Avoiding server downtime from hardware errors in system memory with HP Memory Quarantine

4
Application isolation
HP Memory Quarantine allows MCA Recovery-aware operating systems to avoid complete system
shutdown by isolating the impact of a hard memory error to an application. A hard memory error
occurring in a location used by an application (red x in Figure 2) will result in the operating system
shutting down that application and then restarting it, avoiding the use of the bad memory location.
No other application is affected.
Figure 2: Application isolation with HP Memory Quarantine
Virtual machine isolation
HP Memory Quarantine allows MCA Recovery-aware hypervisors to avoid complete system shutdown
by isolating the impact of a hardware memory error to a VM. A hard memory error occurring in a
location used by a VM (red x in Figure 3) will result in the hypervisor shutting down that VM and then
restarting it, avoiding the use of the bad memory location. No other VM is affected.
Figure 3: Virtual machine isolation with HP Memory Quarantine
Virtual Machine A
Intel Xeon processor with MCA Recovery
BIOS with HP Memory Quarantine
Operating System
App
App
App
Hypervisor with MCA Recovery support
VM B
VM C
VM C
System Memory
X
NOTE: An uncorrectable error can
bring the server down if the fault is
at a memory location used for
certain critical operations by the
hypervisor.
Operating system with MCA Recovery support
Application
System Memory
Intel Xeon processor with MCA Recovery
BIOS with HP Memory Quarantine
Application
Application
Application
X
NOTE: An uncorrectable error can
bring the server down if the fault is
at a memory location used for
certain critical operations by the OS
kernel.