ssh restart gives error in auth.log

tkmsr · 10-29-2010, 02:05 AM

When ever I restart ssh

Quote:

/etc/init.d/ssh restart

I see
following line in auth.log

Quote:

sshd[5678]: error: Bind to port 22 on :: failed: Address already in use.

That is a headless server.
What does the above line signify or tell and why am I seeing that?
Ubuntu 10.04 64 bit server edition

anon237 · 10-29-2010, 02:15 AM

Hi,

This:

Quote:

sshd[5678]: error: Bind to port 22 on :: failed: Address already in use.

tells you that the ssh daemon (or possibly some other program) is still listening on port 22 (the :: tells you it is ipv6) when ssh tries to start.

What does the following command show (after a restart _and_ after a stop):
$ netstat -plan | grep ssh

BTW: Does a /etc/init.d/ssh stop and then /etc/init.d/ssh start work without errors?

Hope this helps.

tkmsr · 10-29-2010, 02:34 AM

Quote:

Originally Posted by druuna

Hi,

This:tells you that the ssh daemon (or possibly some other program) is still listening on port 22 (the :: tells you it is ipv6) when ssh tries to start.

Ok

Quote:

Originally Posted by druuna

What does the following command show (after a restart _and_ after a stop):
$ netstat -plan | grep ssh

Here it is

Quote:

tcp 0 0 0.0.0.0:22 0.0.0.0:* LISTEN 5637/sshd
tcp6 0 0 :::22 :::* LISTEN 5637/sshd

Quote:

Originally Posted by druuna

BTW: Does a /etc/init.d/ssh stop and then /etc/init.d/ssh start work without errors?

Yes this happens without errors.

anon237 · 10-29-2010, 02:46 AM

Hi,

Is the output shown from after a restart?
I'm asking because the output shown seems correct, both ipv4 and ipv6 listen on port 22 and both are started by the same PID (5637). If not: do a restart and repost the output.

Just to make sure, could you post the restart part from the /etc/init.d/ssh script? (all from restart) to the ;;).

tkmsr · 10-29-2010, 04:34 AM

Quote:

Originally Posted by druuna

Just to make sure, could you post the restart part from the /etc/init.d/ssh script? (all from restart) to the ;;).

here it is (that is Ubuntu 10.04 server) I have not modified that script.

Quote:

restart)
check_privsep_dir
check_config
log_daemon_msg "Restarting OpenBSD Secure Shell server" "sshd"
start-stop-daemon --stop --quiet --oknodo --retry 30 --pidfile /var/run/sshd.pid
check_for_no_start log_end_msg
check_dev_null log_end_msg
if start-stop-daemon --start --quiet --oknodo --pidfile /var/run/sshd.pid --exec /usr/sbin/sshd -- $SSHD_OPTS; then
log_end_msg 0
else
log_end_msg 1
fi
;;

here is one more part I think you may be interested in this

Quote:

try-restart)
check_privsep_dir
check_config
log_daemon_msg "Restarting OpenBSD Secure Shell server" "sshd"
set +e
start-stop-daemon --stop --quiet --retry 30 --pidfile /var/run/sshd.pid
RET="$?"
set -e
case $RET in
0)
# old daemon stopped
check_for_no_start log_end_msg
check_dev_null log_end_msg
if start-stop-daemon --start --quiet --oknodo --pidfile /var/run/sshd.pid --exec /usr/sbin/sshd -- $SSHD_OPTS; then
log_end_msg 0
else
log_end_msg 1
fi
;;
1)
# daemon not running
log_progress_msg "(not running)"
log_end_msg 0
;;
*)
# failed to stop
log_progress_msg "(failed to stop)"
log_end_msg 1
;;
esac
;;

Quote:

Originally Posted by druuna

Is the output shown from after a restart?

Yes the above outputs were from restart.

Quote:

Originally Posted by druuna

If not: do a restart and repost the output.

I did it again for the sake of this thread just now
surprisingly I do not see that error message which I reported above with this restart

Code:

Received signal 15; terminating.
Oct 29 14:51:04  sshd[28576]: Server listening on 0.0.0.0 port 22.
Oct 29 14:51:04  sshd[28576]: Server listening on :: port 22.

and here is the output of netstat

Code:

netstat -pal | grep ssh
tcp        0      0 *:ssh                   *:*                     LISTEN      28576/sshd      
tcp        0      0 somedomain.com:ssh         tk.local:36020        ESTABLISHED 25383/2         
tcp6       0      0 [::]:ssh                [::]:*                  LISTEN      28576/sshd

also

Code:

 netstat -plan | grep ssh
tcp        0      0 0.0.0.0:22              0.0.0.0:*               LISTEN      28576/sshd      
tcp6       0      0 :::22                   :::*                    LISTEN      28576/sshd

anon237 · 10-29-2010, 05:06 AM

Hi,

It looks like there is nothing wrong. Maybe the time between the stop and start when executing the restart was too short (still stopping while the start was already initiated), but there is a retry mechanism in place that seems to solve that (although you do see an error when this happens).

Both the output you posted in post #3 and #5 show that all is initiated from the same pid (5637 and 28576).

I wouldn't worry about it, if you see that error again, check to see if both the tcp and tcp6 have the same pid (with the netstat command I supplied).

Hope this helps.