Skip to content
Tutorials Center
My learning

Topics

Linux Command Reference ยท Concepts

Exit codes cheat sheet

A quick table of common Linux exit statuses and signal codes: 0, 1, 2, 126, 127, 128+N, 130, 137, 139, 143, plus tool-specific codes for grep, diff, curl and ssh.

An exit code (exit status) is the number a finished process hands back to its parent: 0 for success, anything from 1 to 255 for failure. The shell keeps the last one in $?. This page is the lookup table; for how to set and test statuses see Bash exit status.

Most codes are conventions, not laws, but a handful are fixed by the shell and the kernel. When a process is killed by a signal, the shell reports 128 + signal number, so a status above 128 usually means "it did not exit by itself". Knowing the table lets you diagnose a failed job from its number alone.

Terminal
some_command
echo "status $?"
kill -l 9     # translate a signal number to a name

Try it

Bash Generate the common codes

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with read or cat), it gets the next line from this box.

Bash Decode a status into plain words

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with read or cat), it gets the next line from this box.

The table

Shell and kernel codes

Code Meaning
0 Success.
1 General error; the default failure of most tools and false.
2 Misuse of a builtin or an invalid option (some shells also use it for syntax errors).
126 Command found but not executable (permissions or wrong file type).
127 Command not found (also an unknown interpreter in a shebang).
128 Invalid argument to exit.
128+N Terminated by signal N.
130 128+2, SIGINT: you pressed Ctrl+C.
131 128+3, SIGQUIT: Ctrl+\, often with a core dump.
137 128+9, SIGKILL: forced kill, frequently the out-of-memory killer or kill -9.
139 128+11, SIGSEGV: segmentation fault.
141 128+13, SIGPIPE: wrote to a closed pipe, e.g. yes | head.
143 128+15, SIGTERM: polite termination request, e.g. kill PID, systemctl stop.
255 Exit status out of range, or a generic fatal error (also ssh connection failure).

Common tool conventions

Tool Codes
grep 0 match found, 1 no match, 2 error.
diff, cmp 0 identical, 1 different, 2 trouble.
test, [ 0 true, 1 false, 2 bad expression.
timeout 124 the command timed out, 125 timeout itself failed, 126 and 127 as above; --preserve-status keeps the command's own.
ssh The remote command's status, or 255 if the connection or authentication failed.
curl 6 host not resolved, 7 connection refused, 22 HTTP error with -f, 28 timeout.
sudo The command's status; 1 if sudo itself refuses.
tar 0 ok, 1 files changed while archiving, 2 fatal.
rsync 23 partial transfer due to error, 24 files vanished, 30 timeout.
make 2 when any recipe failed.

Errors and edge cases: statuses are 8 bits, so exit 256 becomes 0 and exit -1 becomes 255. Statuses above 255 can never be observed.

Usage notes

  • Above 128 means a signal, so subtract 128 and look the number up with kill -l. Conveniently, kill -l 137 prints the name directly.
  • 137 in a container almost always means the process hit its memory limit and was killed; check dmesg or the container runtime's OOM report.
  • Pipelines report the last stage; use set -o pipefail or ${PIPESTATUS[@]} to see the others: Strict mode.
  • Pick your own codes with care. In scripts use 0, 1, 2 and a small range such as 64-113 (the BSD sysexits.h convention starts at 64); avoid 126-255.
  • A service's exit code appears in systemctl status and decides restarts: Restart policies.
  • $? is overwritten by every command, including echo. Save it right away: rc=$?.
  • Test, do not print. if grep -q pat file; then ... is more robust than comparing $? to numbers.