From owner-freebsd-questions Thu Feb 22 22:01:15 1996 Return-Path: owner-questions Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id WAA13636 for questions-outgoing; Thu, 22 Feb 1996 22:01:15 -0800 (PST) Received: from Root.COM (implode.Root.COM [198.145.90.17]) by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id WAA13629 for ; Thu, 22 Feb 1996 22:01:11 -0800 (PST) Received: from localhost (localhost [127.0.0.1]) by Root.COM (8.6.12/8.6.5) with SMTP id WAA00593; Thu, 22 Feb 1996 22:01:22 -0800 Message-Id: <199602230601.WAA00593@Root.COM> X-Authentication-Warning: implode.Root.COM: Host localhost didn't use HELO protocol To: graichen@omega.physik.fu-berlin.de (Thomas Graichen) cc: questions@FreeBSD.ORG Subject: Re: rlogin hangs In-reply-to: Your message of "22 Feb 1996 18:03:51 GMT." <4gib67$1u7@prospero.physik.fu-berlin.de> From: David Greenman Reply-To: davidg@Root.COM Date: Thu, 22 Feb 1996 22:01:22 -0800 Sender: owner-questions@FreeBSD.ORG Precedence: bulk >sreid@edmbbs.iceonline.com wrote: > >: I have a problem, where rlogin works most of the time, but sometimes >: hangs... I think I saw this in the freebsd-isp mailing list a while ago, >: but that seemed to be under more stressful conditions... > >i'm also nerved by this problem - it's absolutely reproducable (only 2 >machines connected via ethernet) - it's 2.1.0 and everytime i rlogin - exit - >rlogin it won't work - hangs a minute then again it works - and it's >_absolutely_ reproducable - if anybody has a solution i may try it out (i >haven't seen any fix in the commits of the last months) This sounds like the 2MSL ("TIME_WAIT") wait is somehow coming into play. I believe that inetd sets the "SO_REUSEADDR" socket option for sockets it manages, so this must be happening on the rlogin client side instead of the server... Can you tell me: Does the problem occur 1) from freebsd to anything, 2) from anything to freebsd, or 3) from freebsd to freebsd ? The problem sounds like (1) and is probably because some sockets that were used previously by someone else are now in a "TIME_WAIT" 2MSL state and this is 60 seconds long. What to do about it, well, that's another thing. The local port is supposed to be dynamically assigned to the next available port number since the local port for rlogin isn't important...HOWEVER, rlogin insists on having the local port be allocated out of the limited number of reserved ports (it asks for port numbers between 512 and 1023), while telnet has no such requirement (it asks for ports >=1024). The problem may very well be related to this and, if so, may also explain why telnet appears to work okay in the same environment. -DG David Greenman Core-team/Principal Architect, The FreeBSD Project