"cg 0: bad magic number" with umass

[Boris Kochergin]

Ari Sovijärvi wrote:
> Hi folks!
>
> I added a 1 terabyte USB harddisk to my Fire V100 for backups. I have 
> recently used the same device with Linux and with i386 FreeBSD. I 
> zeroed the disk and labeled it with sunlabel. However, newfs always 
> dies with the error "cg 0: bad magic number". I also tried to label it 
> into several smaller chunks, no luck.
>
> I also tried to newfs the whole device without a disklabel, but that 
> also ends with the same error.
>
> Here's an example of the outcome, with 1 gigabyte partition:
> # newfs /dev/da0b
> /dev/da0b: 1027.6MB (2104512 sectors) block size 16384, fragment size 
> 2048
>         using 6 cylinder groups of 183.72MB, 11758 blks, 23552 inodes.
> super-block backups (for fsck -b #) at:
>  160, 376416, 752672, 1128928, 1505184, 1881440
> cg 0: bad magic number
>
> Here's bits of the dmesg, showing the drive:
> umass0: <LaCie SA LaCie Hard Drive USB, class 0/0, rev 2.00/0.00, addr 
> 2> on uhub0
> da0 at umass-sim0 bus 0 target 0 lun 0
> da0: <ST310005 28AS > Fixed Direct Access SCSI-2 device
> da0: 1.000MB/s transfers
> da0: 953869MB (1953525168 512 byte sectors: 255H 63S/T 121601C)
>
> Any ideas of what's wrong?
>
I had the same problem on an i386 machine with an onboard USB 1.1 
controller some time ago, using external Seagate disks. I Googled around 
and found that the error message means that newfs tried to read back 
something it wrote near the beginning of the filesystem when it created 
it, but the contents were not what it expected. After making sure that 
it worked with my laptop's ICH6 USB controller--which it did--I 
dismissed the onboard controller as faulty, bought a PCI VT6202 
controller, and it's worked great ever since.

-Boris