System and method for asynchronous reads of old data blocks updated through a write-back cache
View Patent ↗A system for asynchronous reads of old data blocks updated through a write-back cache includes a storage device, a write-back cache, a storage consumer, a storage processing node, and device management software. The device management software may be configured to store a new version of a data block in the write-back cache in response to an update request from the first storage consumer and to then send an update completion notification to the first storage consumer. Some time after the update completion notification has been sent, the device management software may be configured to send a previous version of the updated data block to the storage processing node, where it may be required to perform an operation such as a copy-on write operation or a snapshot update.
1. A system comprising:
a storage device;
a storage consumer configured to generate update requests to write data blocks to the storage device;
a write-back cache configured to temporarily store the data blocks in response to the update requests generated by the storage consumer;
a storage processing node; and
device management software stored on a storage medium and executable to communicate with the storage consumer and the write-back cache,
wherein the device management software is further executable to:
store a new version of a first data block of the storage device in the write-back cache in response to a first update request from the storage consumer, wherein the new version of the first data block is provided by the storage consumer;
send an update completion notification corresponding to the first update request to the storage consumer;
read a previous version of the first data block from the storage device and provide the previous version of the first data block to the storage processing node in response to a detection that the write-back cache has reached a first threshold level of dirty data blocks after the update completion notification has been sent to the storage consumer and before the new version of the first data block is stored in the storage device; and
overwrite the previous version of the first data block in the storage device with the new version of the first data block after the update completion notification has been sent to the storage consumer and after the previous version of the first data block is provided to the storage processing node.
2. The system as recited in claim 1 , wherein the write-back cache is a non-volatile write-back cache.
3. The system as recited in claim 1 , wherein the storage device is a file server.
4. The system as recited in claim 1 , wherein the storage device is a logical volume.
5. The system as recited in claim 1 , wherein the storage device is a physical disk.
6. The system as recited in claim 1 , wherein the storage device is a disk array.
7. The system as recited in claim 1 , wherein the storage device is an object storage device.
8. The system as recited in claim 1 , wherein the storage processing node is configured to store the previous version of the first data block as part of a copy-on-write (COW) operation.
9. The system as recited in claim 1 , wherein the storage processing node is configured to use the previous version of the first data block to maintain a snapshot of contents of the storage device.
10. The system as recited in claim 1 , wherein the device management software is executable to provide the previous version of the data block to the storage processing node in response to a request from the storage processing node.
11. The system as recited in claim 1 , wherein the first threshold level of dirty data blocks is specified by a first configuration parameter, and wherein the device management software is further executable to write dirty data blocks from the write-back cache to the storage device in response to the detection that the write-back cache has reached a first threshold level of dirty data blocks until a number of remaining dirty data blocks is reduced to a level specified by a second configuration parameter.
12. The system as recited in claim 1 , wherein the storage processing node is configured to implement RAID functionality whereby a new parity block is computed by performing an XOR operation on the previous version of the first data block, the new version of the first data block, and a previous version of the parity block.
13. A computer-accessible storage medium for use with a computer system including a storage device, a storage consumer configured to generate update requests to write data blocks to the storage device, a write-back cache configured to temporarily store the data blocks in response to the update requests generated by the storage consumer, and a storage processing node, wherein the storage medium stores program instructions executable to:
store a new version of a first data block of the storage device in the write-back cache in response to a first update request from the storage consumer, wherein the new version of the first data block is provided by the storage consumer;
send an update completion notification corresponding to the first update request to the storage consumer;
read a previous version of the first data block from the storage device and provide the previous version of the first data block to the storage processing node in response to a detection that the write-back cache has reached a first threshold level of dirty data blocks after the update completion notification has been sent to the storage consumer and before the new version of the first data block is stored in the storage device; and
overwrite the previous version of the first data block in the storage device with the new version of the first data block after the update completion notification has been sent to the storage consumer and after the previous version of the first data block is provided to the storage processing node.
14. The computer-accessible storage medium as recited in claim 13 , wherein the write-back cache is a non-volatile write-back cache.
15. The computer-accessible storage medium as recited in claim 13 , wherein the storage device is a file server.
16. The computer-accessible storage medium as recited in claim 13 , wherein the storage device is a logical volume.
17. The computer-accessible storage medium as recited in claim 13 , wherein the storage device is a physical disk.
18. The computer-accessible storage medium as recited in claim 13 , wherein the storage device is a disk array.
19. The computer-accessible storage medium as recited in claim 13 , wherein the storage device is an object storage device.
20. The computer-accessible storage medium as recited in claim 13 , wherein the storage processing node is configured to store the previous version of the first data block as part of a copy-on-write (COW) operation.
21. The computer-accessible storage medium as recited in claim 13 , wherein the storage processing node is configured to use the previous version of the first data block to maintain a snapshot of contents of the storage device.
22. The computer-accessible storage medium as recited in claim 13 , wherein the program instructions are executable to provide the previous version of the data block to the storage processing node in response to a request from the storage processing node.
23. The system as recited in claim 11 , wherein selected ones of the dirty data blocks are each associated with a special tag in the write-back cache to indicate particular handling is required prior to an overwrite, wherein the device management software is further executable to detect the special tag associated with each of the selected ones of the dirty data blocks and to responsively provide a previous version of each of the selected ones of the dirty data blocks to the storage processing node prior to writing each of the selected ones of the dirty data blocks to the storage device.
24. The computer-accessible storage medium as recited in claim 13 , wherein the first threshold level of dirty data blocks is specified by a first configuration parameter, and wherein the program instructions are further executable to write dirty data blocks from the write-back cache to the storage device in response to the detection that the write-back cache has reached a first threshold level of dirty data blocks until a number of remaining dirty data blocks is reduced to a level specified by a second configuration parameter.
25. A method for operating a computer system including a storage device, a storage consumer configured to generate update requests to write data blocks to the storage device, a write-back cache configured to temporarily store the data blocks in response to the update requests generated by the storage consumer, and a storage processing node, wherein the method comprises:
storing a new version of a first data block of the storage device in the write-back cache in response to a first update request from the storage consumer, wherein the new version of the first data block is provided by the storage consumer;
sending an update completion notification corresponding to the first update request to the storage consumer;
reading a previous version of the first data block from the storage device and providing the previous version of the first data block to the storage processing node in response to a detection that the write-back cache has reached a first threshold level of dirty data blocks after the update completion notification has been sent to the storage consumer and before the new version of the first data block is stored in the storage device; and
overwriting the previous version of the first data block in the storage device with the new version of the first data block after the update completion notification has been sent to the storage consumer and after the previous version of the first data block is provided to the storage processing node.
26. The method as recited in claim 25 , wherein the write-back cache is a non-volatile write-back cache.
27. The method as recited in claim 25 , wherein the storage device is a file server.
28. The method as recited in claim 25 , wherein the storage device is a logical volume.
29. The method as recited in claim 25 , wherein the storage device is a physical disk.
30. The method as recited in claim 25 , wherein the storage device is a disk array.
31. The method as recited in claim 25 , wherein the storage device is an object storage device.
32. The method as recited in claim 25 , wherein the storage processing node is configured to store the previous version of the first data block as part of a copy-on-write (COW) operation.
33. The method as recited in claim 25 , wherein the storage processing node is configured to use the previous version of the first data block to maintain a snapshot of contents of the storage device.
34. The method as recited in claim 25 , further comprising providing the previous version of the data block to the storage processing node in response to a request from the storage processing node.
35. The computer-accessible storage medium as recited in claim 24 , wherein selected ones of the dirty data blocks are each associated with a special tag in the write-back cache to indicate particular handling is required prior to an overwrite, wherein the program instructions are further executable to detect the special tag associated with each of the selected ones of the dirty data blocks and to responsively provide a previous version of each of the selected ones of the dirty data blocks to the storage processing node prior to writing each of the selected ones of the dirty data blocks to the storage device.