If I do the following, and the network is down, then the zero case will be executed, which it shouldn't.
case "$(ssh -n $host zfs list -t snapshot -o name -H | grep "tank/fs" | wc -l | awk '{print $1}')" in
0) # do something
;;
1) # do something else
;;
*) # fail
esac
Earlier in the script I check that I can SSH to $host, but today I found this problem, where the network failed right after my check.
If I check the return value from the SSH command, then I will always get the return value from awk as it is executed last.
Question
How do I insure that I actually count zero lines that zfs outputted, and not zero lines from a failed SSH connection?
Say:
set -o pipefail
at the beginning of your script (or before the case statement).
Moreover, check for the return code of the command before executing the case statement:
set -o pipefail
$value=$(ssh -n $host zfs list -t snapshot -o name -H | grep "tank/fs" | wc -l | awk '{print $1}')
if [ $? == 0 ]; then
case $value in
...
esac
fi
From the manual:
pipefail
If set, the return value of a pipeline is the value of the last
(rightmost) command to exit with a non-zero status, or zero if all
commands in the pipeline exit successfully. This option is disabled by
default.
How about this (commands after ssh shouldn't get executed locally):
"$(ssh -n $host 'zfs list -t snapshot -o name -H | grep "tank/fs" | wc -l | awk \'{print $1}\')'"
Note how I single quoted the command for ssh to run on the remote machine. Also escaped the ' for awk like \'. Check the return value of the call and only act when it returns success.
