Make use of virtual hash chunks in reconnects

This is a part of the https://github.com/hiero-ledger/hiero-consensus-node/issues/23970 epic.

Current reconnects are based on individual hashes. Learner provides node hashes to the teacher. Teacher responses if the node is clean or dirty (and includes leaf data for dirty nodes). Recently, hashes have been changed to be stored in chunks on disk. This means, to read multiple hashes in a single chunk, only one disk operation is needed. Reconnects may be improved to leverage this mechanism, on both learner and teacher sides, rather than loading and sending hashes one by one.

Start with epic #23970 and trace the reconnect hash exchange on both learner and teacher sides, then inspect the existing on-disk virtual hash chunk mechanism. Done means reconnects use chunk-based reads and transfers instead of loading and sending hashes one by one; the issue names no specific files or tests.

Written by the indexing model from the issue text.