You raise a good point about one battery pulling down the whole bank. It is a good idea to use a multimeter once in a while to see if all the batteries in a bank are the same voltage. If one is bad, replacing it is a good idea even though the new battery will now die at the same time as the rest of the batteries in the bank.
I've never rotated batteries because, in my case, moving an entire battery bank increases the chances of my doing something stupid like dropping a wrench onto a live battery or breaking a battery case while torquing down terminal bolts. If you ever have reason to take out a battery bank anyway, rotating would be a good idea.
Instead of rotating batteries, it might be a better to put the time into using a more advanced method of making parallel connections. Bus bars on positive and negative are still in their infancy, but they solve problems of jump wire resistance. I think we'll see more of them. Here are a couple of examples:
The trick with bus bars is that the cables from the batteries to the bus bars have to be identical length. You also need a fuse on the positive side of each paralleled battery or battery set. The fuses look like this:
If you want to parallel four batteries (or four sets of two batteries) there is a way to do it that keeps all of your cable resistance the same for the whole battery bank. It is method 4 in the attached link: http://www.smartgauge.co.uk/batt_con.html
L-16 pricing suffers from the "Walmart problem." If you are comparing apples to apples, L-16s and GC2 golf cart batteries from the same vendor will be the same or nearly so on a cost per amp hour basis. I last compared Fulll River brand AGMs and they were within a penny per amp hour of each other. Problem is, you can't stop by Costco or Walmart and pick up a large L-16 industrial battery like you can a golf cart battery. Try as they might, the brick and mortar stores that sell batteries like L-16s can't compete with the Walmarts of the world. The golf cart batteries from Costco or Walmart still work out cheaper per amp hour than where you have to go for an L-16.
There is weight. Golf cart batteries at about 65 pounds aren't that bad. I can lift a 115 pound L-16 by myself, but would lots rather do it with a friend holding the other side.
Problem with the multimeter is you'll probably have to disconnect each pair and let them sit for a while before testing. Otherwise the remaining good batteries will try to charge the bad one. However checking water level frequently is a good place to start.
We had a thread a while ago about trying to come up with a way to remotely monitor each battery in the set... but I'm not sure if anyone actually came up with anything that sounded like it would work.
I can't see much improvement with the bus bar, all the leads going to the bus would have to be the same length and this would just add more wire and resistance. I tried having cables made up once with 3 cables crimped together at one end, but it wasn't really satisfactory. I think just keeping all the connections clean, tight, and coated with anti-corrosive is about as good as any.
I thought about the fuses too, thinking that if a battery went bad the fuse would blow and disconnect that set, but the fuses have to be such high amperage that I'm not sure they'd ever blow. For an inverter each pair of batteries would need say a 100a fuse; even more if the bank is to be available for emergency starting. And if one fuse did blow, the additional load on the remaining pairs would probably cause those fuses to go, too, leaving you DIW. I tried it once but eventually removed them. I do have an external fuse for the inverter, one for the house load, and one for the pumps and emergency equipment which are on a separate circuit.
I remember the monitoring system thread. I think we all decided that a multimeter didn't tell you anything about the parallel batteries, but gave valuable information to compare the voltages of the two 6 volt batteries with series connections to each other. If that's right, as long as you have at least some serial batteries, a multimeter should give good information to compare two batteries in the series sets to each other, even if the batteries are still connected. Maybe that's another reason to favor series connections over parallel connections.
The goal of a bus bar system is that all of the parallel connections have identical resistance. The up side of that is that things stay equal. The down side is that some of the cables are longer than they need to be. If the bus bar is near the center of the battery bank, the parallel cables are all about half the length of the battery bank. That yields shorter cables for the batteries that are farther away than the half way point, but does result in longer cables for the other half of the bank. It shouldn't be that much worse in terms of overall resistance and would give a huge improvement in balancing batteries.
Fuses would indeed be a pain if you had to use a house bank to jump start an engine in a pinch. You're also right that the individual battery fuses won't protect against charge differential between batteries in a bank. The fuses are just there to comply with ABYC rules that require a fuse within 7 inches of the battery for wires that aren't in conduit. It is possible that some surveyors would treat the bus cables like jump wires that don't need fuses. Maybe Ike could tell us more about that.
You guys are REALLY getting anal about this. I n the real world a few inches difference on short cables doesn't mean Diddly Squat. To me. the installations appearance is important. Over length cables create a visual mess.