Backup internal HDD - sleep mode

[LolekUK]

I have installed an additional HDD (spinning platters) in my PVE host, it will be purely used to take weekly backups of VMs and CTs (which currently takes less than 10 minutes).

Would it be possible to put the drive to sleep (and stop the platters spinning) for the remaining time of the week and wake it up just for the backup job?

 

[fschauer]

Yes, that should be possible, but pvestatd keeps the drive active.
If you want to be able to spin it down, you would have to either disable the storage or exclude the drive in /etc/lvm/lvm.conf with

global_filter=["r|/dev/zd.*|", "r|/dev/disk/by-id/<your disk id>"]

To apply the changes either restart PVE or run vgscan to reload the LVM config.
You should then be able to spin down the disk from the host with:

hdparm -y /dev/sdX

If you want it to automatically spin down after some time of inactivity, you can use hd-idle, which is found in the Debian repository.

 

[LolekUK]

Thanks @fschauer, sounds easily achieveable.

I will need to do a bit more research on this as I have made the change, rebooted the host (remotely) and it didn't come back up. I will have to check it (physically) after work when I'm back home.

 

[Dunuin]

Also have a look at Vzdump hook scripts to automate the enabling/disabling of the storage. With that a weekly backup job could enable the storage first and disable it again once the backup job has finished. A disabled storage won't be polled by pvestatd, so it won't be kwpt awake every few seconds.

 

[LolekUK]

Yes, that should be possible, but pvestatd keeps the drive active.
If you want to be able to spin it down, you would have to either disable the storage or exclude the drive in /etc/lvm/lvm.conf with

global_filter=["r|/dev/zd.*|", "r|/dev/disk/by-id/<your disk id>"]

To apply the changes either restart PVE or run vgscan to reload the LVM config.
You should then be able to spin down the disk from the host with:

hdparm -y /dev/sdX

If you want it to automatically spin down after some time of inactivity, you can use hd-idle, which is found in the Debian repository.

update: I've found the missing character | at the end of the command, tested and working Thank you

global_filter=["r|/dev/zd.*|", "r|/dev/disk/by-id/<your disk id>|"]

when I enter disk by-id in lvm.conf, none of the machines (CTs and VMs) start after host reboot. I have checked disk ID by going to the /dev/disk/by-id folder (I have also entered disk ID for PVE GUI, which was WDBlue, same result:

  Invalid separator at end of regex.
  Invalid filter pattern "r|/dev/disk/by-id/ata-WDC_WD10SPCX-60HWST0_WD-WX11AC305006".
  Failed to create global regex device filter
command '/sbin/vgs --separator : --noheadings --units b --unbuffered --nosuffix --options vg_name,vg_size,vg_free,lv_count' failed: exit code 5
  Invalid separator at end of regex.
  Invalid filter pattern "r|/dev/disk/by-id/ata-WDC_WD10SPCX-60HWST0_WD-WX11AC305006".
  Failed to create global regex device filter
command '/sbin/vgscan --ignorelockingfailure --mknodes' failed: exit code 5
  Invalid separator at end of regex.
  Invalid filter pattern "r|/dev/disk/by-id/ata-WDC_WD10SPCX-60HWST0_WD-WX11AC305006".
  Failed to create global regex device filter
TASK ERROR: command '/sbin/lvs --separator : --noheadings --units b --unbuffered --nosuffix --config 'report/time_format="%s"' --options vg_name,lv_name,lv_size,lv_attr,pool_lv,data_percent,metadata_percent,snap_percent,uuid,tags,metadata_size,time' failed: exit code 5

 

[LolekUK]

Fab, thanks @fschauer, works flawlessly

Step 1. in /etc/lvm/lvm.conf

global_filter=["r|/dev/zd.*|", "r|/dev/disk/by-id/<your disk id>|"]

Step 2. install hd-idle and run the following command (to put to standby after 10 minutes/600 seconds)

hd-idle -a sda -i 600

Step 3. add to system startup

systemctl enable hd-idle

Step 4. edit startup hd-idle config file /etc/default/hd-idle, in my scenario it looked like this:

HD_IDLE_OPTS="-a sda -i 600"