More actions
Proxmox High Availability (HA) Guide
Overview
Proxmox High Availability (HA) ensures continuous availability of virtual machines (VMs) within a Proxmox Virtual Environment (PVE) cluster by automatically migrating them to healthy nodes in the event of a node failure. This guide provides an in-depth understanding of HA concepts, configuration options, and best practices.
Requirements
Before setting up Proxmox High Availability (HA), ensure that your environment meets the following requirements:
Proxmox Cluster: HA requires a Proxmox Virtual Environment (PVE) cluster with multiple nodes. Network Connectivity: Ensure reliable network connectivity between cluster nodes for communication. Shared Storage: Use shared storage (e.g., NFS, iSCSI) accessible to all nodes to store VM disks and configuration files. Hardware Compatibility: Verify that your hardware supports features like IPMI for fencing operations. Recommendations
For optimal performance and reliability, consider the following recommendations:
Redundant Networking: Implement redundant network connections to avoid single points of failure. Quorum Configuration: Configure quorum settings appropriately to prevent split-brain scenarios. Fencing: Implement fencing mechanisms to isolate failed nodes and ensure data integrity. Regular Testing: Test HA failover scenarios regularly to verify system behavior and performance. Key Concepts
HA Manager
The HA Manager is a critical component responsible for monitoring the status of nodes within the cluster and orchestrating failover processes when necessary. It ensures seamless migration of VMs to maintain service availability.
Quorum
Quorum refers to the minimum number of healthy nodes required for the HA Manager to make decisions regarding failover. It helps prevent split-brain scenarios and ensures consistent cluster behavior.
Ticket
A ticket is a token granted by the HA Manager to nodes, allowing them to control and manage resources within the cluster. Nodes must possess a valid ticket to participate in the HA process.
Configuration Options
Node Configuration
To enable HA on a node, navigate to the Proxmox GUI or use the CLI and configure the following settings:
Enable HA: Activate HA functionality on the node. HA Manager IP/Port: Specify the IP address and port of the HA Manager for communication. Example CLI command:
pvecm set ha-enabled 1 pvecm set ha-manager-ip 192.168.1.10 pvecm set ha-manager-port 2224
Quorum Configuration
Configure quorum settings to ensure reliable decision-making in the cluster. Adjust the following parameters as needed:
Minimum/Maximum Nodes: Define the minimum and maximum number of nodes required for quorum. Quorum Policy: Specify the policy for determining quorum, such as Node Majority or Node Weight. Example CLI command:
pvecm set min_nodes 3 pvecm set max_nodes 5 pvecm set quorum_policy node majority
Fencing
Fencing is crucial for isolating failed nodes to prevent data corruption. Configure fencing settings based on your environment requirements:
Fencing Method: Choose from power-based fencing, storage-based fencing, or custom methods. Fencing Agent: Specify the agent responsible for executing fencing operations, such as IPMI or SBD. Fencing Delay: Set a delay before initiating fencing operations to allow for graceful shutdowns. Example CLI command:
bash pvecm set no_fencing 0 pvecm set fence_delay 30
Failover Priority
Define failover priorities to ensure critical VMs are migrated first during failover events. Assign priorities to VMs based on their importance to your infrastructure.
Example CLI command:
qm set <VMID> --ha-priority <priority>
Conclusion
Proxmox High Availability (HA) is a robust feature that enhances the reliability and availability of virtualized environments. By understanding the key concepts and properly configuring HA settings, administrators can ensure seamless operation and minimal downtime, even in the face of node failures.