mdadm --monitor --program option not working

Question

I am trying to make mdadm call into a simple bash script which writes a message in the kernel log in case of a state change.

I am running on a VM (qemu). The VM is running debian 12.5, and kernel version 6.1.118.

The script is as follows,

# cat /tmp/test.sh
#!/bin/bash

echo "raid array status change" > /dev/kmsg

I have added the following to the mdadm config file

PROGRAM /tmp/test.sh
ARRAY /dev/md0 UUID=<UUID of md0>

But when I do the test, it says the md0 is being picked up, and the correct program option is being used, but nothing gets printed in dmesg

mdadm --monitor --test /dev/md0 -1

Note that when I manually run the script, it prints the message in kernel log.

My questions

1. Does the above process of calling the program depend on the mail setting also? Because my mail config is not set.
2. Any idea what could be wrong or missing?

Answer 1

^{This is a suggestion to partition the problem. It isn't an answer as such but I can't fit all this into comments. I will delete it afterwards if I haven't turned it into a proper answer}

---

First of all ensure that the functional components work as expected.

1. Ensure that running echo "raid array status change" > /dev/kmsg as root does indeed generate a line that can be picked up by dmesg
2. Ensure that running your code /tmp/test.sh as root - and specifically not bash /tmp/test.sh - does what you want

Assuming those tests succeed, try the following. Create the program /opt/event.sh as follows:

#!/bin/bash
#
event=$1
device=$2
shift 2
extras=("$@")

logger "MDADM EVENT TEST: event=$1, device=$2, extras=(${extras[*]})"
exit 0

and make it executable with chmod a+rx /opt/event.sh.

Edit the file /etc/mdadm/mdadm.conf and ensure there is exactly one PROGRAM line:

# program to run on an event
PROGRAM /opt/event.sh

Restart the MDADM monitoring service:

systemctl restart mdmonitor

Immediately look in the log files for the test event trigger that's generated when the monitoring service (re)starts:

find /var/log -type f -mmin -5 -print0 | xargs -0r grep -h MDADM

You are looking for a line like this:

2025-08-19T09:43:59.087253+00:00 rHost root: MDADM EVENT TEST: event=, device=, extras=()

If you find one the event.sh program has been correctly run. If you don't, it hasn't, and the problem is outside the mdadm subsystem.