[ClusterLabs] Watchdog & Reset
jorge.fabregas at gmail.com
Tue Sep 8 09:29:37 EDT 2015
I'm doing some tests with SBD on SLES 11 SP4 (specially on how it
behaves with the "hardware" watchdog timer) but I went back to a prior
snapshot (fresh installation) where I removed the Intel 6300esb timer
definition from qemu & I disabled nmi_watchdog by putting the proper
argument to GRUB's kernel line. I'm also not running any cluster
services (sbd/corosync,pacemaker etc).
However, if I do "echo c > /proc/sysrq-trigger" the VM will reset in a
couple of seconds. This is a total mystery to me: Who's feeding the
watchdog timer? What or where's the watchdog timer since there's none
defined? And this is not a qemu monitoring thing since it doesn't
happen with other VMs.
If anyone could clarify this for me I'd be very glad.
More information about the Users