Wednesday, 2015-04-08

*** sdake_ has quit IRC00:32
openstackgerritgreghaynes proposed openstack/diskimage-builder: Add generic devuser element
tchaypogreghaynes: bnemec: do you think it’s fair to say that the only changes since bnemec’s original +2 have been stylistic?00:46
tchaypobecause if so I’m happy to +a that right now00:46
openstackgerritDan Sneddon proposed openstack/os-net-config: Add static IP assignment via template tokens in config.yaml
greghaynesI wouldnt mind giving it a day in case somoene replies to the ML with something they want different00:47
*** Marga_ has quit IRC00:53
*** stendulker has joined #tripleo01:01
*** weshay has quit IRC01:03
openstackgerritDan Sneddon proposed openstack/os-net-config: Add static IP assignment via template tokens in config.yaml
openstackgerritgreghaynes proposed openstack/diskimage-builder: Add generic devuser element
*** pradeep has joined #tripleo01:37
*** jehb has quit IRC02:02
*** mgagne has quit IRC02:02
*** adam_g has quit IRC02:02
*** rainya has quit IRC02:02
*** gus has quit IRC02:02
*** larstobi has quit IRC02:02
*** jroll has quit IRC02:02
*** gus has joined #tripleo02:03
*** radez_g0n3 is now known as radez02:05
*** eghobo has quit IRC02:06
*** adam_g has joined #tripleo02:07
*** jroll has joined #tripleo02:07
*** adam_g has quit IRC02:07
*** adam_g has joined #tripleo02:07
*** mgagne has joined #tripleo02:08
*** mgagne is now known as Guest3992602:08
*** rainya has joined #tripleo02:08
*** jroll has quit IRC02:08
*** jroll has joined #tripleo02:08
*** jehb has joined #tripleo02:09
*** stendulker has quit IRC02:11
*** julim has quit IRC02:12
*** yamahata has quit IRC02:16
*** larstobi has joined #tripleo02:19
*** radez is now known as radez_g0n302:23
*** pradeep has quit IRC02:54
*** julim has joined #tripleo03:08
*** eghobo has joined #tripleo03:14
*** eghobo has quit IRC03:14
*** spzala has quit IRC03:22
*** vinkman has joined #tripleo03:25
*** yuanying has joined #tripleo03:28
*** stendulker has joined #tripleo03:37
*** rwsu has quit IRC04:07
*** vinkman has quit IRC04:33
*** tzumainn has quit IRC04:38
openstackgerritMerged openstack/diskimage-builder: Improved apt-sources README
*** Marga_ has joined #tripleo04:45
*** Marga_ has quit IRC04:45
*** Marga_ has joined #tripleo04:45
*** Marga_ has quit IRC04:45
*** Marga_ has joined #tripleo04:46
*** yuanying has quit IRC04:58
*** pradeep has joined #tripleo05:05
*** lazy_prince has joined #tripleo05:05
*** julim has quit IRC05:15
*** ukalifon1 has joined #tripleo05:34
*** adrianopetrich has joined #tripleo05:36
*** masco has joined #tripleo05:44
*** maufart has joined #tripleo05:46
*** pradeep has quit IRC05:47
*** pradeep has joined #tripleo05:51
*** pradeep has quit IRC06:00
*** dasm|afk is now known as dasm06:04
*** mrunge has joined #tripleo06:10
*** jprovazn has joined #tripleo06:39
*** jtomasek has joined #tripleo06:44
*** jcoufal has joined #tripleo06:47
openstackgerritgreghaynes proposed openstack/diskimage-builder: Update cloud-init-datasources README
openstackgerritgreghaynes proposed openstack/diskimage-builder: Add generic devuser element
*** subscope has quit IRC07:04
*** pradeep1 has joined #tripleo07:08
*** mmagr has joined #tripleo07:11
*** shardy has joined #tripleo07:13
*** athomas has joined #tripleo07:15
*** subscope has joined #tripleo07:17
openstackgerritPino Toscano proposed openstack/diskimage-builder: Do not export REG_HALT_UNREGISTER between hook scripts
*** Marga_ has quit IRC07:33
*** yog__ has joined #tripleo07:39
*** bvandenh has joined #tripleo07:42
*** dtantsur|afk is now known as dtantsur07:44
*** mcornea has joined #tripleo07:50
*** jistr has joined #tripleo08:02
*** lucasagomes has joined #tripleo08:05
*** mikal has quit IRC08:22
*** mikal has joined #tripleo08:23
*** gfidente has joined #tripleo08:27
*** pradeep1 has quit IRC08:37
*** pelix has joined #tripleo08:54
*** untriaged-bot has joined #tripleo09:00
untriaged-botNo untriaged bugs so far! \o/09:00
*** untriaged-bot has quit IRC09:00
yog__yog__, Hello everyone, while deploying whole disk is it required to install bootloader since its already a bootable disk,09:06
yog__yog__, diskimage-builder "deploy-ironic" element seems  to do so when boot_option=local09:19
*** pradeep1 has joined #tripleo09:27
*** akrivoka has joined #tripleo09:46
*** stendulker has quit IRC09:52
*** mmagr is now known as mmagr|afk10:04
* gfidente ♥ everyone10:09
*** yog__ has quit IRC10:10
dasmgfidente: good morning10:17
*** pradeep1 has quit IRC10:18
gfidentedasm :)10:18
*** pradeep1 has joined #tripleo10:18
dasmquick question for everyone: do you have working tripleo setup with latest changes? i'm struggling with some problems in between: tripleo-heat-templates + tuskar + tuskar-ui + tripleo10:19
dasmand i don't know, if i'm doing something wrong, or maybe some regressions, or sth.10:19
dasmor, at least: does anyone know about good guide for deployment without tuskar-ui usage?10:19
gfidentedasm, I have one yes but I'm using puppet for the overcloud configuration and I deploy from devtest, not from tuskar10:21
openstackgerritDaneyon Hansen proposed stackforge/kolla: WIP: Adds Neutron Agents Support
*** jprovazn has quit IRC10:23
dasmgfidente: could you provide any link?10:24
gfidentedasm, link t what?10:24
dasmlink to guide with deployment10:24
dasmgfidente: dasm or, at least: does anyone know about good guide for deployment without tuskar-ui usage?10:24
*** jprovazn has joined #tripleo10:24
dasmi thought you're talking about this10:25
gfidenteand if you want to try the puppet version10:25
*** yog__ has joined #tripleo10:26
dasmgfidente: yeah, i've read it. i'm just wondering, if it's possible to deploy overcloud on baremetal, using devtest.10:27
dasmtried to go with happy path10:28
gfidentedasm, so for puppet,
gfidentejprovazn, knows all the things about baremetal10:30
* gfidente ♥ jprovazn as well10:30
jprovazndefinitely not10:30
dasmjprovazn: is it true? wink, wink.10:30
dasmbuu :(10:30
* jprovazn reaqds back10:31
jprovazndasm, aah, baremetal, well I never tried this, sorry :(10:32
*** pdb has quit IRC10:32
jprovaznbut it should work :)10:32
*** pdb has joined #tripleo10:32
dasmso, anyone, anywhere tried to use devtest (for undercloud) and install overcloud on baremetal?10:32
dasmjprovazn: yeah, i know. it even worked around month ago.10:32
dasmafter some changes, updates, i've tried to setup everything from scratch, and right now it seems that I have problems with tripleo-heat-templates.10:33
dasmbut still investigating this problem10:33
gfidentedasm, so jprovazn was a faking a bit but it's definitely hard to find someone who develops on baremetal around I think10:35
*** pblaho has joined #tripleo10:35
gfidentedasm, ideally you can use regular devtest up to having your undercloud deployed10:35
gfidenteand when that is done you mess with the ironic driver until you manage to configure there your baremetal hosts10:36
gfidenteso that ironic can power them up and provision those with a glance image10:36
gfidenteand if you want undercloud on baremetal, same is to be done in the seed10:37
dasmgfidente: thanks. i even succeeded with overcloud baremetal configuration: ironic can power on/off hosts.10:40
dasmeven pxe worked.10:40
gfidentedasm, so you're ahead of many probably10:40
gfidentewhat is failing?10:40
dasmi have a problem with tripleo-heat-templates, i think so.10:40
dasmHTTPBadRequest: ERROR: The specified reference "ObjectStorage" (in SwiftDevicesAndProxyConfig.Properties.object_store_swift_devices) is incorrect.10:41
dasmRecoverable error: ERROR: The specified reference "ObjectStorage" (in SwiftDevicesAndProxyConfig.Properties.object_store_swift_devices) is incorrect.10:41
gfidenteso you see this as output from heat after the stack-create command?10:42
dasmthis one is triggered in tuskar-ui: ERROR:tuskar_ui.infrastructure.overview.forms:ERROR: The specified reference "ObjectStorage" (in SwiftDevicesAndProxyConfig.Properties.object_store_swift_devices) is incorrect.10:43
dasmafter api.heat.Stack.create()10:43
dasmit cannot find stack, so tries to create one.10:44
gfidentethis is correct10:44
gfidentebut I suspect something in the tuskar templates could be broken10:44
gfidentehave you tried deploying with --no-mergepy ?10:44
gfidenteok so that will spawn 1 controller and 2 computes by default10:45
gfidenteI think it's worth trying it to figure if the error in the templates provided by tuskar10:45
gfidenteyou can customize the number of the different node types thougj10:45
dasmi'm playing with tripleo something like 2 months, but still there are many places where i don't know what's happening :)10:46
dasmthanks for tips, i'll try with it.10:46
dasmif this won't work, i'll try to change templates to "old" commit and retry with it10:47
*** spredzy is now known as spredzy|afk10:54
*** mmagr|afk is now known as mmagr11:00
*** yog__ has quit IRC11:12
*** pradeep1 has quit IRC11:12
*** jprovazn has quit IRC11:16
*** morazi has quit IRC11:17
*** jprovazn has joined #tripleo11:19
*** tzumainn has joined #tripleo11:27
* gfidente back soon11:28
*** gfidente has quit IRC11:28
*** yog__ has joined #tripleo11:28
*** dprince has joined #tripleo11:29
*** rhallisey has joined #tripleo11:35
dprincejistr, I tried with F21 locally several times yesterday evening and each time it works for me11:47
dprincejistr: sometimes I see the CI for my patch succeed, but other times it seems like MariaDB fails to startup, or some random DIB package download failures11:48
*** otoolee has quit IRC11:58
*** gfidente has joined #tripleo11:59
*** gfidente has joined #tripleo11:59
*** lucasagomes is now known as lucas-hungry12:00
*** otoolee has joined #tripleo12:02
*** mmagr is now known as mmagr|afk12:04
*** morazi has joined #tripleo12:06
*** pradeep1 has joined #tripleo12:06
*** mmagr|afk is now known as mmagr12:11
*** spredzy|afk is now known as spredzy12:12
jistrdprince: yeah i hit a DIB package download issue just a while ago, and a mariadb startup problem yesterday too12:13
jistrdprince: the test results on the CI fix should be there any minute i think, 3 jobs succeeded, one still pending12:15
*** masco has quit IRC12:15
jistrdprince: yup, all green, i'll merge it12:18
openstackgerritMerged openstack-infra/tripleo-ci: Switch Delorean jobs to F21
jistrso i think we can abandon this
dprincejistr: got it. I just abandoned that one12:20
dprincejistr: Going with F21 is the right decision but I'm still thinking we may have some intermittent failures due to this MariaDB startup issue12:21
dprincejistr: I'll try to hit it again locally today. I'd like to have a closer look12:22
*** thrash|g0ne is now known as thrash12:23
*** trown|outttypeww is now known as trown12:23
*** weshay has joined #tripleo12:25
*** pradeep1 has quit IRC12:25
gfidentedprince, the mariadb issue is where it still attempts to bind on ?12:27
dprincegfidente: I'm not actually sure, but it could be.12:27
dprincegfidente: are you thinking this is perhaps related to our HA proxy config or something?12:28
jistrgfidente: ever hit an issue that CephStorageNodesPostDeployment hangs? everything in my overcloud stack is CREATE_COMPLETE except that one, which is still CREATE_IN_PROGRESS, and it doesn't look like puppet ran on the node12:28
dprincegfidente: the failure seems intermittent though12:28
gfidentedprince, doesn't look like it is related to haproxy, in the mariadb logs instead I can see it failing as is unavailable which is indeed in-use by haproxy12:29
gfidentedprince, but it really is mariadb which shouldn't as we pass to it bind-address to be controller_host12:29
gfidentejistr, you mean that hangs when not cephstorage nodes are supposedly going to be deployed?12:29
gfidentedprince, have a link to a failed job?12:30
jistrgfidente: nope i'm deploying one ceph node (is it ok to deploy just one?). It got created and i can ssh in and os-cloud-config ran there but there were some errors. I'm looking for some details.12:31
gfidentejistr, it's okay to deploy one yes ... yeah if that fails there are other resources depending on it, so creation should fail (not stuck though, just fail)12:32
*** lazy_prince has quit IRC12:32
jistrgfidente: i guess that heat didn't get contacted yet that there was some problem, so it's still CREATE_IN_PROGRESS12:33
gfidentejistr, I just did that with instack though; what is that failed there?12:33
gfidentejistr, indeed if it stays in IN_PROGRESS it will remain stuck12:33
jistri'll post logs12:33
dprincejistr: re, the hanging PostNodesDeployment I've seen it stuck as well12:33
dprincejistr: I was wondering if it just needed more time, or perhaps the signal back to Heat API failed12:34
gfidentedprince, stuck with the ceph node in IN_PROGRESS12:34
dprincegfidente: oh, no. Not Ceph nodes. I've seen the controller hang though12:34
gfidenteso I've seen it hanging when nodes freezed12:35
gfidentebut I never got puppet to succeed or fail without the signal being captured as such12:35
jistrgfidente: here's the log
dprincegfidente: so perhaps MariaDB is getting started before our bind IP gets written out?12:37
gfidentedprince, yeah it looks like12:37
gfidentebut a quick check in the manifest didn't point out anything obvious12:37
gfidentejistr, I don't see errors in there12:38
gfidentecan you actuall post output of 'ceph -s' ?12:38
dprincegfidente: it could be some odd change in the MySQL puppet modules too, although they are generally quite stable12:38
jistrdub 08 11:35:25 ov-6c5aop5jl2-0-bsyhxshbmgm6-cephstorage-iui3wmg2np7n os-collect-config[765]: 2015-04-08 11:35:25.376 765 WARNING os_collect_config.heat [-] No auth_url configured.12:38
jistrdub 08 11:35:25 ov-6c5aop5jl2-0-bsyhxshbmgm6-cephstorage-iui3wmg2np7n os-collect-config[765]: 2015-04-08 11:35:25.377 765 WARNING os_collect_config.request [-] No metadata_url configured.12:38
jistrdub 08 11:35:25 ov-6c5aop5jl2-0-bsyhxshbmgm6-cephstorage-iui3wmg2np7n os-collect-config[765]: 2015-04-08 11:35:25.378 765 WARNING os-collect-config [-] Source [request] Unavailable.12:38
jistrdub 08 11:35:25 ov-6c5aop5jl2-0-bsyhxshbmgm6-cephstorage-iui3wmg2np7n os-collect-config[765]: 2015-04-08 11:35:25.379 765 WARNING os_collect_config.local [-] /var/lib/os-collect-config/local-data not found. Skipping12:38
jistrdub 08 11:35:25 ov-6c5aop5jl2-0-bsyhxshbmgm6-cephstorage-iui3wmg2np7n os-collect-config[765]: 2015-04-08 11:35:25.379 765 WARNING os_collect_config.local [-] No local metadata found (['/var/lib/os-collect-config/local-data'])12:38
jistralso [root@ov-6c5aop5jl2-0-bsyhxshbmgm6-cephstorage-iui3wmg2np7n ~]# ceph -s12:39
gfidentejistr, those aren't errors, it's just the periodic polling from os-collect-config12:39
jistr-bash: ceph: command not found12:39
gfidentejistr, ^^ that is not good :P12:39
jistryeah i wonder if i exported all that needs to be exported etc. :)12:40
gfidentejistr, this change
gfidenteadded into docs the needed vars to build the cephstorage image properly12:41
gfidentebut it's not merged yet and we haven't moved to single image yet either ...12:41
*** prad has quit IRC12:41
gfidente... maybe those are the vars you missed?12:42
gfidenteit's still unclear though how could puppet succeed when ceph was not even installed :)12:42
gfidenteand why heat sees it as stuck IN_PROGRESS12:42
jistrgfidente: yah... forgot the dib vars :)12:42
jistrgfidente: sorry and thanks for the pointer there :)12:43
*** rlandy has joined #tripleo12:43
gfidentejistr, thanks for trying it out actually12:43
gfidenteare you testing your change for libvirt_rbd_secret_key ?12:43
gfidentecool, back soon so we can do together12:43
mark-osping jtomasek ttrifonov12:46
jtomasekmarios: pong12:47
*** athomas has quit IRC12:50
jtomasekmark-os: pong (marios sorry:))12:51
mark-osjtomasek: quick question regarding tuskar / tuskar ui, do I have to have a keystone service for it to work?12:52
mark-osjtomasek: quick question regarding tuskar / tuskar ui, do I have to have a keystone service for it to work?12:52
mark-osdouble post sorry12:52
jtomasekmark-os: yes, I think so. Maybe you can run tuskar-ui without it, but I think Tuskar api requires it12:53
mark-osjtomasek: So I wont be able to access without it?12:53
jtomasekmark-os: I have never tried it, do you get an error when you try to access it?12:54
jtomasekmark-os: what did you use to set openstack (undercloud) up?12:54
mark-osjtomasek: Yes. I received the login screen, logged in and then was given an error.  ServiceCatalogException: Invalid service catalog service: management12:55
mark-osjtomasek: Been setting it up through zypper packages on SLES12:55
mariosmark-os: for the tuskar client, you need keystone afaik. if you are just running tuskar locally isolated, you can get to the api with sthing like curl but is a pita12:55
mariosjtomasek: pong12:55
mark-osmarios: I'm guessing it follows a similar process to the way other services were added to Keystone? If so, it should hopefully be straight forward for me to set up12:56
jtomasekmarios: sorry, I miss-pinged you12:56
jtomasekmark-os: I am not very familiar with the setup as I always used instack to setup undercloud and run UI.12:57
*** athomas has joined #tripleo12:57
mark-osjtomasek: Could you send me the results of $ keystone service-list ; keystone endpoint-list12:58
jistrcan i somehow re-run the puppet apply on an already deployed node? i'd like to run the puppet in debug + verbose mode to see exactly what's happening12:59
mark-osjtomasek: would help me setting up Tuskar UI12:59
jtomasekmark-os: yes, I am at the meeting at the moment. will send it asap12:59
mariosmark-os: yeah, i believe it is registered with keystone 'normally'13:01
mark-osExcellent, thanks marios. Looks like I'm just missing the tuskar service in keystone. jtomasek I don't think I'll need you to post yours as well but thanks for offering to help.13:03
mark-osI'll have a go at setting this up, fingers crossed it should work.13:04
*** lucas-hungry is now known as lucasagomes13:05
*** dasm is now known as dasm|afk13:10
*** julim has joined #tripleo13:11
mark-osmarios: Any reason why you used 8585 for your ports for the tuskar endpoints?13:13
*** noslzzp has quit IRC13:15
*** noslzzp has joined #tripleo13:17
*** yamahata has joined #tripleo13:17
mariosmark-os: who knows, various people have worked on this at various times, i can't remember that particular discussion13:27
mark-osmarios: okay thanks, only ask because it seems like some of the services wont work unless they're a certain port number13:30
*** masco has joined #tripleo13:31
spredzydprince, when you tried the mongodb patch on f21. Did the f21 image had the mongodb package installed ?13:37
spredzyI don't see any limitation on the puppet code that would cause the issue you had13:37
dprincespredzy: I think so. Perhaps I made a silly mistake? The reasons I pulled it locally was I couldn't tell why the CI job failed (if I recall)13:39
dprincespredzy: regardless, if we get a passing CI job I've got no objection13:39
*** Marga_ has joined #tripleo13:39
spredzydprince, ack but I'd be curious to know what happend13:40
*** Marga_ has quit IRC13:40
*** Marga_ has joined #tripleo13:40
*** radez_g0n3 is now known as radez13:40
*** sdake has joined #tripleo13:44
*** lblanchard has joined #tripleo13:46
openstackgerritMartin Mágr proposed openstack/tripleo-heat-templates: Keystone domain for Heat
*** jcoufal has quit IRC13:53
*** sdake_ has joined #tripleo13:53
*** jistr has quit IRC13:54
*** jistr has joined #tripleo13:54
*** yamahata has quit IRC13:55
*** yamahata has joined #tripleo13:55
*** sdake has quit IRC13:58
*** chlong has quit IRC13:58
*** bvandenh has quit IRC13:59
*** pdb has quit IRC13:59
*** pdb has joined #tripleo13:59
greghaynesdprince: Hey, talked with mordred and were all agreed that simple-init should get moved out of DIB14:01
greghaynesmordred: you ok with it getting put on stackforge?14:04
dprincegreghaynes: sounds good to me :) thanks14:05
*** killer_prince has joined #tripleo14:05
*** killer_prince is now known as lazy_prince14:05
*** adrianopetrich has quit IRC14:09
*** pdb has quit IRC14:09
*** pdb has joined #tripleo14:09
*** jistr has quit IRC14:10
*** jistr has joined #tripleo14:12
mark-oscurrently having some trouble, getting a connection error. ConnectionError: ('Connection aborted.', error(111, 'Connection refused'))14:12
mark-osmarios jtomasek have one of you seen this error?14:13
*** pdb has quit IRC14:13
mark-osI'm guessing it's to do with keystone, possibly some port issue where the port I selected isn't currently open14:13
openstackgerritDmitry Teselkin proposed openstack/diskimage-builder: Allow dib to be lauched from venv
jtomasekmark-os: I was getting something similar, I think it was caused by tuskarclient port, try to disable iptables14:14
*** adrianopetrich has joined #tripleo14:15
*** pdb has joined #tripleo14:16
mark-osjtomasek: How would I go about doing that?14:17
jtomasekmark-os: "systemctl stop iptables" on the undercloud machine (the one that runs horizon/tuskar-ui)14:17
mark-osjtomasek: that didn't work14:19
mark-osjtomasek: tools/ ./ runserver , should that not be 8585 instead of 8080 if I used 8585 for my tuskar endpoint?14:19
jtomasekmark-os: no, that's ok14:20
mark-osjtomasek: Failed to stop iptables.service: Unit iptables.service not loaded.14:20
*** spzala has joined #tripleo14:21
jtomasekhmm, tzumainn do you have an idea? I think it is the same problem we were bumping into^14:21
jtomasekmark-os: what os are you on?14:21
mark-osjtomasek: SLES 1214:21
jtomasekmark-os: there might be some other firewall blocking the port, I am not familiar with SLES14:23
jtomasekmark-os: maybe?
greghaynesI have a couple nobrainer DIB bugfixes that have been waiting for review for a while - and
mark-osjtomasek: tried that, was unsuccessful14:24
mark-osjtomasek: thanks for the suggestion14:24
*** pradeep has joined #tripleo14:24
mark-osjtomasek: I think I had this issue with manila whenever I used a port which I shouldn't have14:25
tzumainnjtomasek, mark-os, hm, the issue we originally ran into should have been fixed by , but given that disabling the firewall still doesn't solve the problem, it sounds like this could be something different?14:25
mark-osjtomasek: I think it was listening to the wrong port and so I had to close the wrong port and open the correct one. Though my memory isn't the best and I didn't document the process but it seems like a similar error.14:26
mark-ostzumainn: Possibly. I think it's just the same error as I was having previously with another OpenStack component14:27
*** prad has joined #tripleo14:28
*** mmagr has quit IRC14:30
*** adrianopetrich has quit IRC14:36
*** adrianopetrich has joined #tripleo14:36
*** mcornea has quit IRC14:39
openstackgerritgreghaynes proposed openstack/diskimage-builder: Add a yum-minimal element that just uses yum
*** mmagr has joined #tripleo14:43
*** jistr has quit IRC14:45
*** rhallisey has quit IRC14:48
mark-osjtomasek: is there a command to start tuskar itself?14:49
jtomasekmark-os: no, you just run horizon with infrastructure (tuskar-ui) dashboard enabled
*** jistr has joined #tripleo14:51
mark-osjtomasek: I've done all that, don't understand why it's not working14:53
*** gchamoul has quit IRC14:53
jtomasekmark-os: still the same problem?14:54
mark-osjtomasek: Yeah, unfortunately14:55
mark-osjtomasek: This is what I found out about the error - Generally speaking, when you get Errno 111, it most likely mean that the service you're trying to reach is not running on the port you try to connected to, therefore it most likely mean that the service isn't running at all.14:55
*** untriaged-bot has joined #tripleo15:00
untriaged-botNo untriaged bugs so far! \o/15:00
*** untriaged-bot has quit IRC15:00
*** mmagr has quit IRC15:02
jtomasekmark-os: hmm, I am not quite familiar with how the setup is done, because I use Instack that installs/runs tuskar api for me. You might be right that it might need to get started. jistr could you please help here?15:02
*** rhallisey has joined #tripleo15:05
openstackgerritEmilien Macchi proposed openstack/tripleo-heat-templates: overcloud stepped deployment environment
jistrmark-os, jtomasek: hi, what's the issue and what did you use to install? i've been tweaking my laptop config lately and disconnecting from the env15:06
*** Marga_ has quit IRC15:06
jistrs/from the env/from irc/15:06
*** sseago has left #tripleo15:07
mark-osjistr:  Installed OpenStack using zypper packages15:09
*** jcoufal has joined #tripleo15:11
*** masco has quit IRC15:15
jistrmark-os: ok from the bits i can see back, you have an issue connecting to Tuskar? did you try connecting via CLI, running tuskar <someting> command or how do you connect? does it respond for cURL?15:16
*** jistr is now known as jistr|mtg15:16
shardyAnyone seeing issues with rabbit running via devtest lately?15:18
mark-osjistr|mtg: I can't get this command to work15:19
mark-os$ tuskar-api --config-file etc/tuskar/tuskar.conf15:19
gfidenteshardy, is that with single or multiple controllers?15:19
jistr|mtgmark-os: what's the error it prints? can you paste please?15:20
shardygfidente: single controller, but heat-engine isn't starting because it can't connect to amqp15:21
shardyMessageDeliveryFailure: Unable to connect to AMQP server on after None tries: Exchange.declare: (404) NOT_FOUND - no exchange 'reply_6a54dc328d8c472fab9219d680622aa7' in vhost '/'15:21
gfidente127.0.0.1 seems wrong15:21
*** mrunge is now known as mrunge_away15:22
*** lazy_prince has quit IRC15:24
gfidenteshardy, what is "hiera heat::rabbit_hosts" giving?15:24
gfidente(when run from within the controller)15:26
*** yog__ has quit IRC15:26
shardygfidente: Oh, no this is failing on the seed15:27
gfidenteshardy, I don't redo seed from more than a month :(15:28
*** maufart has quit IRC15:29
*** killer_prince has joined #tripleo15:29
*** killer_prince is now known as lazy_prince15:29
shardygfidente: I had to, to work around bug #144117015:30
openstackbug 1441170 in python-ironicclient "ironicclient "Not Acceptable" error" [High,In progress] - Assigned to Devananda van der Veen (devananda)15:30
shardyor my ironic was too old15:30
shardydoes anyone have another delorean pin, other than current which is known to work?15:30
*** dprince has quit IRC15:30
*** jistr|mtg is now known as jistr15:30
*** dprince has joined #tripleo15:30
*** stendulker has joined #tripleo15:30
*** stendulker_ has joined #tripleo15:35
*** lazy_prince has quit IRC15:36
*** rwsu has joined #tripleo15:37
*** stendulker has quit IRC15:38
*** stendulker has joined #tripleo15:39
devanandashardy: i've got a fix for ^ going into python-ironicclient and will cut another release as soon as it lands15:39
mark-osjistr: will send you what is printed when I use that command15:41
*** stendulker_ has quit IRC15:41
jistrthat should be fine, doesn't look like an error per se. Does it keep running or does it crash when you run tuskar-api like this?15:43
jistrmark-os: can you "curl http://localhost:8585/" ?15:44
jistr(that curl is supposed to be run on the same host as tuskar-api)15:44
*** stendulker_ has joined #tripleo15:44
*** alop has joined #tripleo15:45
*** stendulker has quit IRC15:47
*** rhallisey has quit IRC15:48
*** rhallisey has joined #tripleo15:49
gfidenteshardy, so and / seems right to me in seed though15:49
gfidenteshardy, missing exchage sounds weird to me, are the other services happy with messaging?15:49
*** sdake has joined #tripleo15:50
mark-osjistr: yep can curl it15:50
mark-osjistr: here's the output when I curl it:
mark-osjistr: think it's just curling a redirection page15:51
*** spzala has quit IRC15:52
*** regebro has quit IRC15:52
jistryeah that looks fine, it should redirect to a /v2/ or something. if you run curl with -i, it should also print the headers and show you where the redirect goes15:53
*** sdake_ has quit IRC15:53
jistri need to go (i'm based in europe), but i'll be around tomorrow too15:55
*** killer_prince has joined #tripleo15:55
*** killer_prince is now known as lazy_prince15:55
mark-osjistr: Okay will speak tomorrow!15:55
*** jistr has quit IRC15:56
mark-osjistr: thanks for your help15:56
*** Marga_ has joined #tripleo15:57
*** yamahata has quit IRC15:58
*** yamahata has joined #tripleo15:58
*** Marga_ has quit IRC16:00
*** Marga_ has joined #tripleo16:00
openstackgerritMarios Andreou proposed openstack/tuskar: Wires up constraints and type into returned plan parameters
*** pradeep1 has joined #tripleo16:12
devanandaare ya'll going to be cutting a stable/kilo branch of diskimage-builder?16:12
*** pradeep has quit IRC16:12
*** julim has quit IRC16:14
greghaynesdevananda: Is there a need for one? I think dib is pretty good about not being backwards incompat16:15
greghaynesand AFAIK we dont have any plans to change that16:16
devanandagreghaynes: unfortunately some folks introduced non-backwards-compat things into the  ramdisk image builder recently ...16:16
devanandagreghaynes: new features like whole-disk-image support & creating a local bootloader16:16
devanandagreghaynes: when using a new ramdisk with old ironic, or new ironic with old ramdisk, .... it's bad :(16:16
openstackLaunchpad bug 1441556 in Ironic "Deploying whole disk image with local boot fails" [Medium,Triaged] - Assigned to Lucas Alvares Gomes (lucasagomes)16:17
openstackLaunchpad bug 1422723 in Ironic "Ironic sets instance as active even if bootloader installation fails when boot_option is 'local'" [Critical,Fix committed] - Assigned to Ramakrishnan G (rameshg87)16:17
greghayneshrm, that seems like something we could have / should have avoided16:17
devanandai'm not pleased about how these were added16:17
devanandathe changes came quite late in the cycle for us, and I wasn't able to review them all16:18
*** noslzzp has quit IRC16:18
devanandai would like to tag RC1 of Ironic today, but right now this feature is broken16:19
devanandai've started writing notes on what it will take to fix it here:
greghaynesdo you know what elements youall use when you build?16:19
devanandaelements/deploy-ironic, I belive16:20
devanandayah. the issue is in elements/deploy-ironic/init.d/80-deploy-ironic16:21
devananda160 # If localboot is set, install a bootloader16:21
devananda161 if [ "$IRONIC_BOOT_OPTION" = "local" ]; then16:21
devanandathat block16:21
devanandaIf I were to use a newer version of ironic's PXE driver with an older ramdisk (eg, that doesn't have this block), and i enable local_boot in ironic, the deploy will timeout.16:22
devanandaand if I also pass in a whole disk image, even with a newer ironic and a ramdisk built by the latest DIB/elements/deploy-ironic, it will fail today, because it will overwrite the bootloader that came with the whole disk image16:23
devanandalucasagomes: is ^ an accurate summary of hte issues?16:23
devanandagreghaynes: yeah. that introduced the non-backwards-compatible situation16:24
* lucasagomes reads16:24
devanandathe commit message REALLY should have said something about that :(16:24
lucasagomesdevananda, right, there's 2 things... that block was added recently but the is_whole_disk_image + local boot has been broken before16:25
devanandawell, let me rephrase that slightly. that change itself shouldn't be backwards incompatible, because an older ironic would'nt pass that option, so wouldn't trigger that code.16:26
lucasagomesbecause before the DIB ramdisk would just try to install the bootloader if "boot_option = local" was set16:26
devanandabut the corresponding change in ironic (required to use localboot) doesn't work with older ramdisks16:26
lucasagomesbut it wouldn't try to talk to the ironic api again (which is what causes the deploy to timeout)16:26
lucasagomesyeah, it wouldn't work with old ramdisks16:27
devanandaright. so stable/kilo of Ironic needs to pin its dependency on diskimage-builder to a version that has this code in it16:27
devananda+ what ever code lucasagomes writes to fix
openstackLaunchpad bug 1441556 in Ironic "Deploying whole disk image with local boot fails" [Medium,Triaged] - Assigned to Lucas Alvares Gomes (lucasagomes)16:27
*** stendulker has joined #tripleo16:28
lucasagomesdevananda, +116:28
* greghaynes is still wrapping his head around all the moving parts here16:28
lucasagomesI mean, we can mention in our docs that local boot was introduced in Kilo, and that requires a new ramdisk16:28
lucasagomessame for whole disk image16:28
lucasagomesand maybe avoid pinning?16:29
devanandasure. we need to do that too, for deployers.16:29
greghaynesyea, im really trying to find out if theres a possibe way to do this forward compatibly for the dib element16:30
*** stendulker_ has quit IRC16:30
devanandathe pinning is needed largely for the test environments, so that our stable/kilo jobs do not inadvertently pull in some future change like this one that may break it.16:30
devanandaI think our stable/juno jobs should be fine -- the stable/juno ironic won't pass "localboot" option, so they won't trigger this code in a newer ramdisk16:31
greghayneshrm, so one of the big issues is the callback at the end only works with certain ironic versions, right?16:34
greghaynescan that just version check?16:34
devanandagreghaynes: it could, though the pxe driver's support for localboot isn't an API-visible change, so there's no version for that specifically16:37
devananda*isn't visible in the standard API. it's a driver extension ...16:37
devanandalemme check. that may be discoverable too16:37
*** stendulker has quit IRC16:37
*** Marga_ has quit IRC16:42
*** Marga_ has joined #tripleo16:43
devanandagreghaynes: cheers16:46
*** Marga_ has quit IRC16:46
devanandalucasagomes: curl /v1/nodes/NNN/vendor_psasthru/methods16:46
devanandalucasagomes: if "pass_bootloader_install_info" in result ...16:46
*** Marga_ has joined #tripleo16:47
lucasagomesdevananda, oh that would fix the backward problem16:47
devanandaright :)16:47
lucasagomesthat would be good16:47
lucasagomesdevananda, can we cut the RC on friday maybe?16:47
lucasagomesor it's required to cut it today?16:48
*** radez is now known as radez_g0n316:50
*** julim has joined #tripleo16:54
*** Marga_ has quit IRC16:56
*** Marga_ has joined #tripleo16:57
openstackgerritgreghaynes proposed openstack/diskimage-builder: Only send ironic bootloader info if its supported
greghaynesdevananda: lucasagomes ^ theres an untested fix16:58
greghayneswhich is basically just me writing out your psuedocode ;)16:59
lucasagomesnice, yeah gotta give it a go17:00
*** spzala has joined #tripleo17:01
*** spredzy is now known as spredzy|afk17:03
devanandagreghaynes: cheers17:05
*** mrunge_away is now known as mrunge17:06
*** athomas has quit IRC17:07
*** trown is now known as trown|lunch17:11
*** yamahata has quit IRC17:16
openstackgerritJohn Stafford proposed openstack/diskimage-builder: Only send ironic bootloader info if its supported
*** yamahata has joined #tripleo17:18
*** jdob has quit IRC17:21
*** dtantsur is now known as dtantsur|afk17:22
*** pelix has quit IRC17:24
*** Guest39926 is now known as mgagne17:29
*** mgagne has joined #tripleo17:29
*** jcoufal has quit IRC17:31
*** yamahata has quit IRC17:39
*** thrash is now known as thrash|biab17:50
*** ukalifon1 has quit IRC17:57
*** yamahata has joined #tripleo17:59
*** spzala has quit IRC18:22
*** lucasagomes is now known as lucas-dinner18:26
EmilienMshardy, dprince: for serverspec, I'm thinking about creating an hook in
EmilienMso we can re-use config inputs in serverspec configuration and run tests locally (probably better for large-scale)18:28
*** weshay is now known as weshay_mtg18:29
EmilienMmaybe an elements like "heat-valid-serverspec" in software-config?18:30
*** trown|lunch is now known as trown18:31
EmilienMdprince: ^18:31
dprinceEmilienM: I'm not sure I follow18:32
dprinceEmilienM: can't you just re-use local data that is written in Hiera already?18:33
EmilienMdprince: yes I could18:33
EmilienMand that's what we should do18:33
dprinceEmilienM: or are you saying you prefer the non-hiera (JSON format) data instead18:33
EmilienMno I did not explain correctly what I meant18:34
dprinceEmilienM: if hiera is acceptable, and you want to run something locally then "plugging in" via some sort of generic Heat software deployment might work better18:34
EmilienMI'm looking at creating an interface for service validation18:34
EmilienMdprince: and I wonder which code would run `serverspec`18:35
EmilienMdprince: and also to have a sync : roles/puppet-classes <-> serverspec tests18:36
*** jcoufal has joined #tripleo18:39
*** jdob has joined #tripleo18:46
*** jcoufal has quit IRC18:47
dprinceEmilienM: Hmmm. so your requirement is you need access to the same data as puppet, and you want to run after each puppet apply task?18:47
EmilienMdprince: yes for #1, not required for #218:48
EmilienMyou can run validation after puppet ran everywhere18:48
dprinceEmilienM: okay, so one big validation, instead of several smaller validations along the way?18:49
EmilienMdprince: in my SpinalStack we use to have an association between puppet classes & serverspec tests. So each end of step, we ran serverspec and if not-success, run the step again18:51
dprinceEmilienM: a new hook in would work. And it could give you access to puppet data via hiera. I'm not sure that would be the best way to get at the 'puppet' Heat data directly however since serverspec isn't in the 'puppet' data group.18:51
*** pradeep1 has quit IRC18:52
dprinceEmilienM: okay, so the 'run again' thing here might indicate where the thing driving these validations should live18:52
openstackgerritVictor Lowther proposed openstack/os-cloud-config: Add pxe_drac driver support to os-cloud-config
dprinceEmilienM: I think the best way to have a Heat task "run again" might be to do a stack update if the previous one failed.18:53
dprinceEmilienM: so, each resource could have its own SoftwareDeployment for validation, which if it fails would get detected by heat18:54
dprinceEmilienM: this validation SoftwareDeployment would run after the configuration deployments (puppet, etc.) have finished. We would control that w/ depends_on18:54
dprinceEmilienM: if it fails, you could heat-stack update and it would get another shot...18:55
EmilienMI like the idea18:55
dprinceSo an element to run serverspec would be fine. It could live in heat-templates, sure.18:56
EmilienMdprince: well, i'll start something tonight/night and I'll let you know18:57
*** trown is now known as trown|taxman18:57
dprinceEmilienM: good to run this past shardy too18:57
openstackgerritOpenStack Proposal Bot proposed openstack/diskimage-builder: Updated from global requirements
EmilienMdprince: ack18:58
*** lblanchard has quit IRC19:01
*** lblanchard has joined #tripleo19:03
*** spzala has joined #tripleo19:03
*** Marga_ has quit IRC19:12
*** mmagr has joined #tripleo19:16
*** mmagr has quit IRC19:20
*** akrivoka has quit IRC19:21
*** thrash|biab is now known as thrash19:21
openstackgerritEmilien Macchi proposed openstack/tripleo-puppet-elements: Implement serverspec element
EmilienMdprince: ^19:31
*** jprovazn has quit IRC19:34
dprinceEmilienM: commented19:47
*** trown|taxman is now known as trown19:50
*** sdake_ has joined #tripleo19:55
*** sdake has quit IRC19:58
openstackgerritEmilien Macchi proposed openstack/tripleo-puppet-elements: Implement serverspec element
*** Marga_ has joined #tripleo20:11
*** prad has quit IRC20:19
openstackgerritBen Nemec proposed openstack/diskimage-builder: Make troubleshoot work with dracut ramdisks
openstackgerritBen Nemec proposed openstack/diskimage-builder: Clean up targetcli ramdisk installation
*** sdake_ has quit IRC20:28
*** prad has joined #tripleo20:30
shardyEmilienM, dprince: sounds good, I agree the way to recover from a failed local validation should be a stack-update20:37
shardyI quite like dprince's suggestion of a heat-config hook, but an element works too, I don't have a strong preference20:38
*** Marga_ has quit IRC20:51
*** Marga_ has joined #tripleo20:53
*** Marga_ has quit IRC20:54
*** Marga_ has joined #tripleo20:54
*** jdob has quit IRC20:55
openstackgerritMerged openstack/os-cloud-config: Add support to skip either kernel or ramdisk
*** lblanchard has quit IRC20:57
*** untriaged-bot has joined #tripleo21:00
untriaged-botUntriaged bugs so far:21:00
openstackLaunchpad bug 1441556 in Ironic "Deploying whole disk image with local boot fails" [Medium,In progress] - Assigned to Jim Rollenhagen (jim-rollenhagen)21:00
*** untriaged-bot has quit IRC21:00
*** cwolferh has quit IRC21:05
*** mrunge has quit IRC21:10
openstackgerritMerged openstack/diskimage-builder: Export image properties
*** pblaho has quit IRC21:27
openstackgerritMerged openstack/diskimage-builder: Correctly handle raw type ordering
*** jtomasek has quit IRC21:29
*** dprince has quit IRC21:32
openstackgerritRyan Hallisey proposed stackforge/kolla: Update error message for build-docker-images
*** rhallisey has quit IRC21:34
*** cwolferh has joined #tripleo21:45
*** cwolferh has quit IRC21:50
*** weshay_mtg has quit IRC21:58
*** trown is now known as trown|outttypeww21:59
*** chlong has joined #tripleo22:01
*** cwolferh has joined #tripleo22:01
*** morazi has quit IRC22:02
openstackgerritGiulio Fidente proposed openstack/tripleo-heat-templates: puppet: install Horizon on overcloud-controller
*** gchamoul has joined #tripleo22:20
openstackgerritGiulio Fidente proposed openstack/tripleo-heat-templates: Make memcached bind on ctlplane instead of
*** gchamoul- has joined #tripleo22:24
*** gchamoul has quit IRC22:25
openstackgerritGiulio Fidente proposed openstack/tripleo-heat-templates: Make rabbitmq bind on ctlplane instead of
*** gfidente has quit IRC22:27
*** prad has quit IRC22:27
*** gchamoul has joined #tripleo22:29
*** gchamoul- has quit IRC22:30
openstackgerritEmilien Macchi proposed openstack/tripleo-puppet-elements: Implement serverspec elements
*** Marga_ has quit IRC22:48
*** Marga_ has joined #tripleo22:50
*** lucas-dinner has quit IRC22:56
*** stendulker has joined #tripleo22:56
greghaynesmordred: SpamapS hah! so the vhd-util is corrupting the fs on our .raw's22:57
greghaynesthat was one gnarly bug to track down22:58
*** cwolferh has quit IRC22:58
*** cwolferh has joined #tripleo22:58
SpamapSgreghaynes: wtf!22:59
greghayneshrm, but we cp it22:59
* greghaynes is now very confused22:59
greghaynesI have tracked this down to getting fs errors when I use -t vhd...22:59
SpamapSgreghaynes: possible the raw is not flushed23:00
SpamapSgreghaynes: also possible qemu-img is being clever and forcing it to sync.23:01
*** rlandy has quit IRC23:02
SpamapSgreghaynes: are you using tmpfs?23:02
SpamapSas in, do you have enough RAM it think it will work?23:03
SpamapSthe raw goes on the tmpfs, so not flushed seems.. unlikely.23:04
SpamapSoh wait23:04
SpamapSit's mounted loop isn't it?23:04
SpamapSso could be that it isn't unmounted yet23:04
greghaynesits mounted loop, we copy it while its mounted23:04
greghaynesI think23:04
greghaynesno, we unmount it23:04
SpamapSdon't we umount -l ?23:05
*** chlong has quit IRC23:05
greghaynesumount -fl23:05
SpamapSthats it23:05
SpamapS       -l     Lazy  unmount. Detach the filesystem from the filesystem hierarchy now, and cleanup all references to23:06
SpamapS              the filesystem as soon as it is not busy anymore.  (Requires kernel 2.4.11 or later.)23:06
SpamapSgreghaynes: just for the hell of it, try plugging a 'sync' in before the vhd-util call23:07
greghaynesok, I have a run to re-verify this thats about to end23:07
greghaynesbut then ill do it23:07
greghaynesI spent way too long assuming the issue had to be a result of our block-device.d scripts23:08
*** stendulker has quit IRC23:09
*** spzala has quit IRC23:16
*** alop has quit IRC23:21
SpamapSFYI, these changes likely all need rechecks:
SpamapSI don't have time to recheck them all23:32
SpamapSbut they were all due to a temporary problem in ironicclient23:33
greghaynesin other news, running your own cloud on ssd's = much faster dev cycle for testing23:36
SpamapShah yes23:38
*** gchamoul has quit IRC23:40
*** sdake has joined #tripleo23:40
*** gchamoul- has joined #tripleo23:40
*** sdake_ has joined #tripleo23:42
*** panda|afk has quit IRC23:45
*** panda|afk has joined #tripleo23:45
*** sdake has quit IRC23:45

Generated by 2.14.0 by Marius Gedminas - find it at!