We have multiple Samba servers behind a load balancer that periodically makes TCP connections to the microsoft-ds port (445) in order to check the health of the Samba servers. The load balancer doesn't know how to speak SMBFS or CIFS, so as soon as the TCP connection is established (SYN, SYN+ACK, ACK), it fires back a TCP RST to close the connection. Upon receiving the TCP RST, the Samba servers log: [2006/12/20 16:54:27, 0] lib/util_sock.c:get_peer_addr(1229) getpeername failed. Error was Transport endpoint is not connected [2006/12/20 16:54:27, 0] lib/util_sock.c:get_peer_addr(1229) getpeername failed. Error was Transport endpoint is not connected [2006/12/20 16:54:27, 0] lib/util_sock.c:read_data(534) read_data: read failure for 4 bytes to client 0.0.0.0. Error = Connection reset by peer While I appreciate that Samba is trying to be helpful and let us know a client dropped off unexpectedly, our logs are full of pages and pages and pages of these errors, which makes it extremely difficult to find anything else in the logs. It would be nice to have an option that, when set, caused Samba to treat these messages as debugging messages, and thus not log them unless debugging is enabled. Another good reason to have such an option is because Windows XP Professional clients can generate disconnects that Samba logs in a similar manner, and questions about the log messages seem common; see bug 3480. Thoughts? (I could probably whip up a patch myself, but I don't want to spend the time on it if it wouldn't be accepted.)