bug: Misleading error "The request maybe too large" when secondary replica disk is abnormal

Description

The java-client logs output "The request maybe too large!".
After checking the `handleReplicaException(Request request, DefaultPromise promise, client_operator op, Table table, int timeout)` function in `PegasusTable.java`, it is found that this message is only logged when `ERR_INVALID_DATA` is received.

Then, looking at the server-side code in `replica_2pc.cpp`, specifically the `on_client_write(dsn::message_ex *request, bool ignore_throttling)` function, it is observed that when the server returns `ERR_INVALID_DATA` to the client, it also logs the following message:

"client from {} write request body size exceed threshold, request = [{}], request_body_size = {}, FLAGS_max_allowed_write_size = {}, it will be rejected!"

However, when I run the command:

grep -r "request_body_size" log_directory_path

in the log directory, there is no output.