Method and system for pruned resilvering using a dirty time log
A computer readable medium includes executable instructions for resilvering a hierarchical block tree in a storage pool. Resilvering the hierarchical block tree involves traversing a branch of the hierarchical block tree only if a birth time of a parent block of the branch is greater than or equal to a lowest birth time on a dirty time log (DTL). Resilvering the hierarchical block tree further involves resilvering a child block in the branch only if the DTL comprises a birth time of the child block.
1. A computer readable medium comprising executable instructions for resilvering a hierarchical block tree in a storage pool by:
selectively traversing a branch of the hierarchical block tree when a birth time of a parent block of the branch is greater than or equal to a lowest birth time on a dirty time log (DTL), wherein the DTL is configured to:
track when one or more I/O requests are not successfully completed, and
store birth times for the one or more I/O requests that are not successfully completed, wherein the stored birth times are equivalent to birth times that would have been associated with a successful write to disk; and
resilvering a child block of a plurality of child blocks in the branch when the DTL comprises a birth time of the child block, wherein a birth time of each child block of the plurality of child blocks is less than or equal to the birth time of the parent block,
wherein the parent block comprises the birth time of the child block, and
wherein the DTL is updated only once when more than one I/O request for the child block and the parent block fail at a same birth time.
2. The computer readable medium of claim 1 , wherein the birth time of the parent block corresponds to a transaction group associated with an input/output (I/O) request to store the parent block.
3. The computer readable medium of claim 1 , wherein the birth time of the child block corresponds to a transaction group associated with an input/output (PO) request to store the child block.
4. The computer readable medium of claim 1 , wherein the DTL is associated with a first disk in the storage pool, and wherein resilvering the child block comprises updating the child block using data from a second disk in the storage pool.
5. The computer readable medium of claim 1 , wherein the traversing comprises performing a depth-first traversal of the branch.
6. The computer readable medium of claim 1 , further comprising executable instructions for resilvering a hierarchical block tree in a storage pool by:
resilvering the parent block only if the DTL comprises the birth time of the parent block.
7. A method for resilvering a hierarchical block tree in a storage pool, the method comprising:
obtaining, using a processor, a birth time of each child block of a plurality of child blocks referenced by a parent block in the hierarchical block tree, wherein the birth time of each child block of the plurality of child blocks is less than or equal to a birth time of the parent block; and
for each child block:
a) resilvering, using the processor, the child block when a dirty time log (DTL) comprises the birth time of the child block, wherein the DTL is configured to:
track when one or more I/O requests are not successfully completed, and
store birth times for the one or more I/O requests that are not successfully completed, wherein the birth times are equivalent to birth times that would have been associated with a successful write to disk; and
b) when the birth time of the child block is greater than or equal to a lowest birth time on the DTL:
selectively traversing, using the processor, the hierarchical block tree to obtain a birth time of each grandchild block referenced by the child block, when the child block is an indirect block; and
repeating a) and b) for each grandchild block, if the child block is an indirect block,
wherein the parent block comprises the birth time of the child block, and
wherein the DTL is updated only once when more than one I/O request for the child block and the parent block fail at a same birth time.
8. The method of claim 7 , wherein the birth time of the child block corresponds to a transaction group associated with an input/output (I/O) request to store the child block.
9. The method of claim 7 , wherein the grandchild block is a data block.
10. The method of claim 7 , wherein the DTL is associated with a first disk in the storage pool, and wherein resilvering the child block comprises updating the child block using data from a second disk in the storage pool.
11. The method of claim 7 , wherein the birth time of each child block is obtained only if the birth time of the parent block is greater than or equal to the lowest birth time on the DTL.
12. The method of claim 7 , wherein the hierarchical block tree corresponds to a file in a file system.
13. A computer system comprising:
a processor;
a memory;
a hierarchical block tree in a storage pool, comprising:
a parent block of a branch of the hierarchical block tree; and
a child block of a plurality of child blocks in the branch, wherein a birth time of each child block of the plurality of child blocks is less than or equal to a birth time of the parent block;
a dirty time log (DTL) configured to:
track when one or more I/O requests are not successfully completed, and
store birth times for the one or more I/O requests that are not successfully completed, wherein the stored birth times are equivalent to birth times that would have been associated with a successful write to disk; and
a file system, wherein at least a portion of the file system is resident in the memory and operating under control of the processor, configured to:
selectively traverse the branch when the birth time of the parent block is greater than or equal to a lowest birth time on the DTL; and
resilver the child block when the DTL comprises a birth time of the child block,
wherein the parent block comprises the birth time of the child block, and
wherein the DTL is updated only once when more than one I/O request for the child block and the parent block fail at a same birth time.
14. The computer system of claim 13 , wherein the birth time of the parent block corresponds to a transaction group associated with an input/output (I/O) request to store the parent block.
15. The computer system of claim 13 , wherein the birth time of the child block corresponds to a transaction group associated with an input/output (I/O) request to store the child block.
16. The computer system of claim 13 , wherein the DTL is associated with a first disk in the storage pool, and wherein resilvering the child block comprises updating the child block using data from a second disk in the storage pool.
17. The computer system of claim 13 , wherein the file system is configured to traverse the branch by performing a depth-first traversal of the branch.