Form a engineering standpoint, there's something I'm still missing here: What is the real result you're looking to achieve?

If it's a space limitation, I'd try and deal with it by using higher capacity batteries of the same size.

If it's reliability, you can forget it. Today's batteries are extremely reliable if treated properly certainly WAY more than 99.5% good.... so even if you got absolute maximum benefit from the redundant batteries (which is not even close to possible), you'd only raise a 99.5% reliability to 99.9975% for an improvement of less than 1/2%... Remember, that's the maximum theoretical benefit from redundant batteries, and isn't really attainable in the real world.

So, there'd be very little reliability benefit. BUT and this is more important, most power failures don't happen in the batteries they happen in the connections, and in fitting redundant batteries, you may be tripling the number of connections. So you get minimal improvement in the already reliable part of the system, while you create two or three times as many of failure points where systems usually do have problems.

Oh, and then there's the charging issue... improper charging is also a MUCH bigger cause of failure than actual battery failure even for a single pack. Managing parallel batteries imposes some technical requirements, which means either careful compliance with good practice or adding more electronic gear to manage it for you... again creating greater chance of failure in an already more failure-prone part of the system than the actual batteries.

So, redundant batteries is certainly doable, but if doing it for reliability reasons, personally, I wouldn't.