METHOD FOR TRACKING CHANGES IN VIRTUAL DISKS

• US 20100228913A1
• Filed: 06/23/2009
• Published: 09/09/2010
• Est. Priority Date: 03/06/2009
• Status: Active Grant

• Current Assignee: Vmware LLC (Broadcom, Inc.)
• Original Assignee: VMware, Inc. (Broadcom, Inc.)
• Inventors: CZEZATKE, Christian, Tucker, Andrew, Yadappanavar, Krishna
• Granted Patent: US 8,443,166 B2
• Time in Patent Office: Days
• US Class Current: 711/112
• CPC Class Codes: G06F 11/1451 by selection of backup contents
• CPC Class Codes: G06F 11/1458 Management of the backup or restore process
• CPC Class Codes: G06F 11/1482 by means of middleware or OS functionality
• CPC Class Codes: G06F 11/1484 involving virtual machines
• CPC Class Codes: G06F 12/10 Address translation
• CPC Class Codes: G06F 12/122 of the least frequently used [LFU] type, e.g. with individual count value
• CPC Class Codes: G06F 2201/815 Virtual
• CPC Class Codes: G06F 3/065 Replication mechanisms
• CPC Class Codes: G06F 3/0653 Monitoring storage devices or systems
• CPC Class Codes: G06F 9/45533 Hypervisors; Virtual machine monitors

Abstract

Systems and methods for tracking changes and performing backups to a storage device are provided. For virtual disks of a virtual machine, changes are tracked from outside the virtual machine in the kernel of a virtualization layer. The changes can be tracked in a lightweight fashion with a bitmap, with a finer granularity stored and tracked at intermittent intervals in persistent storage. Multiple backup applications can be allowed to accurately and efficiently backup a storage device. Each backup application can determine which block of the storage device has been updated since the last backup of a respective application. This change log is efficiently stored as a counter value for each block, where the counter is incremented when a backup is performed. The change log can be maintained with little impact on I/O by using a coarse bitmap to update the finer grained change log.

Claims

1. A method for tracking changes made to a virtual disk of a virtual machine running on a computer system, the method comprising:

receiving a request to write data to a first virtual block of the virtual disk, wherein the virtual disk contains a plurality of virtual blocks whose data is stored in blocks of a storage subsystem;

mapping a first write command to the first virtual block to a second write command to a first block of the storage subsystem;

sending, from a virtualization layer, the second write command to the storage subsystem;

receiving, at the virtualization software layer, a write completion response from the storage subsystem; and

based on the write completion response, maintaining tracking information with the virtualization software layer, wherein the tracking information indicates whether each of the plurality of virtual blocks has been written to since an event.

2. The method of claim 1, wherein the blocks of the storage system are physical blocks of physical disks of the storage subsystem or are byte regions within files which in turn are translated to physical blocks by the storage subsystem.

3. The method of claim 1, wherein the tracking information is updated when a write completion response is received at the virtualization software layer from the storage subsystem.

4. The method of claim 1, wherein the tracking information is stored as a bitmap having one bit for each virtual block.

5. The method of claim 1, wherein the tracking information is stored as a bloom filter.

6. The method of claim 1, wherein the event is a backup, copy, move operation on the virtual disk, or fault tolerant failover.

7. The method of claim 1, further comprising:

resetting the tracking information when a new event occurs; and

incrementing an event counter.

8. The method of claim 7, further comprising:

updating a change log when a new event occurs, wherein updating the change log uses the tracking information, and wherein the change log includes, for each virtual block, a change integer having the value of the event counter when the virtual block was last written.

9. The method of claim 8, wherein the change log is stored on a persistent storage device.

10. The method of claim 8, wherein the change log further includes an ID, further comprising:

recording the ID each time a first backup routine is performed for the virtual disk; and

performing a backup of every virtual block when the recorded ID is not equal to the ID included in the change log.

11. The method of claim 10, wherein the change log further includes a clean/dirty bit, further comprising:

upon opening the virtual disk, checking the state of the clean/dirty bit;

when the clean/dirty bit is dirty, initializing the tracking information and the change log; and

creating a new ID and storing the ID in the change log.

12. The method of claim 8, further comprising:

recording the value of the event counter each time a first backup routine is performed for the virtual disk; and

performing a backup of the virtual disk by retrieving the data for each of the virtual blocks that have a change integer higher than the recorded value of the event counter for the last backup performed by the first backup routine.

13. The method of claim 12, wherein performing a backup of the virtual disk is accomplished with a second backup routine that is different than the first backup routine.

14. The method of claim 13, further comprising:

using the second backup routine to perform multiple backups of the virtual disk since the last backup performed by the first backup routine, wherein each backup using the second backup routine retrieves data for each of the virtual blocks that have a change integer higher than the recorded value of the event counter for the last backup performed by the first backup routine.

15. The method of claim 1, further comprising porting the virtual machine to another computer system, wherein porting includes:

replicating data of the virtual disk to another storage subsystem by;

starting a first copy session for replicating the virtual disk on the other storage subsystem;

tracking changes made to the virtual disk from the start of the copy session, wherein the changes are tracked using the tracking information in the kernel;

ending the first copy session;

starting a second copy session that uses the tracking information to replicate virtual blocks that have changed during the first copy session;

tracking changes made to the virtual disk during the second copy session;

continue performing additional copy sessions and tracking of changes made to the virtual disk during these additional copy sessions until a threshold is reached; and

performing a last copy session without tracking changes, wherein the virtual machine is prevented from issuing new write operations while the last copy session is performed.

16. A computer program product comprising a computer readable medium encoded with a plurality of instructions for controlling a processor to perform an operation for tracking changes made to a virtual disk of a virtual machine, the instructions comprising:

receiving a request to write data to a first virtual block of the virtual disk, wherein the virtual disk contains a plurality of virtual blocks whose data is stored in blocks of a storage subsystem;

mapping a first write command to the first virtual block to a second write command to a first block of the storage subsystem;

sending, from a virtualization software layer, the second write command to the storage subsystem;

receiving, at the virtualization software layer, a write completion response from the storage subsystem; and

based on the write completion response, maintaining tracking information in the virtualization software layer, wherein the tracking information indicates whether each of the plurality of virtual blocks has been written to since an event.

17. A method of backing up data on a storage device, the method comprising:

incrementing a counter each time a backup of a plurality of blocks of a storage device is performed;

tracking a first value of the counter when a first backup routine performs a backup of the storage device;

tracking a second value of the counter when a second backup routine performs a backup of the storage device;

for each block, storing the counter value of when data for a block was last changed; and

performing a backup with the first backup routine by retrieving the data for each of the blocks that have a counter value higher than the first value for the last backup performed by the first backup routine.

18. The method of claim 17, wherein the storage device is a virtual disk.

19. The method of claim 17, further comprising:

tracking which blocks are changed during each value of the counter; and

updating the counter value stored for each block by storing the current counter value for each block changed.

20. The method of claim 19, information as to which blocks are changed during each value of the counter is stored as a bitmap having one bit for each block.

21. The method of claim 20, further comprising:

resetting the bitmap when a backup is performed.

22. The method of claim 17, further comprising:

storing a clean/dirty bit for the storage device;

upon opening the storage device, checking the state of the clean/dirty bit;

when the clean/dirty bit is dirty, resetting the counter values for each block; and

creating and storing a new ID associated with the storage device.

23. The method of claim 22, further comprising:

tracking the value of the ID when either backup routine performs a backup of the storage device;

when the tracked ID is not equal to the stored new ID, performing a backup by retrieving the data for all of the blocks of the storage device.