Node Failure Detection Timeout during LPM – [600 seconds]
This setting defines how long the system waits to detect a node failure during a Live Partition Mobility operation.

• A higher timeout (e.g., 600 seconds) is recommended for HACMP clusters and Power9 → Power10 migrations

• Helps prevent false node failure detection during long-running or cross-generation LPM activities

LPM Node Policy – [Unmanage / Manage]

• Unmanage: Temporarily detaches the LPM policy from the node

  • Avoids policy enforcement during HACMP node LPM, troubleshooting, or maintenance

  • Prevents unnecessary validation failures without deleting the policy

• Manage: Re-enables full policy enforcement after migration or maintenance is complete

HACMP Use Case
For HACMP (PowerHA) clustered LPARs, it is a best practice to:

1. Increase the Node Failure Detection Timeout

2. Unmanage the LPM node policy before performing LPM

3. Complete migration/maintenance with no application downtime

4. Re-Manage the LPM policy after successful validation