Methods and systems for establishing connections associated with virtual machine migrations

1. A method of establishing one or more connections between a target host machine and a remote endpoint connected to a virtual machine (VM) running on a source host machine, comprising:

receiving, by a target hypervisor executable on a target host machine, a list of one or more remote endpoints to which a VM executable on a source host machine is connected, the source and target host machines being coupled to an InfiniBand network, and the source host machine and target host machine being remote from the remote endpoint; and

establishing, by a host communication manager executable on the target host machine, a network connection to one or more of the remote endpoints specified in the list before the VM has completed migration from the source host machine to the target host machine, wherein the target host machine includes a first host channel adapter (HCA) including a first queue pair associated with the VM, wherein before the VM has completed migration to the target host machine, the target hypervisor receives the list and the first queue pair is connected to a remote queue pair included in a remote endpoint specified in the list, and wherein the target host machine connects to an InfiniBand fabric of the InfiniBand network via the first HCA.

2. The method of claim 1 , further including:

sending, by the VM executing on the target host machine, a data packet on the connection established to one or more of the remote endpoints specified in the list.

3. The method of claim 2 , wherein the source host machine includes a second host channel adapter (HCA) including a second queue pair associated with the VM, wherein before the VM has completed migration to the target host machine, the second queue pair is connected to the remote queue pair included in the remote endpoint specified in the list, and wherein the source host machine connects to the InfiniBand fabric of the InfiniBand network via the second HCA.

4. The method of claim 3 , wherein the remote endpoint connects to the InfiniBand fabric via a third HCA including the remote queue pair.

5. The method of claim 1 , further including:

identifying, by a source hypervisor executable on the source host machine, the list of one or more remote endpoints;

receiving, by the source hypervisor, an indication that the VM is to be migrated from the source host machine to the target host machine; and

sending, by the source hypervisor, the list of one or more remote endpoints to the target hypervisor before the VM has completed migration from the source host machine to the target host machine.

6. The method of claim 5 , wherein the sending includes during the migration of the VM from the source host machine to the target host machine, sending the list to the target host machine.

7. The method of claim 5 , further including:

detecting, by the source hypervisor, a new connection between the VM and a remote endpoint; and

adding the remote endpoint to the list before the sending.

8. The method of claim 5 , further including:

detecting, by the source hypervisor, a dropped connection between the VM and a remote endpoint specified in the list; and

removing the remote endpoint from the list before the sending.

9. A system for establishing one or more connections between a target host machine and a remote endpoint connected to a VM running on a source host machine, comprising:

a target hypervisor that receives, by one or more processors, a list of one or more remote endpoints to which a VM executable on a source host machine is connected before the VM has completed migration from the source host machine to a target host machine; and

a host communication manager that establishes a network connection to one or more of the remote endpoints specified in the list before the VM has completed migration from the source host machine to the target host machine,

wherein the source and target host machines are coupled to an InfiniBand network, the target hypervisor and host communication manager are executable on the target host machine, and wherein the target host machine includes a first host channel adapter (HCA) including a first queue pair associated with the VM, wherein before the VM has completed migration to the target host machine, the first queue pair is connected to a remote queue pair included in a remote endpoint specified in the list, and wherein the target host machine connects to an InfiniBand fabric of the InfiniBand network via the first HCA.

10. The system of claim 9 , wherein the source host machine is interconnected to the Infiniband fabric of the InfiniBand network.

11. The system of claim 9 , wherein the target hypervisor receives the list before the VM has completed migration to the target host machine, and wherein the source host machine includes a source hypervisor that maintains at least one connection to a remote endpoint specified in the list after the host communication manager initiates a connection to the remote endpoint specified in the list.

12. The system of claim 9 , wherein after the VM has completed migration to the target host machine, the VM sends a data packet on the connection established by the host communication manager, wherein the connection was established before the VM completed migration to the target host machine.

13. The system of claim 9 , wherein the source hypervisor identifies the list of one or more remote endpoints, receives an indication that the VM is to be migrated from the source host machine to the target host machine, and sends the list of one or more remote endpoints to the target hypervisor before the VM has completed migration from the source host machine to the target host machine, wherein the source hypervisor is executable on the source host machine.

14. The system of claim 13 , wherein the source hypervisor sends the list of one or more remote endpoints to the target host machine in response to the indication.

15. The system of claim 13 , wherein the source hypervisor sends the list of one or more remote endpoints to the target host machine before the source hypervisor receives the indication.

16. A non-transitory machine-readable medium comprising a plurality of machine-readable instructions that when executed by one or more processors is adapted to cause the one or more processors to perform a method comprising:

receiving, by a target hypervisor executable on a target host machine, a list of one or more remote endpoints to which a VM executable on a source host machine is connected, the source and target host machines being InfiniBand nodes coupled to an InfiniBand network and

establishing, by a host communication manager executable on the target host machine, a connection to one or more of the remote endpoints specified in the list before the VM has completed migration from the source host machine to the target host machine, wherein the target host machine includes a first host channel adapter (HCA) including a first queue pair associated with the VM, wherein before the VM has completed migration to the target host machine, the target hypervisor receives the list and the first queue pair is connected to a remote queue pair included in a remote endpoint specified in the list, and wherein the target host machine connects to an InfiniBand fabric of the InfiniBand network via the first HCA.

17. The system of claim 9 , wherein the source host machine, target host machine, and remote endpoint are distinct from each other.

18. The system of claim 9 , wherein the source host machine includes a second host channel adapter (HCA) including a second queue pair associated with the VM, wherein before the VM has completed migration to the target host machine, the second queue pair is connected to the remote queue pair included in the remote endpoint specified in the list, and wherein the source host machine connects to the InfiniBand fabric of the InfiniBand network via the second HCA.

19. The non-transitory machine-readable medium of claim 16 , wherein a remote endpoint of the one or more remote endpoints includes a second HCA, the list includes an identifier of the second HCA, and the host communication manager uses the identifier of the second HCA to establish a connection to the remote endpoint specified in the list.