TL;DR version: btrace suggests the culprit is a "[pool]" accessing the disk every 600 seconds. What is it and how do I stop it?
Longer version:
I've been using OMV for a while and I had this working before, but at some point it stopped working (probably after some upgrade or installing some software).
I have 2 WD Red NAS HDDs in RAID1 (using mdadm), that are known for having some issues regarding spin down, so I installed hd-idle as a workaround (as suggested in an older thread here) and I managed to get it to work for while... but now it doesn't work unless I put a very low idle time.
I've increased the smart check interval on OMV web interface to a high number. It currently is 86400 and power mode is set to Standby.
I've set hd-idle to 1800 seconds, but the disks never spindown... If I change that time to very short time (for example 180 seconds) the disk will spin down, but not with 1800.
I've run btrace and there's a "pool" accessing the disk about every 600 seconds. Example:
8,0 1 55 4200.006161055 3626119 D N 0 [pool]
8,0 1 56 4200.006433144 0 C N [0]
8,0 1 57 4200.007231231 3626119 D R 512 [pool]
8,0 1 58 4200.007689361 0 C R [0]
(...)
(there's a few more repeated lines each time and it will repeat every 600 seconds)
This seems to be the culprit...
If I put for example 180 on the idle time for hd-idle the discs will spin down and will remain like that.
[pool] entries still shows up in btrace (now just once every 600 sec) but it doesn't "wake up" the disks.
"ps aux|grep pool" reveals only this:
www-data 788 0.0 0.0 204488 6088 ? S out13 0:00 php-fpm: pool www
www-data 799 0.0 0.0 204488 6104 ? S out13 0:00 php-fpm: pool www
Is this the pool in btrace? How do I stop it from preventing the disks to spin down?
Any insight is most welcome! Thanks.