IRC03:47
*** spzala has quit IRC03:52
*** pradeep1 has quit IRC03:55
*** yog_ has quit IRC03:55
*** untriaged-bot has joined #tripleo04:00
untriaged-botUntriaged bugs so far:04:00
openstackLaunchpad bug 1387487 in tuskar " tripleo-image-elements/elements/tuskar-ui /os-refresh-config /post-configure.d/101-tuskar-ui needs authentication from Keystone" [Undecided,New]04:00
*** untriaged-bot has quit IRC04:00
*** yog_ has joined #tripleo04:01
*** weshay has quit IRC04:03
*** tzumainn has quit IRC04:10
*** pradeep has joined #tripleo04:15
*** yamahata has joined #tripleo04:22
openstackgerritOpenStack Proposal Bot proposed openstack/os-cloud-config: Updated from global requirements
openstackgerritOpenStack Proposal Bot proposed openstack/os-collect-config: Updated from global requirements
*** eghobo has joined #tripleo04:57
openstackgerritOpenStack Proposal Bot proposed openstack/tuskar: Updated from global requirements
*** masco has joined #tripleo05:00
*** clarkb has quit IRC05:00
*** stendulker has joined #tripleo05:13
*** akuznetsov has joined #tripleo05:14
openstackgerritMerged openstack/tripleo-image-elements: Correct typo in the restore-ssh-host-keys element
*** ParsectiX has quit IRC05:26
*** ParsectiX has joined #tripleo05:27
*** pradeep has quit IRC05:32
*** pradeep has joined #tripleo05:33
*** eghobo has quit IRC05:39
*** eghobo has joined #tripleo05:39
*** sdake_ has joined #tripleo05:41
*** sdake__ has joined #tripleo05:44
*** eghobo_ has joined #tripleo05:47
*** sdake_ has quit IRC05:48
*** eghobo has quit IRC05:51
*** shadower has quit IRC05:55
*** shadower has joined #tripleo05:56
*** regebro has joined #tripleo06:01
*** stendulker_ has joined #tripleo06:04
*** stendulker has quit IRC06:06
*** lazy_prince has quit IRC06:08
*** ParsectiX has quit IRC06:08
*** killer_prince has joined #tripleo06:08
*** killer_prince is now known as lazy_prince06:08
*** stendulker has joined #tripleo06:19
*** daneyon has joined #tripleo06:22
*** stendulker_ has quit IRC06:23
*** akuznetsov has quit IRC06:35
*** ishant has joined #tripleo06:48
*** pradeep1 has joined #tripleo07:08
*** akuznetsov has joined #tripleo07:11
*** pradeep has quit IRC07:11
*** pradeep1 has quit IRC07:13
openstackgerritRamakrishnan G proposed openstack/diskimage-builder: Ironic: uefi localboot support
*** Marga_ has quit IRC07:20
*** openstackgerrit has quit IRC07:36
*** openstackgerrit has joined #tripleo07:36
*** ifarkas has joined #tripleo07:37
*** killer_prince has joined #tripleo07:37
*** ifarkas_ has joined #tripleo07:37
*** lazy_prince has quit IRC07:38
*** killer_prince is now known as lazy_prince07:38
*** ifarkas_ has quit IRC07:41
*** rdopiera has joined #tripleo07:42
*** yog_ has quit IRC07:55
*** eghobo_ has quit IRC07:58
openstackgerritMerged openstack/tuskar: Modify tuskar.conf.sample according to the new oslo.db release
*** pradeep has joined #tripleo08:00
openstackgerritDougal Matthews proposed openstack/tuskar: Updated from global requirements
*** akuznetsov has quit IRC08:04
*** athomas has joined #tripleo08:06
*** maufart has joined #tripleo08:10
*** jprovazn has joined #tripleo08:15
*** jtomasek has joined #tripleo08:15
*** akuznetsov has joined #tripleo08:23
*** tchaypo is now known as tchorizo08:25
*** tchorizo is now known as tchaypo08:25
*** brad[] has quit IRC08:44
*** mcornea has joined #tripleo08:44
*** jistr has joined #tripleo08:49
*** akuznetsov has quit IRC08:53
*** ishant is now known as tnashi08:54
*** jcoufal has joined #tripleo08:54
*** tnashi is now known as ishant08:55
*** gfidente has joined #tripleo08:56
*** dtantsur|afk is now known as dtantsur08:58
*** mcornea has quit IRC08:59
*** mcornea has joined #tripleo09:00
*** pblaho has joined #tripleo09:02
*** mcornea_ has joined #tripleo09:03
openstackgerritDougal Matthews proposed openstack/python-tuskarclient: Remove unused exc file
*** mcornea has quit IRC09:05
*** yamahata has quit IRC09:07
*** pdb has joined #tripleo09:14
*** lucasagomes has joined #tripleo09:21
*** pradeep has quit IRC09:24
*** pradeep has joined #tripleo09:25
*** mcornea_ has quit IRC09:33
*** athomas has quit IRC09:33
*** mcornea has joined #tripleo09:34
*** stendulker has quit IRC09:44
*** sdake__ has quit IRC09:44
gfidentespredzy, how's going?09:51
*** MasterPiece has joined #tripleo09:52
spredzygfidente, morning. Good, you ?09:55
gfidentebit slow but okay09:56
gfidentei'd like to get to the point where we can test the pieces later today09:56
gfidenteor tomorrow at last09:56
gfidentesounds reasonable?09:57
spredzygfidente, sure. I am finishing something and will re submit the PR with the proper change so we can go ahead and test it09:57
spredzygive me about 45mins.09:57
gfidenteyeah I was trying to figure if we're missing anything else?09:57
gfidentedoesn't look like but you never know09:58
spredzytest will tell. But at first glance nothing I can think of09:58
*** untriaged-bot has joined #tripleo10:00
untriaged-botUntriaged bugs so far:10:00
openstackLaunchpad bug 1387487 in tuskar " tripleo-image-elements/elements/tuskar-ui /os-refresh-config /post-configure.d/101-tuskar-ui needs authentication from Keystone" [Undecided,New]10:00
*** untriaged-bot has quit IRC10:00
*** yuanying has quit IRC10:05
*** wendar has quit IRC10:06
*** dasm has joined #tripleo10:09
*** MasterPiece has quit IRC10:11
*** jang2 has joined #tripleo10:14
*** masco has quit IRC10:15
*** wendar has joined #tripleo10:16
openstackgerritYanis Guenane proposed openstack/tripleo-heat-templates: Add support for EnableCinderRbdBackend
*** MasterPiece has joined #tripleo10:31
*** athomas has joined #tripleo10:31
*** pelix has joined #tripleo10:42
*** pdb has quit IRC10:51
*** jp_at_hp has joined #tripleo10:51
*** pdb has joined #tripleo10:51
*** lazy_prince is now known as killer_prince10:57
*** rlandy has joined #tripleo11:05
openstackgerritYanis Guenane proposed openstack/tripleo-heat-templates: Add support for EnableCinderRbdBackend
*** mrunge has joined #tripleo11:26
*** ishant has quit IRC11:41
*** ishant has joined #tripleo11:42
*** yog_ has joined #tripleo11:44
openstackgerritSteven Hardy proposed openstack/tripleo-heat-templates: Update puppet post config to enable stepped deployments
shardyshadower: ^^ I think those template changes should make stepped deployments via heat easier11:49
shardystill testing the heat parts atm, feedback welcome on the templates tho :)11:49
* shadower looks11:51
shardyshadower: I'm planning to set break^H^H hooks e.g on *PuppetStep111:51
shadowerI think that should work11:53
shardyshadower: ok, cool, I'll update the undercloud heat to include your patches and try it out11:54
shadower(I mean, if it doesn't, that's a bug but I've tested it)11:54
shardyone thing I wasn't sure is if we can clear breakpoints via a glob yet?11:54
shardyI guess we can make the client do that fairly easily11:55
shadowerwe can't :-(11:55
shardyOk, no worries, I might hack around that in the client if that's OK, e.g get heatclient to walk the list matching a glob11:56
shadowershardy: problem is, if you want to send a bunch of "clear bp on this resource" signals, then `heat breakpoint-clear "Phase1*"` doesn't have enough information11:56
shadoweryou'd have to get the template out of the stack and parse it11:56
shadoweror alternatively have a separate glob clearing signal in the engine11:56
shardyshadower: can't you just do a recursive resource-list and match on that?11:56
shadoweror that11:57
shadowerbut it's not implemented yet11:57
shardyheat breakpoint-clear -R "*PuppetStep1"11:57
shardyor something11:57
shardyshadower: Ok, no worries, I might have a crack at it, if that's OK11:57
shardyif it looks too hard, I'll fall back to specifying it longhand11:58
shadowerno problem at all :-)11:58
shardyalso, it'd be nice to specify a series of environments, so you could do e.g11:58
shadowerit probably shouldn't be too hard but I prioritised the engine work and the basic functionality11:58
shardyheat breakpoint-update -e step2.yaml11:58
shardyor something11:58
shardysince you could start the create with stack-create overcloud -e step1.yaml11:59
shardybut there's no easy way to feed in a step2.yaml atm (AFAICS)11:59
shadowerthough I thought the way this'd be used in tripleo is to specify all steps (via globs) up front and then clear them in waves12:00
shardyyeah, that'd work too12:00
shardyif we have the glob clear-breakpoint thing12:00
shardyI'll get onto that after lunch, thanks!12:00
*** noslzzp has joined #tripleo12:01
*** athomas has quit IRC12:01
*** EmilienM|afk is now known as EmilienM12:04
*** dtantsur is now known as dtantsur|brb12:13
*** athomas has joined #tripleo12:16
*** rhallisey has joined #tripleo12:19
*** akuznetsov has joined #tripleo12:20
openstackgerritYanis Guenane proposed openstack/tripleo-heat-templates: Add support for EnableCinderRbdBackend
*** pradeep has quit IRC12:28
*** ishant has quit IRC12:31
*** weshay has joined #tripleo12:32
*** Marga_ has joined #tripleo12:33
*** jcoufal_ has joined #tripleo12:34
openstackgerritMarios Andreou proposed openstack/tuskar: Adds support for --role-extra files for extra template files
*** jcoufal has quit IRC12:37
*** spredzy is now known as spredzy|afk12:38
*** Marga_ has quit IRC12:40
*** dmellado has quit IRC12:52
openstackgerritJiri Tomasek proposed openstack/tuskar-ui: Required parameters check as part of deployment validation
shardyCan services which have os-svc-install do their source install install from an under-review branch?12:59
shardyLooking at the source I can't quite see what syntax would do it13:00
shardyI want to install heat from source, plus
*** dprince has joined #tripleo13:01
*** dmellado has joined #tripleo13:02
*** athomas has quit IRC13:06
openstackgerritSteven Hardy proposed openstack/os-net-config: Add a --no-activate option to disable device up/down actions
openstackgerritSteven Hardy proposed openstack/os-net-config: Enable renaming of interfaces without reboot
*** tzumainn has joined #tripleo13:12
*** athomas has joined #tripleo13:13
*** morazi has joined #tripleo13:14
openstackgerritRadomir Dopieralski proposed openstack/tuskar-ui: Don't crash when a bulk action is done with empty node list
*** lucasagomes is now known as lucas-hungry13:27
*** Marga_ has joined #tripleo13:28
*** dtantsur|brb is now known as dtantsur13:32
*** yog__ has joined #tripleo13:38
*** spredzy|afk is now known as spredzy13:39
*** dmellado_ has joined #tripleo13:40
*** dmellado has quit IRC13:40
*** dmellado_ is now known as dmellado13:40
*** pradeep has joined #tripleo13:41
*** yog_ has quit IRC13:42
*** Marga_ has quit IRC13:47
*** Marga_ has joined #tripleo13:57
openstackgerritEmilien Macchi proposed openstack/tripleo-heat-templates: puppet/loadbalancer: use puppet-tripleo
*** sdake_ has joined #tripleo14:06
openstackgerritYanis Guenane proposed openstack/tripleo-heat-templates: Add support for EnableCinderRbdBackend
*** leseb has left #tripleo14:07
openstackgerritYanis Guenane proposed openstack/tripleo-heat-templates: Add support for EnableCinderRbdBackend
openstackgerritRadomir Dopieralski proposed openstack/tuskar-ui: Don't crash when a bulk action is done with empty node list
*** lblanchard has joined #tripleo14:18
*** radez_g0n3 is now known as radez14:18
*** Marga_ has quit IRC14:25
*** pradeep1 has joined #tripleo14:30
*** pradeep has quit IRC14:31
jistrdprince: gerrit doesn't like me, i tried re-adding my +A, also tried doing a "reverify", but it still doesn't look like it's going to merge the patch. Can you try giving +A too?
jistri'm not sure what else to do to make the merge happen :D14:33
dprincejistr: okay, I've +A'd again too14:36
dprincejistr: there was a queue issue yesterday in infra. I think we got bit by it here14:36
gfidentedprince, have a problem with packages dependencies, cinder does not strictly depend on python-enum so we don't pull that in by default, but it is needed for the rbd driver14:36
jistrdprince: yeah. thx14:37
dprincegfidente: I guess the question is should Cinder simply include python-enum?14:37
gfidenteyeah I think it should14:37
gfidentemaybe for a few backends, amongst which ceph is included, it'd be better to make sure all the deps are in place14:38
dprincegfidente: push a patch there (via gerrithub)14:38
dprincegfidente: this would require using the latest Cinder build in Delorean though which means we'd have to also fix the Neutron undercloud issue before CI can use it.14:38
dprincegfidente: so long as you build your seed/undercloud from an older Delorean repo (or from source) you could use this package from Delorean current once built though14:39
gfidentedprince, yeah I'm building the OC images from /current14:40
dprincegfidente: then send your patch and we'll land it14:41
dprincegfidente: there is a keystone issue today BTW14:41
dprincegfidente: latest keystone requires python-cryptography (for the new lightweight token stuff)14:41
gfidenteso for testing, I will just mountloop the images and add python-enum14:42
dprincegfidente: I'm working on fixing our deps for that now. It isn't available in F20 :(14:42
dprincegfidente: I think it is time to move to F21....14:42
dprincegfidente: if that works for now yes. I've done that before14:42
* dprince has a hacked local copy of DIB which supports editing14:42
* dprince tried to land that paths year(s) ago and some people didn't like it....14:43
* dprince may try again14:43
*** lucas-hungry is now known as lucasagomes14:43
*** chuckC_ has joined #tripleo14:46
*** chuckC_ has quit IRC14:48
*** rhallisey has quit IRC14:53
*** rhallisey_ has joined #tripleo14:53
*** yog__ has quit IRC14:55
openstackgerritRadomir Dopieralski proposed openstack/tuskar-ui: Don't crash when a bulk action is done with empty node list
*** radez is now known as radez_g0n314:56
*** rhallisey_ is now known as rhallisey14:59
*** prad has joined #tripleo15:09
dprinceslagle: trying to upgrade to Fedora 21 and seem to be hitting some selinux issues with tripleo-image-elements in my seed15:16
dprinceslagle: I really want /mnt/state gone man! and all this SELinux sillyness too15:16
rhalliseydprince, what is the selinux issue you're hitting?15:17
dprincerhallisey: also, mariadb won't start up w/ /mnt/state either :(15:18
dprincerhallisey: so fairly hosed...15:18
*** ChuckC has quit IRC15:19
rhalliseydprince, darn.. I'll take a look at this policy see what's wrong with it..15:19
openstackgerritMerged openstack-infra/tripleo-ci: Explicitly set overcloud stack timeout to lower value
dprincerhallisey: thanks15:20
rhalliseysure :)15:20
*** yamahata has joined #tripleo15:20
shardydprince: hey, I'm hitting some issues with *NodesPostDeployment15:22
shardyis there somewhere I can see the detailed log of why the puppet hook returned 2?15:22
shardyI'm running delorean pinned at f21/0d/97/0d979b220da3c5c809a20d6d9e7afa6b2c246712_772aab0d, with latest master heat15:23
*** jprovazn has quit IRC15:23
shardyalso t-h-t plus
shardyit's quite possible it's my environment, just not sure how to debug15:24
dprinceshardy: usually the stdout is in the os-collect-config logs15:24
dprinceshardy: so 'journalctl -u os-collect-config'15:24
dprinceshardy: however in some peoples environment it could be in /var/log/messages too15:25
shadowerthe puppet stuff still uses occ?15:25
dprinceshadower: yes, that is what invokes the heat puppet hook15:25
dprinceshadower: from heat-elements15:25
*** dpaterson has joined #tripleo15:26
dprinceshadower: there isn't much else though, os-net-config and that is mostly it15:26
dprinceshardy: it would be nice harvest the puppet output and be able to view it via heat somehow15:27
dprinceshardy: probably a bit verbose though15:27
shardydprince: well the deployment captures the stdout so it should be possible15:27
dprinceshardy: I'm not sure what the stdout/stderr split is (with regards to the useful information) but capturing errors would be super useful15:27
dpatersonIs there a way to increase timeouts for os-service-start?15:28
dprinceshardy: because everyone that has tried this eventually hits an error and then asks me where the logs are. It isn't intuitive right now15:28
shardyError: Could not find data item snmpd_readonly_user_password in any Hiera data file and no default supplied15:28
dpatersonos-refresh-config is failing for me continually when installing undercloud.15:29
shardyThat appears to be the only puppet-ish error, I'm not sure how I caused it :)15:29
shardydprince: this is with my stepped deployment patch, maybe there's some intra-deployment dependencies, checking15:30
dprinceshardy: so that setting should be in your /etc/puppet/hieradata files somewhere.15:30
dprinceshardy: sounds like maybe it didn't make it there in time?15:30
dprinceshardy: yep, could be the 'name:' parameters aren't doing ordering correctly and as you mentioned we need to switch to depends_on15:31
shardydprince: this is after the move to depends_on in
shardyso the newly serialized deployments may be causing it15:32
dprinceshardy: oh, I haven't tried that yet. Looks like it passes CI though.15:32
shardydprince: Yeah, it worked for me until I tried to pull latest heat via DIB_INSTALLTYPE_heat=source15:33
dprinceshardy: anyways, snmpd_readonly_user_password metadata is provided by the ControllerConfig resource (for the controller)15:33
dprinceshardy: ah, well there is your problem! :)15:33
shardyheh :)15:33
dprinceshardy: seriously though, it could just be an ordering regressions somewhere15:34
shardyCI heat isn't pinned anymore is it?15:34
openstackgerritMarios Andreou proposed openstack/tuskar: Adds support for --role-extra files for extra template files
openstackgerritMarios Andreou proposed openstack/tuskar: Adds the /role/role_id/extra_data path to the v2 API
gfidentedprince, shardy curious about one thing, that change is meant to run puppet 4 times right?15:34
dprinceshardy: For delorean everything is pinned until we fix the Neutron undercloud issue15:34
dprinceshardy: which I'm trying to work on today15:34
shardydprince: Yeah, I wanted to leave everything pinned but pull in shadowers heat breakpoint patches15:35
dprinceshardy: so the puppet CI is using an old heat (2 weeks old now)15:35
shardyis there any way to pull from a review when using delorean?15:35
dprinceshardy: totally fine to do that. building a source based seed is what I would do as well for this15:35
shardyI assumed I'd have to switch heat to source then cherry-pick into the venv15:35
*** ccrouch-afk is now known as ccrouch15:36
dprinceshardy: that should work fine15:36
shardyOk, will dig into it, thanks!15:37
gfidenteshardy, so it that supposed to produce 4 runs of puppet in a sequence with some additional metadata each?15:37
dprinceshardy: also, if the heat patch you want is up for review you can just set DIB_REPOREF_heat=<refspec>15:37
dprinceshardy: this works since you've set DIB_INSTALLTYPE_heat=source15:37
shardygfidente: we already do 4 runs of puppet, that's just adding explicit dependencies between them15:37
shardydprince: ah, I was wondering about that - is refspec the change id?15:38
dprinceshardy: export DIB_REPOREF_heat=refs/changes/23/146123/6 for example15:38
shardythanks :)15:38
dprinceshardy: if it is a patch series it gets all the relevant changes too (it isn't a cherry pick)15:39
shardygfidente: then, we can use to control each step15:39
shardydprince: Ok, cool, that's really useful15:39
shardyin this case, it's just one patch, but really handy to know15:39
EmilienMdprince, gfidente: I'm looking at spredzy patch ( ) - I think in long term we should patch puppet-tripleo to manage ceph cinder backends, so we can create multiple cinder backends with RBD drivers. His current implementation does not allow it. Thoughts ?15:39
bnemecdprince: slagle is on pto today and tomorrow afternoon.15:40
dprincebnemec: aha, thanks for that15:40
bnemecdprince: We have a trello card open for updating tripleo to f21 by default, but they're not giving us any time to work on it. :-(15:40
gfidenteEmilienM, yes I was thinking to actually always use multibackend15:40
EmilienMgfidente: multibackend is not an option, we all agree - what I think is about the way to create cinder types, etc - I think we need to manage it with a define function in puppet-tripleo, so we can create backends via Hiera only.15:41
*** pradeep1 has quit IRC15:41
*** pradeep has joined #tripleo15:42
EmilienMdprince: I'm trying to see if it matches with your comment on
dprinceshardy, shadower: commented on
dprinceEmilienM: looking....15:42
shardydprince: Thanks, i'm happy to s/Puppet/$generic_name15:43
shardyfrom a breakpoint point of view, it doesn't matter, provided the naming is consistent15:43
shardyand it makes sense to have either a suffix or a prefix to make the glob specification easier15:44
dprinceEmilienM: is a free form JSON blob parameter the best way to do that? (add support for multiple ceph backends)15:44
*** rdopiera has quit IRC15:44
spredzyEmilienM, gfidente dprince . Thing is there is 2 patterns here. #1 Very generic. Use of json blob everywhere with create_resources(). Making it possible to address any possible need upstream modules can cover. #2 More specific. This will make it easier parameter wise but will kind of lock us up in the deployment scenario we provide.15:44
gfidenteEmilienM, yeah it's probably worth having something to manipulate that easily from hiera15:44
EmilienMI like spredzy approach in puppet-cinder > - though we need to manage cinder types, that's what I'm talking about15:46
dprincespredzy: if we go with blobs could we have a reasonable simple case Ceph default that worked by default? I guess this puts the person using the Heat templates in the position of having to be a Puppet (Ceph) expert... or at least power user in order to understand what to provide as an input to these Hiera data (json) blobs.15:47
EmilienMdprince: +115:47
dprincespredzy: and that is new territory...15:47
openstackgerritImre Farkas proposed openstack/tripleo-image-elements: Enable DRAC driver in Ironic
dprincespredzy: I'm not saying we can't do it, only that the free form flexability may make it hard for the simple case. I think we actually may want to support both features: simple case, and extensability15:48
dprincespredzy: so... maybe something like what I did for the os-net-config templates where there are examples in the tree and you can choose one of them.15:49
*** maufart has quit IRC15:49
spredzydprince, I don't see currently how we could provide sane default beyond exposing the mandatory puppet parameters and using the defaults when any. But I can look into a fair solution that might meet both usage15:49
dprincespredzy: so all the net-* templates like this:
*** stendulker has joined #tripleo15:49
*** pdb has quit IRC15:49
*** funzo has quit IRC15:49
* spredzy looking15:50
dprincespredzy: if you could outline the hiera we'd need for the simple case, and also (just a single) more complex hiera case which supported multiple backends I could show you an example of how we could plug these things in.15:50
gfidenteshardy, EmilienM, dprince to me it looks like we should reall have that enable{ceph,iscsi,nfs} boolean which dprince suggested, and work out the multibackend config, together with the volume types, without asking the user for any further input15:51
gfidentein that scenario, I'd always use multibakend, and always start with iscsi by default15:52
gfidenteI am not a fan of letting the user pasting extra json unless we document it well and even in that case, it should only support the scenario where you want to configure the services for a pre-existing deployment15:53
EmilienMthat sounds fair15:53
*** jprovazn has joined #tripleo15:53
gfidentepre-existing ceph deplyment15:53
gfidentemy point is that it is complicated enough to manage that I would try to avoid user misconfiguring15:54
gfidenteand make it more like a boolean choice15:54
dprincegfidente: I think that sounds reasonable. Still I'd like to flesh out this multi-host backend thing a bit too. For more advanced deployments we should support both.15:54
gfidentemulti-host backend?15:54
dprincegfidente: multiple ceph backends, sorry15:55
spredzygfidente, my main concern in the case of cinder is if for some reason multiple RBD backend is needed. I am not sure its a real case scenario but theroically nothing prevents it. I am not sure EnableCeph will be enough. How do you determine the various ceph backend. If a user has two ceph cluster. One already in production and the one we will be installting, and want to put volumes/vms on differents cluster based on business needs. How can this be addres15:55
spredzysed ?15:55
dprincespredzy: so we could have a simple case software deployment which uses the Heat template parameters directly.15:56
dprincespredzy: this would get plugged in via the overcloud-resource-registry by default.15:56
gfidenteyes I can see this won't be easy to achieve but for that I think the json blob can work15:57
dprincespredzy: and then for the more complex case the user could generate or provide their own free form hiera data for multiple backends.15:57
gfidentewhere we just consider each key a setting and don't touch them15:57
gfidentepaste those into a new backend with no further manipulation15:57
dprincespredzy: and again, this would get plugged in by editing (or having tuskar generate) the correct replacement for the overcloud-resource-registry.yaml15:57
*** pdb has joined #tripleo15:58
dprincespredzy: basically very similar to what we are going for with regards to the low level physical network connectivity as well15:58
*** untriaged-bot has joined #tripleo16:00
untriaged-botUntriaged bugs so far:16:00
openstackLaunchpad bug 1387487 in tuskar " tripleo-image-elements/elements/tuskar-ui /os-refresh-config /post-configure.d/101-tuskar-ui needs authentication from Keystone" [Undecided,New]16:00
*** untriaged-bot has quit IRC16:00
*** pradeep has quit IRC16:01
spredzydprince, gfidente so if I get it right we should make our templates to allow both scenarios? #1. Drive a ceph deployment if EnableCeph and configure it automatically based on the input parameters. #2 Have a CephBackends (or whaterver the name is) that would allow someone to pass a json blol for advanced configuration ?16:01
*** pradeep has joined #tripleo16:01
gfidentespredzy, regarding 2), not cephbackends, just a blob where user can paste any key which we paste as a backend16:02
dprincespredzy: yes, and I'd say lets go for the simple case first (like you have)16:02
gfidenteif that is going to be a second ceph or just a new type, we don't care and don't need to, we just paste it in the config16:02
spredzygfidente, ack16:02
spredzydprince, ack16:02
gfidentespredzy, so we call the stanza user_backend and expect each key to represent an actual config key16:03
gfidentewell, an actual puppet class parameter16:04
shardydprince: for some reason, the snmp password is an empty string in the heiradata, but SnmpdReadonlyUserPassword is set in the stack parameters16:04
dprinceshardy: that is weird. Tell you what. Let me try a recent heat build today and get back to you16:05
spredzygfidente, this is what I want this class to let us do :,cm16:05
shardydprince: Ok, thanks!16:05
gfidentespredzy, exactly16:06
dprincestarting to sleet here, moving home... back online in a bit16:06
*** dprince has quit IRC16:07
*** pradeep1 has joined #tripleo16:07
*** funzo has joined #tripleo16:07
EmilienMspredzy: why not creating the type in cinder::backend::rbd ?16:07
EmilienMthat would avoid any patch in puppet-tripleo I think16:08
*** pradeep has quit IRC16:08
openstackgerritImre Farkas proposed openstack/tripleo-image-elements: Enable DRAC driver in Ironic
*** ChuckC has joined #tripleo16:17
*** ChuckC has quit IRC16:18
*** ChuckC has joined #tripleo16:18
*** MasterPiece has quit IRC16:19
*** stendulker has quit IRC16:23
gfidenteEmilienM, spredzy, also if we want to use ceph by default for nova as well, we need to change the default caps, but I am not sure that is a good idea16:23
spredzygfidente, to use ceph for nova ? in the first time16:23
gfidenteideally this is yet another feature we should control with some boolean, using ceph provides for live migrations, not using it should give better performances16:24
gfidentethese are the two ups/downs I can think about as a rule of thumb16:24
gfidentespredzy, patchset #10 switched to using it for nova images as well16:24
spredzygfidente, thought it would be a nice addition but I can leave that for another ps16:25
gfidenteto me starting with persistent storage only sounds less disruptive while providing HA for cinder to start with16:25
gfidentemoving the ephemeral to ceph is more something about which I'd let user decide16:26
gfidentebut to be honest, this is just my guessing16:26
*** Marga_ has joined #tripleo16:27
*** pradeep has joined #tripleo16:29
*** chuckC_ has joined #tripleo16:29
*** chuckC_ has quit IRC16:29
*** pradeep1 has quit IRC16:30
openstackgerritMerged openstack/tripleo-puppet-elements: Install Ceph in the Compute image as well
*** rhallisey has quit IRC16:33
jpeelersdake: did you see these Dockerfiles? might be useful...16:34
*** lxsli has joined #tripleo16:39
lxslijdob: please?16:39
*** mcornea has quit IRC16:41
*** rwsu-afk is now known as rwsu16:43
*** rhallisey has joined #tripleo16:44
*** MasterPiece has joined #tripleo16:46
*** alop has joined #tripleo16:48
*** lxsli has left #tripleo16:49
*** yamahata has quit IRC16:50
EmilienMgfidente: I don't see much logs on my failure ( ) - thoughts ?17:03
*** akuznetsov has quit IRC17:06
*** jistr has quit IRC17:08
*** yamahata has joined #tripleo17:12
*** dtantsur is now known as dtantsur|afk17:16
*** yamahata has quit IRC17:18
*** yamahata has joined #tripleo17:19
*** dprince has joined #tripleo17:26
gfidenteEmilienM, you know I was trying to blame your change but I can't find the issue with it17:39
gfidenteespecially because I don't see any log, it's not just os-collect-config which is empy17:40
gfidenteI only see the swift stuff17:40
gfidenteand the data in host_info looks okay17:42
*** rbrady has quit IRC17:42
*** pdb has quit IRC17:42
EmilienMoh dear, what did I break?17:42
gfidenteI don't think it's the change really17:42
*** mrunge has quit IRC17:47
*** pradeep has quit IRC17:52
*** rbrady has joined #tripleo17:57
*** lazy_prif has joined #tripleo17:59
*** lazy_prif is now known as lazy_prince17:59
*** killer_prince has quit IRC17:59
openstackgerritOpenStack Proposal Bot proposed openstack/tuskar: Updated from global requirements
*** eghobo has joined #tripleo18:08
*** penick has joined #tripleo18:09
*** lucasagomes is now known as lucas-dinner18:09
*** radez_g0n3 is now known as radez18:09
*** jtomasek has quit IRC18:12
openstackgerritImre Farkas proposed openstack/tripleo-image-elements: Enable DRAC driver in Ironic
*** jcoufal_ has quit IRC18:18
*** Marga_ has quit IRC18:22
*** prad has quit IRC18:24
*** prad has joined #tripleo18:24
*** rbrady has quit IRC18:37
*** dtantsur|afk has quit IRC18:43
*** athomas has quit IRC18:43
*** dtantsur has joined #tripleo18:45
*** rebrego has joined #tripleo18:46
*** regebro is now known as Guest856018:46
*** rebrego is now known as regebro18:46
*** Guest8560 has joined #tripleo18:46
*** regebro is now known as Guest6071718:46
*** Guest8560 is now known as regebro18:46
*** penick has quit IRC18:47
*** jrist has joined #tripleo18:47
*** Guest60717 has joined #tripleo18:47
*** regebro has quit IRC18:47
*** Guest60717 is now known as regebro18:47
*** Guest11124 has joined #tripleo18:47
*** regebro is now known as Guest1006118:47
*** Guest11124 is now known as regebro18:47
*** Guest10061 has joined #tripleo18:48
*** regebro has quit IRC18:48
*** Guest10061 is now known as regebro18:48
*** Guest67001 has joined #tripleo18:48
*** regebro has quit IRC18:48
*** Guest67001 is now known as regebro18:48
*** Guest55192 has joined #tripleo18:48
*** regebro has quit IRC18:48
*** Guest55192 is now known as regebro18:48
*** regebro is now known as Guest1382718:49
*** regebro has joined #tripleo18:49
*** Guest13827 has joined #tripleo18:49
*** regebro is now known as Guest314318:49
*** Guest13827 is now known as regebro18:49
*** regebro has quit IRC18:49
*** regebro has joined #tripleo18:49
*** Guest87396 has joined #tripleo18:50
*** regebro has quit IRC18:50
*** Guest87396 is now known as regebro18:50
*** regebro is now known as Guest9927818:50
*** regebro has joined #tripleo18:50
*** Guest99278 has joined #tripleo18:50
*** regebro has quit IRC18:50
*** Guest99278 is now known as regebro18:50
*** regebro has quit IRC18:51
*** regebro has joined #tripleo18:51
*** Guest97717 has joined #tripleo18:51
*** regebro is now known as Guest9683018:51
*** Guest97717 is now known as regebro18:51
*** rbrady has joined #tripleo18:51
*** regebro is now known as Guest2829718:51
*** regebro has joined #tripleo18:51
*** Guest28297 has joined #tripleo18:52
*** regebro has quit IRC18:52
*** Guest28297 is now known as regebro18:52
*** regebro has quit IRC18:52
*** regebro has joined #tripleo18:52
*** Guest23052 has joined #tripleo18:52
*** regebro is now known as Guest5382818:52
*** Guest23052 is now known as regebro18:52
*** regebro has quit IRC18:53
*** regebro has joined #tripleo18:53
*** regebro is now known as Guest6891818:53
*** Guest74350 has joined #tripleo18:53
*** Guest68918 has quit IRC18:53
*** Guest74350 is now known as regebro18:53
*** jerryz_ has quit IRC18:53
*** regebro is now known as Guest3269718:53
*** regebro has joined #tripleo18:53
*** Guest32697 has joined #tripleo18:54
*** regebro has quit IRC18:54
*** Guest32697 is now known as regebro18:54
*** regebro is now known as Guest5413718:54
*** regebro has joined #tripleo18:54
*** Guest54137 has joined #tripleo18:54
*** regebro has quit IRC18:54
*** Guest54137 is now known as regebro18:54
*** regebro is now known as Guest1045718:55
*** regebro has joined #tripleo18:55
*** Guest10457 has joined #tripleo18:55
*** regebro has quit IRC18:55
*** Guest10457 is now known as regebro18:55
*** regebro is now known as Guest1602218:55
*** regebro has joined #tripleo18:55
*** Guest16022 has joined #tripleo18:56
*** regebro is now known as Guest9749018:56
*** Guest16022 is now known as regebro18:56
*** regebro has quit IRC18:56
*** regebro has joined #tripleo18:56
*** regebro is now known as Guest5319918:56
*** Guest25068 has joined #tripleo18:56
*** Guest53199 has quit IRC18:56
*** Guest25068 is now known as regebro18:56
*** regebro is now known as Guest3631318:57
*** regebro has joined #tripleo18:57
openstackgerritDan Prince proposed openstack/tripleo-image-elements: Create Fedora 21 softlink for /var/run/mariadb
openstackgerritDan Prince proposed openstack/tripleo-image-elements: Work around 20-neutron-selinux issues on F21
openstackgerritPetr Blaho proposed openstack/python-tuskarclient: Replaces _create calls with _post in PlanManager
openstackgerritImre Farkas proposed openstack/tripleo-image-elements: Enable DRAC driver in Ironic
*** athomas has joined #tripleo19:04
*** penick has joined #tripleo19:05
*** jp_at_hp has quit IRC19:08
*** sdake_ has quit IRC19:09
*** sdake_ has joined #tripleo19:10
*** Marga_ has joined #tripleo19:11
*** jprovazn has quit IRC19:17
dprincebnemec: have you looked at this neutron issue yet?19:18
dprincebnemec: specifically the one we reverted here
bnemecdprince: I haven't, but I thought it was fixed.  Are we still reverting it?19:20
dprincebnemec: we are still reverting it. and it isn't fixed19:20
dprincebnemec: seems to be *the* issue at hand. #1 priority IMO at this point as its been over a week now19:20
bnemecdprince: Shoot, I thought it was fixed by
dprincebnemec: the motiviation was to fix Neutron DVR, which makes me wonder if we switched to DVR if it would work. I'm not sure that matters for a seed though19:21
dprincebnemec: let me try reverting the revert19:22
openstackgerritPetr Blaho proposed openstack/python-tuskarclient: Strenghten tests in {Plan,Role}ManagerTest
*** pelix has quit IRC19:23
lifelessdvr shouldn't matter on the seed today since we don't use neutron routing at all at that layer (today - no containers)19:23
dprincebnemec: when I tried it locally just now I see my nova instances stuck in BUILD/spawning, and things keep hitting neutron for port lookups19:23
dprincelifeless: agree, just pointing out that this patch broke us (the simple case) and we don't know why yet19:24
dprincelifeless: the DVR thing was a "just wondering"...19:24
dprincelifeless: if you have time to take a gander at this though :)19:25
dprincelifeless: you are probably faster here than me...19:25
*** Guest36313 has quit IRC19:25
openstackgerritDan Prince proposed openstack-infra/tripleo-ci: Drop the Neutron temprevert
dprincebnemec,lifeless: lets see if that passes. My guess is it won't19:26
bnemecIt was passing at one point.19:29
dprincebnemec: i see, hmm. well lets try it again19:35
openstackgerritDan Prince proposed openstack-infra/tripleo-ci: Remove neutron temprevert
dprincebnemec: ^^^ rebased Derek's and marked mine as a WIP.19:44
bnemecdprince: Sounds good19:45
lifelessdprince: sorry, was doing pre-kindy routine with cynthia20:08
dprincelifeless: pre-kindy is fun20:09
*** MasterPiece has quit IRC20:09
*** MasterPiece has joined #tripleo20:17
*** Marga_ has quit IRC20:20
*** Marga_ has joined #tripleo20:21
*** openstackgerrit has quit IRC20:25
*** openstackgerrit has joined #tripleo20:26
*** penick has quit IRC20:31
openstackgerritSteven Hardy proposed openstack/tripleo-incubator: Fall back to generated snmp password with --no-undercloud
shardydprince: ^^ FYI that fixes the error I mentioned earlier20:40
shardyI'm not sure why it started failing now, maybe something changed in the puppet implementation?20:40
shardyanyway, not passing a zero-length password solves it for me20:40
dprinceshardy: let me see if I have one set locally perhaps, wierd that CI was always passing before though....20:43
shardydprince: Yeah, it's odd20:43
dprincegfidente, EmilienM: just discovered when I asked in infra that this isn't landing due to depends-on
*** penick has joined #tripleo20:44
dprincegfidente, EmilienM: Honestly, would say lets try and keep our use of that to sparingly20:44
dprincegfidente, EmilienM: if we end up needing it all the time we aren't doing something right20:44
*** david-lyle_afk has joined #tripleo20:45
openstackgerritDan Prince proposed openstack/tripleo-heat-templates: Puppet: First support Ceph
openstackgerritMerged openstack/tripleo-heat-templates: Puppet: First support Ceph
EmilienMdprince: ok20:45
dprincegfidente, EmilienM: Depends-on nuked, and sent ^^^ :)20:46
*** david-lyle_afk has quit IRC20:47
*** david-lyle_afk has joined #tripleo20:47
*** penick has quit IRC20:53
openstackgerritSteven Hardy proposed openstack/tripleo-heat-templates: Update puppet post config to enable stepped deployments
*** david-lyle_afk is now known as david-lyle21:03
*** rlandy has quit IRC21:08
*** ifarkas has quit IRC21:10
*** Marga_ has quit IRC21:10
*** Marga_ has joined #tripleo21:11
*** mjturek1 has joined #tripleo21:17
mjturek1hey, I'm currently using upstream diskimage-builder and trying to build a ubuntu image with the 'ubuntu' and 'baremetal' elements. The image gets created but when I mount it, the kernel and ramdisk are absent from the /boot/ directory. Has anyone run into this before?21:19
rhalliseyccrouch, you should push your changes to keystone/glance containers up stream I'm hitting the same race condition21:20
rhalliseyso I'm borrowing your containers :p21:20
ccrouchwill do later today21:21
lifelessmjturek1: thats what baremetal does21:21
lifelessmjturek1: how do you want to use the image?21:21
mjturek1lifeless, sorry I let me clarify21:21
mjturek1so I tried with both elements first, that failed. So as a test I tried with just the ubuntu element and mounted the image to see if they were there for extraction21:22
mjturek1lifeless, this is for an ironic deployment21:23
lifelessmjturek1: what do you mean by 'failed' ?21:25
mjturek1I don't have the error on hand but DIB was unable to extract the kernel and ramdisk.21:26
lifelessmjturek1: I think I'd need to see a pastebin of the output. The expected behaviour is that you end up with three files in your output location - image, kernel, ramdisk21:27
mjturek1lifeless ok no worries, I'm doing a run right now to get the output21:27
openstackgerritMerged openstack-infra/tripleo-ci: Remove neutron temprevert
*** penick has joined #tripleo21:30
openstackgerritSteven Hardy proposed openstack/tripleo-incubator: Fall back to generated snmp password with --no-undercloud
*** dprince has quit IRC21:42
openstackgerritDan Prince proposed openstack-infra/tripleo-ci: Switching back to delorean current
*** Marga_ has quit IRC21:50
*** Marga_ has joined #tripleo21:50
*** thrash is now known as thrash|g0ne21:56
*** untriaged-bot has joined #tripleo22:01
untriaged-botUntriaged bugs so far:22:01
*** untriaged-bot has quit IRC22:01
openstackLaunchpad bug 1428726 in tripleo "ceilometer_snmpd_password unset with --no-undercloud" [Undecided,In progress] - Assigned to Steven Hardy (shardy)22:01
openstackLaunchpad bug 1387487 in tuskar " tripleo-image-elements/elements/tuskar-ui /os-refresh-config /post-configure.d/101-tuskar-ui needs authentication from Keystone" [Undecided,New]22:01
openstackgerritSteven Hardy proposed openstack/tripleo-incubator: Fall back to generated snmp password with --no-undercloud
*** lblanchard has quit IRC22:10
*** penick has quit IRC22:25
*** eghobo has quit IRC22:33
openstackgerritSteven Hardy proposed openstack/tripleo-heat-templates: WIP overcloud stepped deployment environment
*** penick has joined #tripleo22:36
*** penick has quit IRC22:37
*** Marga_ has quit IRC22:45
*** Marga_ has joined #tripleo22:45
*** jang2 has quit IRC22:46
openstackgerritGiulio Fidente proposed openstack/tripleo-incubator: Add support for building Ceph nodes and configure Cinder to use it
*** daneyon_ has joined #tripleo22:56
*** daneyon has quit IRC22:59
*** rhallisey has quit IRC23:03
*** prad has quit IRC23:03
*** yuanying has joined #tripleo23:11
*** ChuckC is now known as ChuckC_23:19
*** rhallisey has joined #tripleo23:23
*** mjturek1 has quit IRC23:23
*** daneyon has joined #tripleo23:24
*** ChuckC_ is now known as ChuckC23:26
*** daneyon_ has quit IRC23:28
*** openstack has joined #tripleo23:53
*** rhallisey has quit IRC23:56

Generated by 2.14.0 by Marius Gedminas - find it at!