On Thu, 7 Jun 2007, John Lagrue wrote:
Paul Heinlein wrote:
Try a different tack. Get a root shell on the machine in question
and fire up sshd in no-fork mode under strace:
strace -o /tmp/ssh.trace sshd -D -p 2222
Then try to ssh into that machine on port 2222 (or whatever you
choose). Assuming it fails, close down the temporary sshd and
point $PAGER at /tmp/ssh.trace. The failure point will likely be
about 90% of the way through the file. (The end of the file will
be related to closing down sshd).
I tried that, and monitored the ssh.trace file with tail-f in
another window. There were no obvious signs of error, only what
looks like a child starting and then exiting. Mind you, it's been a
fair few years since I last looked at a trace file!
Ugh. Sorry. Try -d instead of -D, so sshd won't fork... You can use up
to three -d switches to increase verbosity.
--
Paul Heinlein <> heinlein@xxxxxxxxxx <> http://www.madboa.com/
_______________________________________________
CentOS mailing list
CentOS@xxxxxxxxxx
http://lists.centos.org/mailman/listinfo/centos
