From owner-freebsd-questions@FreeBSD.ORG Wed May 7 07:35:35 2003 Return-Path: Delivered-To: freebsd-questions@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 5790637B401 for ; Wed, 7 May 2003 07:35:35 -0700 (PDT) Received: from web.cs.ndsu.nodak.edu (web.cs.ndsu.NoDak.edu [134.129.125.7]) by mx1.FreeBSD.org (Postfix) with ESMTP id 9300343FDF for ; Wed, 7 May 2003 07:35:34 -0700 (PDT) (envelope-from tinguely@web.cs.ndsu.nodak.edu) Received: from web.cs.ndsu.nodak.edu (localhost [127.0.0.1]) by web.cs.ndsu.nodak.edu (8.12.9/8.11.4) with ESMTP id h47EZXWr007324 for ; Wed, 7 May 2003 09:35:33 -0500 (CDT) (envelope-from tinguely@web.cs.ndsu.nodak.edu) Received: (from tinguely@localhost) by web.cs.ndsu.nodak.edu (8.12.9/8.12.8/Submit) id h47EZXNK007323 for freebsd-questions@freebsd.org; Wed, 7 May 2003 09:35:33 -0500 (CDT) (envelope-from tinguely) Date: Wed, 7 May 2003 09:35:33 -0500 (CDT) From: mark tinguely Message-Id: <200305071435.h47EZXNK007323@web.cs.ndsu.nodak.edu> To: freebsd-questions@freebsd.org Subject: Re: HELP: TCP stack problem in 4.8-R with rsh connects (second channel) X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 07 May 2003 14:35:35 -0000 If I remember from your original posting that you are doing sequential rsh and eventually it starts giving time outs. I saw something like this starting in FreeBSD 3.0 when we use rsh to do sequential remote dumps (especially noticable on larger partitions). It acted like FreeBSD stayed in CLOSING state too long. A sleep (I chose 60 seconds because it worked) would let the previous rsh command to close before starting the next one. Obviously, you do not want to do that. I am just throwing this idea out to see if we can define the conditions that cause your problem. I don't have your original post, but I thought it was the one where netstat on each side had one side of the connection in a closing state but the other side had already remove the connection. It would be interesting to get the netstat of both sides when you are using the FreeBSD <-> Solaris and FreeBSD <-> FreeBSD. Does this problem existing in the FreeBSD <-> Solaris case when the role of FreeBSD and Solaris are reversed? --Mark Tinguely.