ONTAP Automatic Giveback: Commands and Parameter Interactions
An HA partner normally gives storage back after the offline node returns, reaches the required state and the configured delay expires. Non-default settings interact, so review both nodes and the matching ONTAP release before changing behavior.
Commands from the source
| Purpose | Command |
|---|---|
| Enable ordinary automatic giveback | storage failover modify -node nodename -auto-giveback true |
| Disable ordinary automatic giveback | storage failover modify -node nodename -auto-giveback false |
| Disable automatic giveback after panic | storage failover modify -node nodename -auto-giveback-after-panic false |
| Set the minimum automatic-giveback delay | storage failover modify -node nodename -delay-seconds seconds |
Replace nodename/seconds with verified values. The documented defaults in this source are true for auto-giveback, onreboot, onpanic and auto-giveback-after-panic; delay-seconds defaults to 600. The source identifies the panic-specific option as advanced privilege. Disabling auto-giveback alone does not disable the independently enabled panic-specific path.
Non-panic takeover matrix
| onreboot | auto-giveback | Trigger | Automatic giveback in the referenced table |
|---|---|---|---|
| true | true | ONTAP reboot | Yes |
| true | true | ONTAP halt / Service Processor power cycle | Yes |
| true | false | ONTAP reboot | Yes |
| true | false | ONTAP halt / Service Processor power cycle | No |
| false | true | ONTAP reboot | N/A: no takeover for this reboot path |
| false | true | ONTAP halt / Service Processor power cycle | Yes |
| false | false | ONTAP reboot | No |
| false | false | ONTAP halt / Service Processor power cycle | No |
The last halt/power-cycle row was absent from the stored Chinese table and is restored from its exact official English counterpart. If onreboot=true and reboot triggers takeover, the documented reboot path automatically gives back even when auto-giveback=false. With onreboot=false, a reboot does not trigger that takeover path, so changing auto-giveback cannot restore the missing takeover; clients can be disrupted.
Panic-related source matrix
| onpanic | auto-giveback | auto-giveback-after-panic | Automatic giveback after panic in source matrix |
|---|---|---|---|
| true | true | true | Yes |
| true | true | false | Yes |
| true | false | true | Yes |
| true | false | false | No |
| false | Either | Either | No takeover/giveback through the panic-triggered path |
This reproduces the source’s five-case matrix, also present on the official English overview. The CLI reference describes the panic-specific setting as independent of auto-giveback. Confirm the exact release/event classification and active settings rather than treating the matrix as a reason to ignore either parameter.
Heartbeat failure is a different trigger
The source distinguishes a received panic notification from loss of communication/heartbeat. If the partner never receives a panic packet, it may classify the incident as failure rather than panic. The source overview says that such behavior is controlled by onfailure; more precisely, the CLI defines -onfailure as the switch that enables takeover on partner failure. It is not a separate generic giveback command. Correlate the actual takeover reason and appropriate automatic-giveback path.
English references: official automatic-giveback overview; storage failover modify CLI reference; NetApp automatic-giveback KB.