rlogin-related problems

[Paul Manno]

[]

We are experiencing a minor problem with our remote login connections
under BRL (4.2) Unix on VAXen using all DEC DEUNAs.  Often times, a
process is left hanging on LAST_ACK / FIN_WAIT_2 trying to close the
connection.  After several (5-10) of these accumulate, I find that a
significant amount of CPU and bandwidth is being wasted.  The only
remedy I've found is to re-address one of the ends, wait a minute, then
set the address back correctly.  Is there a known solution to this
problem that does not break rcp (as indicated in BRL's code)?

	Paul Manno
	Medical Systems Development Corp.
UUCP:	{akgua, gatech, ihnp4, mcnc}!msdc!paul
USPS:	P.O. Box 26022 /  Atlanta, GA  30335
VOICE:  "Uh, Paul?"
QUOTE:  "Jelly Baby?"