
When assuming subsequent update/upgrade of deployments in the context: * Place a note for overcloud deploy commands to always contain full environments. * Note as well an alternative resource for the OS::TripleO::NodeUserData is the OS::TripleO::NodeExtraConfigPost. Change-Id: Iec834320c04be509761ec45483fa9c9708505f21 Signed-off-by: Bogdan Dobrelya <bdobreli@redhat.com>
1.2 KiB
Scaling overcloud roles
If you want to increase or decrease resource capacity of a running overcloud, you can start more servers of a selected role or delete some servers if capacity should be decreased. To set the capacity for the compute role, first an environment file should be created:
$ cat ~/environment.yaml
parameter_defaults:
ComputeCount: 5
Then following command can be used to deploy it:
openstack overcloud deploy --templates [templates dir] \
-e <full environment> -e ~/environment.yaml
Note
It is especially important to remember that you must include all environment files that were used to deploy the overcloud. Make sure you pass those in addition to your customization environments at the end (environment.yaml).
Note
Scaling out assumes that newly added nodes has already been registered in Ironic.
Note
When scaling down random servers of specified role will be deleted,
how to delete specific nodes is described in delete_nodes
.
Note
The different scale parameters can be seen in the output of:
openstack help overcloud deploy