curious if anyone has gotten spindown to work on 25.10 with sas drives and disk shelves (check sig for what i have). setting apm to 1 (minimum power with spindown) and spindown to 5 minutes and completely leaving the pool alone (no shares, tasks, etc associated and closing webui to avoid any heath/stat checks) the drives still wont go to sleep. using the spindown script works but causes the shelf to throw an error light on the individual drives when they spin down which is not promising. also trying to check any drive in the shelf with
with the last 5 bits shown sometimes changing, while doing the same to a non shelf drive (sata or nvme) gives a proper result, which also doesnt inspire confidence.
and yes, i know spindown bad, the shelf would only be for periodic backups and be completely idle the other ~6.75 days a week so i would much rather have it idle at 50w than 250 for that period.
I’m looking for more automation not less, plus it saves on the hassle of having to manually import and export the pool and turning configurations back on that automatically shut off when you export.
Which hassle?
Backup server boots and run a PULL replication job; no need to explicitly import or export anything. If the server has IPMI, you can automate a wake up call for it. That leaves “shut down after replication is complete” as last problem to automate.
If the server does not have IPMI, you might be able to do this anyway, though it’s rather hacky, and I haven’t tested it personally. Set your BIOS to turn the machine on when it’s powered on, then plug the machine into a smart plug (I use a Tapo device, about $20). Then you can have the machine shut itself off when it’s done with your backup job, and set the Tapo to turn off power some safe time after that. Then set it to turn the power back on at your desired backup time. If you want to be really fancy, you can control it with HomeAssistant.
As a bonus, the smart plug will also let you easily monitor your system’s power consumption.
there is no secondary server, there is only one server computer with internal main pool hdds and the external disk shelf i want to use for backing up the main pool attached to the one and only server via qsfp/minisas hd, so i do need to import and export the disk shelf so truenas doesnt have a conniption when the whole pool up and disappears, and deal with the configs every time too.
I have the same problem. apm set to 1 and spin down to 30min and Scale 25.10 is in constant cycle of sleeping and waking up disks. after 30min inactivity disk are going to sleep and something is wake them up in random intervals. After few days fighting with this reverted back to 25.04 and everything works ok.
is not this.
I’m using drivetemp kermel module for long time (I think more than 2 years) in my script to control fans and this will never wake up disks. that kernel module is reporting disk temp even in STANDBY state.
I finally decided to rack my whole setup and in doing so I swapped the sas mg04s for my sata hc520s in the ds4246 (no interposers) and went from a 9300-16i+8i8e adapter to a 9305-i16e, otherwise it was exactly the same. Now hdparm -c reports the status correctly and the drives do actually spin down when set to on their own using the inbuilt apm settings.
So either it was the sas drives, the 8i8e adapter or the 9300 itself. I was going to test them individually but I forgot before I redid everything, so oh well.
