When enabling NTP (Network Time Protocol) on a RAC (Real Application Clusters) database running on Solaris 11, downtime is not strictly necessary, but it's highly recommended to minimize potential issues. Here's why:
1. *Node synchronization*: NTP ensures all nodes in the RAC cluster have synchronized clocks. If you enable NTP without downtime, the nodes might not sync immediately, potentially causing issues with database operations.
2. *Clock adjustments*: When NTP is enabled, the system clock might be adjusted to match the reference clock. This adjustment can cause issues with database processes, especially if they're sensitive to time changes.
3. *RAC node evictions*: If the clock adjustments are significant, RAC nodes might be evicted from the cluster, leading to downtime and potential data inconsistencies.
To minimize risks, consider the following:
1. *Schedule downtime*: Plan a maintenance window to enable NTP, ensuring all nodes are updated and synchronized simultaneously.
2. *Use the `ntpd` command*: Instead of enabling NTP through the `svcadm` command, use `ntpd` with the `-r` option to restart the NTP service. This allows for a smoother transition.
3. *Monitor the cluster*: Closely monitor the RAC cluster during and after NTP enablement to quickly address any issues that arise.
In summary, while downtime is not mandatory, it's highly recommended to ensure a smooth transition when enabling NTP on a RAC database running on Solaris 11.
