SSH session not active

Fed · 07-02-2013, 10:22 AM

I've been putting one of our new servers through it's paces today. It's running RHEL6.4 and all of the testing is occurring from a secondary server, which is going in via SSH to the RHEL server.

Quite often, it's failing the testing with the following error message:

Code:

SSH session not active

From what I understand, this usually indicates some kind of SSH timeout. In response to this, I uncommented the ClientAliveInterval in /etc/ssh/sshd_config and gave it a value of 400.

Unfortunately this hasn't solved this issue hence why I'm writing this thread. Could anyone suggest something I may have missed? This is particularly frustrating as we have another server, running 6.3, which works flawlessly and I can't see any difference in configuration.

On my latest run, the following was outputted (EDIT: On the client): 10054 which apparently means "Connection reset by peer", or "An existing connection was forcibly closed by the remote host". /var/log/message (EDIT: On the server) is listing: "imuxsock lost X messages from PID x due to rate-limiting" and "imuxsock begins to drop messages from pid X due to rate-limiting".

smeuth · 07-02-2013, 12:11 PM

Quote:

Originally Posted by Fed

From what I understand, this usually indicates some kind of SSH timeout. In response to this, I uncommented the ClientAliveInterval in /etc/ssh/sshd_config and gave it a value of 400.

Did you set that varible on the server or the client?
Did you bounce the service after making the change?

Fed · 07-03-2013, 02:26 AM

Quote:

Originally Posted by smeuth

Did you set that varible on the server or the client?
Did you bounce the service after making the change?

Thanks for replying, smeuth. The value (ClientAliveInterval) is set on the sever and the service was restarted.

I don't think the issue is the client, given it's solid performance with the other servers running 6.3.

Also, the firewall does allow port 22.

slipstreamed · 07-03-2013, 11:01 AM

Maybe this'll work, edit the sshd_config and insert the following lines

Code:

ServerAliveInterval x
ServerAliveCountMax y

These settings will make the SSH server send a null packet to the other side every <x> seconds, and give up if it doesn't receive any response after <y> tries, at which point the connection is likely to have been discarded anyway.

Fed · 07-03-2013, 04:02 PM

Quote:

Originally Posted by slipstreamed

Maybe this'll work, edit the sshd_config and insert the following lines

Code:

ServerAliveInterval x
ServerAliveCountMax y

These settings will make the SSH server send a null packet to the other side every <x> seconds, and give up if it doesn't receive any response after <y> tries, at which point the connection is likely to have been discarded anyway.

Thanks for the reply. I imagine the lower X and Y are, the more alive it's likely to be?

slipstreamed · 07-04-2013, 02:13 AM

Example

Code:

ServerAliveInterval 600
ServerAliveCountMax 5

here ssh server will send a null packet to client(s) every 600 seconds and will drop the connection if it fails to receive a response after 5 tries
http://patrickmylund.com/blog/how-to...-ssh-sessions/

Quote:

On my latest run, the following was outputted (EDIT: On the client): 10054 which apparently means "Connection reset by peer", or "An existing connection was forcibly closed by the remote host". /var/log/message (EDIT: On the server) is listing: "imuxsock lost X messages from PID x due to rate-limiting" and "imuxsock begins to drop messages from pid X due to rate-limiting".

http://bugs.contribs.org/show_bug.cgi?id=7470
http://www.rsyslog.com/tag/rate-limiting/