[CX] Tolerate Corrupted System Checkpoint Files

Issue was encountered by a customer.

Currently during node startup, we attempt to read the system latest checkpoint file to determine the system state and complete the startup based on that state. However, if all checkpoints files found are corrupted, we currently attempt to bootstrap the node as a new node and as a result of that, all data is deleted. We should distinguish between the case where there is no checkpoint files at all (real bootstrap) and the case where we couldn't successfully read any of the checkpoints files, and on that case, we should force recovery to start from the start of the available transaction logs.