Comment 4 for bug 909424

Revision history for this message
rduke15 (rduke15) wrote : Re: Parted reports ext3 partition as hfs+

Thank you for looking into this.

Unfortunately, I cannot remember 100% how I prepared this disk a few years ago. But most probably, it would have been with cfdisk in Debian 4.

Maybe this WD external disk was originally delivered with a Mac partition? The disk was at a client's site, where they have many Macs. So it is also possible someone once connected it to a Mac. But they certainly didn't "initialize" or "erase" it (to use Macs terminology), because the disk still has all it's data and works as expected as an ext3 partition.

So apparently, parted gets mislead in it's partition identification code, in some special case. Hopefully my sample will help correct this glitch. I rely a lot on parted since it seems to be the only tool that allows me to entirely script a partitioning of new disks, and I started to need to do that a lot.

In the meantime, maybe it is safer to first zero out the first block(s)? I sometimes do it for the first 512 bytes. Should I do it for all the first 16K?