I know there are lots of discussions about it but i need you help with ssh remote command exit codes. I have that code:
(scan is a script which scans for viruses in the given file)
for i in $FILES
do
RET_CODE=$(ssh $SSH_OPT $HOST "scan $i; echo $?")
if [ $? -eq 0 ]; then
SOME_CODE
The scan works and it returns either 0 or (1 for errors) or 2 if a virus is found. But somehow my return code is always 0. Even, if i scan a virus.
Here is set -x output:
++ ssh -i /home/USER/.ssh/id host 'scan Downloads/eicar.com; echo 0'
+ RET_CODE='File Downloads/eicar.com: VIRUS: Virus found.
code of the Eicar-Test-Signature virus
0'
Here is the Output if i run those commands on the "remote" machine without ssh:
[user#ws ~]$ scan eicar.com; echo $?
File eicar.com: VIRUS: Virus found.
code of the Eicar-Test-Signature virus
2
I just want to have the return Code, i dont need all the other output of scan.
!UPDATE!
It seems like, echo is the problem.
The reason your ssh is always returning 0 is because the final echo command is always succeeding! If you want to get the return code from scan, either remove the echo or assign it to a variable and use exit. On my system:
$ ssh host 'false'
$ echo $?
1
$ ssh host 'false; echo $?'
1
$ echo $?
0
$ ssh host 'false; ret=$?; echo $ret; exit $ret'
1
$ echo $?
1
ssh returns the exit status of the entire pipeline that it runs - in this case, that's the exit status of echo $?.
What you want to do is simply use the ssh result directly (since you say that you don't want any of the output):
for i in $FILES
do
if ssh $SSH_OPT $HOST "scan $i >/dev/lull 2>&1"
then
SOME_CODE
If you really feel you must print the return code, that you can do that without affecting the overall result by using an EXIT trap:
for i in $FILES
do
if ssh $SSH_OPT $HOST "trap 'echo \$?' EXIT; scan $i >/dev/lull 2>&1"
then
SOME_CODE
Demo:
$ ssh $host "trap 'echo \$?' EXIT; true"; echo $?
0
0
$ ssh $host "trap 'echo \$?' EXIT; false"; echo $?
1
1
BTW, I recommend you avoid uppercase variable names in your scripts - those are normally used for environment variables that change the behaviour of programs.
Related
I have task to create quite complicated bash script, which at first part is checking the connection to host:port with telnet and- if telnet would fail or would not be installed, try the connection using netcat.
I have problem with loop, where it will skip netcat if telnet would connect to the host and also, if both- telnet and netcat would fail- script would finish with error message.
The script:
#!/bin/bash
echo Type host IP address
read REMOTEHOST
echo Type port number
read REMOTEPORT
TIMEOUT=5
echo quit | timeout --signal=9 5 telnet $REMOTEHOST $REMOTEPORT
if nc -w $TIMEOUT -z $REMOTEHOST $REMOTEPORT; then
echo "I was able to connect to ${REMOTEHOST}:${REMOTEPORT}"
else
echo "Connection to ${REMOTEHOST}:${REMOTEPORT} failed. Exit code from Netcat was ($?)."
fi
You can use the $? variable to get the exit code from the last command.
I found that your original telnet command exits with error code 1 on my system because the escape character is ^]. When I telnet manually I need to hit ctrl-] to enter the telnet prompt, then I can enter 'quit'.
The trick here is you cannot just type ^], you have to type ctrl-v ctrl-]
ctrl-v tells the system to capture the next ctrl character.
The following gives me an exit code of 0, and you can verify by running it manually with echo $? at the command line
-- remember to use ctrl-v ctr-]
$ (echo ^]; echo quit) | timeout --signal=9 5 telnet <REMOTEHOST> <REMOTEPORT>
$ echo $?
Then you can use this in your script:
#!/bin/bash
echo Type host IP address
read REMOTEHOST
echo Type port number
read REMOTEPORT
TIMEOUT=5
(echo ^]; echo quit) | timeout --signal=9 5 telnet $REMOTEHOST $REMOTEPORT > /dev/null 2>&1
TELNET_EXIT_CODE=$?
if [[ $TELNET_EXIT_CODE -ne 0 ]]; then
nc -w $TIMEOUT -z $REMOTEHOST $REMOTEPORT > /dev/null 2>&1
NC_EXIT_CODE=$?
fi
if [[ $TELNET_EXIT_CODE -eq 0 ]] || [[ $NC_EXIT_CODE -eq 0 ]]; then
echo "success"
else
echo "fail"
fi
Tested on Ubuntu 20.04.04, GNU bash version 5.0.17, Telnet version 0.17-41
I'm pretty new to bash scripting and I'm attempting to write a script that does some basic operations.
I want to check certain conditions and if they are met, terminate the script. So for example, I want to check whether the zip of files was successful:
echo "Zipping file..."
for file in $fileList;
do
echo $file | zip -v $archive -#
if [[ $? != 0 ]];
then
echo "Error creating zip"
exit 1
fi
done
What happens though is that the exit 1 signal causes the ssh connection to terminate as well:
Zipping file...
Command 'zip' not found, but can be installed with:
sudo apt install zip
Error creating zip
Connection to 3.137.7.52 closed.
What's the correct way to terminate a script without also disconnecting from the server?
If you wrap it all in a script with shebang #!/bin/bash than exit 1 will be fine
but if you run this as a oneliner directly in console then this exit 1 means exit from console, and that would break ssh connection obvy
cat > ziper.sh << \EOF
#!/bin/bash
echo "Zipping file..."
for file in $fileList;
do
echo $file | zip -v $archive -#
if [[ $? != 0 ]];
then
echo "Error creating zip"
exit 1
fi
done
EOF
./ziper.sh
In oneliner use break
I am trying to execute a script, In the midway of the script, it executes a part of the code in a remote shell and then again comes back to the same local shell.
However, my question is about the exit codes of the command executed in the remote shell. I am trying to do decision in remote with if-then-else based on the exit codes. When it didn't execute as expected I tried to narrow down the problem.
Here are my findings,
The output is always IP Found, as $? is always equal to 0.
ssh root#<remote server> bash -c "'
host www.google.com123
if [ $? != 0 ]
then
echo "Invalid Host"
else
echo "IP Found"
fi
'"
Also, in this case, the same output is expected
ssh root#<remote server> bash -c "'
host www.google.com
if [ $? != 0 ]
then
echo "Invalid Host"
else
echo "IP Found"
fi
'"
Ref: I got to know about this remote execution method from here
Please Help me understand this behavior of remote shell execution. Also, if there are any other ways to execute a part of a shell-script in remote through ssh please suggest.
The outermost quotes are double quotes, so $? is expanded locally before you ever run bash remotely. Instead, use
ssh root#<remote server> 'bash -c "
host www.google.com123
if [ \$? != 0 ]
then
echo \"Invalid Host\"
else
echo \"IP Found\"
fi
"'
Simpler is to avoid checking $? explicitly at all:
ssh root#<remote server> 'bash -c "
if ! host www.google.com123
then
echo \"Invalid Host\"
else
echo \"IP Found\"
fi
"'
Even simpler is not run a second unnecessary shell.
ssh root#<remote server> '
if ! host www.google.com123
then
echo "Invalid Host"
else
echo "IP Found"
fi
'
I have a file named "host.txt" with listing IP addresses of two systems.
~] cat hosts.txt
10.1.1.10
10.1.1.20
Using below script I am trying to login to each system, check status of a service and print the output of each system. The script prompts to login, however does not continue to execute the /opt/agent.sh status command. Can someone please help fix this script?
#!/bin/bash
for HOST in `cat hosts.txt`
do
ssh root#$HOST
STATUS=`/opt/agent.sh status | awk 'NR==1{print $3 $4}'`
echo $STATUS
if [ $STATUS! == "isrunning" ]; then
echo "$host == FAIL"
else
echo "$host == PASS"
fi
You script does not continue until the ssh command completes, which does not happen the interactive shell on $HOST that you started with ssh exits. Instead, you want to execute a script on $HOST.
(Also, note the correct way to iterate over the contents of hosts.txt.)
#!/bin/bash
while read HOST; do
do
if ssh root#$HOST '
STATUS=`/opt/agent.sh status | awk 'NR==1{print $3 $4}'`
[ "$STATUS" = "isrunning" ]
'; then
echo "$HOST == FAIL"
else
echo "$HOST == PASS"
fi
done < hosts.txt
The remote script simply exits with the result of comparing $STATUS to "isrunning". An if statement on the local host outputs a string based on the that result (which is the result of the ssh command itself). This saves the trouble of having to pass the value of $HOST to the remote host, simplifying the quoting required for the remote script.
Below is my snippet of shell script in which I am executing scp command to copy the files from machineB to machineA.
for element in ${x[$key]}; do # no quotes here
printf "%s\t%s\n" "$key" "$element"
if [ $key -eq 0 ]
then
scp david#machineB:/data/be_t1_snapshot/20131215/t1_$element_5.data /data01/primary/.
fi
done
I have a very simple question which is mentioned below -
If the above scp command in my shell script gives me this error for whatever reason - No such file or directory
then I need to try doing scp from machineC and for that scp command will be like this, only machine will be different and everything else will be same -
scp david#machineC:/data/be_t1_snapshot/20131215/t1_$element_5.data /data01/primary/.
So my question is how to check the output of the above scp command in my shell script and then decide whether I need to call scp command from machineC or not? Is there any status kind of thing which I can use to check and if it got failed for whatever reason, then I can call scp command on machineC?
Is this possible to do in shell script?
Here you go:
for element in ${x[$key]}; do # no quotes here
printf "%s\t%s\n" "$key" "$element"
if [ $key -eq 0 ]
then
scp david#machineB:/data/be_t1_snapshot/20131215/t1_$element_5.data /data01/primary/. || scp david#machineB:/data/be_t1_snapshot/20131215/t1_$element_5.data /data01/primary/
fi
done
Well-behaving commands exit with "success" (exit code = 0) if the operation was successful, or otherwise with an exit code != 0. You can chain commands together like this:
cmd && echo successful || echo failed
cmd && keep going || do something else
The exit code is also stored in the $? variable, so this is equivalent:
cmd; if $? = 0; then echo successful; else echo failed; fi
Not only this is possible, the status code of commands is extremely important in shell scripting. Consider these two examples:
./configure && make && make install
./configure; make; make install
The first one will execute the chain of commands if all are successful. The second will execute all of them always, even if an earlier command failed.
scp returns 0 only wen it succeeds.
so you can write like this:
scp machineB:/path/toyourfile .
if [ $? -ne 0 ]
then
scp machineC:/path/to/your/file .
fi
a shorter way is:
scp machineB:/path/toyourfile .
[ $? -eq 0 ] || scp machineC:/path/to/your/file .
or
scp machineB:/path/toyourfile .
[ $? -ne 0 ] && scp machineC:/path/to/your/file .
personally I prefer the even shorter way, and the scp output is of no use in script:
scp -q machineB:/path/to/your/file . || scp -q machineC:/path/to/your/file .
and remember to use ${element} instead of $element