SSH through SSH
 

I have inherited a test station at work that runs bash script. I am new to linux but have managed to get most of the software to work with this exception:

I am having a problem with the following line:

ssh -q -o "ConnectTimeout 10" root@$OSR_PC "sshpass -p default ssh -q -o \"ConnectTimeout 10\" root@$OSR_IP \"export lm_ctrl_channel_bw=$CHNL_BW; export lm_ctrl_dl_freq=$DL_CF; export lm_ctrl_afs=0; configdb -S lm_ctrl; reboot; exit\""

The test setup is as follows:

OBR_PC is linked to OSR_PC (address 172.18.0.20)
The OSR_PC is linked to OSR_IP(address: 169.254.101.2)

The software is to send the commands to the OSR_IP through the OSR_PC from the OBR_PC but I cannot get a link with this line and nether can my IT tech.
He wants to create an entire new configuration.
I can ssh into the OSR_PC and then ssh onto the OSR_IP.

Can I make the above script work without rewriting the entire program which is quite extensive?

Thanx in advance

WimaxTech

Where is it failing?

I don't know about that command, but I can run
and it successfully makes it through the chain and spits out the contents of comp3.

It appears fo fail after "sshpass -p default ssh -q -o \"ConnectTimeout 10\" root@$OSR_IP \"
There's a lot going on with this script that I do not know yet.

If I were in your shoes, for the sake of testing where the failure is, I would take out the variables that reference known IP's, I would also setup ~/.ssh/config files on each machine that you are ssh'ing into to pre configure the "connecttimeout" options, and if possible I would even configure a passwordless ssh key for root@$OSR_IP to get rid of the "sshpass" command (BTW I had never heard of that till your post). That should leave you with a very minimalist command. Assuming that all works, you could back out of those changes to find the one that breaks.

If you are a tester you should be aware of the fact that debug info
is crucial. Can you please tell us 'how" it fails, what output are
you getting?


Cheers,
Tink

The software is setting frequency and bandwidth parameters on two systems to establish communication between the two.
The OBR unit, controlled from OBR_PC, is not a problem and works fine.
The software control of the OSR_IP does not report any error. The program continues to run but the parameters on OSR_IP are not changed. This results in a "Test Failure".
Entering the command line in Terminal,

ssh -q -o "ConnectTimeout 10" root@$OSR_PC "sshpass -p default ssh -q -o \"ConnectTimeout 10\" root@$OSR_IP"

I am unable to reach the command prompt of OSR_IP.
I have tried symplfing the line to : ssh root@$OSR_PC "sshpass -p default ssh root@$OSR_IP"
The return shows that the OSR_IP address has been added but nothing more happens.
I have been searching for a reason for this. As I have said earlier, I can ssh into OSR_PC and then ssh into OSR_IP and from what I know so far – this should work.
Is there a Good reference document that would help my understanding of using ssh in this manner?

Well ... I'd start hunting for sshpass; I've not come across that command in
over 15 years of using Linux. What does it do, where does it live?


Cheers,
Tink

sshpass is not recommended for the reasons stated on the man page http://linux.die.net/man/1/sshpass.
I would indeed go with post#4; simplify the code (at least for debugging the issue) and use ssh auth-keys instead of sshpass.

sshpass is a utility designed for running ssh using the mode referred to as "keyboard-interactive" password authentication, but in non-interactive mode.

Security is a non-issue as the entire system on which it's run is closed and there can be no outside connections.

I know what its used for; I read the page :)
Nonetheless, simplification would be better.