Rebooting a device shuts down all its processes and then restarts them. You can reboot a device from the Manager or from the device's CLI. By default, when you reboot a device, it restarts the entire system. So, until the device is fully up again, there is a break in the device's functions. For example, in case of an inline Sensor with no fail-open kit, the traffic flow is interrupted until the Sensor is up again.
From Trellix IPS 9.1, two reboot options are available:
- Full reboot: This is the default reboot option where the entire system, including the hardware, restarts.
- Hitless reboot: This option restarts only those software processes of a device that require to be restarted. When you do a hitless reboot, the device goes into Layer 2 pass-through mode and restarts the required processes, but the data path continues to pass traffic. After restarting the required processes, it automatically comes out of Layer 2 pass-through mode. This reduces the reboot time and also prevents traffic interruption during the reboot.
If you have enabled autorecovery, a Sensor attempts to recover automatically without loss of traffic, when it determines internal errors or process failures. If it is unable to recover from the error, it goes into Layer 2 pass-through mode or goes through a full reboot. Functionally, hitless reboot is to some extent similar to the autorecovery feature, but they are two independent features. That is, you can use one without enabling the other. One difference is, you can trigger hitless reboot, whereas autorecovery is triggered by the Sensor itself, provided you have this feature enabled. Also, hitless reboot cannot recover a Sensor from internal errors because the Sensor must be in good health for it to undergo hitless reboot.
Notes regarding hitless reboot:
- Hitless reboot is supported only on the following devices:
- M-3050
- M-4050
- M-6050
- M-8000
- Hitless reboot is supported on NS-series Sensors.
Note
For NS-Series Sensors, hitless reboot is not supported when SSL decryption is enabled.
- A Sensor must be in good health for hitless reboot. If the health turns bad during a hitless reboot, it undergoes a full reboot.
- If for some reason, a Sensor is unable to undergo a hitless reboot, a fault is raised in the Manager and the Sensor undergoes a full reboot.
- Hitless reboot is supported only if Layer 2 pass-through monitoring is enabled. Though not mandatory, Trellix recommends that you also enable autorecovery and Sensor watchdog features when using the hitless reboot feature. For information on how to enable these features, see the IPS Administration section.
For the relevant devices, the hitless reboot option is available from the Manager and from the Sensor CLI. This section provides information on how to reboot a device from the Manager. For information on how to reboot from the CLI, see the CLI commands section.