root@merlin.bhpmrl.oz (any of ianh) (04/24/89)
Has anyone noticed telnet hanging while an attempt is being made to establish a connection ??? For example, when using telnet from my Macintosh to our 3500 I occasionally get the following: ------------------------------------- # telnet merlin Trying... Connected to merlin. Escape character is '^]'. Domain/OS sr10 (merlin) ------------------------------------- it gets this far, and then just sits there (and sits there and sits there and ....). What has happened to the login prompt ??? I can do a kill -1 1 on init to reinitialize things but to no avail. ian
ianh@merlin.bhpmrl.oz (Ian Hoyle) (04/24/89)
In article <859@merlin.bhpmrl.oz>, root@merlin.bhpmrl.oz (any of ianh) writes: > Has anyone noticed telnet hanging while an attempt is being > made to establish a connection ??? For example, when . . > there and ....). What has happened to the login prompt ??? I can do a > kill -1 1 on init to reinitialize things but to no avail. > Sorry, but I didn't realise that this message was sent while I wasn't me !!! ian -- Ian Hoyle /\/\ Computer Systems Superintendent / / /\ BHP Melbourne Research Laboratories / / / \ 245 Wellington Rd, Mulgrave, 3170 / / / /\ \ AUSTRALIA \ \/ / / / \ / / / Phone : +61-3-560-7066 \/\/\/ ACSnet : ianh@merlin.bhpmrl.oz.au Internet: ianh%merlin.bhpmrl.oz.au@uunet.uu.net
pha@CAEN.ENGIN.UMICH.EDU (Paul H. Anderson) (04/25/89)
Date: 24 Apr 89 08:08:02 GMT From: root%merlin%bruce%murtoa.cs.mu.oz.au%munnari.uucp@uunet.uu.net Subject: telnet hanging (SR10.1) Has anyone noticed telnet hanging while an attempt is being made to establish a connection ??? For example, when using telnet from my Macintosh to our 3500 I occasionally get the following: ------------------------------------- # telnet merlin Trying... Connected to merlin. Escape character is '^]'. Domain/OS sr10 (merlin) ------------------------------------- it gets this far, and then just sits there (and sits there and sits there and ....). What has happened to the login prompt ??? I can do a kill -1 1 on init to reinitialize things but to no avail. ian This is a symptom of a bug in the standard release of the SR10.1 tcp daemon. You will notice a telnetd sitting there running continuously, eating up CPU time, but not doing anything. Any attempted connections that happen to try using that /dev/pty? will leave a telnetd spinning the same exact way. There are two temporary fixes - first is to remove the /dev/pty and /dev/ttyp for the corresponding tty, then kill the processes that were spinning on them. The other fix is to restart the tcpd after killing it, recreating all the ptys. The real fix may be coming in 10.2, but don't hold your breath. Note that this problem will cause other related problems, since the connection actually stays open, even if the user terminates the connection from the telnet command at their end. This winds up causing a machine with a finite number of connection slots to be filled up, thereby denying service to not only the screwed up Apollo, but also the thing that has a bunch of open ghost connections. Paul Anderson CAEN
wescott@LNIC1.HPRC.UH.EDU (Andrew M. Wescott) (05/10/89)
Don't be discouraged, I see this on lots of machines. Just do a ^] and start over again. The world has bigger problems. Andrew Wescott University of Hous ton Department of Chemical Engineering