recur

...is a tiny utility for running recurring tasks on Linux or Unix-like systems. It has no dependencies aside from libc and produces a static build of less than 50 KB, making it suitable for use in containers or low-resource environments.

Features

The functionality provided by recur is very simple:

• It runs a given command repeatedly, at a set interval;
• It stops if the command fails at some point, and returns its exit code;
• Optionally, it can use a file to remember the last time the command successfully executed.

The last feature is useful in cases where you don't want a restart to cause the command to rerun too early, e.g. for heavier or less frequent tasks like renewing TLS certificates or performing backups.

Building

Install musl-gcc (on Ubuntu, you will need the gcc and musl-tools packages) and run ./build.sh. This will produce a static binary in build/recur.

Alternatively, compile recur.c with a C compiler of your choice.

Command-line interface

./recur -i <interval in seconds> [-f <file to store last execution time>] -- <command>

For example:

1. Execute /bin/echo 123 every 30 seconds:

   ./recur -i 30 -- /bin/echo 123
   

2. Execute /bin/date every 5 minutes, and remember the last execution time in /tmp/last_execution:

   ./recur -i 300 -f /tmp/last_execution -- /bin/date
   

   If you interrupt and restart this command while it is waiting, it should still execute /bin/date 5 minutes after the last execution.
3. Execute a command that fails with an exit code and capture the exit code:

   ./recur -i 10 -- /bin/sh -c "exit 123"
   echo $?
   

   recur should stop and it should return with exit code 123.

License

Copyright (c) 2025 Valentin Dimov

This project is licensed under the zero-clause BSD license, see LICENSE.