There is something wrong with struct.unpack.

import struct
print struct.calcsize('<BBBBBHB')
8
print struct.calcsize('>BBBBBHB')
8
print struct.calcsize('BBBBBHB')
9

Looks fine to me.  The third one uses native size and alignment (see the "Byte Order, Size, and Alignment" section of the struct docs).  After the first 5 B's, a pad byte is inserted so that the next H is properly aligned (to a 2-byte boundary).  That makes 6 bytes so far.  H then takes 2 bytes, making 8 bytes so far.  The final B adds the 9th byte.  As the docs say, < and > do no padding, so no pad byte is added in the first two cases.

sorry my bad.