hdparm: SG_IO: bad/missing sense data

[tristank]

There is currently bug in the kernel which breaks tools like hdparm and hddtemp.

$ hdparm -y /dev/sdb


/dev/sdb:

 issuing standby command

SG_IO: bad/missing sense data, sb[]:  f0 00 01 00 50 00 00 0a 80 00 00 00 00 1d 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

AFAICS, this has now been reverted upstream with fa0db8e56878 ("Revert
"ata: libata-scsi: Honor the D_SENSE bit for CK_COND=1 and no error"")
in 6.11-rc4 and the revert was as well backported to stable releases
(5.15.165, 6.1.106, 6.6.47 and 6.10.6)

Source

Is there any way to backport the changes to the current stable kernel 6.8.12-2-pve or when will the fix be available in Proxmox?

Related links:

- https://forum.manjaro.org/t/lm-sens...t-showing-temperature-of-sata-drives/166492/7
- https://forums.unraid.net/bug-reports/stable-releases/61212-harddrives-cant-be-spun-down-r3191/
- https://forum.manjaro.org/t/lm-sens...t-showing-temperature-of-sata-drives/166492/7

 

[phaze75]

I am not sure if this is a kernel issue. I have the same problem. But if I run the hdparm commands on the same HDD in my Proxmox environment's CLI, it works flawlessly. The issue seems to be connected to the "passthrough" of the HDD to the virtual machine. As the hdparm -I command produces hardly any information in my virtualized Ubuntu guest system, this doesn't seem to be a "full passthrough", but some sort of abstract assignment of the HDD to the virtual machine. Did you try that?

Can anybody explain how a passthrough is handled by Proxomox? I am new here, coming from a Xen environment, where passthroughs are "full passthroughs". Thanks in advance!

 

[phaze75]

Actually, I just found the answer to our issue. We need to configure the spin down on the PVE side.

Read this post: Any easy way to spindown ?

 

[mrish]

I am not sure if this is a kernel issue. I have the same problem. But if I run the hdparm commands on the same HDD in my Proxmox environment's CLI, it works flawlessly. The issue seems to be connected to the "passthrough" of the HDD to the virtual machine. As the hdparm -I command produces hardly any information in my virtualized Ubuntu guest system, this doesn't seem to be a "full passthrough", but some sort of abstract assignment of the HDD to the virtual machine. Did you try that?

Can anybody explain how a passthrough is handled by Proxomox? I am new here, coming from a Xen environment, where passthroughs are "full passthroughs". Thanks in advance!

Facing the exact same issue. The commands to set standby time are working if the disk is not setup as a storage in Proxmox host. But entry in /etc/hdparm.conf for spindown time is ignored. I have also updated filter for the HDD in /etc/lvm/lvm.conf.

So, we could run a (start-once) service that starts on boot and sets the standby time using the command.