Shell Programming and Scripting

Hello all,

been writing a little inventory script (hpux 11.23, ksh Version 11/16/88) which takes our hostname from our inventory db and tests them with ssh connection.

I need to be able to differentiate a timeout (host not reachable on main net link could try adm prod back links) and a permission denied (which would basicly be a bad key).

Thing is the return code can't help me. Anybody has a tip on how i would be able to handle such case?

Code:

[username@checkinghost]/var/opt/username> ssh -o "BatchMode=yes" -o "StrictHostKeyChecking=no" -o "ConnectTimeout=3" host1
ssh: connect to host host1 port 22: Connection timed out
[username@checkinghost]/var/opt/username> echo $?
255

Code:

[username@checkinghost]/var/opt/username> ssh -o "BatchMode=yes" -o "StrictHostKeyChecking=no" -o "ConnectTimeout=3" host2
Permission denied (publickey,gssapi-keyex,gssapi-with-mic,password).
[username@checkinghost]/var/opt/username> echo $?
255

The only lead i've tried is redirecting the standard error into standard output and greping both error message. Works on the command line but i wonder how it would fare into a script. There must be a simpler way.

Thank you for your help!

The reason it doesn't give exit status for this is because it's supposed to give you the exit status of whatever happened on the remote end. This makes it hard to tell the difference between the code ssh gave you, and the code the remote host (if any) gave you.

The same is true for error output actually, since something on the other end could conceivably print that message on stderr instead of the local program.

Try ssh ... -E logfile instead of ssh ... 2>logfile. This definitively separates errors ssh prints from things that happen on the remote end.

I understand now. Yes this will be harder than i tought.

The -E looked like a promising lane until....

Code:

ssh: illegal option -- E

Code:

man ssh
HP-UX Secure Shell is based on OpenSSH version 5.9p1