CODEWITHSUNDEEP
linuxbashshell
capser
capser

Reputation: 2635

return status of command sent to host via ssh

I have a script that open up cores on machine It works fine, but if I enhance it, like if there are not cores on the machine, I would like it to tell me, not just kick out an error command. I would like to replace the error with something So I am trying to ge the exit status of the last command and base the if statement off of that.

#!/bin/bash
rc=$?
for i in server5432foo server5438foo server5344foo server5430foo;
do
echo $i ;
ssh -q -T $i " ls -ltr  /local/data/log/core*; echo $? ";
echo $rc

if [[ $rc == 0 ]] ; then
    ssh -q -T $i " chmod ugo+rw  /local/data/log/core* ";
    else
    echo " No core files on $i /local/data/log/ "
    fi
done

This is what the script returns where there are no cores in the file specified. Even though the ls is unsuccessful it still returns a 0 and does not run the else command.

ls: cannot access /local/data/log/core*: No such file or directory
0
0
chmod: cannot access `/local/data/log/core*': No such file or directory

I am trying to get the return code on the commnand sent to that box vias ssh. In the script it seems that the SSH is successful, so it returns a zero. However when I run the command from the shell - I get a return code other than zero if it is unsuccessful.

casper@server0170foo:~/walt/restart_superprocesss$ ssh -q -T server5432foo "ls -ltr   /local/data/log/core00001" ; echo $?
ls: cannot access /local/data/log/core00001: No such file or directory
2
casper@server0170foo:~/walt/restart_superprocesss$ ssh -q -T server5432foo "touch    /local/data/log/core00001" ; echo $?
0
casper@server0170foo:~/walt/restart_superprocesss$ ssh -q -T server5432foo "ls -ltr    /local/data/log/core00001" ; echo $?
-rw-r--r-- 1 casper casper 0 Oct 13 11:38 /local/data/log/core00001
0

How come the return code is different from when I run a command directly from the command line, which are accurate as opposed to running the command from a script - where I get a 0 everytime

Upvotes: 0

Views: 432

Answers (2)

Cole Tierney
Cole Tierney

Reputation: 10324

I would try the following:

for i in server5432foo server5438foo server5344foo server5430foo; do
    echo $i;

    if ssh -q -T $i " ls -ltr /local/data/log/core* > /dev/null; then
        ssh -q -T $i "chmod ugo+rw  /local/data/log/core*"
    else
        echo "No core files on $i /local/data/log/"
    fi
done

Upvotes: 1

zerodiff
zerodiff

Reputation: 1700

If you want to check the return status of ssh (which should be return status of the remotely executed command or some higher number indicating ssh itself failed), I think you have a bug in your script, and you want to remove the echo in the ssh statement, because it will exit with 0 (and because, with double-quotes, the $? will be the return value of the command before ssh, not the return value of the ls):

#!/bin/bash
# This line isn't really helpful
#rc=$?
for i in server5432foo server5438foo server5344foo server5430foo;
do
echo $i ;
ssh -q -T $i "ls -ltr /local/data/log/core*";
# This is the new line to make sure $rc has the ssh exit status
rc=$?
echo $rc

if [[ $rc == 0 ]] ; then
    ssh -q -T $i "chmod ugo+rw  /local/data/log/core*";
    else
    echo " No core files on $i /local/data/log/ "
    fi
done

From the man page:

EXIT STATUS ssh exits with the exit status of the remote command or with 255 if an error occurred.

Upvotes: 1

Related Questions