[Starlingx-discuss] Controller recovery question

Perez Carranza, Jose jose.perez.carranza at intel.com
Tue Oct 9 15:01:06 UTC 2018


I'm running a test scenario where a critical process is killed and I have to verify that Controller is recovered, with below steps I'm seeing that swat happens correctly but controller with the process killed keeps rebooting and process killed is not regenerated, my question is if the scenario is valid and the sm script should be regenerated.


Perform the following steps on the standy controller:

a) Locate a critical process that is pmon managed:

-> cd /etc/pmon.d
-> grep -r "severity = critical"

b) E.g. cat /etc/pmon.d/sm.conf
c) Locate the name of the daemon pid file or startup script file that is configured in the conf file:
e.g. server.pid-file = "/var/run/lighttpd.pid" or script = /etc/init.d/sm
c) Rename the startup script file so that the critical process is not respawned after it is killed:
e.g. mv /etc/init.d/sm /etc/init.d/sm~

1) Kill the critical process on the standby controller
sudo pkill sm

****    EXPECTED *********
1) Verify that the controller goes for a reset and recovers.


More information about the Starlingx-discuss mailing list