Anne Wilson wrote: > Hi, Mikkel. I'm running testdisk at the moment, to see how much is > recoverable. Originally it had a FAT32 partition at the beginning, (Win98, > at one time) and the extended partition shows up as FAT32, both, seemingly, > type Oc. The first windows partition may not be recoverable, by the look of > it, and maybe some of the linux partitions inside the extended partition. > They are partitions that I don't care about anyway, so it doesn't really > matter. When the intense search has finished I'll tell it to mark for > recovery all the bits it thinks it can recover, then write a new partition > table. If that works fdisk should be able to see the drive again, and I'll > take it from there. > > Anne > Hi Anne, Let me know how testdisk does. On drives that have been repartitioned a couple of times, it will sometimes find the 'old' partitions. This is because the signature for the old partition is still there. But it will usually show up as non-recoverable. But that is why the program asks for user input before trying to recover partitions. The user usually has an idea of how the drive was partitioned, and can make a better 'guess' about what partitions are valid. The partition types for an extended partition and a FAT32 partition are not the same. But that information was in the partition table, and that is gone. So what testdisk is going by is the 'signature'. But it sounds like it is finding more then one 'signature' for the same chunk of drive. This is because changing the partition table only affects a few sectors on the drive, and even formatting a partition does not erase most of the old data from the drive. Depending on the format, it will only overwrite the data when writing the FATs, or the Superblocks. Mikkel -- Do not meddle in the affairs of dragons, for thou art crunchy and taste good with Ketchup!
