Maintaining high availability during network partitions for virtual machines stored on distributed object-based storage

Patent metadata

Field	Value
Publication number	US-10417032-B2
Application number	US-201615352046-A
Country	US
Kind code	B2
Filing date	Nov 15, 2016
Priority date	Jun 27, 2014
Publication date	Sep 17, 2019
Grant date	Sep 17, 2019

Abstract

Techniques are disclosed for maintaining high availability (HA) for virtual machines (VMs) running on host systems of a host cluster, where each host system executes a HA module in a plurality of HA modules and a storage module in a plurality of storage modules, where the host cluster aggregates, via the plurality of storage modules, locally-attached storage resources of the host systems to provide logical data store, and where persistent data for the VMs is stored across the locally-attached storage resources comprising the logical data store.

First claim

What is claimed is:

1. A method for managing virtual machine (VM) high availability (HA) in a host cluster comprising a plurality of host systems, the method comprising:
executing, by each host system in the plurality of host systems, a HA module in a plurality of HA modules and a storage module in a plurality of storage modules, wherein the host cluster aggregates, via the plurality of storage modules, locally-attached storage resources of the plurality of host systems to provide a logical data store, wherein persistent data for VMs running on the plurality of host systems is stored across the locally-attached storage resources comprising the logical data store, wherein the locally-attached storage resources of said each host system are housed in said each host system, and wherein the storage module executing on said each host system receives VM I/O requests directed to the logical data store and routes the VM I/O requests to one or more host systems in the host cluster that house one or more locally-attached storage resources storing data associated with the VM I/O requests;
invoking, by a first HA module running on a first host system in the host cluster, an application programming interface (API) for retrieving HA protection states for one or more VMs that are accessible by the first host system, wherein the HA protection states indicate that the one or more VMs should be restarted on an active host system in case of a failure in the host cluster, and wherein the API is exposed by a storage layer comprising the plurality of storage modules, the storage layer being configured to manage the logical data store;
determining, by the storage layer in response to the invoking, the HA protection states by consulting an in-memory database residing on the first host system; and
returning, by the storage layer, the HA protection states to the first HA module.

2. The method of claim 1 wherein the first HA module is a master HA module in the host cluster that has been elected in response to a failure or partitioning of a previous master HA module.

3. The method of claim 1 wherein the determining is performed by the storage layer without accessing the logical data store.

4. The method of claim 1 wherein the in-memory database comprises cached metadata for storage objects of the one or more VMs.

5. The method of claim 4 wherein the cached metadata comprises, for each of the one or more VMs, metadata for a namespace object of the VM, the namespace object including a representation of a file system configured to store the VM's files.

6. The method of claim 5 wherein determining the HA protection states comprises, for each of the one or more VMs:
extracting the VM's HA protection state from the metadata for the VM's namespace object.

7. The method of claim 1 further comprising:
re-invoking, by the first HA module, the API in response to identifying a change in accessibility of one or more per-VM storage objects in the logical data store.

8. A non-transitory computer readable storage medium having stored thereon computer software executable by a processor of a host system, the computer software embodying a method for managing virtual machine (VM) high availability (HA) in a host cluster comprising a plurality of host systems, the method comprising:
executing, by each host system in the plurality of host systems, a HA module in a plurality of HA modules and a storage module in a plurality of storage modules, wherein the host cluster aggregates, via the plurality of storage modules, locally-attached storage resources of the plurality of host systems to provide a logical data store, wherein persistent data for VMs running on the plurality of host systems is stored across the locally-attached storage resources comprising the logical data store, wherein the locally-attached storage resources of said each host system are housed in said each host system, and wherein the storage module executing on said each host system receives VM I/O requests directed to the logical data store and routes the VM I/O requests to one or more host systems in the host cluster that house one or more locally-attached storage resources storing data associated with the VM I/O requests;
invoking, by a first HA module running on a first host system in the host cluster, an application programming interface (API) for retrieving HA protection states for one or more VMs that are accessible by the first host system, wherein the HA protection states indicate that the one or more VMs should be restarted on an active host system in case of a failure in the host cluster, and wherein the API is exposed by a storage layer comprising the plurality of storage modules, the storage layer being configured to manage the logical data store;
determining, by the storage layer in response to the invoking, the HA protection states by consulting an in-memory database residing on the first host system; and
returning, by the storage layer, the HA protection states to the first HA module.

9. The non-transitory computer readable storage medium of claim 8 wherein the first HA module is a master HA module in the host cluster that has been elected in response to a failure or partitioning of a previous master HA module.

10. The non-transitory computer readable storage medium of claim 8 wherein the determining is performed by the storage layer without accessing the logical data store.

11. The non-transitory computer readable storage medium of claim 8 wherein the in-memory database comprises cached metadata for storage objects of the one or more VMs.

12. The non-transitory computer readable storage medium of claim 11 wherein the cached metadata comprises, for each of the one or more VMs, metadata for a namespace object of the VM, the namespace object including a representation of a file system configured to store the VM's files.

13. The non-transitory computer readable storage medium of claim 12 wherein determining the HA protection states comprises, for each of the one or more VMs:
extracting the VM's HA protection state from the metadata for the VM's namespace object.

14. The non-transitory computer readable storage medium of claim 8 wherein the method further comprises:
re-invoking, by the first HA module, the API in response to identifying a change in accessibility of one or more per-VM storage objects in the logical data store.

15. A host system comprising:
a processor; and
a non-transitory computer readable medium having stored thereon program code for managing virtual machine (VM) high availability (HA) in a host cluster comprising a plurality of host systems, the program code comprising code that, when executed by the processor, causes the processor to:
executing, a HA module in a plurality of HA modules and a storage module in a plurality of storage modules, wherein the host cluster aggregates, via the plurality of storage modules, locally-attached storage resources of the plurality of host systems to provide a logical data store, wherein persistent data for VMs running on the plurality of host systems is stored across the locally-attached storage resources comprising the logical data store, wherein the locally-attached storage resources of said each host system are housed in said each host system, and wherein the storage module executing on said each host system receives VM I/O requests directed to the logical data store and routes the VM I/O requests to one or more host systems in the host cluster that house one or more locally-attached storage resources storing data associated with the VM I/O requests;
invoke, via the HA module, an application programming interface (AP

Assignees

• Vmware Inc

Inventors

• Sevigny Marc
• Farkas Keith
• Karamanolis Christos

Classifications

• G06F11/1484

  involving virtual machines · CPC title

• G06F11/1438

  Restarting or rejuvenating · CPC title

• G06F11/202

  where processing functionality is redundant (redundant communication control functionality G06F11/2005, redundant storage control functionality G06F11/2089) · CPC title

• G06F11/16

  Error detection or correction of the data by redundancy in hardware · CPC title

• G06F9/542

  Event management; Broadcasting; Multicasting; Notifications · CPC title