Description                                                    Configure fail-safe automatic recovery.

Syntax                                                                  [no] fail-safe
fpga-buff-recovery-threshold 256-buffer-units |
hw-error-recovery-timeout minutes |
session-memory-recovery-threshold percentage |
sw-error-monitor-enable |
sw-error-recovery-timeout minutes |
total-memory-size-check Gb {kill | log}




Minimum required number of free (available) FPGA buffers. If the num­ber of free buffers remains below this value until the recovery timeout, fail-safe software recovery is triggered.

You can specify 1-10 units. Each unit contains 256 buffers.

The default is 2 units (512 buffers).


Disables fail-safe monitoring and recovery for hardware errors.

This is enabled by default.


Enables fail-safe monitoring and recovery for hardware errors.

This is enabled by default.

hw-error-recovery-timeout minutes 

Number of minutes fail-safe waits after a hardware error occurs to reboot the ACOS device. You can specify 1-1440 minutes.

The default is 0 (not set).

session-memory-recovery-threshold percentage

Minimum required percentage of system memory that must be free. If the amount of free memory remains below this value long enough for the recovery timeout to occur, fail-safe software recovery is triggered.

You can specify 1-100 percent. The default is 30 percent.


Enables fail-safe monitoring and recovery for software errors.

This is disabled by default.

sw-error-recovery-timeout minutes 

Number of minutes (1-1440) the software error condition must remain in effect before fail-safe occurs:

  If the system resource that is low becomes free again within the recovery timeout period, fail-safe allows the ACOS device to continue normal operation. Fail-safe recovery is not triggered.

  If the system resource does not become free, then fail-safe recovery is triggered.

The default timeout is 3 minutes.

total-memory-size-check Gb
{kill | log}

Amount of memory the device must have after booting.

  Gb - Minimum amount of memory required.

  kill – Stops data traffic and generates a message. However, the management port remains accessible.

  log – Generates a log message but does not stop data traffic.

Default                                                                By default, fail-safe automatic recovery is enabled for hardware errors and disabled for soft­ware errors. You can enable the feature for hardware errors, software errors, or both. When you enable the feature, the other options have the default values described in the table above.

Mode                                                                   Configuration mode

Usage                                                                  Fail-safe hardware recovery also can be triggered by a “PCI not ready” condition. This fail-safe recovery option is enabled by default and can not be disabled.

Table of Contents