I realized that my external disk (western digital 8tb) sleeps after a while when there is no activity on disk.
i’m wondering if this is a problem for farming?
Second,
How can i prevent my external disk for sleeping in ubuntu desktop 20.04?
Thanks.
1 Like
Yes, long wakeup time will be very bad for harvester proof times, which must be under 30 seconds, and are “supposed to be” under 5 seconds.
It’s kind of janky but use this KeepAliveHD program in Windows… I have it set to write a file every 60 seconds to the drive, then delete it.
I assume there are similar programs or scripts for Ubuntu.
I initially had it reading a file every 30 seconds but that actually wasn’t working; I could hear one of the external drives, which I had read set on, waking up from time to time!
5 Likes
it seems you have found a working solution for you which is cool.
I’m looking for alternatives for ubuntu 
1 Like
You could easily write a small shell script to copy and then delete a small file to your external HDD and setup a cron job to execute that script every 30 seconds
2 Likes
thanks for the solution. i will try to do that. it is a dirty way but if it works, why not 
1 Like
If you want to wake up your drives, query the SMART info.
For PowerShell:
Run this from a .vbs on your Task Scheduler every N minutes set to run under the SYSTEM account (Administrative privileges)
Get-PhysicalDisk -ErrorAction Ignore | Get-StorageReliabilityCounter -ErrorAction Ignore
Windows/Linux/macOS will not cache the SMART info and the on-board controller will always report it and guarantee spin-up of the disk if it is able to do so.
5 Likes
hmm thats an interesting idea.
i found a tool in ubuntu smartmontools it gives smart status of hdd; change sdc1 with your drive.
sudo apt install smartmontools
sudo smartctl --all /dev/sdc1
the result is as below there is no smart support for my hdd but i think even getting below output will prevent from sleeping, right?
=== START OF INFORMATION SECTION ===
Vendor: WD
Product: Elements 25A3
Revision: 1030
Compliance: SPC-4
User Capacity: 8.001.529.315.328 bytes [8,00 TB]
Logical block size: 512 bytes
Physical block size: 4096 bytes
LU is fully provisioned
Rotation Rate: 5400 rpm
Serial number: VGJJT9LG
Device type: disk
Local Time is: Mon May 3 00:59:30 2021 +03
SMART support is: Unavailable - device lacks SMART capability.
1 Like
Glad you solved your problem my friend, my response time is 0.00300. Do I have to use it?
for ubuntu systems,
hd-idle solved my issue and my hdd is not sleeping anymore.
here was the solution.
4 Likes
does this keepaliveHD program only works for Western digital?
It works for any drive. I set it to write file and then delete, I found that just reading alone wasn’t keeping the drives awake (due to caching, perhaps).
1 Like
For Linux, I used this cron job posted at Ask Ubuntu. I modified it to 4 minutes rather than every minute:
*/4 * * * * /bin/bash -c 'dd if=/dev/disk/by-uuid/THE-UUID of=/dev/null count=1 skip=$RANDOM'
The post linked goes into more detail.
My problem wasn’t the drive but the ORICO enclosure that has an annoying “smart sleep” feature.
2 Likes
I use USBDLM to fixed drive letter and set time for DiskKeepAlive function, work well without needed to write file to hard drive as its only read sector 0 of every drive.
Basically, if I set
hd-idle -i 0
, my external HDD would never sleep, right?
yes you can try first that solution.
if it does not work, you can create a cron job which basically copies any file to hdd every x minutes. that also works too.
2 Likes
does it solves same problem for seagate external hdds?
If using windows 10 you can untick this in device manager. Works like a charm!
I found out that the bullet proof solution is to use crontab -e. It’s the simplest and cleanest solution, doesn’t mess with your system settings in any way, also works 100% regardless your HDD model.
To use crontab, open the terminal and type crontab -e. Then, for instance, you can tab your HDD every 5 mins so it won’t go to sleep by adding this line into your crontab:
*/5 * * * * /bin/touch /media/user/hdd_label/.stayawake &> /dev/null
Please note that /media/user/hdd_label/.stayawake should correspond to your HDD directory path that point to .stayawake file (you have to create an empty file called .stayawake first).
2 Likes
+1, this is the best approach I’ve tried ever since.
Thanks for the idea. It was the only thing that worked.
Note hd-idle -i 0 does not work at all.
and for my Toshiba USB drives, hdparm -S 0 or -B 255 do not work, nor does sdparm or smartctl.
I have created the script like this so that I don’t have to add in new paths when I make more mount points / add more drives:
[root@fedora plots]# pwd
/mnt/plots
[root@fedora plots]# ls -l
total 32
-rwxr-xr-x. 1 root root 98 Jun 21 23:15 stayawake.sh
drwxrwxrwx. 3 carl carl 4096 Jun 22 01:46 toshiba_4tb_usb_10
drwxrwxrwx. 3 carl carl 4096 Jun 22 15:19 toshiba_4tb_usb_11
drwxrwxrwx. 3 carl carl 4096 Jun 22 13:18 toshiba_4tb_usb_12
drwxrwxrwx. 3 carl carl 4096 Jun 21 23:12 toshiba_4tb_usb_13
drwxrwxrwx. 3 carl carl 4096 Jun 21 23:12 toshiba_4tb_usb_14
drwxrwxrwx. 3 carl carl 4096 Jun 21 23:12 toshiba_4tb_usb_15
drwxrwxrwx. 3 carl carl 4096 Jun 21 23:12 toshiba_4tb_usb_9
[root@fedora plots]# cat stayawake.sh
#!/bin/bash
paths=`ls -d /mnt/plots/*/`
for path in ${paths[@]}; do
touch $path/.stayawake
done
[root@fedora plots]# cat /var/spool/cron/root
* * * * * /mnt/plots/stayawake.sh
1 Like