You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
hd-idle (faulty): 1.21+ds-1+b8 (from the Debian 13 "Trixie" repository)
Host OS: Proxmox VE 9 (based on Debian 13)
Setup Specifics: Three SATA HDDs are passed through to a Debian VM using "Raw Disk Passthrough". hd-idle is running on the Proxmox host, as the guest VM cannot control the physical power state of the drives.
Bug Description
A specific regression is present in hd-idle version 1.21. After restarting the hd-idle service, the drives are correctly put into standby after the specified timeout. However, once the drives are woken up by any access, hd-idle fails to put them back into standby after they become idle again.
The core issue appears to be that the daemon does not correctly recognize or process the new "active" state of the disk after the first wake-up event. Consequently, the idle timer is not reset, and no further spindown commands are sent.
This behavior does not occur with the older version 1.05, which handles repeated spin-down cycles reliably under the exact same configuration.
Expected Behavior
After a disk wakes up, hd-idle should detect the activity, reset its idle timer, and successfully spin the disk down again after the timeout period expires.
Actual Behavior
hd-idle only spins the disks down once after the service starts. After the first wake-up, the disks remain permanently in the active/idle state, as confirmed by hdparm -C. A service hd-idle restart is required to trigger the spin-down cycle one more time.
Additional Context and Troubleshooting
Workaround: Downgrading to hd-idle version 1.05 completely resolves the issue.
Troubleshooting Attempt: Adding the -c ata parameter to the HD_IDLE_OPTS in version 1.21 was attempted but had no effect. The behavior remained unchanged.
Environment
Bug Description
hd-idleversion 1.21. After restarting thehd-idleservice, the drives are correctly put into standby after the specified timeout. However, once the drives are woken up by any access,hd-idlefails to put them back into standby after they become idle again.Expected Behavior
hd-idleshould detect the activity, reset its idle timer, and successfully spin the disk down again after the timeout period expires.Actual Behavior
hd-idleonly spins the disks down once after the service starts. After the first wake-up, the disks remain permanently in theactive/idlestate, as confirmed byhdparm -C. Aservice hd-idle restartis required to trigger the spin-down cycle one more time.Additional Context and Troubleshooting
hd-idleversion 1.05 completely resolves the issue.-c ataparameter to theHD_IDLE_OPTSin version 1.21 was attempted but had no effect. The behavior remained unchanged.Configuration and Logs
/etc/default/hd-idle Configuration:
journalctl -u hd-idle(shows the first, successful cycle):(Note: After this initial cycle, no further
spindownorspinuoevents are logged, even though the disks remain active)./etc/lvm/lvm.conf(proxmox additional):