Performing Cluster Maintenance
Feature History
Feature Name | Release Information | Description |
---|---|---|
Support for Performing Cluster Maintenance | Cisco Operations Hub, Release 23.1 | Cisco Operations Hub supports cluster maintenance tasks when there’s cluster node maintenance or node failure. |
This section provides information about how to handle an Operations Hub cluster when there’s cluster node maintenance or node failure.
Maintenance activities for cluster nodes can be classified as:
-
Planned maintenance: This activity is performed when the cluster node is still working but needs to be removed from the Kubernetes cluster for maintenance. Maintenance activity in this case drains the node, remove all the pods running on the node, and marks the node as not schedulable.
-
Unplanned maintenance: This activity is performed when the node is dead. The node state is Not Available, and its IP address is unreachable. In this case, since the node is present in the vCenter but isn’t running, it’s deleted and re-created during maintenance.
Cluster Node Maintenance through Autodeploy Tool
Use the following steps to perform maintenance of cluster through the Autodeploy tool.
-
To trigger cluster maintenance, use the --maintenance option as an argument with the Autodeployer.
~/opshub-installer-<release-version-tag>$ ./deploy -c <config_file> --maintenance
-
Cluster maintenance provides you two options. Select one of the two available options:
-
Node Maintenance: Use this option to push one or more nodes into maintenance. After selecting this option, Autodeployer shows a list of node names and prompts the user to enter a list of nodes to be maintained in a comma-separated way.
-
Host Maintenance: Use this option to push all the nodes running on the ESXi host to maintenance. Typically used when an upgrade is planned for ESXi. After selecting this option, Autodeployer shows a list of ESXi host IPs and FQDNs and prompts the user to select one of the ESXI host details. After selecting ESXi host IP/FQDN, all nodes running on top are pushed into maintenance.
-
Note |
|
Resuming Cluster Node from Maintenance through Autodeploy Tool
Use the following steps to resume cluster nodes from maintenance through Autodeployer.
-
To resume cluster nodes from maintenance, use the --resume-maintenance option as an argument with the Autodeployer.
~/opshub-installer-<release-version-tag>$ ./deploy -c <config_file> --resume-maintenance
This ensures that all the nodes under maintenance are added back to the cluster. If nodes are deleted due to unplanned maintenance, those nodes are recreated.