Linux script fails.

Usage and Config
1

Seems like a simple script, but it fails with a -127. Any ideas would be great.

thanks

createfile until end
#!/bin/sh
/tmp/SEAClient_lin_State_of_Ohio_SOCC.bin
end
//move __appendfile /tmp/InstallSEA.sh
move __createfile /tmp/InstallSEA.sh
wait chmod a+x /tmp/InstallSEA.sh
if {exists match (case insensitive regex “(linux|esx)”) of name of operating system}
wait sh /tmp/InstallSEA.sh > /tmp/seaout.txt 2>&1 &
//wait /tmp/SEAClient_lin_State_of_Ohio_SOCC.bin
else
wait ksh /tmp/InstallSEA.sh
endif

2

Try running it as

wait /bin/sh -c “/tmp/InstallSEA.sh > /tmp/seaout.txt 2>&1”

Oh, also, try deleting the file from /tmp before copying or moving over that path. Bigfix will not replace the file if it already exists, and will fail with an error message if that is the case.

3

Well, here is the out put of when I ran it from the script;

/tmp/InstallSEA.sh: line 2: 21780 Segmentation fault (core dumped) /tmp/SEAClient_lin_State_of_Ohio_SOCC.bin

But if I run exactly the same command on the console, it works fine. I can’t figure out a way around it.

4

In that case it may have to do with the environment. The BigFix session is not running though a login shell and so may not have some environment variables populated that maybe your script or the binary depends upon.

You might try forcing it to run as a login shell (which sources the root account’s dot-files to build the environment) via

wait /bin/sh --login -c “/tmp/InstallSEA.sh > /tmp/seaout.txt 2>&1”

Also, check whether you have SELinux, fstab options, or any other security package that might block running executables from the /tmp path.

5

That was the ticket! THANK YOU!

6

You’re very welcome…which part was it? The login shell, or execute rights on /tmp ?

7

login shell. I knew there was a difference in the environment that had something to do with the login, but I didn’t know how to overcome it.

Thanks again

1 Like
8

Jason, I may have to put you on the payroll, but I have one last question; How do I cause the client to restart on a failure on linux. I could do it if it were systemd, but it has that init file and i am not that smart.

9

If you are trying to restart the client from an Action/Task, the

client restart

command will restart the client, and works cross-platform.

Outside of the BES Client, from a root shell on Red Hat, you can use

service besclient restart

or

/etc/init.d/besclient restart

10

I understand that. If you were using a straight up service file, the option of “Restart=on-failure” is availble. Inittab would make a “respawn” work. But if you have this mesh of the two, is there a ready made option or do you have to like run a cron job or something.

Thanks