In the system messages customer get the following error repeating.
Error Message: Pgm nhiRftIn. Unable to receive a message from another process (connection reset by peer).
The "connection reset by peer" error messages make sense. Every time we try to send/receive a message and find that the underlying TCP connection is torn down we display a message such as below and re-connect:
Thursday, April 21, 2011 21:05:54 Error nhiRmtIn Pgm nhiRmtIn: Unable to receive a message from another process (Connection reset by peer).
As long as such errors are not accompanied by timeouts or long running cluster commands or distributed reports, these can be ignored. These messages are normal.