Trixie, mdadm, and systemd
Hi all,
Does anyone know if the defaults for mdadm make sense anymore?
In particular, the three options for:
AUTOCHECK, AUTOSCAN, and START_DAEMON
It would *appear* that systemd has timers for these things now.
I'm confused as to whether systemd, mdadm, or both need configuring here.
Does systemd use the information in /etc/default/mdadm to determine what
to do?
What about arrays that get hotplugged - does systemd pick them up too?
or do I need mdadm for that? If so, how does that interact with systemd?
I used dpgk-reconfigure mdadm to disable the two checks, and the systemd
timers still exist as per:
root@orlith:/var/www/nextcloud# systemctl list-timers
NEXT LEFT
LAST PASSED UNIT ACTIVATES
Thu 2024-11-21 10:24:20 GMT 19h Wed 2024-11-20 03:19:50
GMT - mdcheck_continue.timer mdcheck_continue.service
Thu 2024-11-21 12:13:25 GMT 21h Wed 2024-11-20 14:17:45
GMT - mdmonitor-oneshot.timer mdmonitor-oneshot.service
Sun 2024-12-01 12:28:51 GMT 1 week 3 days Tue 2024-11-12 05:26:06
GMT - mdcheck_start.timer
I noticed that when reconfiguring mdadm, the package config recommended
having it email root if it found inconsistencies. Does the systemd stuff
take on this role now? or should I leave that option enabled?
What is the proper way to set this up? It seems to me that whilst
systemd might be very clever, none of this stuff is documented properly.
-Ian
Reply to: