There are two configurations needed for O2CB cluster stack to work properly. The first kernel setting is panic_on_oops, which you must enable to change a kernel oops into a panic. If a kernel thread required for O2CB crashes, the system is reset to prevent a cluster hang.
The other kernel setting is panic, which specifies the number of seconds after a panic that the system is automatically reset. The default setting is zero, which disables auto-reset, in which case the cluster requires manual intervention. The recommended setting is 30 seconds, but you can set it higher for large systems.
To manually enable panic_on_oops and set a 30-second timeout for reboot on panic:
# echo 1 > /proc/sys/kernel/panic_on_oops
# echo 30 > /proc/sys/kernel/panic
To preserve settings across a system reboot, add entries to a file created in the /etc/sysctl.d/ directory. For example:
# vi /etc/sysctl.d/99-sysctl.conf
kernel.panic_on_oops = 1
kernel.panic = 30
Use the sysctl -p command to cause the changes made to /etc/sysctl.d/99-sysctl.conf to take effect immediately. For example:
# sysctl –p
kernel.panic_on_oops = 1
kernel.panic = 30
Recent Comments