On 03/13/2012 01:08 AM, Peter Stuge wrote:
Igor Mammedov wrote:
not even sure we should halt at all. May be warn the user and continue
BIOSes usually stop in case of error
That doesn't mean that it's a good idea.
You might be right, however I'd call it expected behavior.
Imagine VM that need processing power of several cpus to provide adequate service. During its runtime several cpus were hot-plugged and it worked like a charm. Then during scheduled maintenance VM was soft-rebooted and it seems that VM is fully up (nobody noticed that cpus # are less then expected) and promoted to servicing usual load. If I were admin of this vm, I'd rather fix easy to spot problem during maintenance time than find out later that VM is under-powered and deal with angry customers.
//Peter