Previously we could not feed the IP of an instance into its own
Metadata because of circular dependency problems. 0.0.0.0 was used
with the heat-localip element to work around this problem. This
caused problems though, as heat-localip would edit the source local
heat metadata, which would make cfn-hup and/or os-collect-config think
that the Metadata had changed, causing it to restart everything every
time we query the Metadata, which was about every five minutes.
Now we can just query this inert LaunchConfiguration resource to get
its' Metadata once it is ready. This resource will only change when
legitimate things are changed in the stack, and so we won't restart
everything every five minutes anymore. Note that when the native Heat
DSL lands, the OS::Heat::SoftwareConfig resource is meant for exactly
this purpose.
After this is merged we no longer need the heat-localip element.
Fixes bug #1202322
Change-Id: Id06323ba43203570eeebfa5b3d03fa56c16c0c10
Os-collect-config is meant to replace cfn-hup from heat-cfntools. It
allows pulling from multiple metadata sources and runs continuously.
Fixes bug #1211289
Change-Id: Ia4e9127fb79048bd1022b32a37272f8463a774ae
In nova-compute-instance.yaml, we need to use resource names that are
unique within the entire set of resources that may be merged or
included. However, we need the instance resource name to _match_ the
one in overcloud-source.yaml so that its own access policies can
function.
Without this change we will not have unique users and Metadata access
keys/policies for compute and controller.
Change-Id: Iebde7e6adede4984f4f693cf2d57b6fadb8be558
Also change default to '' as null is not technically allowed as a
string default in cloudformation (Heat accidentally allows this).
Os-apply-config, via pystache, will still treat this as false for
the purposes of checking if it is set or not.
Change-Id: Ia02dbcf619bdc92647f1d21157fa4a8e3f749de3
The overcloud control plane needs to bring up an ovs bridge for the
ovs plugin hot-plugging and floating ip logic to work. The previous
in-instance script didn't migrate IP address and route information. We
can do that by using the core ovs script we wrote for baremetal setups
- triggered by setting the physical interface parameter.
Change-Id: I6d6b09140ee751371607c0963dc76cc5b95f7674
This is supplied to neutron etc via API calls - having it in Heat
leads to having redundancy, confusion and tricky scripts. Incubator
now has scripts to configure this and instances won't consult these
keys.
Change-Id: Ia45d9bf5bb326ec88f9f4a48c5520570cf5b6d72
Several users with slightly older hardware report the PXE deploy
process aborting deploys that were progressing fine. By increasing the
default PXE deploy timeout we should cater for this. A separate bug
has been filed in Nova/Ironic about the timeout being too simple.
Change-Id: I9e340fc7f928b32b1146ce758cda7180888d6cc9
Fixed merge code populating properties from yaml files which was
setting only Type attribute for certain attribues related to Image
For example, default value for notcomputeImage is populated properly
in overcloud.yaml
Fixes bug #1202884
Change-Id: I9ea270dd19e9ebdb997be5d07b04ee2df28e3721
Per bug 1201656 (which this doesn't fix) nova-compute can only run one
bm arch from a single compute host.
Change-Id: I0eabe6e85ed388af111ef503882dfdf7959d8d97
The undercloud (and boot-stack in general) no longer create baremetal
flavors or register baremetal ramdisks, so passing in keys for that
isn't needed.
Change-Id: I8c20e95f3a57418d7ae160a296a092cd71254e3d
These values were missed in the initial overcloud-source.yaml.
nova.metadata-proxy is required to allow vms to communicate with the
EC2-style metadata service.
quantum.ovs.fixed_range constraints DHCP allocation for the undercloud,
but in the overcloud we do not want this constraint because the individual
subnets we define will do that. quantum.ovs.ovs_range configures the
range of ips that can be assigned to the tenant subnets.
Change-Id: If7191fa8baf1209ae33b9e8200f52ea0ade97810
Previously these would have to be set for overcloud at runtime, but we
have good sane defaults now, and for nova-compute-instance we are able
to feed them in using the appropriate references with Fn::GetAtt.
Change-Id: Idfff2885bf2afa58b2dec84f06639198e411eae2
We will need to replace the use of Parameters in the included template
with actual values in many cases.
Change-Id: I95fc9116dc2bba74c31d5570851c5c4eb476291a
We now have an undercloud specific template, so we can use
bootstack-vm specifically for the overcloud control plane.
Change-Id: I06eb2b1d00e4d2211b632a06880aed51f01762ff
The current code uses dictionary sort order which is defined as being
an undefined order. Since humans review .yaml files having them sorted
is a good thing.
Change-Id: I43a31530bc042750448ac904c0c22f9a59b09b4d
Also add Makefile stanza to produce overcoud.yaml from bootstack-vm
and nova-compute-group. This should allow independent usage of
nova-compute-group without having to manually update overcloud.
Change-Id: Id2bd20596e6be887b55019fdad31e6b0f4762eb7
Heat must download nested stack templates from a static URL. We
have chosen a reasonable default, the official repository for
tripleo-heat-templates.
Change-Id: I9f8f1f3d3ec6492cf3891fab4fe793e05e703eb5
Recent changes to the postfix element added new parameters so
the nagios3 template needed to updated to ensure postfix works
as expected.
Change-Id: I0a9eb8e0988470888f35adcc426ccd314f93b301
Authored-by: Chris Krelle <nobodycam@gmail.com>
This resource was remove in
6504e6bd09b8b3b14ce56490ea3e17daf8cc0a93
"Simplify keystone into one instance."
Change-Id: Iea1502e4f00539abc4b8d04ea37e55c2133f7ae7
The local_ip value of 127.0.0.1 breaks clusters. Using 0.0.0.0 will
have the right value when the heat-localip element is used.
Change-Id: Id0da5d46c60e808944a40bcaf05f1072be55eed7
