Save output of command in array

Question

I am executing this command

ar=($( ssh -i id_rsa -T -y [email protected] sh -c "id;whoami;ps aux")) 

I need to save the output of each command executed into the array and call the array, but i am not able to get the desired result as ps aux or other command when calling the array as

echo ${ar[2]}

its just showing me the first line of ps aux that is

USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND 

and not the full result of ps aux

Answer 1

Try (bash4.4+)

readarray -td '' a < <( ssh -i id_rsa -T -y [email protected] ' id; printf "\0" whoami; printf "\0" ps aux; printf "\0"' ) id_output=${a[0]} whoami_output=${a[1]} ps_output=${a[2]} 

ssh already invokes a shell (the login shell of that root user there) to interpret the command line. To do that it concatenates the arguments. So what ends up being run in your approach is that shell with -c and sh -c id;whoami;ps aux as arguments, and that shell will then run sh -c id, whoami and ps aux. The sh -c is superfluous here and can only cause confusion.

Beware the array elements will contain the full output of each command including the trailing newline character.

If you want to print them, use:

printf %s "$id_output" 

To remove the newline character, you can do:

id_output=${id_output%$'\n'} 

Answer 2

Your command is storing each whitespace-separated string of the ssh command in an array. So, since you are sshing and then running id, whoami and ps aux, all of their output is added to the array, splitting on whitespace (the default value of the $IFS variable). You can see this with declare -p ar:

$ ar=($( ssh localhost sh -c "id;whoami;ps aux")) $ declare -p ar | head -c500 declare -a ar=([0]="uid=1000(terdon)" [1]="gid=1000(terdon)" [2]="groups=1000(terdon),3(sys),7(lp),10(wheel),14(uucp),56(bumblebee),84(avahi),96(scanner),209(cups),995(plugdev)" [3]="terdon" [4]="USER" [5]="PID" [6]="%CPU" [7]="%MEM" [8]="VSZ" [9]="RSS" [10]="TTY" [11]="STAT" [12]="START" [13]="TIME" [14]="COMMAND" [15]="root" [16]="1" [17]="0.0" [18]="0.0" [19]="174456" [20]="11996" [21]="a" [22]="b" [23]="f" [24]="R" [25]="Ss" [26]="Jun23" [27]="7:06" [28]="/sbin/init" [29]="root" [30]="2" 

As you can see there, each whitespace-delimited string of the output of each of the commands run is stored in its own array element.

If you want to have an array with only three elements, one per command, you need to use a different character to split on. One way to do this is to edit your commands so that they print a unique character after executing and then use mapfile to read the array, telling it to split on that unique character. For example, \0:

$ mapfile -d '' < <( ssh localhost sh -c "id; printf '\0'; whoami; printf '\0'; ps aux") ar $ for i in 0 1 2; do echo "$i: ${ar[i]}"; done | head 0: uid=1000(terdon) gid=1000(terdon) groups=1000(terdon),3(sys),7(lp),10(wheel),14(uucp),56(bumblebee),84(avahi),96(scanner),209(cups),995(plugdev) 1: terdon 2: USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND root 1 0.0 0.0 174456 11996 ? Ss Jun23 7:07 /sbin/init root 2 0.0 0.0 0 0 ? S Jun23 0:00 [kthreadd] root 3 0.0 0.0 0 0 ? I< Jun23 0:00 [rcu_gp] root 4 0.0 0.0 0 0 ? I< Jun23 0:00 [rcu_par_gp] root 6 0.0 0.0 0 0 ? I< Jun23 0:00 [kworker/0:0H-kblockd] 

Answer 3

The expensive part of ssh is the authentication process. I would use the ControlMaster option to allow several separate ssh sessions to share a single authenticated connection.

ssh_get () { ssh -o ControlMaster=auto -o ControlPersist=5 -i id_rsa -T -y [email protected] "$@" } for cmd in id whoami "ps aux"; do ar+=("$(ssh_get "$cmd")") done 

The first call will do the authentication, but after it exits, the ssh connection will remain in the background for up to 5 seconds (per the ControlPersist option), allowing the next call to use it without reauthenticating. The overhead of running ssh itself is minimal.