Should reset committed index after replay logs

[lzydmxy]

RaftKeeper version

2.1.0

Description

2024.06.20 18:12:28.761597 [ 53213 ] <Information> LastCommittedIndexManager: Read last committed index 318789221
2024.06.20 18:12:55.554352 [ 53213 ] <Information> KeeperStateMachine: Replayed log batch [318647616, 318654803)

We replay log to 318654803 which is lower than 318789221, we should reset our log index to 318654803.

After the node starts, the replay log should reset the node's committed index to the actual replay log index to prevent startup anomalies caused by local log file loss or write failures.

How to reproduce

[JackyWoo]

In case of local log file loss we should terminate the starting, because we can not service with the inconsistence data.