Warnings in logs during replication

[lvca]

After testing, it was found that:

1. Set Darcadedb. typeDefault Buckets to 32, and when 10659 data entries are written, the above logs will begin to appear;
2. Set Darcadedb. typeDefault Buckets to 4, and when 657 data entries are written, the above log will appear as follows:
   “ WARNI [ReplicationLogFile] <arcadedb_9> Found a jump (2) in message numbers. Last was 657 and now receiving 659. Skip saving this entry (threadId=160)”;
3. It seems unrelated to the configuration parameters of JMeter, as no matter what write rate is used, problems will occur when the data volume reaches the above values. The main parameters of HTTP requests are as follows:
   http
   /api/v1/command/ytb_db
   true
   POST
   true
   ....
4. Another phenomenon is that when the above-mentioned problem logs appear, executing the operation of creating a new database will cause two Replica nodes to undergo a state transition of OFFLINE-JOINING-ONLINE, resulting in the failure of database creation. The log is as follows:
   “2025-04-17 14:34:49.162 WARNI [HAServer] <arcadedb_9> Timeout waiting for quorum (3) to be reached for request 1086
   2025-04-17 14:34:49.166 INFO [PostServerCommandHandler] <arcadedb_9> Error on command execution (PostServerCommandHandler): Timeout waiting for quorum (3) to be reached for request 1086”

Originally posted by @yang1-tb in #2163 (reply in thread)