This section instructs on how to correctly start an MCP OpenStack environment after it was powered off.
If you have Ceph in your MCP cluster, start it as described in Start a Ceph cluster.
To start an MCP OpenStack environment:
Start the kvm01
node hosting the dbs01
node.
Start the kvm02
node hosting the dbs02
node.
Start the kvm03
node hosting the dbs03
node.
Note
By default, the autostart
option is enabled for the host
nodes. Therefore, all VMs they host will start automatically.
Otherwise, start the database server nodes in the following strict order:
dbs01
node.dbs02
node.dbs03
node.Verify that the cluster is up. In case of issues, restore the cluster as described in Restore a Galera cluster.
Start the RabbitMQ server msg
nodes.
Verify that the cluster is up. In case of issues, proceed with Troubleshoot RabbitMQ.
Start the gtw
nodes.
Start the OpenStack controller ctl
nodes.
Start the bare metal bmt
nodes.
For the OpenContrail deployments, start the nal
and ntw
nodes.
Verify that the cluster is up. In case of issues, proceed with Troubleshoot OpenContrail.
Start the proxy prx
nodes.
Start the StackLight LMA mon
, mtr
, and log
nodes.
Start the StackLight OSS cid
nodes.