*** maddtux has joined #heat | 00:57 | |
*** ricolin has joined #heat | 01:03 | |
*** mhen has quit IRC | 01:22 | |
*** mhen has joined #heat | 01:27 | |
*** higgins has quit IRC | 02:37 | |
*** higgins has joined #heat | 02:40 | |
gregwork | what does it mean when stack create shoots back: ERROR: Property error: : resources.aggregate-hci-0.properties.hosts[0]: : 'Client' object has no attribute 'hosts' | 02:40 |
---|---|---|
gregwork | OS::Nova::HostAggregates most definitely has the attribute hosts which the spec says is a list | 02:41 |
*** ramishra has joined #heat | 02:42 | |
gregwork | eg. | 02:42 |
gregwork | https://pastebin.com/b6gmKZvu | 02:42 |
gregwork | this is more or less out of the github.com hot templates examples repo | 02:43 |
*** cliffparsons has joined #heat | 02:48 | |
gregwork | if i omit the hosts attribute the stack crate succeeds but it is created but when i do an openstack aggregate show 1rsp-lab-az-0 | 02:52 |
gregwork | i see the hosts: field is empty | 02:52 |
*** iokiwi has quit IRC | 02:54 | |
*** higgins has quit IRC | 02:57 | |
*** higgins has joined #heat | 02:57 | |
gregwork | i think this is busted | 03:05 |
gregwork | https://pastebin.com/1616Ymg2 | 03:05 |
*** gkadam has joined #heat | 03:31 | |
*** gkadam has quit IRC | 03:40 | |
*** jdwidari has quit IRC | 04:10 | |
*** jdwidari has joined #heat | 04:11 | |
*** skramaja has joined #heat | 04:13 | |
*** jdwidari has quit IRC | 04:15 | |
ramishra | gregwork: yes, the host constraint check seems boken after https://github.com/openstack/python-novaclient/commit/1d88717e751e743ca7edd978346b95b6e7b3f235#diff-4ef50ccc65413eb1dda481e0484e831b. Feel free to a raise a bug and/or submit a patch | 04:21 |
gregwork | i did open a ticket with redhat about this earlier, was curious if it was something i was doing or an actual bug | 04:22 |
gregwork | ramishra: could i ask you a question about instance placement | 04:22 |
ramishra | gregwork: it would be aquestion for #nova channel :) | 04:23 |
gregwork | well its using heat | 04:23 |
ramishra | yeah, but heat does not do any placement related logic, just passes the hints to nova | 04:24 |
gregwork | ramishra: given a stack deploying 3 sets of instances, 3 of type-a, 3 of type-b, and 6 of type-c .. given 3 compute nodes, how could i gaurantee that each type-a server got placed on each of the 3 compute nodes (for ha purposes) | 04:24 |
gregwork | well im trying to figure out what options i have with heat | 04:24 |
gregwork | im deploying openshift on my openstack lab, and i want to gaurantee that the 3 master instances dont find each other on the same compute node | 04:25 |
gregwork | i was thinking about using scheduler_hints: different_server: [ ... ] | 04:26 |
gregwork | but im not sure how to populate the contents of that list dynamically | 04:26 |
gregwork | im using a resource group to iterate the deployment of these nodes | 04:27 |
gregwork | so the first master node that gets deployed .. different_server would be empty as no other masters exist | 04:27 |
ramishra | I think you can use anti-afinity rules with server groups https://github.com/openstack/heat/blob/master/heat/engine/resources/openstack/nova/server_group.py#L55 | 04:27 |
ramishra | so the nodes in a server group won't be placed together | 04:28 |
gregwork | soft vs. hard anti-affinity ? | 04:28 |
ramishra | your choice, if no nodes are available for placement then soft will tolerate, else scheduling would fail.. | 04:29 |
gregwork | ah | 04:29 |
gregwork | another question, is there a method to look up the value of a previously created resource in a stack as it is deploying | 04:30 |
ramishra | lookup by what? name/id? | 04:30 |
gregwork | instance id | 04:30 |
*** cliffparsons has quit IRC | 04:31 | |
gregwork | for my earlier example anti-affinity will work nicely, for the master nodes, but with the type-c i cant use anti affinity and i need to decide on what is a balanced deployment | 04:31 |
gregwork | so ideally 2 of type-c instances per compute node | 04:32 |
ramishra | you can just do a resource list -n <depth> and then grep the instance_id, there is no better way to check in a nested stack where the server is deep somewhere | 04:32 |
gregwork | so id probably fiddle with different_servers | 04:32 |
gregwork | in scheduler hints, but i need to look up instance id's to fill that with | 04:32 |
ramishra | oh, you mean have some lookup in a template to provide hints, templates are static..so not possible | 04:34 |
gregwork | well i am using a resource group that calls another build.yaml | 04:35 |
gregwork | im thinking the build.yaml would be able to look up resources in the parent | 04:35 |
gregwork | for example | 04:36 |
gregwork | https://pastebin.com/r5wszG8s | 04:36 |
gregwork | so when ocp-node.yaml gets called as this rg iterates, is there no way for it to look up and query what other resources were created | 04:37 |
gregwork | to populate hints | 04:38 |
ramishra | gregwork: no not in the template itself..but you can possibly have logic in your build script to update the template with the instances created as part of last stack deployment | 04:40 |
gregwork | you are suggesting i cant do this in heat alone and should use something outside of heat? | 04:42 |
ramishra | gregwork: yes | 04:42 |
*** spsurya has joined #heat | 05:02 | |
gregwork | do projects like tripleo use mistral to create the workflow of template deployment during the overcloud provisioning since heat itself cant do implicit ordering of resources at deployment ? | 05:16 |
gregwork | i mean there isnt way in 1 template to say create resource A first, then create B, then create C ? | 05:17 |
ramishra | gregwork: you can use 'depends_on' for that kind of dependencyy in a template https://docs.openstack.org/heat/latest/template_guide/hot_spec.html#resources-section | 05:23 |
ramishra | https://docs.openstack.org/heat/latest/template_guide/hot_spec.html#hot-spec-resources-dependencies has more details | 05:25 |
gregwork | interesting | 05:25 |
ramishra | it's always a good idea to go through the template guide in detail before composing templates, we spend good amount of time documenting stuff:) | 05:28 |
*** mlycka has joined #heat | 05:58 | |
*** rcernin has quit IRC | 06:18 | |
*** jtomasek has joined #heat | 07:01 | |
*** joxyuki has joined #heat | 07:14 | |
joxyuki | hello folks, would you kindly review below patches? https://review.opendev.org/#/c/669789/ https://review.opendev.org/#/c/673403/ | 07:40 |
joxyuki | Now I am working on https://storyboard.openstack.org/#!/story/2006372. These patches needs to be merged in advance. | 07:40 |
*** mlycka has quit IRC | 08:12 | |
*** ivve has joined #heat | 08:19 | |
*** mikecmpbll has joined #heat | 08:25 | |
*** e0ne has joined #heat | 09:04 | |
*** skramaja has quit IRC | 09:10 | |
*** maddtux has quit IRC | 09:47 | |
*** dmellado has quit IRC | 09:53 | |
*** dmellado has joined #heat | 09:55 | |
*** ricolin_ has joined #heat | 12:05 | |
*** ricolin has quit IRC | 12:07 | |
*** jmlowe has quit IRC | 12:43 | |
*** cliffparsons has joined #heat | 12:48 | |
*** jcoufal has joined #heat | 12:54 | |
*** jcoufal has quit IRC | 13:02 | |
*** cliffparsons has quit IRC | 13:03 | |
*** jcoufal has joined #heat | 13:06 | |
*** ekultails has joined #heat | 13:14 | |
*** sm806 has quit IRC | 13:32 | |
*** sm806 has joined #heat | 13:32 | |
*** mchlumsky has joined #heat | 13:41 | |
*** ricolin_ is now known as ricolin | 13:55 | |
*** jcoufal has quit IRC | 13:55 | |
*** cliffparsons has joined #heat | 13:56 | |
*** jcoufal has joined #heat | 14:35 | |
*** spsurya has quit IRC | 15:34 | |
*** ricolin_ has joined #heat | 15:43 | |
*** ricolin_ has quit IRC | 15:45 | |
*** ricolin has quit IRC | 15:46 | |
*** ricolin_ has joined #heat | 15:46 | |
*** ricolin_ is now known as ricolin | 15:47 | |
*** beekneemech has joined #heat | 16:16 | |
*** e0ne has quit IRC | 16:16 | |
*** beekneemech has quit IRC | 16:16 | |
*** mikecmpbll has quit IRC | 16:34 | |
*** e0ne has joined #heat | 17:05 | |
*** ricolin_ has joined #heat | 17:16 | |
*** ricolin has quit IRC | 17:19 | |
*** ricolin_ is now known as ricolin | 17:32 | |
*** gmoro has quit IRC | 17:34 | |
*** e0ne has quit IRC | 17:35 | |
*** mchlumsky has quit IRC | 18:39 | |
*** ramishra has quit IRC | 19:02 | |
*** e0ne has joined #heat | 19:16 | |
*** ivve has quit IRC | 19:29 | |
*** jcoufal has quit IRC | 19:40 | |
*** jtomasek has quit IRC | 20:31 | |
*** ivve has joined #heat | 20:32 | |
*** e0ne has quit IRC | 20:37 | |
*** jcoufal has joined #heat | 20:53 | |
*** vishwanathj has joined #heat | 21:31 | |
*** jtomasek has joined #heat | 21:37 | |
*** jcoufal has quit IRC | 22:02 | |
*** kei-ichi has quit IRC | 22:03 | |
*** jsquare has quit IRC | 22:07 | |
*** jsquare has joined #heat | 22:07 | |
*** jsquare has quit IRC | 22:08 | |
*** rcernin has joined #heat | 22:53 | |
*** ricolin_ has joined #heat | 23:02 | |
*** ricolin has quit IRC | 23:05 | |
*** ekultails has quit IRC | 23:16 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!