I know the issue was already discovered to be a wiring mistake, but I thought I'd join the others who believe your charging and cutout voltages are a bit high. I would recommend floating at 3.5v/cell (so set 56.0v on your charge controller), and cutting out at 3.6v/cell (set on the BMS for battery protection). You don't want to ever exceed (or spend much time at) 3.65v, and there's no guarantee that the BMS's voltage calibration is spot on. You go overvoltage just a little bit on a cell, and it will expand rapidly. So, while the peak of 57v you saw is within the absolute maximum pack voltage of 58.4v, what likely went wrong was a couple cells got full before the rest and exceeded their maximum rated voltage of 3.65v because other cells in series with them were still at 3.55v, having not yet reached full charge.
Additionally, several people have done capacity tests at various target charge/float voltages, and the capacity loss at 3.5v is negligible—the biggest loss when running that voltage over the cells' absolute maximum voltage is the extra time it takes being held there to reach full charge as the current drops off quicker, slowing down the charging process. For solar systems that are expected to float multiple hours a day, this is no big deal, and the lower voltage will help mitigate the battery life impact caused by floating at 100% for hours each day. Furthermore, floating at 56.0v (instead of 57v) gives your cells that much more imbalance headroom to avoid cells that get full first exceeding their maximum cell voltage and needing the BMS cut-out (which failed in your case, due to the wiring error) to save them in the first place.
What would've been particularly interesting would be to see a screenshot of your BMS cell voltage figures right before you shut off power. That would be crucial information to have—I betcha you would have seen some cells over 3.7v (they tend to go overvoltage really quickly when it happens).