#5115 Hard disk Standby timeout not working at all

[ullix]

(Milestone 1.881 actually)
I tried to set a Standby timeout for my harddisks, but no disk ever went to sleep or any other standby. In my new webmin controlled server I have 4 disks (SSD, HD 160GB, HD 2TB) from 3 manufacturer (Intel, Samsung, Seagate).

In my desktop I also experienced difficulties with hdparm, and solved it by using hd-idle, which so far has worked with all disks.

There isn't by any chance a hd-idle module or some other integration into webmin?

[ullix]

n.b.: all disks spin down immediately with hdparm -y /dev/sdX

[Jamie Cameron]

If you SSH in as root and run hdparm -S 123 /dev/sda1 (replace 123 with the desired timeout), do disks ever spin down?

[ullix]

Now even with a 5th disk in the server I logged in locally and did 'hdparm -S 2 /dev/sdX' on each of them and all went to sleep after 10sec, even the SSD with the OS on it (though woke up pretty much right away).

So, the disk behave exactly as they should.

I tried to set the timeout to 10min (setting 120) via webmin, and this was the answer:

Performing : hdparm -a 256 -d 0 -r 0 -k 0 -u 0 -m 0 -c 1 -A 1 -K 0 -P 0 -X 0 -W 0 -S 120 /dev/sdb

/dev/sdb:
setting fs readahead to 256
setting 32-bit IO_support flag to 1
setting multcount to 0

Looks ok to me, but now the disks do a spindown even faster (!) than the orginal 10sec, not more than 1 sec, if at all. Is it possible to readout the -S timeout value from the disk? (not to my knwledge) I am using 'hdparm -C /dev/sdX' to see the status (and my ears).

I then manually set 'hdparm -S 120 /dev/sdX' , and all disks properly went to sleep at 10min.

Something from webmin reached the disks, or the 10sec wouldn't have been reduced to ~1sec, but not the intended value of 10min.