What Does 'set -e' Mean in a Bash Script?

TLDR

The set -e command in a Bash script tells the shell to exit immediately if any command returns a non-zero (error) status. This helps catch failures early and prevents scripts from continuing after an error, which is especially useful in automation and CI/CD pipelines. However, it can sometimes cause scripts to exit unexpectedly if not used carefully.

Why Use set -e?

By default, Bash scripts keep running even if a command fails. This can lead to subtle bugs or unwanted side effects, especially in deployment or automation scripts. Adding set -e at the top of your script makes it safer by stopping execution as soon as something goes wrong.

Here's a simple example:

#!/bin/bash
set -e

mkdir /tmp/mydir
cd /tmp/mydir
cp /nonexistent/file .
echo "This line will not run if the copy fails."

What happens here?

• If cp /nonexistent/file . fails (because the file doesn't exist), the script exits immediately.
• The echo line is never reached.

Without set -e, the script would keep running, which could cause problems later on.

How Does It Work?

When you use set -e, Bash checks the exit status of each command. If any command (that isn't part of an if or while test, or a command in a pipeline that is handled) fails, the script stops right there.

The overall flow looks like this:

+-------------------+
|  Script starts    |
+-------------------+
          |
          v
+-------------------+
|  Run command 1    |
+-------------------+
          |
          v
+-------------------+
|  Command fails?   |--Yes--> Exit script
+-------------------+
          |
         No
          v
+-------------------+
|  Run next command |
+-------------------+

Common Pitfalls and Caveats

While set -e is helpful, it can sometimes cause confusion:

• Pipelines: By default, set -e only checks the exit status of the last command in a pipeline. To make the script exit if any command in a pipeline fails, add set -o pipefail as well.
• Commands in if or while: If a command fails inside an if or while test, the script does not exit.
• Subshells and functions: Errors in subshells or functions may not always cause the main script to exit, depending on how they're called.

Example with pipefail:

#!/bin/bash
set -e
set -o pipefail

grep "foo" file.txt | sort | tee output.txt

If grep fails (e.g., file.txt doesn't exist), the script exits immediately, not just if tee fails.

When Not to Use set -e

There are cases where you might not want your script to exit on every error:

• When you expect some commands to fail and want to handle those failures manually.
• In scripts that need to clean up resources even after an error (use trap for cleanup).

You can temporarily disable set -e with set +e and re-enable it with set -e:

set +e
some_command_that_might_fail
set -e

Conclusion

Using set -e in your Bash scripts is a simple way to make them more robust and catch errors early. Just be aware of its limitations and test your scripts to avoid surprises. For even safer scripts, combine it with set -o pipefail and handle expected errors explicitly.