Disks locked with seemingly no way to fully format or unlock

[NilsS]

I recently got a set of new (to me) used drives, Hitachi HUSMM141 EMC1600's, and I've been having some very weird issues trying to get them up and running. In normal use I can read SMART, they show up in storcli, and in the GUI both under multipath and disks. However, if I try to do anything to them I get permission denied. I did a bit of digging and found the sysctl kern.geom.debugflags=16 which did allow me to do some things to the drives, but not all.

Listed below are what I have tried and its results after setting the debug flag, and at the bottom is the hardware I'm using.

• dd:
  Running
  Code:

  dd if=/dev/zero of=/dev/da10

  with or without
  Code:

  bs=512

  and with either side of the multipath, I get /dev/da10: Invalid argument.
• fdisk:
  fdisk says
  Code:

  could not detect sector size

• zpool import
  Just checking to see if these disks had a zpool on them that prevented them from being wiped (checked this before setting the flags and doing any wipes that worked)
• sg_readcap
  Now this can actually see the disks, and reports the correct size and block size.
• sg_format
  This can also actually format the disks, or at least it claims it can, and I did see the block size go from 520 to 512 when I had sg_format do that format, changing the block size didn't fix the issues with any of the other commands that I was trying to do, but it is one of the few things I've tried that has some effect on the disks.
• wipe disk from GUI
  Whichever type of wipe I try to do I get
  Code:

  Error: [Errno 22] Invalid argument

• Clear SED Password from GUI
  Just goes, doesn't seem to have any issue, but also doesn't seem to have any effect.
• SMART test from GUI
  Short tests run just fine which is rather odd, but long tests give
  Code:

  Long (extended) offline self test failed [unsupported field in scsi command]

• Creating a pool from GUI
  Get an error ending with
  
  Code:

    File "/usr/local/lib/python3.9/site-packages/middlewared/plugins/disk_/format.py", line 21, in format    raise CallError(f'Failed to wipe disk {disk}: {job.error}')
  middlewared.service_exception.CallError: [EFAULT] Failed to wipe disk multipath/disk26: [Errno 22] Invalid argument

  
  I also got the same error earlier when it was trying to use the daX labels instead of the multipath

My setup is a Supermicro X10DRFR-NT with an LSI 9400-16e HBA running firmware 24.00.00.00 in IT mode. The 9400 has 2 cables, one going to each controller on an MD1420, which then has in turn a cable going to an MD1200. MD1200 is filled with 12x HGST H7280B520SUN8.0T's and the MD1420 already had 8x Toshiba PX02SMQ160's. Both sets of drives have had their own pool and have been functioning fine and continue to while I have issues with these Hitachi disks. This is running on TrueNAS Core, 13.0-U4.


I've seen some people talk about something like "protectionlevel" possibly locking the drives and it being a variable that I would need to clear, but I haven't seen anything online about how I would go about doing that.

Any insight would be greatly appreciated.

 

[NilsS]

Well, I did eventually get it fixed. Ended up taking the drives out of my enclosure, and putting them into another server I have lying around, booting it up with GpartedLive and doing an nwipe on them. For anyone else who ends up with this situation it does seem to be some sort of software flag that can be fixed by doing that low level full disk wipe, I would also recommend using sg_format to make sure it's set as 512 bit blocks before you do that wipe.