Bug 1143

Summary:	connections with "sshd: root@notty" is established but not closed
Product:	Portable OpenSSH	Reporter:	sujit <knsujit>
Component:	Kerberos support	Assignee:	Assigned to nobody <unassigned-bugs>
Status:	CLOSED WORKSFORME
Severity:	critical
Priority:	P2
Version:	3.9p1
Hardware:	ix86
OS:	Linux

Description sujit 2006-01-05 08:00:18 AEDT

There is connection leak between my twom production servers. One server shows number of sshd processes in sleep mode with name "sshd:root@notty". This server has ssh version 3.9 installed.

Another server on other hand shows same number of ssh processes with name "/usr/bin/ssh -x -oFallBackToRsh no -lroot 172.25.32.105 scp -t /var/log/gtwlogs/gtw40.log". This server has ssh version 2.9pl installed.I can not upgrade this ssh version as of now as this is planned sometimes later.

Please help me knowing the exact cause behind this.

Regards
Sujit

Comment 1 Darren Tucker 2006-01-11 21:50:53 AEDT

Sounds like the client is either stalling or completing and not closing the connection afterward.  There ought to be corresponding TCP connections shown in netstat, what's the SendQ on those on both client and server?

You can try enabling ClientAliveInterval in sshd_config on the server running 3.9.  You can try killing the ssh processes on the client, if that's the only thing that helps and you really can't upgrade ssh then you'll probably have to script the kill and put it in cron.

Comment 2 Damien Miller 2006-07-06 21:09:45 AEST

6 months, no reply == no bug

Comment 3 Darren Tucker 2006-10-07 11:44:03 AEST

Change all RESOLVED bug to CLOSED with the exception of the ones fixed post-4.4.