A question about block sizes on NetApp disks.

My impression is that disk access occurs in blocks. So if I write 1 byte to a disk, at some level the OS has a block in cache, updates the block, then writes that block out to disk. If I write 20K, then it probably affects multiple blocks.

If this is true, then it might be useful to know the underlying block size - because if I could ensure that my application wasn't spanning multiple blocks when I didn't need to be, I could give the physical disks a bit less of a hard time.

Does anyone know how to find this out? Or is this all wrong, and the access doesn't occur in blocks?

Thanks,

Edward.