During which two operations is the NVRAM replayed by the high-availability (HA) partner? (Choose two.)

NVRAM replay is a critical process in an ONTAP high-availability (HA) pair to ensure data consistency and prevent data loss after a takeover. The surviving partner node replays the content of the failed node's NVRAM to commit any client-acknowledged writes that were not yet written to disk.

This replay occurs in both negotiated (graceful) takeovers, initiated by an administrator, and non-negotiated (unplanned) takeovers. A sudden power loss is a primary cause of a non-negotiated takeover, mandating an NVRAM replay to recover in-flight writes. Therefore, both a negotiated takeover and a power loss event trigger the partner node to perform an NVRAM replay.

A. A sync core is a diagnostic tool that causes a panic and subsequent takeover, but the replay is part of the takeover operation, not the sync core itself.

B. During a clean halt, the node flushes all data from its NVRAM to disk before shutting down, so no replay is required by the partner.

1. NetApp TR-4705: ONTAP High-Availability Technical Overview, Page 11, Section: "Takeover Operations".

The document states: "In both negotiated and non-negotiated takeovers, the surviving node replays the NVRAM log of the failed node to disk. This action ensures that any data that was acknowledged to the client but not yet written to disk is committed, preventing data loss." This directly supports both chosen answers.

2. NetApp ONTAP 9 Documentation, "High-availability concepts > How takeover and giveback works".

The documentation for unplanned takeovers states: "In an unplanned takeover (for example, due to a hardware or software failure, or power loss on a node)... The partner's NVRAM/NVMEM is replayed on the healthy node to commit any uncommitted write operations to the partner's disks." This confirms that a power loss (D) triggers a replay. The same principles of ensuring consistency apply to negotiated takeovers (C).