Disk spin-down not working on Seagate IronWolf Pro

simon-77 · September 2, 2023, 8:39am

Hey, I had an issue on my Rockstor system regarding a HDD not spinning down and I would like to share my findings and workaround.

I have 2 drives in my new Rockstor system:

Seagate IronWolf Pro
Toshiba N300 NAS

Both are 18TB in capacity, SATA-III and are fairly new models (bought in 2023).

For the Toshiba N300 NAS drive I was able to configure the spin-down settings via Rockstor Web-UI and everything worked as expected.

But the Seagate IronWolf Pro drive was not configurable via Rockstor Web-UI.
The APM column only shows ???, the Power Status is unknown and I can’t press the hour-glass symbol:

And also querying the APM level with hdparm manually showed that this is not supported:
Screenshot_20230902_100244

Doing some search on the internet revealed many people with similar issues on Seagate IronWolf & Seagate IronWolf Pro HDDs.
It seems that Seagate has its own tool for power management: SeaChest Utility

openSeaChest:
There is also a open-source version of this tool: openSeaChest
I found this blog entry helpful as guidance for my workaround on Rockstor: Spinning down Seagate Hard Drives in Ubuntu 20.04 - André Jacobs

My workaround:

I was not keen on installing openSeaChest from source on a “production” system, so I downloaded the pre-built binaries form the latest release: Releases · Seagate/openSeaChest · GitHub
File: openSeaChest-v23.03.1-linux-x86_64-portable.tar.xz

copy tar-ball to Rockstor via SSH ($ scp openSeaChest-... root@hostname:)
extract tar-ball on Rockstor (# tar xf openSeaChest-...)
enter the new directory to execute the binaries for setup.
./openSeaChest_Info --scan - list all devices
./openSeaChest_PowerControl -d /dev/sg0 --showEPCSettings - show the timers for drive sg0
./openSeaChest_PowerControl -d /dev/sg0 --idle_c 600000 --standby_z 1200000 - modify the timers for drive sg0 (idle_c: 60 minutes; standby_z - aka spin-down: 20 minutes)

To check the current power state of all HDDs use ./openSeaChest_PowerControl -d all --checkPowerMode

With:
Device is in the PM1: Idle state = spinning disk
Device is in the PM2: Standby state = not spinning disk

Although some manual steps are involved, this now works for me and these PowerControl settings also survive a reboot.

Cheers Simon

phillxnet · September 2, 2023, 11:28am

@simon-77 Thanks for sharing your findings on this one. And a nice bit of investigation there.

I did have some much smaller Seagate IronWolf (non pro) drives, 4 TB I think they were, but no longer have access to those drives unfortunately. They did work as expected, so likely this is a more recent ‘thing’. Shame.

Yes, this is our cue to leave well alone and disable the Web-UI options in this area for given drive:

github.com

rockstor/rockstor-core/blob/master/src/rockstor/system/osi.py#L1618-L1635


      
          def get_disk_power_status(dev_byid):
              """
              When given a disk name such as that stored in the db ie /dev/disk/by-id
              type we return it's current power state via hdparm -C
              /dev/disk/by-id/<dev_byid> Possible states are:
              unknown - command not supported by disk
              active/idle - normal operation
              standby - low power mode, ie drive motor not active ie -y will do this
              sleeping - lowest power mode, completely shut down ie -Y will do this
              N.B. -C shouldn't spin up a drive in standby but has been reported to wake
              a drive from sleeping but we aren't going to invoke sleeping as pretty much
              any request will wake a fully sleeping drive.
              Drives in 'sleeping' mode typically require a hard or soft reset before
              becoming available for use, the kernel does this automatically however.
              :param dev_byid: disk name as stored in db / Disk model ie without path
              :return: single word sting of state as indicated by hdparm -C /dev/<disk>
              and if we encounter an error line in the output we return unknown.
              """

And we set the disk spindown via:

github.com

rockstor/rockstor-core/blob/4d60a2c683140aaa10a8e5b201157163cc9a898b/src/rockstor/system/osi.py#L1694-L1714


      
          def set_disk_spindown(
              dev_byid, spindown_time, apm_value, spindown_message="no comment"
          ):
              """
              Takes a value to be used with hdparm -S to set disk spindown time for the
              device specified.
              Executes hdparm -S spindown_time and ensures the systemd script to do the
              same on boot is also updated. Note we do not restart the systemd service
              to enact these changes in order to keep drive intervention to a minimum.
              :param dev_byid: The name of a disk device as used in the db ie by-id type
              without a path.
              :param spindown_time: Integer received from settings form ie 240
              :param apm_value: value to be used with hdparm's -B parameter to set the
              drives APM level. Should be between 1-255 and will be ignored if not. When
              ignored there will be no hdparm -B executed and no -B switch added to the
              relevant systemd line.
              :param spindown_message: message received from drop down as human presented
              selection, used later in systemd script to retrieve previous setting.
              :return: False if an hdparm command was not possible ie inappropriate dev,
              or an error was return by the command, True otherwise.
              """

Which in turn configures/reconfigures the rockstor-hdparm systemd service to enable consistency over power cycling.

Just noting the above in case a little development in the testing channel takes your fancy :).

Thanks again for the heads up on this one. And for sharing your fix/work-around. Indicentally have you ensured that the settings, thus applied, survive a power-cycle. I think hdparm settings survive a reboot but not a power cycle (from memory at least). Hence our use of our dedicated systemd service to re-establish them on boot.

simon-77 · September 2, 2023, 7:31pm

I just checked it:
this setting also survives a power-cycle (and this setting is also supposed to be persistent).

aremiaskfa · September 2, 2023, 7:34pm

Thanks for posting your findings here!

@phillxnet I also have a ironwolf pro with the same affliction. Posting just in case you need testers in the future