My rambling thoughts:

Note that your mSATA drive is a different size (capacity) than the PA25-128 drive.

And the PA25-128 might also be a different size (capacity) than your "master" drive. So doing a bit copy from, say, a larger master drive to a slightly smaller PA25-128 could result in the access errors you see. The PA25-128 simply does not have as many sectors as the software thinks it has.

03:04: rw=0, want=1972874128, limit=117113850

In this case, the capacity of the erring /dev/hda4 appears to be 117113850 blocks, where a block is two sectors. 117113850 * 2 = 234227702 sectors for that specific partition. This of course matches what fdisk also reports (234227700), since the two numbers come from exactly the same place.

Most of the errors show "want" blocks that are believable but slightly too large for the partition, but a few of the later ones are quite obviously WAYYY out of range.

I suggest perhaps you should run filesystem checks on /dev/hda4 on the failing drive to see what it says. Something there really looks mucked up from the diskcopy, or ??

Another way things could get messed up is if the old kernel on the empeg somehow has decided to use CHS addressing rather than LBA. That could also confuse the situation. Unfortunately, that kernel doesn't appear to print any indicator of which type of addressing it is using. Except for drives larger than 128GB where it will print LBA48 in the log.

EDIT: I suppose we can rule out CHS addressing, because it reduces the addressable capacity by QUITE A LOT more that what we are seeing here. So that part is likely just fine.