Disk spin-down: What program to use and best practice today?

English Install/Boot/Login
1

Hi,
Disk spin-down: What program to use and best practice today?
There are threads about udisk2, hdparm, hd-idle (malcolmlewis](https://forums.opensuse.org/member.php/740-malcolmlewis)). What is the best with Leap 15.3 and how to use it?

Thanks

2

Nobody to advise me?

Use case:
I switch on my desktop every day. I have 4 HDD, each is used for different application. I don’t use all the applications every day, so I don’t need access to all the HDD every day. I would be glad to boot the PC without starting the 4 HDD.
I was thinking of an utility that could reach my need;
An applet that I could add in the panel with for each HDD a select option for starting it manually.
In a setting window for each HDD:

  • Starting the HDD: “at boot(as usual)/manual (with the applet in the panel)/on program demand (when an application want to access the HDD)”. This selection is persistent so when manual or on program demand is selected, the corresponding HDD doesn’t start at boot.
  • Idle delay time for each HDD.
    Improvement,
  • Memory size on system SSD (persistent), to memorize the tree structure and directories/files name when looking at the idle HDD with a file manager. It allow to see the n top directories and files names without starting the HDD. The structure and name is recorded just before the HDD is changing to idle.

Is it possible to make this type of utility, does it exist?
But the first step is to find the good way to spin-down the HDD (my first request).

3

hd-idle starts on boot and manages the HDDs on my machines:

**erlangen:~ #** journalctl -b -1 -o short-monotonic -u hd-idle.service  
    3.464288] erlangen systemd[1]: Started hd-idle disk spindown service. 
    3.465527] erlangen hd-idle[762]: hd-idle starting in nodaemon mode 
    3.465527] erlangen hd-idle[762]:   disk: sdb timeout: 300 
    3.465527] erlangen hd-idle[762]:   default timeout: 0 
  393.469635] erlangen hd-idle[762]: spindown: sdb 
 7863.959542] erlangen hd-idle[762]: spinup: sdb 
 8463.966742] erlangen hd-idle[762]: spindown: sdb 
[11244.411224] erlangen systemd[1]: Stopping hd-idle disk spindown service... 
[11244.425346] erlangen systemd[1]: hd-idle.service: Deactivated successfully. 
[11244.425450] erlangen systemd[1]: Stopped hd-idle disk spindown service. 
**erlangen:~ #**

Configuration is easy:

**erlangen:~ #** cat /etc/default/hd-idle  
# hd-idle command line options 
# Options are: 
#  -a <name>               Set device name of disks for subsequent idle-time 
#                          parameters (-i). This parameter is optional in the 
#                          sense that there's a default entry for all disks 
#                          which are not named otherwise by using this 
#                          parameter. This can also be a symlink 
#                          (e.g. /dev/disk/by-uuid/...) 
#  -i <idle_time>          Idle time in seconds. 
# 
# Options not exactly useful here: 
#  -t <disk>               Spin-down the specfified disk immediately and exit. 
#  -d                      Debug mode. This will prevent hd-idle from 
#                          becoming a daemon and print debugging info to 
#                          stdout/stderr 
#  -h                      Print usage information. 
#  -l <logfile>            Name of logfile (written only after a disk has spun 
#                          up). Please note that this option might cause the 
#                          disk which holds the logfile to spin up just because 
#                          another disk had some activity. This option should 
#                          not be used on systems with more than one disk 
#                          except for tuning purposes. On single-disk systems, 
#                          this option should not cause any additional spinups. 
# 
# spin down all disks after 180 seconds 
#HD_IDLE_OPTS="-i 180" 
# 
# only spin down /dev/sdb after 180 seconds 
#HD_IDLE_OPTS="-i 0 -a sdb -i 180" 
**HD_IDLE_OPTS="-i 0 -a /dev/disk/by-id/scsi-SATA_WDC_WD40EZRX-22S_WD-WCC4E2FYXSNV -i 300" **
**erlangen:~ #**
4

Hi
You need to test them all to see what works best for your setup. You could try powering off the device, but does the system motherboard support SATA hotswap?

5

Almost done but…

~> cat /etc/default/hd-idle
...
# sda
#HD_IDLE_OPTS="-i 0 -a /dev/disk/by-id/scsi-SATA_WDC_WD40EFRX-68N_WD-WCC7K1URKCCA -i 180"
HD_IDLE_OPTS="-i 0 -a sda -i 180"
# sdb
#HD_IDLE_OPTS="-i 0 -a /dev/disk/by-id/scsi-SATA_WDC_WD40EFRX-68N_WD-WCC7K3FP1KRD -i 180"
HD_IDLE_OPTS="-i 0 -a sdb -i 180"
# sdc
#HD_IDLE_OPTS="-i 0 -a /dev/disk/by-id/scsi-SATA_WDC_WD40EFRX-68W_WD-WCC4E1XXZ0TY -i 180"
HD_IDLE_OPTS="-i 0 -a sdc -i 180"
# sdd
#HD_IDLE_OPTS="-i 0 -a /dev/disk/by-id/scsi-SATA_WDC_WD40EFRX-68N_WD-WCC7K1LDVZ6H -i 180"
#HD_IDLE_OPTS="-i 0 -a sdd -i 180"

~> sudo journalctl -b -o short-monotonic -u hd-idle.service
[sudo] password for root: 
-- Logs begin at Mon 2022-03-14 17:46:44 GMT, end at Tue 2022-03-15 15:48:38 GMT. --
    6.175759] localhost systemd[1]: Started hd-idle disk spindown service.
    6.176789] localhost hd-idle[1118]: hd-idle starting in nodaemon mode
    6.176789] localhost hd-idle[1118]:   disk: sdc timeout: 180
    6.176789] localhost hd-idle[1118]:   default timeout: 0

As you can see only the last HDD in the list is taken into consideration, confirmed by few tries, by label or by id.

Global idle is working fine.

HD_IDLE_OPTS="-i 180"

Is it a bug?

6

What do you mean by “powering off the device”:

  • Hardware switch to open the +V wires? :frowning:
  • Software power on (off should be done by hd-idle). Is it possible? How?
    ASRock Steel Legend allows hotsplug (I’d guess it’s hotswap equivalent)
7

Hi
That’s expected, only one variable used, you specify all as one…


HD_IDLE_OPTS="-i 0 -a sda -i 180 -a sdb -i 180 -a sdc -i 180 -a sdd -i 180"

{Following should spin all down}
HD_IDLE_OPTS="-i 180"
8

My mistake, this is a variable. :shame:
Working fine.
Thanks

This is a first step, I’d be glad to reach what I explained in post #2
Maybe it’s not possible by software as at boot the power supply provides voltage to HDD to spin before any software command.

9

Hi
I use the following on another box (Intel M/B Leap 15.3) to just power off the devices and unplug from a SATA disk caddy…


if (( $# != 1 )); then
   echo "Usage: $0 <Drive to remove /dev/sNX>" >&2
   exit 1
fi


if test -h "$1"
then
   disk=$(chase "$1")
else
   disk="$1"
fi


echo $disk


if test -b "$disk"
then
   echo "Syncing and unmounting $disk"
   sync
   umount /dev/$disk
   sync
   echo "Powering off $disk"
   echo 1 >/sys/block/$(basename "$disk")/device/delete
   echo "Disk $disk safe to remove"
else
   echo "$0: not a block device: $1" >&2
   exit 1
fi
10

Your script is good for a USB HDD but the issue with a SATA HDD is the power on/booting, the HDD start to spin.
Maybe it’s not possible to avoid this by software as at boot the power supply provides voltage to HDD to spin before any software command.
I tried to connect a HDD with only power connector, it spins. This is the setting by default to gain time. Is it possible to have “do not spin and wait the command” by default?

11

Hi
No, that script is used with SATA disks connected to the motherboard SATA connectors (The disks are in a 4xSATA Caddy) and once powered down disconnect (unplug), but they don’t need to be, just turn the power back on via another script, but in my case when plugging back in come back to life. The motherboard I have supports real hot swapping.

12

I learned something today. HDD power switches are for sell.
I didn’t know that.
So this is the only way.

For information, when HDD spin is off by hd-idle, it restarts after suspend to RAM and restart just before power off the PC.

13

Disk spins down and stays down:

**erlangen:~ #** journalctl -b -1 -o short-monotonic -u hd-idle.service -u systemd-suspend.service              
    3.038747] erlangen systemd[1]: hd-idle.service: Installed new job hd-idle.service/start as 224
    3.302810] erlangen hd-idle[757]: hd-idle starting in nodaemon mode 
    3.302810] erlangen hd-idle[757]:   disk: sdb timeout: 300 
    3.302810] erlangen hd-idle[757]:   default timeout: 0 
    3.303622] erlangen systemd[1]: hd-idle.service: Passing 0 fds to service
    3.303968] erlangen systemd[1]: hd-idle.service: About to execute /usr/sbin/hd-idle -n "\$HD_IDLE_OPTS"
    3.304666] erlangen systemd[1]: hd-idle.service: Forked /usr/sbin/hd-idle as 757
    3.305034] erlangen systemd[1]: hd-idle.service: Changed dead -> running
    3.305139] erlangen systemd[1]: hd-idle.service: Job 224 hd-idle.service/start finished, result=done
    3.305241] erlangen systemd[1]: Started hd-idle disk spindown service. 
    3.306532] erlangen systemd[757]: hd-idle.service: Executing: /usr/sbin/hd-idle -n -i 0 -a /dev/disk/by-id/scsi-SATA_WDC_WD40EZRX-22S_WD-WCC4E2FYXSNV -i 300
    4.763593] erlangen systemd[1]: systemd-suspend.service: Collecting.
    5.088142] erlangen systemd[1]: systemd-suspend.service: Collecting.
   22.518632] erlangen systemd[1]: systemd-suspend.service: Collecting.
   22.878986] erlangen systemd[1]: systemd-suspend.service: Collecting.
  333.306337] erlangen hd-idle[757]: spindown: sdb 
  541.545341] erlangen systemd[1]: systemd-suspend.service: Collecting.
  541.551879] erlangen systemd[1]: systemd-suspend.service: Installed new job systemd-suspend.service/start as 1726
  541.558094] erlangen systemd[1]: systemd-suspend.service: Passing 0 fds to service
  541.558380] erlangen systemd[1]: systemd-suspend.service: About to execute /usr/lib/systemd/systemd-sleep suspend
  541.558731] erlangen systemd[1]: systemd-suspend.service: Forked /usr/lib/systemd/systemd-sleep as 2652
  541.559310] erlangen systemd[1]: systemd-suspend.service: Changed dead -> start
  541.559465] erlangen systemd[1]: Starting System Suspend... 
  541.559652] erlangen systemd[2652]: systemd-suspend.service: Executing: /usr/lib/systemd/systemd-sleep suspend
  541.599605] erlangen systemd-sleep[2654]: INFO: Skip running /usr/lib/systemd/system-sleep/grub2.sleep for suspend 
  541.602814] erlangen systemd-sleep[2652]: Entering sleep state 'suspend'... 
  553.898986] erlangen systemd-sleep[2711]: INFO: Skip running /usr/lib/systemd/system-sleep/grub2.sleep for suspend 
  553.905301] erlangen systemd[1]: systemd-suspend.service: Failed to read oom_kill field of memory.events cgroup attribute: No such file or directory
  553.905345] erlangen systemd[1]: systemd-suspend.service: Child 2652 belongs to systemd-suspend.service.
  553.905380] erlangen systemd[1]: systemd-suspend.service: Main process exited, code=exited, status=0/SUCCESS (success)
  553.905456] erlangen systemd[1]: systemd-suspend.service: Deactivated successfully. 
  553.905502] erlangen systemd[1]: systemd-suspend.service: Service will not restart (restart setting)
  553.905536] erlangen systemd[1]: systemd-suspend.service: Changed start -> dead
  553.905574] erlangen systemd[1]: systemd-suspend.service: Job 1726 systemd-suspend.service/start finished, result=done
  553.905614] erlangen systemd[1]: Finished System Suspend. 
  553.905765] erlangen systemd[1]: systemd-suspend.service: Consumed 114ms CPU time.
  553.906093] erlangen systemd[1]: systemd-suspend.service: Control group is empty.
  553.907054] erlangen systemd[1]: systemd-suspend.service: Collecting.
 1194.987125] erlangen systemd[1]: systemd-suspend.service: Collecting.
 1199.966308] erlangen systemd[1]: systemd-suspend.service: Collecting.
 1199.979277] erlangen systemd[1]: hd-idle.service: Installed new job hd-idle.service/stop as 2292
 1200.025407] erlangen systemd[1]: hd-idle.service: Changed running -> stop-sigterm
 1200.025539] erlangen systemd[1]: Stopping hd-idle disk spindown service... 
 1200.036316] erlangen systemd[1]: hd-idle.service: Failed to read oom_kill field of memory.events cgroup attribute: No such file or directory
 1200.036389] erlangen systemd[1]: hd-idle.service: Child 757 belongs to hd-idle.service.
 1200.036416] erlangen systemd[1]: hd-idle.service: Main process exited, code=killed, status=15/TERM (success)
 1200.036459] erlangen systemd[1]: hd-idle.service: Deactivated successfully. 
 1200.036489] erlangen systemd[1]: hd-idle.service: Service restart not allowed.
 1200.036518] erlangen systemd[1]: hd-idle.service: Changed stop-sigterm -> dead
 1200.036542] erlangen systemd[1]: hd-idle.service: Job 2292 hd-idle.service/stop finished, result=done
 1200.036566] erlangen systemd[1]: Stopped hd-idle disk spindown service. 
 1200.036706] erlangen systemd[1]: hd-idle.service: Consumed 10ms CPU time.
 1200.047303] erlangen systemd[1]: hd-idle.service: Control group is empty.
**erlangen:~ #**
14

I try to understand the script as I am wandering if I can use the same.
After search on the web, I can’t find explanations for:

  • disk=$(chase “$1”) > What does chase?
  • echo 1 >/sys/block/$(basename “$disk”)/device/delete > What exactly does /delete to the device (software, hardware)? Is it possible to restart the device after that, how.

malcolmlewis, I’d be glad to get the second script to start the devices.
Thanks

15

This is what happen even when I don’t use any HDD. System is in nvme SSD.

~> sudo journalctl -b -o short-monotonic -u hd-idle.service -u systemd-suspend.service
[sudo] password for root: 
-- Logs begin at Mon 2022-03-14 17:46:44 GMT, end at Wed 2022-03-16 12:05:19 GMT. --
    6.237117] localhost systemd[1]: Started hd-idle disk spindown service.
    6.238452] localhost hd-idle[1125]: hd-idle starting in nodaemon mode
    6.238452] localhost hd-idle[1125]:   disk: sdd timeout: 300
    6.238452] localhost hd-idle[1125]:   disk: sdc timeout: 300
    6.238452] localhost hd-idle[1125]:   disk: sdb timeout: 300
    6.238452] localhost hd-idle[1125]:   disk: sda timeout: 300
    6.238452] localhost hd-idle[1125]:   default timeout: 0
  366.241048] ASR5600G hd-idle[1125]: spindown: sda
  366.566703] ASR5600G hd-idle[1125]: spindown: sdb
  367.307255] ASR5600G hd-idle[1125]: spindown: sdc
  367.710989] ASR5600G hd-idle[1125]: spindown: sdd
 1808.945861] ASR5600G hd-idle[1125]: spinup: sda
 1808.945861] ASR5600G hd-idle[1125]: spinup: sdb
 1808.945861] ASR5600G hd-idle[1125]: spinup: sdc
 1808.945861] ASR5600G hd-idle[1125]: spinup: sdd
 2138.948133] ASR5600G hd-idle[1125]: spindown: sda
 2139.225745] ASR5600G hd-idle[1125]: spindown: sdb
 2139.495551] ASR5600G hd-idle[1125]: spindown: sdc
 2139.906531] ASR5600G hd-idle[1125]: spindown: sdd
 3610.172364] ASR5600G hd-idle[1125]: spinup: sda
 3610.172364] ASR5600G hd-idle[1125]: spinup: sdb
 3610.172364] ASR5600G hd-idle[1125]: spinup: sdc
 3610.172364] ASR5600G hd-idle[1125]: spinup: sdd
 3910.174305] ASR5600G hd-idle[1125]: spindown: sda
 3910.174863] ASR5600G hd-idle[1125]: spindown: sdb
 3910.174863] ASR5600G hd-idle[1125]: spindown: sdc
 3910.174988] ASR5600G hd-idle[1125]: spindown: sdd
 5410.185095] ASR5600G hd-idle[1125]: spinup: sda
 5410.185095] ASR5600G hd-idle[1125]: spinup: sdb
 5410.185095] ASR5600G hd-idle[1125]: spinup: sdc
 5410.185095] ASR5600G hd-idle[1125]: spinup: sdd
 5710.187060] ASR5600G hd-idle[1125]: spindown: sda
 5710.196890] ASR5600G hd-idle[1125]: spindown: sdb
 5710.196890] ASR5600G hd-idle[1125]: spindown: sdc
 5710.196890] ASR5600G hd-idle[1125]: spindown: sdd
 7210.197720] ASR5600G hd-idle[1125]: spinup: sda
 7210.197720] ASR5600G hd-idle[1125]: spinup: sdb
 7210.197720] ASR5600G hd-idle[1125]: spinup: sdc
 7210.197720] ASR5600G hd-idle[1125]: spinup: sdd

Is there a log where I could know why this happen, according to the timestamp? How to read it?

16

The above shows plain and utter blunder. hd-idle does what it advertises. It spins down the disk(s) after specified idle time and monitor both spin up and spin down:

**erlangen:~ #** systemctl status hd-idle.service  
**●** hd-idle.service - hd-idle disk spindown service 
     Loaded: loaded (/usr/lib/systemd/system/hd-idle.service; enabled; vendor preset: disabled) 
     Active: **active (running)** since Wed 2022-03-16 05:48:08 CET; 7h ago 
   Main PID: 759 (hd-idle) 
      Tasks: 1 (limit: 4915) 
        CPU: 217ms 
     CGroup: /system.slice/hd-idle.service 
             └─759 /usr/sbin/hd-idle -n -i 0 -a /dev/disk/by-id/scsi-SATA_WDC_WD40EZRX-22S_WD-WCC4E2FYXSNV -i 300 

Mar 16 05:48:08 erlangen systemd[1]: Started hd-idle disk spindown service. 
Mar 16 05:48:08 erlangen hd-idle[759]: hd-idle starting in nodaemon mode 
Mar 16 05:48:08 erlangen hd-idle[759]:   disk: sdb timeout: 300 
Mar 16 05:48:08 erlangen hd-idle[759]:   default timeout: 0 
Mar 16 05:53:38 erlangen hd-idle[759]: spindown: sdb 
Mar 16 06:08:08 erlangen hd-idle[759]: spinup: sdb 
Mar 16 06:13:08 erlangen hd-idle[759]: spindown: sdb 
**erlangen:~ #**

**erlangen:~ #** journalctl -b -u hd-idle.service  
Mar 16 05:48:08 erlangen systemd[1]: Started hd-idle disk spindown service. 
Mar 16 05:48:08 erlangen hd-idle[759]: hd-idle starting in nodaemon mode 
Mar 16 05:48:08 erlangen hd-idle[759]:   disk: sdb timeout: 300 
Mar 16 05:48:08 erlangen hd-idle[759]:   default timeout: 0 
Mar 16 05:53:38 erlangen hd-idle[759]: spindown: sdb 
Mar 16 06:08:08 erlangen hd-idle[759]: spinup: sdb 
Mar 16 06:13:08 erlangen hd-idle[759]: spindown: sdb 
**erlangen:~ #**

Setting LogLevel=debug and rebooting results in additional messages as already shown in post #13.

**erlangen:~ #** grep LogLevel /etc/systemd/system.conf 
**LogLevel**=debug
**erlangen:~ #**
17

Hi
I need to find and rebuild the chase command, it’s for following symlinks. The system I used has been semi retired since I got my HP Z440 workstation…

To show clock time, don’t use the -o short-monotonic

I would also check out this since you have WD drives…

https://build.opensuse.org/package/show/utilities/idle3-tools

18 
~> sudo systemctl status hd-idle.service
[sudo] password for root: 
● hd-idle.service - hd-idle disk spindown service
     Loaded: loaded (/usr/lib/systemd/system/hd-idle.service; enabled; vendor preset: disabled)
     Active: active (running) since Wed 2022-03-16 10:04:57 GMT; 3h 57min ago
   Main PID: 1125 (hd-idle)
      Tasks: 1 (limit: 4915)
     CGroup: /system.slice/hd-idle.service
             └─1125 /usr/sbin/hd-idle -n -i 0 -a /dev/disk/by-id/scsi-SATA_WDC_WD40EFRX-68N_WD-WCC7K1URKCCA -i 300 -a /dev/disk/by->

Mar 16 13:10:03 ASR5600G hd-idle[1125]: spindown: sdc
Mar 16 13:10:03 ASR5600G hd-idle[1125]: spindown: sdd
Mar 16 13:35:03 ASR5600G hd-idle[1125]: spinup: sda
Mar 16 13:35:03 ASR5600G hd-idle[1125]: spinup: sdb
Mar 16 13:35:03 ASR5600G hd-idle[1125]: spinup: sdc
Mar 16 13:35:03 ASR5600G hd-idle[1125]: spinup: sdd
Mar 16 13:40:03 ASR5600G hd-idle[1125]: spindown: sda
Mar 16 13:40:03 ASR5600G hd-idle[1125]: spindown: sdb
Mar 16 13:40:03 ASR5600G hd-idle[1125]: spindown: sdc
Mar 16 13:40:03 ASR5600G hd-idle[1125]: spindown: sdd

More and more spinuo/spindown even when I don’t use the PC
I changed the loglevel to debug. I’ll reboot now.

Could you point me what it does, how to use it and how to install it as it is not in my repos?

19

Hi
So I built chase;


lsblk

NAME        MAJ:MIN RM   SIZE RO TYPE MOUNTPOINT
sda           8:0    1 931.5G  0 disk 
└─sda1        8:1    1 931.5G  0 part 
sdb           8:16   1   7.6G  0 disk 
├─sdb1        8:17   1   260M  0 part /boot/efi
├─sdb2        8:18   1   512M  0 part /boot
└─sdb3        8:19   1   6.9G  0 part [SWAP]
nvme0n1     259:0    0 232.9G  0 disk 
├─nvme0n1p1 259:1    0    40G  0 part /
└─nvme0n1p2 259:2    0 192.9G  0 part /var

hotswap 
Usage: /usr/sbin/hotswap <Drive to remove /dev/sNX>

hotswap /dev/sda

/dev/sda
Syncing and unmounting /dev/sda
umount: /dev//dev/sda: no mount point specified.
Powering off /dev/sda
Disk /dev/sda safe to remove

lsblk

NAME        MAJ:MIN RM   SIZE RO TYPE MOUNTPOINT
sdb           8:16   1   7.6G  0 disk 
├─sdb1        8:17   1   260M  0 part /boot/efi
├─sdb2        8:18   1   512M  0 part /boot
└─sdb3        8:19   1   6.9G  0 part [SWAP]
nvme0n1     259:0    0 232.9G  0 disk 
├─nvme0n1p1 259:1    0    40G  0 part /
└─nvme0n1p2 259:2    0 192.9G  0 part /var

I can unplug the drive, when plugged back in it reappears.

Look at http://idle3-tools.sf.net/

Just install direct from the utilities repo link? Or manually download and install the rpm with zypper.

20

Idle3-tools can only read and set the idle timer. Not useful as hd-idle run fine.

After running the script for one HDD, is it possible to mount, start it back without unplug and re-plug it, only by software?
I ask that because at this time, HDDs spinup/spindown at interval. If I unmount them it could avoid this behaviour.

BTW did you try autofs to self mount and unmount disks? I use it with a shared HDD with NFS connected to router USB, working fine.