blob id reused before replayed logs commited

[Besroy]

When all replicas are restarted simultaneously, the most recently committed LSN and its corresponding blob id may not have been persisted. During the replay of logs, leader will updating blob id in the commit phase. However, if new blob requests arrive before the replayed logs have been fully committed, the latest blob id may not have been updated yet. This can result in the reuse of the original blob id, leading to a conflict in blob ids.

More details in here