Tried to upgrade Core to Scale and received DIF error

Hi all,

Just attempted to upgrade my Core install to Scale. After the upgrade, and logging in, my main Pool won’t mount and there is an error that states, " Disk(s): sdg, sdi, sdh, sdc, sdd, sde, sdb, sdf are formatted with Data Integrity Feature (DIF) which is unsupported."

Is this a show stopper? Did a quick search, but everything seems to suggest reformatting the drives?

I may be talking crap here - but

DIF is 520 byte sectors, rather than 512 byte sectors. This is configured / used by RAID cards (as far as I was aware), and not by “ZFS friendly” devices which suggests to me that you have a potentially major issue here - and not just with your upgrade to Scale.

Can you post your hardware please - which you should do anyway. Be careful to detail how the existing disks are connection to the NAS, via what card / device.

I could be wandering down the wrong rabbit hole here

Heya NugentS,

So the drives are Hitachi 0F29530 HUH721212AL5200 12TB SAS.

Capacity: 12TB
Speed: 7.2K RPM
Interface Types: SAS
Form Factor: 3.5" LFF Hard Drive
Sector Size: 512 / 512e
Sustained Throughput: 243/255
Average Seek Time: 8.0/8.6
Electrical Interface: SAS-3 12Gbps
On-Board Cache: 256MB

The storage controller is a SAS3008.

Been running those on TrueNAS Core for years now with no issue.

Not a show stopper, have a few disks with this myself. You cannot add new DIF-formatted disks without reformatting them, but the ones added will continue to work.

Well there doesn’t seem to be anything wrong there.

Just reformat them

Why do they work with TrueNAS Core though? What changed? Is it just the move from FreeBSD to Linux?

But the ones added don’t continue to work, they won’t mount.

New disks needs to be reformatted. All my old disks that was already in a pool works as expected.

Can confirm that doesn’t work. TrueNAS Scale refused to mount the pool at all.

How can I tell if my disks are setup this way? I don’t think they are, but I’m not sure…

These instructions should help: tales in IT from the help desk.: Formatted with Type 2 Protection, huh?

But if I understand the OP correctly the drives contain a live pool with valuable data and work just fine in CORE. So formatting to make them work in CE might not be a desireable option.

The question the is: why doesn’t CORE have a problem with these disks?

A possible strategy depending on the level of redundancy of the pool and the willingness to take a risk (and the availability of a backup) could be:

  • remove a single disk
  • format it without protection in another system
  • readd the formatted disk to the pool and resilver
  • assuming that worked repeat for all disks one at a time
  • then upgrade to CE

HTH,

Patrick

From another Forum post it may be that CE (and any scale version before) don’t have the kernel module needed for DIF to work, but it’s in the bsd kernel of core. At least that’s how i understood it from

Actually I hadn’t thought of this idea… I might try it… will take a while though :rofl: