(6.12.3) Writes to ZFS pools constantally

is zfs.master plugin, the tab was not present today. I removed it and once installed writes started again, removed and them stops. 😞 I'll write in plugin post  

I set zfs_master refresh interval to "no refresh" and that did seem to help.

can you try removing it and check the disk space used before and after few hours that is removed?

I am on 6.12.4, but I also was having an issue with a single ZFS disk on my otherwise XFS array.

 

It only happened when I would open the main tab on the WebGUI. It did not bother me too much, because it was only when I was on the WebGUI, and I knew it was not spinning up other times.

 

I tried setting ZFS master to no-refresh, but it did not seem to help. I removed it entirely and it seems to have taken care of it.

 

I read about the spin up every 30 minutes or so issue, so I am going to leave it for a few hours and see if I have it as well.

I have uninstalled ZFS master, and have found the same activity, my disk will spin up every 30 minutes or so.

• Stop all your docker containers and then start then 1 by 1 until you find the devil.
• Check your settings for 'folder caching' 
• In several cases I found some containers like 'Dash dot' running in privileged caused disk spin up.
• Also check plex scanning.
• Install turbo write so not all drives need to spin up to write to parity.

2 hours ago, dopeytree said:

• Stop all your docker containers and then start then 1 by 1 until you find the devil.
• Check your settings for 'folder caching' 
• In several cases I found some containers like 'Dash dot' running in privileged caused disk spin up.
• Also check plex scanning.
• Install turbo write so not all drives need to spin up to write to parity.

So I also have this issue, if I close my browser the disks stay spun down.

I tested with 1 disk converted to XFS and it does not spin back up at all. This is something with ZFS needing access to the disks. Seeing as there is no real benefit from having individual disks in ZFS I will just convert them back to XFS and call it a day. I have my cache pool and a dedicated ZFS pool in ZFS and they are NVME and SSD's respectively so they don't need to be spun down.

EDIT:

I also tried turning off the refresh interval in ZFS Master plugin and it stops the spin up of the drives too.

Edited January 5 by jbohbot

I have 2x zfs pools (one is 2x nvme mirror & the ether is 4x 12TB drives in z1) and they have stopped randomly spinning up after following the steps above.

 

I also removed zfs master plugin but may add it again. basically if you open the MAIN tab it spins up all zfs drives.

 

Hope you get it sorted mate.

I have this exact same problem, I will get continous small bursts of writes of around 400Kb/s.

 

I have no dockers/VMs etc running, only an array, 2 cache drives and a few plugins.

 

I disabled the cache dirs plugin and zfs master refresh to see if it was those and I still got the continuous small bursts.

 

Only thing that iotop shows is that kworker is triggering when the writes happen.

 

Any ideas?

Edited January 8 by xreyuk

I had the same problem, and it wasn't caused by ZFS master, docker or VMs. But it turned out, that I had a scrub process running on the ZFS drive that is taking forever, which had never been a problem for XFS or BTRFS. It wasn't just that it was constantly reading/writing, it was also slowing down all other processes such as parity checks.