The Sensor often reboots due to crashes in the management module processes, causing a downtime. Using this command, you can enable the restart of the management module processes, which restores the Sensor to normal functionality without going through the complete reboot cycle.
If any of the following four processes have been crashed/killed, the system tries to restart all of them automatically.
Zebra (Individual process - takes 3-5 seconds approx.)
Process Group : If any one of the processes (either control channel, correlation engine, or pktlog) goes down, all the three are stopped and restarted in the following sequence:
Correlation Engine (Process group - takes 15 -20 seconds approx.)
Packet Log Channel (Process group)
Control Channel (Process group)
Syntax:
set mgmtprocessrestart enable
Use this command to enable the restart of the management module processes.
set mgmtprocessrestart disable
Use this command to disable the restart of the management module processes.
In case the Control Channel group restarts, the Alert and PktLog channels flap at the Manager. The reason for this flap is unknown at the Manager, as the process which maintains the means of communication (that is, the alert channel) to the Manager is being restarted. These are not logged at the Sensor in the events.log file.
The process restart is visible only from the Sensor's default logs and not on the CLI.
If the process is crashed/killed more than 3 times in an hour, the Sensor goes for an auto-recovery or warm reboot depending on the configuration.
This feature is enabled by default.
For debugging management issues, you can keep the watchdog off. Sensor remains in bad health and does not try to recover.
Applicable to:
NS-series Sensors