Managing the migration of virtual machines in the presence of uncorrectable memory errors
Techniques for migrating virtual machines (VMs) in the presence of uncorrectable memory errors are provided. According to one set of embodiments, a source host hypervisor of a source host system can determine, for each guest memory page of a VM to be migrated from the source host system to a destination host system, whether the guest memory page is impacted by an uncorrectable memory error in a byte-addressable memory of the source host system. If the source host hypervisor determines that the guest memory page is impacted, the source host hypervisor can transmit a data packet to a destination host hypervisor of the destination host system that includes error metadata identifying the guest memory page as being corrupted. Alternatively, if the source host hypervisor determines that the guest memory page is not impacted, the source host hypervisor can attempt to read the guest memory page from the byte-addressable memory in a memory exception-safe manner.
1. A method comprising, at a time of executing a live migration of a virtual machine (VM) from a source host system to a destination host system:
determining, by a source hypervisor running on the source host system, that a guest memory page of the VM is mapped to a physical memory page of the source host system having an uncorrectable memory error; and
in response to the determining, transmitting, by the source hypervisor to a destination hypervisor running on the destination host system, error metadata identifying the guest memory page as being at least partially corrupted.
2. The method of claim 1 wherein the determining is performed prior to a pre-copy phase of the live migration, and
wherein the transmitting is performed during an initial iteration of the pre-copy phase.
3. The method of claim 1 wherein the determining is performed during a pre-copy phase of the live migration, before the guest memory page has been copied to the destination host system as part of an initial iteration of the pre-copy phase, and
wherein the transmitting is performed during the initial iteration.
4. The method of claim 1 wherein the determining is performed during a pre-copy phase of the live migration, after the guest memory page has been copied to the destination host system as part of an initial iteration of the pre-copy phase,
wherein the guest memory page is dirtied by a guest operating system (OS) of the VM after the initial iteration, and
wherein the transmitting is performed during a subsequent iteration of the pre-copy phase after the initial iteration.
5. The method of claim 1 wherein the determining is in response to a consumed memory exception raised during a pre-copy phase of the live migration.
6. The method of claim 1 wherein the determining is in response to a consumed memory exception raised during a page-in phase of the live migration, and
wherein the error metadata is transmitted to the destination host system as part of a remote page fault return message.
7. The method of claim 1 wherein upon completion of the live migration of the VM from the source host system to the destination host system, the destination hypervisor simulates the uncorrectable memory error for a guest OS of the migrated VM.
8. A non-transitory computer readable storage medium having stored thereon program code executable by source hypervisor of a source host system, the program code embodying a method comprising, at a time of executing a live migration of a virtual machine (VM) from the source host system to a destination host system:
determining that a guest memory page of the VM is mapped to a physical memory page of the source host system having an uncorrectable memory error; and
in response to the determining, transmitting, to a destination hypervisor running on the destination host system, error metadata identifying the guest memory page as being at least partially corrupted.
9. The non-transitory computer readable storage medium of claim 8 wherein the determining is performed prior to a pre-copy phase of the live migration, and
wherein the transmitting is performed during an initial iteration of the pre-copy phase.
10. The non-transitory computer readable storage medium of claim 8 wherein the determining is performed during a pre-copy phase of the live migration, before the guest memory page has been copied to the destination host system as part of an initial iteration of the pre-copy phase, and
wherein the transmitting is performed during the initial iteration.
11. The non-transitory computer readable storage medium of claim 8 wherein the determining is performed during a pre-copy phase of the live migration, after the guest memory page has been copied to the destination host system as part of an initial iteration of the pre-copy phase,
wherein the guest memory page is dirtied by a guest operating system (OS) of the VM after the initial iteration, and
wherein the transmitting is performed during a subsequent iteration of the pre-copy phase after the initial iteration.
12. The non-transitory computer readable storage medium of claim 8 wherein the determining is in response to a consumed memory exception raised during a pre-copy phase of the live migration.
13. The non-transitory computer readable storage medium of claim 8 wherein the determining is in response to a consumed memory exception raised during a page-in phase of the live migration, and
wherein the error metadata is transmitted to the destination host system as part of a remote page fault return message.
14. The non-transitory computer readable storage medium of claim 8 wherein upon completion of the live migration of the VM from the source host system to the destination host system, the destination hypervisor simulates the uncorrectable memory error for a guest OS of the migrated VM.
15. A source host system comprising:
a source hypervisor;
a memory; and
a non-transitory computer readable medium having stored thereon program code executable by the source hypervisor at a time of performing a live migration of a virtual machine (VM) from the source host system to a destination host system, the program code causing the source hypervisor to:
determine that a guest memory page of the VM is mapped to a physical memory page of the memory having an uncorrectable memory error; and
in response to the determining, transmit, to a destination hypervisor running on the destination host system, error metadata identifying the guest memory page as being at least partially corrupted.
16. The source host system of claim 15 wherein the determining is performed prior to a pre-copy phase of the live migration, and
wherein the transmitting is performed during an initial iteration of the pre-copy phase.
17. The source host system of claim 15 wherein the determining is performed during a pre-copy phase of the live migration, before the guest memory page has been copied to the destination host system as part of an initial iteration of the pre-copy phase, and
wherein the transmitting is performed during the initial iteration.
18. The source host system of claim 15 wherein the determining is performed during a pre-copy phase of the live migration, after the guest memory page has been copied to the destination host system as part of an initial iteration of the pre-copy phase,
wherein the guest memory page is dirtied by a guest operating system (OS) of the VM after the initial iteration, and
wherein the transmitting is performed during a subsequent iteration of the pre-copy phase after the initial iteration.
19. The source host system of claim 15 wherein the determining is in response to a consumed memory exception raised during a pre-copy phase of the live migration.
20. The source host system of claim 15 wherein the determining is in response to a consumed memory exception raised during a page-in phase of the live migration, and
wherein the error metadata is transmitted to the destination host system as part of a remote page fault return message.
21. The source host system of claim 15 wherein upon completion of the live migration of the VM from the source host system to the destination host system, the destination hypervisor simulates the uncorrectable memory error for a guest OS of the migrated VM.