2010-11-09 02:22 PM
If you have an HA pair (a system clustered together) and you don't want the other controller to take over, from the CLI you type: halt -f
That will give you a clean shutdown. Please note, DO NOT type that until you are ready for your system to completely stop serving IO. In an HA pair you will need to type that from each controller.
2010-11-09 02:29 PM
I don't think there is a way to completely power it down. Also check FilerView. halt is the method I use, then we pull the power and unrack or do whatever we need to do to the hardware that required a shutdown. :-)
To my knowledge FAS2020's do not have RLM so if you halt and are not physically there to power cycle the machine and let it boot up then you should have a plan for this. Cause once you halt there is no going back on the 2020....
2010-11-09 08:30 PM
if it is a pair (cluster) then it would help to disable clustering before halting (cf disable). This way you will not have a failover when the one node is halted.
2010-11-09 08:49 PM
There is a BMC which is similar to RLM with the some of the same commands "system console" , etc. use "bmc setup" to set it up.
I like "halt -f" since people often forget to "cf enable" later... so leaving the cluster enabled and using halt -f, after reboot the cluster is still enabled.