Hard disk spins up every 30 minutes

J

Joshua M

New Member
Feb 13, 2024
21
3
3
I had PVE and PBS co-hosted and running fine for a few months. Yesterday, the computer locked-up and I had to restart it.

After re-starting, the hard disks spin up every 30 minutes only to sleep again shortly afterwards.

All installation and files are on SSDs, the HDDs are only there to receive backups which run once per day. Previously the drives remained in sleep and only woke up for the backup. I don't want to keep the drives spinning due to noise, and energy usage.

I'm trying to figure out what could have changed. I have followed this (but shouldn't be relevant as I have no LVM and use ZFS): https://forum.proxmox.com/threads/pvestatd-doesnt-let-hdds-go-to-sleep.29727/

And also disabled and stopped the pvestatd.service to make sure. I also stopped proxmox-backup-proxy.service to rule this out. I also killed rrdcached and commented out the journal file. At first I suspected the journal commits as these happen every 30 mins and around the same time, but after the commenting out of the journal entry and stopping the rrdcached service, I still kept getting the wake-ups.

monitoring journalctl doesn't show anything obvious at the wake-up times.

I'm running out of ideas. anyone have an idea or did something change in proxmox/pbs the last few months which would cause this and which only triggered on the reboot? any help welcome.
 
I'm in the same situation. In my case, it seems that smart is the culprit.

Code: 
systemctl stop smartmontools
appears to have stopped it. I think it should be possible to change the configuration in /etc/smartd.conf to improve it, potentially some kind of self test that is triggered every 30 minutes, but I haven't tried that yet.

I noticed that
Code: 
journalctl -u smartmontools
says "Device: /dev/sda [SAT], no ATA CHECK POWER STATUS support, ignoring -n Directive", that's probably why the drives are spinning up.
 
  • Like
Reactions: Joshua M
Thanks. I think you may be right. I do see:

Dec 14 12:59:36 vm5 proxmox-backup-proxy[2205]: rrd journal successfully committed (25 files in 0.019 seconds)

in logs exactly when the drives spin up, but smartd logs appear right after that.

Now that you mention it, I do remember stopping smartmontools, but maybe I didn't disable it. I will do so now. It also makes sense as one drive does not spin up and it is a newer one which presumable can do the self-check without spin up. I couldn't figure out why it would not and the others were spinning up as they were part of the same RAID array but now it makes sense if it is SMART.

Thanks! I've been tearing my hair out trying to locate this!
 
Last edited:
You must log in or register to reply here.
Top Bottom