HDS real cache is only mirror the write data but not the read data.
as some of the experts said that we should actually worried about read response but not the write response.
In this case, if there is any controller failure, how does HDS manage the read response??
If read data is mirrored, we dont have to re-promote the blocks again from the disks to cache and hence can improve the read response right? I may be missing something here. But wanted to know how HDS manages read response in this case?? Any help experts?