CERN

Cold reset BMC = failback

If the BMC has failovered to another interface, it’s not going to go back to the dedicated BMC-interface. You can reboot the machine without effect. You can shut it down, without effect. Resetting the BMC via the web interface doesn’t work either. Previously I pulled the power of the unit physically, and this worked. Now, however, I discovered that ‘ipmitool’ can send cold resets, and this actually works.

root@cn012:~# modprobe ipmi-devintf
root@cn012:~# modprobe ipmi-si
root@cn012:~# ipmitool mc reset cold
Sent cold reset command to MC
root@cn012:~# ping -c1 cn012-bmc
PING cn012-bmc.internal (10.162.64.23) 56(84) bytes of data.
64 bytes from cn012-bmc.internal (10.162.64.23): icmp_seq=1 ttl=63 time=0.359 ms
 
--- cn012-bmc.internal ping statistics ---
1 packets transmitted, 1 received, 0% packet loss, time 0ms
rtt min/avg/max/mdev = 0.359/0.359/0.359/0.000 ms