Arm64 - Failure (SCSI response not ok) about hd-idle HOT 15 CLOSED

As an update: I managed to install a arm64 version of ubuntu and I reproduced the error.
Now I "only" have to fix it.

from hd-idle.

I've added an entry in the readme to this wiki topic.

from hd-idle.

I can confirm that the issue is related to ata, since scsi works fine as seen below.

$ sudo systemctl status hd-idle
* hd-idle.service - hd-idle - spin down idle hard disks
   Loaded: loaded (/lib/systemd/system/hd-idle.service; enabled; vendor preset: enabled)
   Active: active (running) since Tue 2019-09-10 11:55:51 CEST; 3min 32s ago
     Docs: man:hd-idle(8)
 Main PID: 2109 (hd-idle)
    Tasks: 5 (limit: 3350)
   Memory: 4.7M
   CGroup: /system.slice/hd-idle.service
           `-2109 /usr/sbin/hd-idle -i 0 -a /dev/disk/by-id/ata-ST6000DM003-2CY186_WF306G8F -i 180

Sep 10 11:55:51 hostname systemd[1]: Started hd-idle - spin down idle hard disks.
Sep 10 11:55:51 hostname hd-idle[2109]: symlinkPolicy=0, defaultIdle=0, defaultCommand=scsi, debug=false, devices=[{sda /dev/disk/by-id/ata-ST6000DM003-ID 180 scsi}], logFile=
Sep 10 11:59:19 hostname hd-idle[2109]: sda spindown

This is not something that bother me as everything actually seems to be fine with scsi 👍

$ sudo systemctl status hd-idle
* hd-idle.service - hd-idle - spin down idle hard disks
   Loaded: loaded (/lib/systemd/system/hd-idle.service; enabled; vendor preset: enabled)
   Active: active (running) since Tue 2019-09-10 12:09:02 CEST; 3min 21s ago
     Docs: man:hd-idle(8)
 Main PID: 2345 (hd-idle)
    Tasks: 5 (limit: 3350)
   Memory: 4.7M
   CGroup: /system.slice/hd-idle.service
           `-2345 /usr/sbin/hd-idle -i 0 -s 0 -l /var/log/hd-idle.log -a /dev/disk/by-id/ata-ST4000DM004-ID -i 180 -a /dev/disk/by-id/ata-ST6000DM003-ID -i 120

Sep 10 12:09:02 hostname systemd[1]: Started hd-idle - spin down idle hard disks.
Sep 10 12:09:02 hostname hd-idle[2345]: symlinkPolicy=0, defaultIdle=0, defaultCommand=scsi, debug=false, devices=[{sdc /dev/disk/by-id/ata-ST4000DM004-ID 180 scsi} {sda /dev/disk/by-id/ata-ST6000DM003-ID 120 scsi}], logFile=/var/log/hd-idle.log
Sep 10 12:11:24 hostname hd-idle[2345]: sda spindown
Sep 10 12:12:16 hostname hd-idle[2345]: sdc spindown

from hd-idle.

I was hoping that everything was fine, but it seems that my HDDs spinup without hd-idle noticing and won't get spun down until it recognises that example I play a video from a HDD.

From systemctl status hd-idle

Sep 10 18:01:29 hostname hd-idle[4875]: sdb spindown
Sep 10 18:01:30 hostname hd-idle[4875]: sda spindown

But the HDDs are actually not standby as seen below.

$ sudo smartctl -i -n standby /dev/disk/by-id/ata-ST4000DM004-ID
smartctl 6.6 2017-11-05 r4594  (local build)
Copyright (C) 2002-17, Bruce Allen, Christian Franke, www.smartmontools.org

Device is in SLEEP mode, exit(2)

$ sudo smartctl -i -n standby /dev/disk/by-id/ata-ST6000DM003-ID
smartctl 6.6 2017-11-05 r4594  (local build)
Copyright (C) 2002-17, Bruce Allen, Christian Franke, www.smartmontools.org

=== START OF INFORMATION SECTION ===
...
...
...
SMART support is: Enabled
Power mode was:   IDLE_B

Update

I installed the armhf package instead of the arm64 and now everything is back to normal. So it seems that the arm64 package has issues related to the use of ata.

from hd-idle.

Hi @V33m

I agree with you on blaming the arm64 package. I cross compiled from my x86_64 desktop to create all the packages (amd64, i386, armhf and arm64). Weird that the arm64 failed but not the armhf.
Sadly I don't have any arm64 hardware where I can test. I guess I should start considering getting one.
On what hardware are you running?

from hd-idle.

Khadas VIM3, however, there are other great arm64 SBCs to choose as example Raspberry Pi 4 Model B, ODROID-N2 etc.

from hd-idle.

Sadly I don't have any arm64 hardware where I can test. I guess I should start considering getting one.

Have you had any time to look into arm64? If you have done any changes which you think have sorted the issue, let me know so I can give it a try 👍

from hd-idle.

I got a raspberry 4 but I didn't manage to get a 64bit linux on it yet. The pi is always headless and I have no way to debug when ssh is not built in.
I keep trying every now and then with different options.

The 1.8 release has better error handling, but it won't solve the arm64 issue anyway. So no real need for trying it out.

from hd-idle.

Which OSes have you tried? I don't have a RPi4, but you should give the following a try:

• Balena
• Gentoo
• Manjaro (Arch)
• Raspbian with 64-bit kernel (32-bit userland)
• Ubuntu MATE 18.04.2 (Bionic)

I've never had any RPis, but no ssh built in sounds odd for me.. Most of my arm SBCs are headless, so built in ssh has always been a requirement to use a OS for me. Let me know when you want to do some testing down the road.. 👍

from hd-idle.

So far I don't have any good news.
I based the ata sgio command on hdparm's implementation. I ran the command sudo hdparm -S 10 /dev/sda to see if it works on arm64 and got the following error:

/dev/sda:
 setting standby to 10 (50 seconds)
SG_IO: bad/missing sense data, sb[]:  70 00 01 00 00 00 00 0a 00 00 00 00 00 1d 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

This is exactly the same error hd-idle throws.

So I will have to re-think (online research) the whole ata sgio command for arm64 :(

from hd-idle.

Yeah, we probably have to do some online research to figure this out. The positive aspect of hdparm having the same issue is that we can report it and ask for input. Which version of hdparm did you test? 9.58+ds-1? 9.58+ds-4? The maintainer is Alexandre Mestiashvili as seen from the patch notes: Debian changelogs and Ubuntu changelogs.

A quick google search didn't give me too more, yet this thead could be a start - https://askubuntu.com/questions/768373/hard-drive-error-bad-missing-sense-data

from hd-idle.

After some digging I ended up in, what it looks, the description of this problem and how to solve it: SAT-with-UAS-Linux
It is an interesting reading but the TL;DR is: ATA commands might not work if UAS is enable.

I was able to solve it following the manual.
@V33m Please let me know if it works for you too.

from hd-idle.

Glad to hear that! Did you disable UAS or overwrite SAT detection for the drive? Which arm64 device do you have, Rpi4?

from hd-idle.

I temporary disable UAS.

cat /sys/module/usb_storage/parameters/quirks
2109:0711:u

I used Rpi4 with Ubuntu arm64.

from hd-idle.

I'm thinking about the same, given the decrease in write/read speed is not an issue for me as none of the drives are SSD. I consider the issue fixed and one should probably mention it in the README

from hd-idle.