When a defective compute node needs to be replaced or repaired, or when a compute node is retired in favor of a newer model with higher capacity and better performance, it is highly recommended that you deprovision the compute node before removing it from the appliance rack. Deprovisioning ensures that all configuration entries for a compute node are removed cleanly, so that no conflicts are introduced when a replacement compute node is installed.
Deprovisioning a Compute Node for Repair or Replacement
Log into the Oracle VM Manager user interface.
For detailed instructions, see Section 5.2, “Logging in to the Oracle VM Manager Web UI”.
Migrate all virtual machines away from the compute node you wish to deprovision. If any VMs are running on the compute node, the deprovision command fails.
Using SSH and an account with superuser privileges, log into the active management node, then launch the Oracle PCA command line interface.
# ssh root@10.100.1.101 root@10.100.1.101's password: root@ovcamn05r1 ~]# pca-admin Welcome to PCA! Release: 2.3.2 PCA>
Lock provisioning to make sure that the compute node cannot be reprovisioned immediately after deprovisioning.
PCA> create lock provisioning Status: Success
Deprovision the compute node you wish to remove. Repeat for additional compute nodes, if necessary.
PCA> deprovision compute-node ovcacn29r1 ************************************************************ WARNING !!! THIS IS A DESTRUCTIVE OPERATION. ************************************************************ Are you sure [y/N]:y Shutting down dhcpd: [ OK ] Starting dhcpd: [ OK ] Shutting down dnsmasq: [ OK ] Starting dnsmasq: [ OK ] Status: Success
When the necessary compute nodes have been deprovisioned successfully, release the provisioning lock. The appliance resumes its normal operation.
PCA> delete lock provisioning ************************************************************ WARNING !!! THIS IS A DESTRUCTIVE OPERATION. ************************************************************ Are you sure [y/N]:y Status: Success
When the necessary repairs have been completed, or when the replacement compute nodes are ready, install the compute nodes into the rack and connect the necessary cables. The controller software detects the new compute nodes and automatically launches the provisioning process.