I have been having issues with my FileServer recently.

See “rcu_sched detected stalls on CPUs/tasks” - jiffies - ESXi Ubuntu 16 FileServer Guest

I just rebuilt it since I couldn't get any guidance on the issue. Unfortunatley, its resyncing my array again.

Seeing how this has been done multiple time in the last few days, I'm fairly confident that its "in sync".

I have a few questions to that regard.

  • When it resync's what is happening, is it just doing read operations on every piece of data? I feel i'm really abusing my HDD's at the moment
  • Can I cancel resync? (I still want to be able to use the MDADM Though!)
  • So Can I force it to "clean" state and remount?
  • Is resync automatic when you mount a new MDADM (This is a new VM), or is there an error in the metadata that is forcing a resync and I NEED to let it finish. (It takes about 24 Hours and my server is extremely slow during that time)


Update: There are 4x 5tb HDDs in raid5 passed through via mapped raw lun to a VM guest which then exports nfs and smb.

This question is mostly about if the commands i want exist for mdadm and what sync does.

  • 2
    I'm not saying it's unusual for a VM to have multiple "disks" (e.g. give it a virtual disk for boot/OS and one or more virtual disks disks for data storage), just unusual for the VM to manage its own RAID.
    – cas
    Commented Mar 14, 2018 at 0:28
  • anything in the kernel logs indicating problems with either the disk controller or any of the drives?
    – cas
    Commented Mar 14, 2018 at 0:28
  • 1
    Esxi doesnt support software raid or being an nfs/smb server Commented Mar 14, 2018 at 0:29
  • then why use esxi rather than kvm?
    – cas
    Commented Mar 14, 2018 at 0:29
  • 1
    I think these questions are better posted on my question that I linked. This question is purely about mdadm, it doesn't make a difference that it's a VM. Commented Mar 14, 2018 at 0:30

1 Answer 1


You can't stop a "parity check" as even though all the data is there, mdadm's metadata isn't correct. You have to let it finish.

Side Note: If you are having this issue repeatedly look into hardware failures as the culprit and solve asap!

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .