| *** dsneddon has quit IRC | 00:15 | |
| *** dsneddon has joined #tripleo | 00:15 | |
| *** cwolferh has quit IRC | 00:16 | |
| openstackgerrit | Merged stackforge/kolla: Sets debug logging to false https://review.openstack.org/173072 | 00:24 |
|---|---|---|
| *** cwolferh has joined #tripleo | 00:26 | |
| openstackgerrit | Dan Prince proposed openstack/tripleo-image-elements: Add an element to pin to Qemu 2.1.3-3 on Fedora 21 https://review.openstack.org/172986 | 00:32 |
| *** cwolferh has quit IRC | 00:35 | |
| openstackgerrit | Dan Prince proposed openstack-infra/tripleo-ci: Pin qemu-img on Fedora 21 jobs. https://review.openstack.org/172989 | 00:35 |
| openstackgerrit | Steve Baker proposed openstack/diskimage-builder: Mount with -o nouuid for XFS base images https://review.openstack.org/173138 | 00:39 |
| *** daneyon has joined #tripleo | 01:27 | |
| *** daneyon has quit IRC | 01:31 | |
| *** eghobo has joined #tripleo | 01:37 | |
| *** julim has quit IRC | 01:38 | |
| openstackgerrit | greghaynes proposed openstack/diskimage-builder: Add functional smoke test for disk-image-create https://review.openstack.org/169167 | 01:48 |
| openstackgerrit | greghaynes proposed openstack/diskimage-builder: Fix image size to fit filesystem journal https://review.openstack.org/169166 | 01:50 |
| openstackgerrit | greghaynes proposed openstack/diskimage-builder: Add functional smoke test for disk-image-create https://review.openstack.org/169167 | 01:50 |
| *** wuhg has joined #tripleo | 01:53 | |
| *** jerryz has joined #tripleo | 02:02 | |
| *** eghobo has quit IRC | 02:04 | |
| openstackgerrit | Dan Prince proposed openstack/tripleo-image-elements: An element to force install Qemu 2.1.3-3 on F21 https://review.openstack.org/172986 | 02:07 |
| openstackgerrit | Dan Prince proposed openstack-infra/tripleo-ci: Pin qemu-img on Fedora 21 jobs. https://review.openstack.org/172989 | 02:09 |
| *** cwolferh has joined #tripleo | 02:21 | |
| *** yamahata has quit IRC | 02:23 | |
| *** thrash is now known as thrash|g0ne | 02:26 | |
| *** cwolferh has quit IRC | 02:40 | |
| *** spzala has quit IRC | 02:41 | |
| *** tzumainn has quit IRC | 02:48 | |
| *** eghobo has joined #tripleo | 02:48 | |
| *** regebro has joined #tripleo | 03:01 | |
| *** wuhg has quit IRC | 03:06 | |
| *** aberdine has quit IRC | 03:06 | |
| *** cwolferh has joined #tripleo | 03:08 | |
| *** wuhg has joined #tripleo | 03:13 | |
| *** aberdine has joined #tripleo | 03:13 | |
| *** daneyon has joined #tripleo | 03:16 | |
| *** eghobo has quit IRC | 03:20 | |
| *** daneyon has quit IRC | 03:21 | |
| *** yamahata has joined #tripleo | 03:27 | |
| *** rlandy has joined #tripleo | 03:34 | |
| *** sdake has joined #tripleo | 03:56 | |
| *** tzumainn has joined #tripleo | 03:57 | |
| *** sdake_ has quit IRC | 03:58 | |
| *** eghobo has joined #tripleo | 04:00 | |
| *** rlandy has quit IRC | 04:01 | |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Prepare compose for custom namespace https://review.openstack.org/173044 | 04:01 |
| *** sdake_ has joined #tripleo | 04:02 | |
| *** eghobo_ has joined #tripleo | 04:03 | |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Update developer documentation https://review.openstack.org/172998 | 04:04 |
| *** sdake has quit IRC | 04:04 | |
| *** eghobo has quit IRC | 04:06 | |
| *** tzumainn has quit IRC | 04:29 | |
| openstackgerrit | Martin André proposed stackforge/kolla: Fix validate-all-yaml.sh to check for files with yml extension https://review.openstack.org/173168 | 04:46 |
| *** pradeep has quit IRC | 04:50 | |
| *** jcoufal has joined #tripleo | 04:53 | |
| *** masco has joined #tripleo | 04:57 | |
| *** maufart has joined #tripleo | 04:58 | |
| *** daneyon has joined #tripleo | 05:04 | |
| *** maufart has quit IRC | 05:06 | |
| *** daneyon has quit IRC | 05:09 | |
| *** ishant has joined #tripleo | 05:14 | |
| *** pradeep has joined #tripleo | 05:16 | |
| *** pradeep has quit IRC | 05:22 | |
| *** eghobo_ has quit IRC | 05:36 | |
| *** eghobo has joined #tripleo | 05:38 | |
| *** pradeep has joined #tripleo | 05:40 | |
| *** jtomasek has joined #tripleo | 05:54 | |
| *** jprovazn has joined #tripleo | 05:57 | |
| *** yamahata has quit IRC | 05:58 | |
| *** yamahata has joined #tripleo | 05:59 | |
| *** jprovazn has quit IRC | 06:03 | |
| *** devananda has quit IRC | 06:10 | |
| *** devananda has joined #tripleo | 06:17 | |
| *** spredzy|afk is now known as spredzy | 06:24 | |
| *** Marga_ has quit IRC | 06:33 | |
| *** maufart has joined #tripleo | 06:47 | |
| *** ukalifon has joined #tripleo | 06:50 | |
| *** maufart has quit IRC | 06:53 | |
| *** daneyon has joined #tripleo | 06:53 | |
| *** daneyon has quit IRC | 06:58 | |
| *** jprovazn has joined #tripleo | 07:02 | |
| openstackgerrit | Yanis Guenane proposed openstack/tripleo-heat-templates: puppet: implement MongoDB on controller nodes https://review.openstack.org/165474 | 07:05 |
| *** dmellado has quit IRC | 07:10 | |
| *** eghobo has quit IRC | 07:13 | |
| *** mcornea has joined #tripleo | 07:24 | |
| *** dmellado has joined #tripleo | 07:24 | |
| *** chlong has quit IRC | 07:27 | |
| *** mmagr has joined #tripleo | 07:28 | |
| *** jistr has joined #tripleo | 07:29 | |
| *** yamahata has quit IRC | 07:30 | |
| *** dtantsur|afk is now known as dtantsur | 07:35 | |
| openstackgerrit | Jan Provaznik proposed openstack/tripleo-common: Add basic project structure https://review.openstack.org/172907 | 07:38 |
| *** mmagr is now known as mmagr|afk | 07:38 | |
| *** ifarkas has joined #tripleo | 07:42 | |
| openstackgerrit | Jan Provaznik proposed openstack/tripleo-common: Add basic project structure https://review.openstack.org/172907 | 07:50 |
| *** bvandenh has joined #tripleo | 07:51 | |
| *** mmagr|afk is now known as mmagr | 07:54 | |
| *** yog__ has joined #tripleo | 07:56 | |
| *** mcornea has quit IRC | 07:57 | |
| *** yog__ has quit IRC | 07:57 | |
| *** mcornea has joined #tripleo | 07:58 | |
| *** wuhg has quit IRC | 07:59 | |
| spredzy | jistr, unfortunately, in 6 different run I couldnt reproduce the rabbitmq issue. Could you paste me the puppet log of step 2 about RabbiMQ resources ? | 08:01 |
| jistr | spredzy: atm i don't have it deployed, trying to deploy a potential fix | 08:02 |
| spredzy | jistr, ack. What angle did you take for the fix ? | 08:02 |
| jistr | i found this http://rabbitmq.1065348.n5.nabble.com/Rabbitmq-boot-failure-with-quot-tables-not-present-quot-td24494.html | 08:02 |
| jistr | which simply suggests that we cannot start all nodes at the same time | 08:03 |
| *** gfidente has joined #tripleo | 08:03 | |
| *** gfidente has joined #tripleo | 08:03 | |
| *** pradeep has quit IRC | 08:03 | |
| jistr | i'm trying to add one more step which will deploy RabbitMQ on one node first, and then in a next step it will deploy on the rest | 08:03 |
| spredzy | I though about that yesterday, and suggested the master get deployed in step 2 the other on step 3 | 08:04 |
| spredzy | but it randomly works, so is it just orchestration luck | 08:04 |
| spredzy | like master got enough time to do the job before salves went up | 08:04 |
| spredzy | Because in puppet-openstack-cloud we didn't have this issue, because during step1 (which RabbitMQ is part), the provisioning is done sequentially, not in // | 08:05 |
| openstackgerrit | Yanyan Hu proposed openstack/os-collect-config: Use non-versioned auth_url for keystone https://review.openstack.org/138246 | 08:05 |
| *** jcoufal_ has joined #tripleo | 08:06 | |
| jistr | ah right, i was wondering if you hit that previously in the other composition layer... that explains it | 08:06 |
| openstackgerrit | Jan Provaznik proposed openstack/tripleo-common: Add basic project structure https://review.openstack.org/172907 | 08:06 |
| *** pradeep has joined #tripleo | 08:08 | |
| *** jcoufal has quit IRC | 08:09 | |
| *** derekh has joined #tripleo | 08:10 | |
| gfidente | jistr, spredzy we can't make resources of the same resourcegroup depend on each other | 08:15 |
| gfidente | if we're down to that I'd add I'd rather a random sleep so one node will get there first | 08:16 |
| gfidente | *I | 08:16 |
| *** athomas has joined #tripleo | 08:16 | |
| gfidente | hu, good keyboarding everyone | 08:16 |
| gfidente | I meant I'd rather add | 08:16 |
| * gfidente ♥ | 08:17 | |
| jistr | ok i've just had my second successful deployment using the step split approach... | 08:17 |
| gfidente | where the other nodes come up in the second step? | 08:18 |
| spredzy | jistr, can you show us the diff. | 08:18 |
| jistr | yes | 08:18 |
| spredzy | gfidente, not sure if the sleep will be accepted upstream | 08:18 |
| gfidente | ah that's nice from me, I forgot we could do that :) | 08:18 |
| gfidente | spredzy, don't let me start on the puppet-corosync module ... | 08:18 |
| gfidente | ... it actually prevents you from using udpu on Red Hat forcing to pass params which can't go toggether | 08:19 |
| jistr | http://paste.fedoraproject.org/210663/28999568 | 08:19 |
| gfidente | spredzy, so the sleep is the worst idea possible | 08:20 |
| jistr | i added one more step because i wanted all rabbitmq nodes to be ready before the OpenStack services start to get configured | 08:20 |
| jistr | the issue there is that i cannot do: | 08:21 |
| jistr | step 2 - include rabbit if i'm bootstrap node | 08:21 |
| jistr | step 3 - include rabbit always | 08:21 |
| gfidente | jistr, can we shift one node to step1 ? | 08:21 |
| jistr | because i cannot cut out the rabbit part out for inclusion, we have to put everyting into one file the way it is now | 08:21 |
| *** sdake has joined #tripleo | 08:22 | |
| hewbrocca | I don't guess the rabbit resource agent helps with this bootstrapping stuff at all? | 08:22 |
| jistr | it might | 08:22 |
| spredzy | jistr, I think to keep it shorter gfidente is right. We could change the if with [Step 1 AND bootstrap node] OR [step2+] | 08:25 |
| spredzy | we do nothing in step 2 that requires rabbitmq | 08:25 |
| spredzy | so initiating it in the bootstrap node on step 1 won't hurt | 08:25 |
| *** lucasagomes has joined #tripleo | 08:26 | |
| spredzy | this way the change stay concise, only puppet related | 08:26 |
| *** sdake_ has quit IRC | 08:26 | |
| spredzy | hewbrocca, in the euro timezone ? : | 08:26 |
| spredzy | :) | 08:26 |
| jistr | spredzy, gfidente: do you know why we currently initialize HAproxy in its own step outside everything else? do we need it in step 2? | 08:27 |
| * jistr looks further into the code | 08:27 | |
| hewbrocca | spredzy: I am -- I have moved permanently to Brno :) | 08:28 |
| gfidente | jistr, stuff in step 2 relies on the VIP | 08:29 |
| *** mcornea has quit IRC | 08:29 | |
| *** sdake has quit IRC | 08:30 | |
| gfidente | jistr, I think the ::mysql might have attempted sync | 08:30 |
| spredzy | jistr, yes all the X::db::mysql on step 2 relies on the vip | 08:31 |
| spredzy | hewbrocca, nice, big change though. | 08:31 |
| jistr | gfidente, spredzy: ack, thanks. I've updated the code for step 1/2 instead of 2/3, testing now. | 08:32 |
| hewbrocca | spredzy: yep! | 08:34 |
| *** yog__ has joined #tripleo | 08:34 | |
| *** mcornea has joined #tripleo | 08:40 | |
| *** daneyon has joined #tripleo | 08:42 | |
| *** mmagr is now known as mmagr|afk | 08:43 | |
| *** shardy has joined #tripleo | 08:46 | |
| *** daneyon has quit IRC | 08:47 | |
| *** mmagr|afk is now known as mmagr | 08:53 | |
| openstackgerrit | Jiri Stransky proposed openstack/tripleo-heat-templates: Initialize RabbitMQ on bootstrap node first https://review.openstack.org/173222 | 08:58 |
| jistr | ^^ gfidente spredzy | 08:58 |
| jistr | hmm i think we should get to a green multi-node controller deployment as soon as we can. RabbitMQ will start now but the deployment still fails later on. | 08:59 |
| spredzy | jistr, what the following issue ? | 09:00 |
| *** untriaged-bot has joined #tripleo | 09:00 | |
| untriaged-bot | Untriaged bugs so far: | 09:00 |
| untriaged-bot | https://bugs.launchpad.net/diskimage-builder/+bug/1443706 | 09:00 |
| *** untriaged-bot has quit IRC | 09:00 | |
| openstack | Launchpad bug 1443706 in diskimage-builder "Failed to mount: Filesystem has duplicate UUID" [Undecided,In progress] - Assigned to Steve Baker (steve-stevebaker) | 09:00 |
| jistr | DB syncs | 09:00 |
| jistr | i tried with the Galera patch too but it failed on Ceilometer DB sync too there | 09:00 |
| spredzy | jistr, which one ? | 09:00 |
| *** pradeep has quit IRC | 09:02 | |
| jistr | spredzy: this one https://review.openstack.org/#/c/164983/ and the dependent one in puppet-tripleo. I now didn't deploy with it to get rid of any side effects though, so i don't have the error at hand right now. | 09:02 |
| jistr | i think we should get to a green HA deployment as soon as we can, no matter if we have to remove some features initially (from HA only). Right now when the deployment constantly fails, it's time consuming to verify if the breakage is due to my change or if its "the normal breakage"... | 09:04 |
| jistr | which reminds me of an issue that's even higher priority i'd say -- our non-HA CI is broken again | 09:05 |
| hewbrocca | jistr: +1 | 09:05 |
| spredzy | jistr, dprince pushed a patch to tripleo-ci | 09:05 |
| spredzy | https://review.openstack.org/#/c/172989/ | 09:05 |
| spredzy | it needs to be approved for CI to be back in shape | 09:06 |
| *** dtantsur is now known as dtantsur|brb | 09:06 | |
| openstackgerrit | Derek Higgins proposed openstack/tripleo-image-elements: An element to force install Qemu 2.1.3-3 on F21 https://review.openstack.org/172986 | 09:06 |
| jistr | spredzy: thx for the info | 09:08 |
| openstackgerrit | Derek Higgins proposed openstack-infra/tripleo-ci: Pin qemu-img on Fedora 21 jobs. https://review.openstack.org/172989 | 09:08 |
| jistr | hmm the CI failed on it though | 09:08 |
| spredzy | jistr, for the Galera I'll be interested in knowing the issue you've found. Just re run one and got the stack up and running | 09:08 |
| spredzy | jistr, check derekh last change ^ | 09:09 |
| *** pelix has joined #tripleo | 09:10 | |
| derekh | jistr: ya, the puppet job is currently failing because of a regression in the qemu package, were trying to get CI to use an old version | 09:10 |
| *** Goneri has joined #tripleo | 09:10 | |
| jistr | derekh, spredzy: thanks, i've +2'd both patches, we'll see what Jenkins will post | 09:12 |
| derekh | jistr: k,thanks, the first shouldn't be merged its only there so we can cherrypick it, I'll WIP it now | 09:13 |
| openstackgerrit | Merged openstack/tripleo-image-elements: An element to force install Qemu 2.1.3-3 on F21 https://review.openstack.org/172986 | 09:13 |
| derekh | sh*t, that was the oposite to what I wanted to do | 09:14 |
| jistr | derekh: :D | 09:14 |
| *** bus-104_ has quit IRC | 09:15 | |
| jistr | nvm it won't break anything | 09:15 |
| derekh | nope, we can remove it again later, stupid stupid stupid | 09:15 |
| jistr | yeah we can prepare a revert patch and remove it anytime | 09:15 |
| derekh | still gotta wait on cI to see what happens | 09:16 |
| * derekh goes to wake himself up properly | 09:16 | |
| *** pradeep has joined #tripleo | 09:17 | |
| *** akrivoka has joined #tripleo | 09:26 | |
| spredzy | jistr, derekh where are 'precise' and 'nonha' defined for the CI ? I suspect its on toci_gate_test.sh but I am wondering where is TOCI_JOBTYPE coming from | 09:31 |
| derekh | spredzy: TOCI_JOBTYPE is coming from the infra job definition here http://git.openstack.org/cgit/openstack-infra/project-config/tree/jenkins/jobs/tripleo.yaml | 09:33 |
| derekh | spredzy: the next question would be where does {type} come from, one sec and I'll find it | 09:33 |
| spredzy | derekh, hehe yes indeed | 09:34 |
| derekh | spredzy: that first file is a template for all the tripleo jobs, which is used here | 09:35 |
| derekh | http://git.openstack.org/cgit/openstack-infra/project-config/tree/jenkins/jobs/projects.yaml#n315 | 09:35 |
| *** yuanying has quit IRC | 09:35 | |
| spredzy | derekh, thanks. great. So noramlly adding a f20puppet-ha job is a simple as creating an extra stanza in this file. Since f20puppet and ha are already part of toci_gate_test.sh things will picked up autimatically, right ? | 09:37 |
| spredzy | well apparently there is something else to do since 'ironic-vlan-precise-nonha' is in the file but not enabled, so I should also enable it properly | 09:39 |
| derekh | spredzy: yup, instead of adding an extra entry, I'm thinking we should replace overcloud-precise-ha with overcloud-f20puppet-ha because the precise one is consistently failing and nobody is maintaining it | 09:40 |
| derekh | spredzy: ahh, yes you would also need to define which projects to run the new job against, looking for a link | 09:41 |
| spredzy | ok I ll make the review. I'll specify that we overide precise-ha due to the fact that precise-ha is unmaintained | 09:41 |
| derekh | spredzy: cool, here is the file where you define what projects to run against http://git.openstack.org/cgit/openstack-infra/project-config/tree/zuul/layout.yaml#n2628 | 09:43 |
| * spredzy checking | 09:44 | |
| spredzy | derekh, thanks | 09:45 |
| derekh | spredzy: gfidente was looking into doing the same thing I think, but was waiting until the puppet ha job would pass | 09:45 |
| derekh | spredzy: if its not expected to pass yet, make sure its left as non voting, we can then switch it to voting once its good | 09:45 |
| gfidente | derekh, yeah I was a little concerned that during migration it will not that much useful | 09:45 |
| gfidente | but non-voting is bettet than not having | 09:45 |
| spredzy | derekh, that was my idea, so we can see jistr patch passing and then probably enable it | 09:46 |
| spredzy | currently with jistr patch I have a succesful deployment | 09:46 |
| derekh | spredzy: ack | 09:46 |
| spredzy | with 3 nodes | 09:46 |
| spredzy | controller nodes I meant | 09:46 |
| gfidente | derekh, I was also thinking why we insist on 2 computes | 09:47 |
| gfidente | that seems to me waste of resources, we're not testing the nova scheduler | 09:47 |
| gfidente | I'd drop all jobs to 1 ctrl and 1 compute | 09:47 |
| gfidente | and use those resources to play a bit with the block storage scaling, or ceph nodes | 09:47 |
| gfidente | ceph probably more useful than cinder scaling | 09:48 |
| gfidente | can I attempt that after spredzy switches to puppetha? | 09:48 |
| *** cody-somerville has quit IRC | 09:49 | |
| *** cody-somerville has joined #tripleo | 09:49 | |
| openstackgerrit | Derek Higgins proposed openstack/tripleo-image-elements: Mask yum downgrade qemu errors https://review.openstack.org/173236 | 09:50 |
| openstackgerrit | Derek Higgins proposed openstack-infra/tripleo-ci: Pin qemu-img on Fedora 21 jobs. https://review.openstack.org/172989 | 09:51 |
| derekh | gfidente: sounds ok to me, I'm trying to remember why we originally kicked off 2 compute nodes but can't think of why | 09:53 |
| *** lifeless1 has joined #tripleo | 09:54 | |
| *** mandre_ has joined #tripleo | 10:01 | |
| *** cwolferh has quit IRC | 10:02 | |
| *** dsneddon has quit IRC | 10:02 | |
| *** lifeless has quit IRC | 10:02 | |
| *** mandre has quit IRC | 10:02 | |
| openstackgerrit | Derek Higgins proposed openstack-infra/tripleo-ci: Pin qemu-img on Fedora 21 jobs. https://review.openstack.org/172989 | 10:07 |
| *** dtantsur|brb is now known as dtantsur | 10:09 | |
| *** pblaho has joined #tripleo | 10:11 | |
| *** pblaho has quit IRC | 10:12 | |
| *** pblaho has joined #tripleo | 10:12 | |
| gfidente | derekh, https://review.openstack.org/#/c/172989 seems like one of the needed changes merged | 10:13 |
| derekh | gfidente: thanks for pointing that out, I merged it by accident, when I intended to -2 it .... the comment in that patch needs to be updated to reference a different patch, I'll sort it out once I see what ci does | 10:15 |
| gfidente | but the code is using the qemu-pinned element | 10:17 |
| gfidente | you need to remove that too? | 10:17 |
| derekh | gfidente: no we need to use the qemu-pinned element, maybe I'm missunderstanding what your asking, the qemu-pinned element merged but now we are pulling in an update to it https://review.openstack.org/#/c/173236/ | 10:19 |
| openstackgerrit | Giulio Fidente proposed openstack/tripleo-incubator: Switch default COMPUTESCALE setting down to 1 from 2 https://review.openstack.org/173248 | 10:19 |
| gfidente | ah the update is what I missed | 10:19 |
| gfidente | so you wanted to -2 initially to make the update there from the start | 10:20 |
| gfidente | correct? | 10:20 |
| derekh | gfidente: yup, I should never have merged the qemu-pinned element | 10:21 |
| *** daneyon has joined #tripleo | 10:31 | |
| spredzy | https://review.openstack.org/173255 | 10:34 |
| spredzy | derekh, gfidente jistr ^ | 10:34 |
| openstackgerrit | Derek Higgins proposed openstack-infra/tripleo-ci: Pin qemu-img on Fedora 21 jobs. https://review.openstack.org/172989 | 10:34 |
| *** daneyon has quit IRC | 10:36 | |
| derekh | spredzy: looks fairly close to me, comments added | 10:41 |
| gfidente | derekh, looking https://github.com/openstack-infra/tripleo-ci/blob/master/toci_gate_test.sh | 10:44 |
| gfidente | given USE_MERGEPY is set in _variables.sh to 0 I think all jobs are actually running without merge at this point? | 10:44 |
| ukalifon | shardy: Hello. Can I ask you about the stepwise installation which uses the heat breakpoints? I want to know where the breakpoints will be placed on the stack, so I can test the feature similarly to how it will be used. Which tests will run in each phase ? | 10:44 |
| spredzy | derekh, thank you. updated. | 10:45 |
| openstackgerrit | Giulio Fidente proposed openstack/tripleo-puppet-elements: Add pacemaker, pacemaker_remote and pcs packages to images https://review.openstack.org/172565 | 10:46 |
| openstackgerrit | Giulio Fidente proposed openstack/tripleo-heat-templates: Perform basic setup of pacemaker cluster on controllers https://review.openstack.org/172572 | 10:46 |
| derekh | gfidente: on ubuntu we are setting USE_MERGEPY to 1 in that file and nothing is changing it back to 0 | 10:47 |
| derekh | gfidente: its only set it devtest_variables if it hasn't been set already | 10:47 |
| openstackgerrit | Merged openstack/tuskar-ui: Fix node list https://review.openstack.org/172399 | 10:47 |
| gfidente | derekh, ah so this is actually setting it | 10:47 |
| derekh | yup | 10:48 |
| gfidente | I was looking at how combine things to get only a single overcloud-precise-nonha job use merge.oy | 10:48 |
| shardy | ukalifon: Hi, currently the breakpoints will be placed in the *NodesPostDeployment resources, e.g: | 10:49 |
| shardy | https://github.com/openstack/tripleo-heat-templates/blob/master/overcloud-resource-registry-puppet.yaml#L9 | 10:49 |
| shardy | https://github.com/openstack/tripleo-heat-templates/blob/master/puppet/controller-post-puppet.yaml#L27 | 10:49 |
| shardy | ukalifon: EmilienM is working on serverspec tests which will test each *StepN of the configuration | 10:49 |
| *** ishant has quit IRC | 10:50 | |
| openstackgerrit | Giulio Fidente proposed openstack-infra/tripleo-ci: Use merge.py only for overcloud-precise-nonha job https://review.openstack.org/173263 | 10:52 |
| ukalifon | EmilienM: Can you outline what in general the scripts will test in every step? What is installed and ready for test in step1, what in step2 etc... | 10:57 |
| *** lsmola_ has joined #tripleo | 10:58 | |
| gfidente | spredzy, derekh the ha job, instead of adding it to tripleo-check-jobs I was thinking if we shouldn't run it instead only were very pertinent | 11:00 |
| gfidente | for instance, not sure what is benefit of running it against diskimage-builder, we test that on f20 already | 11:00 |
| gfidente | I'm even unsure about tripleo-image-elements given it runs on puppet now | 11:01 |
| spredzy | gfidente, which project would be pertinent ? puppet-tripleo, tripleo-heat-templates and what else ? | 11:01 |
| gfidente | spredzy, yeah I would take the route of only adding it were makes sense | 11:01 |
| gfidente | so I would add it to puppet-tripleo for instance, exactly, remove from dib | 11:02 |
| spredzy | gfidente, can you -1 it dropping this comment | 11:02 |
| derekh | gfidente: spredzy yup shaving the list of projects down a bit would be a good idea | 11:02 |
| spredzy | there, so we have 'history' and I'll amend it | 11:02 |
| gfidente | done | 11:03 |
| spredzy | derekh, gfidente ok this is a proposition. put f20puppet-ha (remove it from check-tripleo-jobs) and add it on tripleo-heat-templates. Would that make sense ? (total projects would be: puppet-tirpleo, tripleo-heat-templates, tripleo-puppe-element) | 11:05 |
| spredzy | sorry put f20puppet-ha in check-tripleo-puppet-jobs* | 11:05 |
| gfidente | spredzy, those three plus tripleo-incubator | 11:06 |
| gfidente | so four in total, like they already are, only reshuffled :P | 11:06 |
| gfidente | derekh, the CONTROLSCALE thing, why is it there? | 11:08 |
| gfidente | we don't enforce COMPUTESCALE so why shall we enforce CONTROLSCALE? | 11:08 |
| gfidente | honestly was only an attempt to cleanup, I think either we enforce both or none of the two and rely on defaults | 11:09 |
| *** masco has quit IRC | 11:09 | |
| spredzy | gfidente, thx. updated. Does it look more accurate to you ? | 11:11 |
| gfidente | spredzy, needs to be added to tripleo-puppet-elements and puppet-tripleo? | 11:12 |
| spredzy | gfidente, puppet-tripleo yes, to test changes to the tripleo::loadbalancer class | 11:14 |
| *** thrash|g0ne is now known as thrash | 11:14 | |
| spredzy | and other stuff we might be adding to it in relation w/ HA | 11:14 |
| *** aberdine has quit IRC | 11:15 | |
| gfidente | it's missing in the change I mean | 11:15 |
| spredzy | gfidente, nop. the job is now part of check-tripleo-puppet-jobs | 11:15 |
| spredzy | and both puppet-tripleo and tripleo-puppet-elements run checks in that group | 11:15 |
| gfidente | AH, sorry missed that, thanks | 11:15 |
| gfidente | +1 | 11:15 |
| derekh | gfidente: I guess it was put there to be explicit about what we are doing | 11:17 |
| *** rhallisey has joined #tripleo | 11:17 | |
| derekh | gfidente: I know git blame will tell you it was me that did it, but git blame is wrong ;-) | 11:18 |
| gfidente | ahahaha | 11:18 |
| derekh | gfidente: I took the config from one repo and moved it to another | 11:18 |
| derekh | gfidente: https://review.openstack.org/#/c/107468/ | 11:18 |
| gfidente | so I dunno, it's a stupid thing to ask but, for the glory of coherence, would you add COMPUTESCALE or remove CONTROLSCALE ? | 11:19 |
| *** aberdine has joined #tripleo | 11:20 | |
| derekh | gfidente: I think any variable we set in CI at all in some cases we should set in all cases, that would mean leaving CONTROLSCALE the way it is so it obvious what we are doing when looking at that file | 11:21 |
| derekh | for COMPUTE scale if it is alway going to be default, then I have no problem with leaving it out of the file | 11:21 |
| gfidente | derekh, ack, but I don't like leaving COMPUTESCALE depend on default then :P | 11:21 |
| derekh | but I'm also happy top see it in there mainly for consistency | 11:22 |
| gfidente | in another change though, if we want | 11:22 |
| derekh | yup | 11:22 |
| *** spredzy is now known as spredzy|afk | 11:27 | |
| openstackgerrit | Giulio Fidente proposed openstack-infra/tripleo-ci: Use merge.py only for overcloud-precise-nonha job https://review.openstack.org/173263 | 11:30 |
| openstackgerrit | Giulio Fidente proposed openstack/tripleo-heat-templates: Perform basic setup of pacemaker cluster on controllers https://review.openstack.org/172572 | 11:30 |
| *** yog__ has quit IRC | 11:41 | |
| *** weshay has joined #tripleo | 11:43 | |
| jistr | spredzy|afk: tried deploying again with the galera patch, no failed db syncs reported it seems, but neutron fails to start because of a DB state issue: http://paste.fedoraproject.org/210753/42901174/ | 11:47 |
| * jistr -> english class | 11:47 | |
| *** jistr is now known as jistr|class | 11:47 | |
| *** lucasagomes is now known as lucas-hungry | 11:49 | |
| *** trown|outttypeww is now known as trown | 11:53 | |
| *** mcornea has quit IRC | 11:57 | |
| *** Marga_ has joined #tripleo | 11:57 | |
| *** ukalifon has quit IRC | 11:58 | |
| *** mcornea has joined #tripleo | 11:58 | |
| *** panda|afk is now known as panda | 11:59 | |
| derekh | that qemu pin on the puppet job worked https://jenkins04.openstack.org/job/check-tripleo-ironic-overcloud-f20puppet-nonha/196/ , back in a few plase merge when the f20 job passes | 12:01 |
| *** ukalifon has joined #tripleo | 12:03 | |
| *** morazi has joined #tripleo | 12:06 | |
| *** dprince has joined #tripleo | 12:07 | |
| openstackgerrit | Jan Provaznik proposed openstack/tripleo-common: Scale out heat stack https://review.openstack.org/173283 | 12:12 |
| *** daneyon has joined #tripleo | 12:20 | |
| openstackgerrit | Giulio Fidente proposed openstack/tripleo-heat-templates: Preformat node IPs list in the template https://review.openstack.org/173287 | 12:24 |
| openstackgerrit | Merged openstack-infra/tripleo-ci: Pin qemu-img on Fedora 21 jobs. https://review.openstack.org/172989 | 12:25 |
| *** daneyon has quit IRC | 12:25 | |
| dprince | derekh: ^^ thanks | 12:25 |
| dprince | derekh: FWIW, I was testing yesterday with a seed VM and my qemu pin'ing patches worked there. | 12:27 |
| dprince | derekh: then we made changes around the DIB up-to-date patch. A long road to simply get a package 'pinned'. | 12:29 |
| dprince | derekh: seems like it shouldn't have to be this hard to do something so simple :( | 12:29 |
| * hewbrocca notes Pulp has versioned repos | 12:32 | |
| *** regebro has quit IRC | 12:34 | |
| *** spredzy|afk is now known as spredzy | 12:42 | |
| *** rlandy has joined #tripleo | 12:44 | |
| *** mmagr is now known as mmagr|afk | 12:44 | |
| *** lucas-hungry is now known as lucasagomes | 12:46 | |
| *** jdob has quit IRC | 12:46 | |
| *** jdob has joined #tripleo | 12:47 | |
| openstackgerrit | Merged openstack/tuskar-ui: Only show charts for provisioned nodes https://review.openstack.org/172406 | 12:50 |
| slagle | shardy: hi, got a few mins to help me with getting a software deployment DELETE action to execute? | 12:51 |
| spredzy | jistr|class, hmmm never had that :/ | 12:52 |
| spredzy | Let's try to get the f20puppet-ha in so we can troubleshoot from a common setup | 12:53 |
| *** tzumainn has joined #tripleo | 12:53 | |
| spredzy | https://review.openstack.org/#/c/173255/ | 12:53 |
| spredzy | jistr|class, dprince ^ | 12:53 |
| *** ukalifon has quit IRC | 12:57 | |
| dprince | spredzy: thanks, looks good. | 12:59 |
| dprince | spredzy: I gotta say... after the latest round of Fedora 21 breakages I'm about to say lets rip all the Fedora 21 jobs out and go with Centos instead | 12:59 |
| dprince | spredzy: but for now this is fine I think | 13:00 |
| dprince | spredzy: what are your thoughts on running ceph? | 13:01 |
| dprince | spredzy: do we want to do that in the HA job? Or perhaps in the normal fedora20 job instead. | 13:01 |
| dprince | spredzy: I've got a couple patches in the works to move towards that... | 13:01 |
| spredzy | dprince, for CentOS from my point of view I'll be fine with it. The less the CI is broken, the faster we will move on | 13:02 |
| derekh | dprince: yup, it should have been easier, btw I didn't mean to merge the t-e-i patch, hit the wrong button when I meant to -2 | 13:02 |
| spredzy | dprince, for Ceph, even if it might be overkill I tend to think that ceph should be tested in both ha and nonha setup | 13:03 |
| *** jistr|class is now known as jistr | 13:03 | |
| dprince | derekh: oh, right. I hadn't noticed. | 13:03 |
| spredzy | it would be sad to miss a patch that break the setup in a really specific case if we could avoid it | 13:03 |
| dprince | spredzy: right, so for the HA setup I think we are going to have to have breakpoints | 13:03 |
| spredzy | this way we will be confident in the CI | 13:03 |
| shardy | slagle: Hi, sure | 13:04 |
| dprince | spredzy: for Ceph, without breakpoints I have to enable this: https://review.openstack.org/#/c/172597/ | 13:04 |
| * spredzy checks | 13:04 | |
| dprince | spredzy: otherwise we have a race between when Cinder recognizes how much Ceph storage there actually is | 13:05 |
| dprince | spredzy: what I'm getting at is we've had really flaky HA jobs up until this point. I'd like to suggest we use breakpoints to guarantee things come up in an expected order | 13:06 |
| spredzy | dprince, what's the actual status of breakpoints ? | 13:07 |
| dprince | spredzy: feature wise it is all landed I think. We just need to outfit devtest_overcloud.sh to support it | 13:07 |
| shardy | spredzy: all the heat stuff has landed, I'm working on making it easier to use via devtest_overcloud right now | 13:07 |
| dprince | spredzy: yup, shardy is on it | 13:08 |
| hewbrocca | shardy: excellent! | 13:08 |
| spredzy | dprince, shardy awesome. | 13:10 |
| spredzy | dprince, So the question is which scenario we want to test in nonha. It seems that you need this patch if you want Ceph on the controller. but what about a setup with a CEPHSTORAGE node ? | 13:12 |
| spredzy | we shouldn't heat the issue you're trying to bypass with the patch you mentioned | 13:12 |
| dprince | spredzy: SO I had issues in both cases. | 13:12 |
| dprince | spredzy: because there is a race with Cinder volume's periodic task | 13:13 |
| dprince | spredzy: if cinder-volume gets started before the Ceph cluster is fully up CI is hosed | 13:13 |
| dprince | spredzy: because it would register 0.0 space | 13:13 |
| dprince | spredzy: and thus volume's fail to get scheduled via cinder scheduler | 13:13 |
| dprince | spredzy: if we put a storage on the controller (in step2) we guarantee there is some storage available, and thus cinder-scheduler is happy for the CI case | 13:14 |
| dprince | spredzy: again, the right solution here is breakpoints | 13:14 |
| *** jprovazn has quit IRC | 13:14 | |
| dprince | spredzy: but for the simple All-in-one testing case having Ceph storage on the controller is useful too (and matches what we already do w/ Swift) | 13:15 |
| *** regebro has joined #tripleo | 13:15 | |
| spredzy | dprince, ack. Then IMO lets work toward making CI relying on it first and then improve the setup deployed in the CI in both ha and nonha scenario, once orchestration won't be an issue anymore | 13:15 |
| dprince | spredzy: note that I'm also suggesting we add a flag to disable Swift on the controller here as well: https://review.openstack.org/#/c/172616/ | 13:15 |
| dprince | spredzy: exactly | 13:15 |
| dprince | spredzy: the reason I'm proposing these things is they are required to make devtest.sh (using Ceph) actually run to completion | 13:16 |
| dprince | spredzy: using these patches I am able to enable RBD for Nova, Cinder, Glance and we are good to go | 13:16 |
| *** julim has joined #tripleo | 13:17 | |
| slagle | shardy: hi, so i'm not seeing the deployments that have the DELETE action get run when i delete my stack | 13:18 |
| slagle | shardy: i added a deployment for DELETE here: https://review.openstack.org/#/c/172065/ | 13:18 |
| slagle | shardy: was wondering if i was missing something somewheres, like a wait condition perhaps? | 13:20 |
| spredzy | dprince, those 2 patch are usefull even w/o breakpoint. It allows to test a all-in-one setup | 13:20 |
| shardy | slagle: No, that looks OK to me, maybe a bug, let me try to reproduce | 13:21 |
| dprince | spredzy: exactly, Swift storage is unfortunately already enabled by default on the controller though so I set the default there to true | 13:21 |
| shardy | slagle: provided you don't specify NO_SIGNAL in the deployment, things should get correctly serialized | 13:22 |
| shardy | slagle: with SoftwareDeployment resource, WaitConditions are generally not needed, because there's signalling built-in | 13:22 |
| slagle | shardy: ok | 13:22 |
| spredzy | dprince, so first iteration CI ha setup w/o ceph, when breakpoints gets fully supported in devtest we will apply the necessary change to have ceph deployed in the CI in both setup | 13:23 |
| dprince | spredzy: exactly | 13:23 |
| slagle | shardy: i also tried moving the deployment to not be part of ExtraConfig, but that didnt seem to make any difference | 13:23 |
| spredzy | dprince, deal | 13:24 |
| spredzy | CI is back to green on reviews. Thanks dprince derekh and all | 13:28 |
| dprince | bnemec, helped alot too. A team effort... and a lesson in frustration (for me at least) | 13:33 |
| *** lblanchard has joined #tripleo | 13:33 | |
| spredzy | and bnemec :) | 13:33 |
| *** mcornea has quit IRC | 13:33 | |
| *** sdake has joined #tripleo | 13:35 | |
| *** mcornea has joined #tripleo | 13:36 | |
| *** pradeep has quit IRC | 13:38 | |
| *** sdake has quit IRC | 13:40 | |
| spredzy | dprince, do you know any one core on openstack-infra I should add as a reviewer on the f20puppet-ha review ? | 13:41 |
| dprince | spredzy: I'd ask on IRC | 13:43 |
| *** weshay is now known as weshay_mtg | 13:45 | |
| dprince | spredzy: since it is in the projects repo maybe anteaya, and AJaeger | 13:46 |
| spredzy | dprince, ok thank you. | 13:46 |
| dprince | spredzy: they are all quite helpful (but busy too) | 13:46 |
| *** jprovazn has joined #tripleo | 13:53 | |
| *** jprovazn has quit IRC | 13:53 | |
| *** jprovazn has joined #tripleo | 13:53 | |
| openstackgerrit | Yanis Guenane proposed openstack/tripleo-heat-templates: puppet: implement MongoDB on controller nodes https://review.openstack.org/165474 | 13:55 |
| *** sdake has joined #tripleo | 13:57 | |
| openstackgerrit | Merged openstack/tripleo-puppet-elements: Add newline to end of source-repository-puppet-modules https://review.openstack.org/172973 | 13:59 |
| *** mmagr|afk is now known as mmagr | 14:02 | |
| *** daneyon has joined #tripleo | 14:09 | |
| jistr | spredzy: thx for tackling the HA CI, that's going to be extremely helpful | 14:12 |
| *** daneyon has quit IRC | 14:13 | |
| hewbrocca | it will also be a thing we have never had before | 14:14 |
| hewbrocca | yay! | 14:14 |
| gfidente | hewbrocca, not true, we're indeed switching the elements based job to puppet | 14:16 |
| hewbrocca | gfidente: ahh, sure, fair point | 14:18 |
| gfidente | hewbrocca, jprovazn could be around so ... | 14:18 |
| gfidente | dprince, if you worry about the number of nodes in CI, can you check this https://review.openstack.org/#/c/173248/ | 14:19 |
| gfidente | and also I think the sooner we switch this the bette https://review.openstack.org/#/c/173263/ | 14:19 |
| hewbrocca | jankomat :D | 14:19 |
| dprince | gfidente: I'm not too worried about that now. | 14:20 |
| dprince | gfidente: for the HA job I think we can adjust it | 14:20 |
| gfidente | dprince, why you want the ceph node on controllers then? | 14:20 |
| dprince | gfidente: but for the normal default I think two computes is fine (it allows you to test migrations for example) | 14:20 |
| gfidente | dprince, except we never tested it | 14:21 |
| dprince | gfidente: I want ceph on the controller so with a smaller setup we can use ceph | 14:21 |
| gfidente | and continue spawning 2 computes for each job | 14:21 |
| dprince | gfidente: 1 controller (with ceph), 1 compute | 14:21 |
| dprince | gfidente: I want flexability | 14:21 |
| dprince | gfidente: people can argue about defaults... | 14:21 |
| dprince | gfidente: also, see my notes to spredzy about why (functionally) Ceph on the controller today is required | 14:22 |
| gfidente | you mean the ceph on controller default or the number of computes default? | 14:22 |
| * jprovazn reads back | 14:22 | |
| dprince | gfidente: without breakpoints to orchestrate things coming online in an orderly fashion devtest will fail | 14:22 |
| dprince | gfidente: using these patches was the *only* way I could get devtest (with ceph) actually running | 14:22 |
| dprince | gfidente: I'll push a tripleo-ci patch soon which uses these things... | 14:23 |
| dprince | gfidente: but these would be a requirement | 14:23 |
| gfidente | dprince, so the problem here is that devtest attempts to start a demo gues | 14:23 |
| gfidente | before cinder-volume realized there is some amount in the ceph storage | 14:24 |
| dprince | gfidente: that isn't a problem. That is a requirement | 14:24 |
| shardy | dprince: doh, I just hacked out a patch for adding custom environments before seeing https://review.openstack.org/#/c/172257/ ;) | 14:24 |
| gfidente | which it checks every minute, did i get it right? | 14:24 |
| dprince | shardy: I'm working on making the images customizable too FWIW | 14:24 |
| shardy | I was thinking it'd be good to enable any number of additional custom environment files | 14:24 |
| dprince | shardy: oh, cool. I could go for that as well | 14:25 |
| shardy | dprince: my patch does that, I may push it to see if there's a combined version we could agree on | 14:25 |
| jprovazn | gfidente, hewbrocca: apparently I joined in the middle of some chat, not sure what was the question, but you made a good point with jankomat ;) (LFMF - never ever lend money to jcoufal_ ) | 14:25 |
| shardy | your patch looks fine tho so we could just land that for now :) | 14:25 |
| jcoufal_ | jprovazn: what? do you owe me some money? :) | 14:26 |
| jprovazn | jcoufal_, almost correct | 14:26 |
| dprince | shardy: where is your patch? | 14:27 |
| dprince | shardy: I want to get this right, so if merging efforts makes sense lets do it now | 14:27 |
| openstackgerrit | Steven Hardy proposed openstack/tripleo-incubator: Allow additional heat environment files https://review.openstack.org/173346 | 14:28 |
| shardy | dprince: ^^ | 14:28 |
| shardy | your interface is a bit cleaner, I took the easy option of requiring -e in the variable string | 14:28 |
| shardy | we could add a loop which just adds that though I guess | 14:29 |
| spredzy | jistr, if I want to see how pacemaker is used in staypuft/quickstack what link should I look at ? | 14:29 |
| jistr | spredzy: here https://github.com/redhat-openstack/astapor/tree/master/puppet/modules/quickstack/manifests/pacemaker | 14:29 |
| dprince | shardy: either way works. A bash array with a loop could be nice | 14:29 |
| spredzy | jistr, thanks | 14:30 |
| jistr | spredzy: in quickstack it's an essential part of the deployment process so there's *a lot* of puppet code dealing somehow with pacemaker | 14:30 |
| shardy | dprince: since yours is already reviewed, I guess we just land that, I can add loop & support for multiple files after | 14:30 |
| jistr | spredzy: e.g. because quickstack deploys in a single step, but it still needs to somehow perform ordering between the nodes, pacemaker properties are used to do that synchronization/"orchestration" | 14:31 |
| spredzy | jistr, ok, looking into it | 14:32 |
| hewbrocca | jistr: currently quickstack uses the Galera resource agent to handle that orchestration, correct? | 14:33 |
| dprince | shardy: I'm okay w/ that I guess. Although if we are going to change the variable name we should perhaps do that now | 14:33 |
| jistr | hewbrocca: hmm i don't think so, lemme dig up what i meant | 14:34 |
| *** Marga_ has quit IRC | 14:34 | |
| shardy | dprince: I'm fine with OVERCLOUD_CUSTOM_HEAT_ENV provided we can make that allow multiple space-seperated paths later | 14:34 |
| shardy | OVERCLOUD_CUSTOM_HEAT_ENV="firstboot_params.yaml post_deploy_config.yaml" etc | 14:35 |
| dprince | shardy: So in that case: would OVERCLOUD_CUSTOM_HEAT_ENVIRONMENTS make more sense? | 14:35 |
| jistr | https://github.com/redhat-openstack/astapor/blob/master/puppet/modules/quickstack/manifests/pacemaker/glance.pp#L149-L160 | 14:35 |
| jistr | https://github.com/redhat-openstack/astapor/blob/master/puppet/modules/quickstack/templates/ha-all-in-one-util.erb#L18-L67 | 14:35 |
| dprince | shardy: longer, but more specific | 14:35 |
| jistr | hewbrocca, spredzy ^^ | 14:35 |
| shardy | dprince: yeah possibly - I went for "extra" due to the historical connection with extra config | 14:36 |
| jistr | hewbrocca, spredzy: that's essentially a workaround for not being able to do multi-step deployment, and the resulting ordering is even more strict than what we have in the steps. The quickstack manifests are quite complex and have a lot of explicit ordering, but as a result the deployment is pretty predictable. | 14:39 |
| jistr | (not suggesting we should follow the same path though ;) ) | 14:40 |
| hewbrocca | Then there is also the method that jprovazn / jankomat implemented in the image elements | 14:40 |
| jistr | jprovazn: can you please point me where i could find the important bits re ^ ? | 14:42 |
| jprovazn | jistr, hewbrocca: TIE script is more complex because it doesn't expect multi step deployment, the important bits are here - since the orig patch this has been rewritten by better experts (so "implemented by jankomat is not true ;): | 14:44 |
| jprovazn | https://github.com/openstack/tripleo-image-elements/blob/master/elements/rabbitmq-server/os-refresh-config/post-configure.d/51-rabbitmq | 14:44 |
| jprovazn | but I suggest to at least read through to get better idea what all pitfalls you can hit | 14:45 |
| *** noslzzp has joined #tripleo | 14:45 | |
| jistr | thx | 14:45 |
| jistr | hmm that's pretty complex logic there | 14:47 |
| jprovazn | jistr, in case of multi step deployment it might be easier, nice thing about the above script is that it will make a cluster of nodes started in various order and it will deal with failed nodes | 14:51 |
| *** pradeep has joined #tripleo | 14:53 | |
| gfidente | spredzy, regarding the pacemaker change, resource-agents are installed as deps so I didn't include it explicitly | 14:53 |
| * jistr rebuilding images with the pacemaker bits included | 14:55 | |
| gfidente | jistr++ | 14:56 |
| gfidente | jistr, spredzy so the reason I had to use broadcast is that there is a problem in puppet-corosync as to use udpu you're supposed to pass unicast_addresses AND quorum_members | 14:56 |
| gfidente | jistr, spredzy but if you do that the resulting config turns invalid (for RH based only, where votequorum is enabled) | 14:57 |
| gfidente | I think we should fix this at some point though and use udpu | 14:57 |
| * spredzy back from meeting, reading backlog about pacemaker | 14:59 | |
| *** untriaged-bot has joined #tripleo | 15:00 | |
| untriaged-bot | No untriaged bugs so far! \o/ | 15:00 |
| *** untriaged-bot has quit IRC | 15:00 | |
| *** noslzzp_ has joined #tripleo | 15:06 | |
| *** noslzzp has quit IRC | 15:08 | |
| openstackgerrit | Giulio Fidente proposed openstack-infra/tripleo-ci: Enforce COMPUTESCALE to 1 for CI, except HA jobs https://review.openstack.org/173372 | 15:09 |
| *** noslzzp_ has quit IRC | 15:11 | |
| *** pdb has quit IRC | 15:20 | |
| *** pdb has joined #tripleo | 15:21 | |
| *** wfoster has quit IRC | 15:23 | |
| openstackgerrit | Derek Higgins proposed openstack-infra/tripleo-ci: Remove heat-templates cherrypick https://review.openstack.org/172937 | 15:23 |
| *** yamahata has joined #tripleo | 15:24 | |
| *** mmagr is now known as mmagr|afk | 15:28 | |
| *** mmagr|afk is now known as mmagr | 15:30 | |
| *** wfoster has joined #tripleo | 15:32 | |
| *** jerryz has quit IRC | 15:39 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Cinder container https://review.openstack.org/170965 | 15:41 |
| *** dprince has quit IRC | 15:46 | |
| *** rhallisey has quit IRC | 15:48 | |
| *** dtantsur is now known as dtantsur|afk | 15:48 | |
| *** ifarkas has quit IRC | 15:48 | |
| *** rhallisey has joined #tripleo | 15:50 | |
| openstackgerrit | James Slagle proposed openstack/tripleo-heat-templates: Enable passing optional first-boot user-data https://review.openstack.org/170137 | 15:50 |
| openstackgerrit | James Slagle proposed openstack/tripleo-heat-templates: Example of post-deploy hook doing rhel registration https://review.openstack.org/172065 | 15:50 |
| openstackgerrit | James Slagle proposed openstack/tripleo-heat-templates: Add hooks for extra post-deployment config https://review.openstack.org/172050 | 15:50 |
| *** jistr has quit IRC | 15:51 | |
| *** spredzy is now known as spredzy|afk | 15:52 | |
| slagle | shardy: so in https://review.openstack.org/#/c/172065/6 i attempted moving the delete deployments to the top level stack | 15:52 |
| *** eghobo has joined #tripleo | 15:53 | |
| slagle | shardy: didnt seem to help though | 15:53 |
| *** mmagr has quit IRC | 15:55 | |
| *** daneyon has joined #tripleo | 15:58 | |
| shardy | slagle: Ok, thanks, I'm trying to reproduce on devstack with some minimal examples | 16:01 |
| *** spzala has joined #tripleo | 16:01 | |
| *** bvandenh has quit IRC | 16:02 | |
| *** daneyon has quit IRC | 16:02 | |
| *** jcoufal_ has quit IRC | 16:11 | |
| *** sdake_ has joined #tripleo | 16:11 | |
| *** Marga_ has joined #tripleo | 16:12 | |
| *** sdake_ has joined #tripleo | 16:12 | |
| *** Marga_ has quit IRC | 16:14 | |
| *** Marga_ has joined #tripleo | 16:15 | |
| *** noslzzp has joined #tripleo | 16:15 | |
| *** sdake has quit IRC | 16:15 | |
| *** Marga_ has quit IRC | 16:18 | |
| *** Marga_ has joined #tripleo | 16:19 | |
| *** noslzzp has quit IRC | 16:20 | |
| *** mcornea has quit IRC | 16:22 | |
| *** Marga_ has quit IRC | 16:23 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Cinder container https://review.openstack.org/170965 | 16:27 |
| mordred | greghaynes: our new image works on both vexxhost and runabove | 16:31 |
| greghaynes | how many distros? | 16:32 |
| mordred | greghaynes: I'm going to kick off a new end-to-end test with all of the clouds and regions with your latest patch | 16:32 |
| mordred | greghaynes: only tried ubuntu for now | 16:32 |
| greghaynes | \O/ | 16:32 |
| greghaynes | ah, ok | 16:32 |
| greghaynes | I verified fedora gets networking and then the others bulid | 16:32 |
| mordred | woot | 16:32 |
| greghaynes | before this last round of changes | 16:32 |
| mordred | fedora and ubuntu cover most of teh crazy choices | 16:32 |
| greghaynes | although they were only rebases for !simple-init | 16:32 |
| mordred | greghaynes: well, I haven't tested since you fixed the source-repositories depend | 16:36 |
| mordred | greghaynes: so I';m removing my hackaroudn | 16:36 |
| greghaynes | ok, I was able to build and boot those | 16:36 |
| greghaynes | didnt verify networking and such | 16:36 |
| mordred | greghaynes: my build/boot script does a "wait_for: port=22" | 16:37 |
| mordred | greghaynes: so if it doesn't get networking and a working ssh - it errors | 16:37 |
| greghaynes | hah, I havent added ssh in | 16:37 |
| greghaynes | do you have an element that adds it or something? | 16:37 |
| *** Goneri has quit IRC | 16:38 | |
| openstackgerrit | Giulio Fidente proposed openstack/tripleo-heat-templates: Preformat node IPs list in the template https://review.openstack.org/173287 | 16:39 |
| mordred | greghaynes: github.com:emonty/ansible-build-image | 16:41 |
| mordred | greghaynes: is what I'm using to test | 16:41 |
| mordred | greghaynes: it requires the infra element from https://review.openstack.org/#/c/140840/ | 16:42 |
| mordred | greghaynes: which I've just updated with my latest playbook | 16:42 |
| *** derekh has quit IRC | 16:43 | |
| *** noslzzp has joined #tripleo | 16:44 | |
| *** Marga_ has joined #tripleo | 16:46 | |
| *** Marga_ has quit IRC | 16:46 | |
| *** Marga_ has joined #tripleo | 16:51 | |
| *** Marga_ has quit IRC | 16:58 | |
| *** sdake has joined #tripleo | 17:07 | |
| *** Marga_ has joined #tripleo | 17:08 | |
| *** sdake_ has quit IRC | 17:11 | |
| *** sdake_ has joined #tripleo | 17:14 | |
| *** dprince has joined #tripleo | 17:18 | |
| *** sdake has quit IRC | 17:18 | |
| openstackgerrit | Giulio Fidente proposed openstack/tripleo-heat-templates: Preformat node IPs list in the template https://review.openstack.org/173287 | 17:21 |
| *** spredzy|afk is now known as spredzy | 17:22 | |
| *** trown is now known as trown|lunch | 17:26 | |
| openstackgerrit | Giulio Fidente proposed openstack/tripleo-heat-templates: Preformat node IPs list in the template https://review.openstack.org/173287 | 17:26 |
| *** sdake has joined #tripleo | 17:31 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Cinder container https://review.openstack.org/170965 | 17:31 |
| *** pelix has quit IRC | 17:32 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Cinder container https://review.openstack.org/170965 | 17:34 |
| mordred | greghaynes: qemu-img: Could not open '/tmp/image.64ZHgR8r/image.raw': Could not open '/tmp/image.64ZHgR8r/image.raw': No such file or directory | 17:34 |
| *** sdake_ has quit IRC | 17:35 | |
| *** Marga_ has quit IRC | 17:35 | |
| *** Marga_ has joined #tripleo | 17:35 | |
| *** Marga_ has quit IRC | 17:36 | |
| *** Marga_ has joined #tripleo | 17:36 | |
| mordred | greghaynes: hrm - maybe that wasn't the real error ... | 17:38 |
| greghaynes | mordred: huh, is https://review.openstack.org/#/c/169911/ in your ancestry? | 17:38 |
| greghaynes | oh | 17:38 |
| shardy | slagle: sorry for the slow response, I've reproduced your problem - it's a heat bug, I've got a fix, just working on tests now | 17:38 |
| mordred | greghaynes: nope. it is not | 17:39 |
| mordred | greghaynes: rebased - trying again | 17:39 |
| *** lucasagomes is now known as lucas-dinner | 17:42 | |
| *** sdake_ has joined #tripleo | 17:44 | |
| *** sdake has quit IRC | 17:45 | |
| *** sdake_ has quit IRC | 17:48 | |
| *** sdake has joined #tripleo | 17:48 | |
| openstackgerrit | Merged openstack/tripleo-heat-templates: Make rabbitmq bind on ctlplane instead of 0.0.0.0 https://review.openstack.org/171843 | 17:50 |
| *** sdake has quit IRC | 17:52 | |
| *** sdake has joined #tripleo | 17:52 | |
| openstackgerrit | Merged openstack/tripleo-heat-templates: puppet: implement MongoDB on controller nodes https://review.openstack.org/165474 | 17:52 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Make glance work reliably https://review.openstack.org/173113 | 17:56 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: CentOS 7.1 image doesn't work with libvirt https://review.openstack.org/173440 | 17:56 |
| openstackgerrit | Merged stackforge/kolla: CentOS 7.1 image doesn't work with libvirt https://review.openstack.org/173440 | 18:01 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Make glance work reliably https://review.openstack.org/173113 | 18:14 |
| *** daneyon has joined #tripleo | 18:16 | |
| shardy | slagle: https://review.openstack.org/#/c/173446/1 | 18:16 |
| shardy | should fix your DELETE deployment issue | 18:16 |
| openstackgerrit | Yanis Guenane proposed openstack/tripleo-heat-templates: puppet: Ceilometer use MongoDB datastore backend https://review.openstack.org/165480 | 18:16 |
| openstackgerrit | Yanis Guenane proposed openstack/tripleo-heat-templates: Add support for Redis configuration https://review.openstack.org/167538 | 18:20 |
| openstackgerrit | Yanis Guenane proposed openstack/tripleo-heat-templates: Enable use of coordination_url in ceilometer https://review.openstack.org/171996 | 18:20 |
| *** daneyon_ has joined #tripleo | 18:20 | |
| openstackgerrit | Steven Dake proposed stackforge/kolla: remove wait_for from glance-api https://review.openstack.org/173447 | 18:21 |
| *** yamahata has quit IRC | 18:22 | |
| openstackgerrit | Yanis Guenane proposed openstack/tripleo-heat-templates: galera: Allow installer to setup galera cluster https://review.openstack.org/164983 | 18:22 |
| *** daneyon has quit IRC | 18:23 | |
| *** sdake_ has joined #tripleo | 18:25 | |
| *** sdake has quit IRC | 18:29 | |
| *** Marga_ has quit IRC | 18:29 | |
| slagle | shardy: cool, i'll give it a go | 18:31 |
| *** trown|lunch is now known as trown | 18:32 | |
| openstackgerrit | Giulio Fidente proposed openstack/tripleo-heat-templates: Do not downcase IP addresses https://review.openstack.org/173454 | 18:36 |
| *** sdake has joined #tripleo | 18:39 | |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Update developer documentation https://review.openstack.org/172998 | 18:41 |
| *** sdake_ has quit IRC | 18:43 | |
| *** yamahata has joined #tripleo | 18:45 | |
| *** Marga_ has joined #tripleo | 18:50 | |
| *** rhallisey has quit IRC | 18:53 | |
| *** Goneri has joined #tripleo | 18:53 | |
| *** rhallisey has joined #tripleo | 18:54 | |
| *** pradeep has quit IRC | 18:57 | |
| *** rhallisey has quit IRC | 18:59 | |
| gfidente | has 3x +2 but was failing CI https://review.openstack.org/#/c/171840/ :) | 19:01 |
| *** rhallisey has joined #tripleo | 19:01 | |
| gfidente | please | 19:01 |
| *** rhallisey has quit IRC | 19:02 | |
| openstackgerrit | Merged openstack/tripleo-heat-templates: Make memcached bind on ctlplane instead of 0.0.0.0 https://review.openstack.org/171840 | 19:03 |
| gfidente | jdob++ | 19:04 |
| jdob | haha | 19:05 |
| jdob | didnt think you were still around, but i'm keeping that HA etherpad up to date | 19:05 |
| gfidente | jdob, nice :) | 19:06 |
| *** bvandenh has joined #tripleo | 19:10 | |
| *** mmagr has joined #tripleo | 19:10 | |
| spredzy | jdob, can I get your review in this one https://review.openstack.org/#/c/165480/ | 19:10 |
| spredzy | please. Just removed the conflict | 19:10 |
| gfidente | spredzy, ! | 19:11 |
| gfidente | : | 19:11 |
| gfidente | :) | 19:11 |
| *** Goneri has quit IRC | 19:11 | |
| spredzy | o/ still around :p | 19:12 |
| *** pradeep has joined #tripleo | 19:12 | |
| *** jprovazn has quit IRC | 19:12 | |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Add a conf-novanet to configure nova https://review.openstack.org/170306 | 19:15 |
| *** akrivoka has quit IRC | 19:19 | |
| *** rhallisey has joined #tripleo | 19:19 | |
| tchaypo | gfidente: I’m curious | 19:19 |
| tchaypo | gfidente: does the downcasing cause problems, or is it just silly? | 19:19 |
| tchaypo | gfidente: but also - is it required or useful for ipv6 addresses? | 19:19 |
| gfidente | tchaypo, absolutely just me being silly :) | 19:20 |
| gfidente | tchaypo, I was trying to make the list propagate into .pp as an array preformatting it in the .yaml | 19:20 |
| gfidente | I was able to, with str_replace | 19:20 |
| gfidente | but I still had to use any2array in the .pp | 19:20 |
| gfidente | so I abandoned that which seemed not worth | 19:20 |
| gfidente | but noticed there were these downcase things :) | 19:21 |
| gfidente | tchaypo, btw, how are you doing? :) | 19:21 |
| tchaypo | aside from being up at 5am for a meeting? | 19:21 |
| spredzy | gfidente, it is totally worth it. Spent an hour trouble shooting an issue where the hostname from one side was xxx-Controller-xxx and from the fact it was xxx-controller-xxx | 19:21 |
| *** mmagr has quit IRC | 19:21 | |
| gfidente | spredzy, it is totally needed | 19:22 |
| spredzy | so w/o the downcase the equality test was failing | 19:22 |
| tchaypo | ouch | 19:22 |
| gfidente | I remember we discussed that before | 19:22 |
| spredzy | and it never got into the if it was supposed to | 19:22 |
| gfidente | but for hostnames, not ips | 19:22 |
| tchaypo | spredzy: could you comment on the review? | 19:22 |
| gfidente | hostnames are based on the resource name "Controller" | 19:22 |
| spredzy | tchaypo, sorry which review are yu guys talking about ? :) | 19:22 |
| spredzy | just catched the question | 19:22 |
| gfidente | that's why we need the downcase | 19:22 |
| tchaypo | https://review.openstack.org/#/c/173454 | 19:22 |
| tchaypo | wait, that’s different | 19:23 |
| spredzy | well on ips its not necessary | 19:23 |
| spredzy | only on hostname X_names stuff | 19:23 |
| gfidente | spredzy, so to avoid the downcase I was thinking to switch rabbit into ips as well | 19:24 |
| gfidente | to avoid the Controller trap | 19:24 |
| spredzy | gfidente, ack. My point was mainly be careful when using _names/hostname, and do not ofrget the downcase() function | 19:25 |
| gfidente | spredzy, exactly, I remember we hit this before | 19:25 |
| spredzy | but if it could be replaced by IPs sure. | 19:25 |
| *** weshay_mtg has quit IRC | 19:29 | |
| jdob | spredzy: looking | 19:34 |
| slagle | shardy: worked :) | 19:34 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Add a conf-novanet to configure nova https://review.openstack.org/170306 | 19:35 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Add a conf-novanet to configure nova https://review.openstack.org/170306 | 19:36 |
| openstackgerrit | James Slagle proposed openstack/tripleo-heat-templates: Example of post-deploy hook doing rhel registration https://review.openstack.org/172065 | 19:36 |
| *** sdake_ has joined #tripleo | 19:37 | |
| *** sdake has quit IRC | 19:41 | |
| *** sdake has joined #tripleo | 19:51 | |
| *** sdake_ has quit IRC | 19:54 | |
| openstackgerrit | Merged openstack-infra/tripleo-ci: Remove heat-templates cherrypick https://review.openstack.org/172937 | 20:03 |
| openstackgerrit | Merged openstack-infra/tripleo-ci: Filter session-XXX.scope files from logs https://review.openstack.org/169670 | 20:04 |
| *** sdake_ has joined #tripleo | 20:05 | |
| openstackgerrit | Daneyon Hansen proposed stackforge/kolla: Adds Neutron to Start/Stop Scripts https://review.openstack.org/173494 | 20:06 |
| *** sdake has quit IRC | 20:07 | |
| openstackgerrit | Daneyon Hansen proposed stackforge/kolla: Adds Neutron to Start/Stop Scripts https://review.openstack.org/173494 | 20:08 |
| *** gfidente has quit IRC | 20:13 | |
| openstackgerrit | Thomas Bechtold proposed openstack/diskimage-builder: openSUSE update https://review.openstack.org/173498 | 20:15 |
| *** lucas-dinner has quit IRC | 20:17 | |
| openstackgerrit | Daneyon Hansen proposed stackforge/kolla: Adds Neutron to Start/Stop Scripts https://review.openstack.org/173494 | 20:23 |
| *** sdake has joined #tripleo | 20:23 | |
| *** sdake_ has quit IRC | 20:25 | |
| *** Marga_ has quit IRC | 20:28 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Add cinder into the docker compose setup https://review.openstack.org/173507 | 20:31 |
| *** lblanchard has quit IRC | 20:31 | |
| *** mmagr has joined #tripleo | 20:37 | |
| *** lifeless1 is now known as lifeless | 20:37 | |
| *** MasterPiece has joined #tripleo | 20:38 | |
| openstackgerrit | Daneyon Hansen proposed stackforge/kolla: Adds a Status Script https://review.openstack.org/173511 | 20:43 |
| *** mmagr has quit IRC | 20:43 | |
| *** Goneri has joined #tripleo | 20:48 | |
| *** jsidhu has joined #tripleo | 20:49 | |
| openstackgerrit | Merged openstack/diskimage-builder: Do not export REG_HALT_UNREGISTER between hook scripts https://review.openstack.org/162239 | 20:50 |
| *** jtomasek has quit IRC | 20:51 | |
| *** untriaged-bot has joined #tripleo | 21:00 | |
| untriaged-bot | Untriaged bugs so far: | 21:00 |
| untriaged-bot | https://bugs.launchpad.net/tripleo/+bug/1444068 | 21:00 |
| openstack | Launchpad bug 1444068 in tripleo "Rabbitmq_plugin[rabbitmq_management]: Could not evaluate: execution expired"Puppet deployment fails with "" [Undecided,Triaged] | 21:00 |
| *** untriaged-bot has quit IRC | 21:00 | |
| *** pblaho has quit IRC | 21:02 | |
| openstackgerrit | Daneyon Hansen proposed stackforge/kolla: Adds a Remove Script https://review.openstack.org/173520 | 21:05 |
| *** trown is now known as trown|outttypeww | 21:05 | |
| shardy | slagle: thanks for the confirmation, chasing some reviews so we can hopefully land the fix soon | 21:09 |
| *** dprince has quit IRC | 21:14 | |
| *** dsneddon has joined #tripleo | 21:18 | |
| *** cwolferh has joined #tripleo | 21:18 | |
| *** rhallisey has quit IRC | 21:20 | |
| *** daneyon has joined #tripleo | 21:30 | |
| *** daneyon_ has quit IRC | 21:32 | |
| *** shardy is now known as shardy_z | 21:33 | |
| openstackgerrit | Merged stackforge/kolla: Adds Neutron to Start/Stop Scripts https://review.openstack.org/173494 | 21:35 |
| *** Goneri has quit IRC | 21:35 | |
| *** weshay has joined #tripleo | 21:35 | |
| *** daneyon has quit IRC | 21:44 | |
| *** tzumainn has quit IRC | 21:47 | |
| *** bvandenh has quit IRC | 21:52 | |
| *** julim has quit IRC | 21:52 | |
| *** Marga_ has joined #tripleo | 22:21 | |
| openstackgerrit | wes hayutin proposed openstack/diskimage-builder: update the dib centos7 baseurl to use a mirrorlist https://review.openstack.org/172615 | 22:28 |
| *** dsneddon has quit IRC | 22:34 | |
| *** dsneddon has joined #tripleo | 22:38 | |
| *** rhallisey has joined #tripleo | 22:39 | |
| *** Marga_ has quit IRC | 22:49 | |
| *** jsidhu has quit IRC | 22:52 | |
| *** jsidhu has joined #tripleo | 22:56 | |
| openstackgerrit | Steven Dake proposed stackforge/kolla: genenv generated invalid variable https://review.openstack.org/173590 | 22:56 |
| *** yuanying has joined #tripleo | 22:57 | |
| *** jsidhu has quit IRC | 22:58 | |
| *** Marga_ has joined #tripleo | 23:07 | |
| *** chlong has joined #tripleo | 23:10 | |
| *** regebro has quit IRC | 23:38 | |
| *** jerryz has joined #tripleo | 23:42 | |
| *** chlong has quit IRC | 23:43 | |
| *** rlandy has quit IRC | 23:44 | |
| *** chlong has joined #tripleo | 23:45 | |
| *** alop has quit IRC | 23:53 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!