Why Clients Trust Beechglen
We were recently working with a client with multiple systems in a Serviceguard environment with a shared EVA4400 array and the array is still under HP warranty. The array had a fairly minor warning condition and a call was logged with HP on how to clear the warning. HP’s answer was a long drawn out procedure involving shutting down an array controller via Command View, pulling power cords, pulling the controller, removing the batteries to let NVRAM drain, and then putting everything back together. HP also wanted the client to execute this procedure immediately at 2 o’clock on a Tuesday afternoon while all systems in the cluster were actively accessing the array.
While that technically may have been a valid procedure to follow, we recommended against it because the risk was much too high to solve a problem that was merely a warning. There was just no way this was going to happen until a maintenance window could be obtained. Instead we went with a soft reboot of the controller during a scheduled outage which cleared the warning. The point of this is we solved the problem within the framework of what should be done, not what can be done.