Redeploy cinder/compute nodes if new controller added
Addition of new controller affects cluster messaging,
particularly cinder/computes nodes which is reliant on
rabbitmq_hosts settings
In current orchestration model - the only way to
fix configuration on nodes is to redeploy them.
Configuration will be applied from /etc/astute.yaml and
it is not enough to simply add additional PostDeployment
method in astute
Introduced additional setting for roles_metadata:
- update_required
should store list of roles that dependends on this role
On deployment stage:
- make update_required list for the whole cluster
- select ready nodes without pending_roles and deploy them
No migration added to keep behaviour on old clusters as it is
Reviewed: https:/ /review. openstack. org/132549 /git.openstack. org/cgit/ stackforge/ fuel-web/ commit/ ?id=1d3982f606b 5d0d0cba78b52a5 806abee6a97918
Committed: https:/
Submitter: Jenkins
Branch: master
commit 1d3982f606b5d0d 0cba78b52a5806a bee6a97918
Author: Dima Shulyak <email address hidden>
Date: Mon Nov 3 13:37:37 2014 +0200
Redeploy cinder/compute nodes if new controller added
Addition of new controller affects cluster messaging,
particularly cinder/computes nodes which is reliant on
rabbitmq_hosts settings
In current orchestration model - the only way to
fix configuration on nodes is to redeploy them.
Configuration will be applied from /etc/astute.yaml and
it is not enough to simply add additional PostDeployment
method in astute
Introduced additional setting for roles_metadata:
- update_required
should store list of roles that dependends on this role
On deployment stage:
- make update_required list for the whole cluster
- select ready nodes without pending_roles and deploy them
No migration added to keep behaviour on old clusters as it is
DocImpact
Closes-Bug: 1368445
Change-Id: I1735a8b0653101 8b1240726f5faa4 f7ce6e6a631