Method and system for data consistency across failure and recovery of infrastructure
A method and system for data consistency across failure and recovery of infrastructure. In one embodiment of the method, copies of first data blocks stored in a source memory are sent to a target site via a data link. While sending one or more of the copies of the first data blocks to the target site, source hashes for second data blocks stored in the source memory are calculated, wherein the first data blocks are distinct from the second data blocks. While sending one or more of the copies of the first data blocks to the target site, target hashes of data blocks stored in a target memory of the target site are received. While sending one or more of the copies of the first data blocks to the target site, the source hashes are compared with the target hashes, respectively. After sending the first data blocks to the target site via the data link, copies of only those second data blocks are sent to the target site with source hashes that do not compare equally with respective target hashes.
1. A method comprising:
during a first period of time, sending copies of n data blocks stored in a source memory to a target site via a data link, wherein n is an integer greater than two;
while sending copies of the n data blocks to the target site, calculating source hashes for some but not all of m data stored in the source memory, wherein m is an integer greater than two, wherein no data blocks are common between the n data blocks and the m data blocks;
receiving target hashes of data blocks stored in a target memory of the target site;
comparing the calculated source hashes with respective target hashes;
during a second period of time, sending copies of only those blocks of the m data blocks to the target site with calculated source hashes, respectively, that do not compare equally to respective target hashes;
wherein the first and second periods of time are non-overlapping;
wherein the second period of time is after the first period of time.
2. The method of claim 1 wherein each of the target hashes were calculated for a respective data block of the data blocks stored in the target memory.
3. The method of claim 1 wherein each of the source and target memories store Z contiguous data blocks, wherein m+n=Z.
4. The method of claim 3 wherein the data blocks in the target memory are contiguous with each other.
5. The method of claim 4 wherein the m consecutive data blocks correspond to m consecutive data blocks, respectively, of the target memory.
6. The method of claim 5 further comprising replacing n data blocks of the target memory with the copies of the n consecutive blocks, respectively.
7. A non-transitory computer readable medium (CRM) comprising instructions that are executable on one or more processors to implement a method, the method comprising:
during a first period of time, sending copies of n data blocks stored in a source memory to a target site via a data link, wherein n is an integer greater than two;
while sending copies of the n data blocks to the target site, calculating source hashes for some but not all of m data stored in the source memory, wherein m is an integer greater than two, wherein no data blocks are common between the n data blocks and the m data blocks;
receiving target hashes of data blocks stored in a target memory of the target site;
comparing the calculated source hashes with respective target hashes;
during a second period of time, sending copies of only those blocks of the m data blocks to the target site with calculated source hashes, respectively, that do not compare equally to respective target hashes;
wherein the first and second periods of time are non-overlapping;
wherein the second period of time is after the first period of time.
8. The non-transitory of claim 7 wherein each of the target hashes were calculated for a respective data block of the data blocks stored in the target memory.
9. The non-transitory CRM of claim 7 wherein each of the source and target memories store Z contiguous data blocks, wherein m+n=Z.
10. The non-transitory CRM of claim 9 wherein the data blocks in the target memory are contiguous with each other.
11. The non-transitory CRM of claim 10 wherein the m consecutive data blocks correspond to m consecutive data blocks, respectively, of the target memory.
12. The non-transitory CRM of claim 11 wherein the method further comprises replacing n data blocks of the target memory with the copies of the n consecutive blocks, respectively.
13. A computer system comprising:
one or more processors;
a computer-readable storage medium coupled to the one or more processors; and
a plurality of instructions, encoded in the computer-readable storage medium, wherein a method is implemented in response to executing the instructions, the method comprising:
during a first period of time, sending copies of n data blocks stored in a source memory to a target site via a data link, wherein n is an integer greater than two;
while sending copies of the n data blocks to the target site, calculating source hashes for some but not all of m data stored in the source memory, wherein m is an integer greater than two, wherein no data blocks are common between the n data blocks and the m data blocks;
receiving target hashes of data blocks stored in a target memory of the target site;
comparing the calculated source hashes with respective target hashes;
during a second period of time, sending copies of only those blocks of the m data blocks to the target site with calculated source hashes, respectively, that do not compare equally to respective target hashes;
wherein the first and second periods of time are non-overlapping;
wherein the second period of time is after the first period of time.
14. The computer system of claim 13 wherein each of the target hashes were calculated for a respective data block of the data blocks stored in the target memory.
15. The computer system of claim 13 ,
wherein each of the source and target memories store Z contiguous data blocks,
wherein n+m=Z;
wherein the Z data blocks are contiguous with each other in the target memory;
wherein the m consecutive data blocks correspond to m data blocks, respectively, of the target memory.