Disk unusable as pool member

phillxnet · June 1, 2022, 8:57pm

@ITstaff Hello again.
Re:

Agreed, nice find. That’s a pain.

Cheers. This is a tricky one. As real independant disks always have unique serial numbers and we have this as a hard dependency.

And that log snippet is consistent with Rockstor fighting with independatn drives having the same serial. We basically have to ‘cover’ each duplicate found with a fake-uuid serial that is temporary and changed on each refresh. See the following doc for some background:

The fact that there are sda sdb type names here for the drives concerned pretty much guarantees that hpsa is already being used. Plus we have:

OK, we may be stuffed here with the current raid controller function, as is. But take a look at exactly how we extract the serial using udev here:

github.com

rockstor/rockstor-core/blob/81a0d5f5e0a6dfe5a922199828f66eeab0253e65/src/rockstor/system/osi.py#L1203-L1295


      
          def get_disk_serial(device_name, device_type=None, test=None):
              """Returns the serial number of device_name using udevadm to match that
              returned by lsblk. N.B. udevadm has been observed to return the following:-
              ID_SCSI_SERIAL  rarely seen
              ID_SERIAL_SHORT  often seen
              ID_SERIAL        thought to always be seen (see note below)
              N.B. if used in this order the serial is most likely to resemble that shown
              on the device label as well as that returned by the lsblk. ID_SERIAL seems
              always to appear but is sometimes accompanied by one or both of the others.
              When ID_SERIAL is accompanied by ID_SERIAL_SHORT the short variant is
              closer to lsblk and physical label. When they are both present the
              ID_SERIAL appears to be a combination of the model and the ID_SERIAL_SHORT
              --------- Additional personality added for md devices ie md0p1 or md126,
              these devices have no serial so we search for their MD_UUID and use that
              instead.
              :param device_name: eg /dev/sda as per lsblk output used in scan_disks()
              :param device_type: the lsblk TYPE for the given device eg: disk, crypt.
              The equivalent to the output of lsblk -n -o TYPE device_name. Defaults to
              None as an indication that the caller cannot provide this info.
              :param test: When not None this parameter's contents is substituted for the

This file has been truncated. show original

Basically it calls

/usr/bin/udevadm info --name=device_name_here

Where the device name is the canonical one such as sdX. And parses the output to gather a serial that looks most like what we used to use exclusively from lsblk but it started to fail a lot in reporting serials and so we moved to failing through udevadm. See also:

github.com

rockstor/rockstor-core/blob/81a0d5f5e0a6dfe5a922199828f66eeab0253e65/src/rockstor/system/osi.py#L260


      
          dnames = {}  # Working dictionary of devices.
          disks = []  # List derived from the final working dictionary of devices.
          serials_seen = []  # List tally of serials seen during this scan.
          # Stash variables to pass base info on root_disk to root device proper.
          root_serial = root_model = root_transport = root_vendor = root_hctl = None
          # flag to indicate bcache backing device found.
          bdev_flag = False
          # To use udevadm to retrieve serial number rather than lsblk, make this
          # True N.B. when lsblk returns no serial for a device then udev is used
          # anyway.
          always_use_udev_serial = False
          device_names_seen = []  # List tally of devices seen during this scan
          for line in o:
              # skip processing of all lines that don't begin with "NAME"
              if re.match("NAME", line) is None:
                  continue
              # setup our line / dev name dependant variables
              # easy read categorization flags, all False until found otherwise.
              is_root_disk = False  # base dev that / is mounted on ie system disk
              is_partition = is_btrfs = False
              dmap = {}  # to hold line info from lsblk output eg NAME: sda

And you could start by simply changing the “False” to “True” (and then restarting Rockstor) as from memory I think we only fall through to udev if lsblk can give us no serials at all.

Let see the output of a couple of different disks on this controller that otherwise show up as the same serial. It may be a tweak there could make this controller work in it’s current configuration. I’m assuming there is no other way to configure it to pass drives throught.

The hope I see here is that udev has created by-id names of:

lrwxrwxrwx 1 root root 9 mag 30 2022 scsi-0HP_LOGICAL_VOLUME_00000000 -> …/…/sdb
...
lrwxrwxrwx 1 root root 9 mag 30 2022 scsi-0HP_LOGICAL_VOLUME_01000000 -> …/…/sda
lrwxrwxrwx 1 root root 9 mag 30 2022 scsi-0HP_LOGICAL_VOLUME_02000000 -> …/…/sdc
lrwxrwxrwx 1 root root 9 mag 30 2022 scsi-0HP_LOGICAL_VOLUME_03000000 -> …/…/sdd
lrwxrwxrwx 1 root root 9 mag 30 2022 scsi-0HP_LOGICAL_VOLUME_04000000 -> …/…/sde

Which looks like artificially assigned serials of 01000000, 02000000, 03000000, 04000000. But if we can see these in the udev output then we may well be able to use them as the serials instead. Long shot but may workout. However before changing anything in the serial finder/parser try that False to True thing first as it may just be we are being initially mislet by lsblk’s serial reporting. For more info on this could you post the output of:

lsblk -P -p -o NAME,MODEL,SERIAL,SIZE,TRAN,VENDOR,HCTL,TYPE,FSTYPE,LABEL,UUID

If it’s dishing out those non unique serials of 5001438010043670 then the False → True change to force udevadm serial retrieval on our part may this may get us a little further along the way.

Hope that helps, and thank for working through this with us.

Nice.