Greetings;
I put the following in my sshd_config file believing that any session
would time out in 2 minutes if the server could not communicate with
the client.
ClientAliveCountMax 8
ClientAliveInterval 15
However I found that if the server continually attempts to send some data
to the client this does not work. The connection stays up for hours.
I started sshd with the -d -d -d switches to collect some trace. I
looked in the trace for the entry that shows the heartbeat msg being sent
to the client
but it does not appear in the trace:
debug2: channel 0: request keepalive@openssh.com << I never see trace
entry this for this scenario
[ to break the connection I pull the ethernet cable between the network
switch and the SSH Client workstation]
[ to create some continuous traffic to send I execute the command "ping
localhost" ]
Otherwise the ClientAlive mechanism seems to work :
I found that if there is a period of 15 minutes with NO activity and the
client is
connected the trace shows:
debug2: channel 0: request keepalive@openssh.com
debug1: Got 100/13 for keepalive
If the server is not attempting to send data to the client and I pull the
cable between the network switch and the client I get a timeout after 2
minutes
as expected:
debug2: channel 0: request keepalive@openssh.com
debug2: channel 0: request keepalive@openssh.com
debug2: channel 0: request keepalive@openssh.com
debug2: channel 0: request keepalive@openssh.com
debug2: channel 0: request keepalive@openssh.com
debug2: channel 0: request keepalive@openssh.com
debug2: channel 0: request keepalive@openssh.com
debug2: channel 0: request keepalive@openssh.com
Disconnecting: Timeout, your session not responding.
debug3: channel 0: close_fds r 9 w 9 e 11
debug1: do_cleanup
debug1: PAM: cleanup
debug3: PAM: sshpam_thread_cleanup entering
Any suggestions?
OpenSSH level: SSH-2.0-OpenSSH_3.8p1
Distro SuSe SLES 9.3
Thanks, Rick
blasiak@us.ibm.com
