Re: FS (dm-0): device supports 4096 byte sectors (not 512)

[Eric Sandeen <sandeen@sandeen.net>]

On 3/23/23 5:45 AM, Johnatan Hallman wrote:
> Hello List,
> 
> I get this error when I try to mount an XFS partition.
> Fortunately there is no critical data on it as it is just a backup but I would still like to mount it if it's possible.
> 
> I have tried with various Linux distros with kernels ranging from 5.6 to 6.1 it's the same result.
> 
> xfs_info /dev/mapper/test
> meta-data=/dev/mapper/test       isize=256    agcount=32, agsize=30523559 blks
>          =                       sectsz=512   attr=2, projid32bit=0
>          =                       crc=0        finobt=0, sparse=0, rmapbt=0
>          =                       reflink=0    bigtime=0 inobtcount=0 nrext64=0
> data     =                       bsize=4096   blocks=976753869, imaxpct=5
>          =                       sunit=0      swidth=0 blks
> naming   =version 2              bsize=4096   ascii-ci=0, ftype=0
> log      =internal log           bsize=4096   blocks=476930, version=2
>          =                       sectsz=512   sunit=0 blks, lazy-count=1
> realtime =none                   extsz=4096   blocks=0, rtextents=0
> 
> mount -t xfs -o ro /dev/mapper/test  /mnt/
> mount: /mnt: mount(2) system call failed: Function not implemented.
>        dmesg(1) may have more information after failed mount system call.

So I assume dmesg contained the error in $SUBJECT:

FS (dm-0): device supports 4096 byte sectors (not 512)

It seems that the filesystem was created with 512-byte sectors - at that time, the device
must have supported them. Perhaps something about the devicemapper target changed from
a 512 device to a 4k device? I'm not sure what might cause that to happen, but IMHO
it should never happen... did the dm device recently get reconfigured?

As a last resort, I think you could dd the filesystem (all 3T) to a file, and use a
loopback mount to access the files.

Alternatively, I wonder if we could relax the sector size check for a read-only
mount (that does not require log replay) - I'm not sure about that though.

-Eric