* Boyd Stephen Smith Jr. <[EMAIL PROTECTED]> [2008 Nov 14 20:40 -0600]: > On Friday 14 November 2008, ghe <[EMAIL PROTECTED]> wrote about 'Re: udev > causing data loss?': > >Florian Kulzer wrote: > >The names of internal disks change without notice. This, IMHO, is an > >extremely lame idea. I know better now, but it was an expensive lesson. > > That could happen with devfs, or static device nodes, too. I suppose it > was rarer, but then device scanning was non-asynchronous then. Udev > finally gives us a way to attach a name to the device vs. depending on > quirks of the device scanning code to stay the same. > > Udev didn't cause the data loss. fdisk did!
If one wasn't careful, such a data loss could happen in the DOS days (not sure about Win32). IIRC, consider a drive formatted as two partitions, C: and D: where C: is a primary partition and D: was an extended DOS (logical) partition. Add a new hard drive to the mix and fdisk it for two partitions. One could be forgiven for thinking the original drive would remain C: and D: and the new drive might be E: and F:, however, one would be wrong. Assuming IDE drives on the same cable (Master/Slave), the primary partition of the Master was C:, as expected, the primary of the Slave was D:, the extended partition of the Master was E: and the extended partition of the Slave was F:. In short, the system would assign all the primary partitions drive letters in physical order and then all of the extended partitions on the first drive followed by all the extended partitions on the second drive and so on. That required some explanation when a larger drive was added to a system prior to FAT32/NTFS. I thought I'd forgotten that arcane knowledge/experience. - Nate >> -- "The optimist proclaims that we live in the best of all possible worlds. The pessimist fears this is true." Ham radio, Linux, bikes, and more: http://n0nb.us/index.html -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
