greg (07/30/82)
I have found the hard way that dd'ing raw file systems does not point out errors in the transfer. I use bs=60k (the maximum allowed by minphys() in the hp driver. The really unpleasent matter is that fsck didn't show up any problems on a freshly dd'ed file system which had a clobbered directory file 6 levels deep. The "." and ".." pointers went into ozone. Part of the problem may be the hp drivers handling of ECC code. We are running 4.1BSD. Has anyone out there had any problems with the hp drivers ECC code??? Anyone have a dd like program which can handle bad disk blocks (from drivers without ECC?). I was prematurely elated at the prospect of being able to dd an entire 300 Meg RM05 pack in 27 minutes, but if I can't detect errors? Greg Hidley ucbvax!sdcsvax!dcdwest!greg or decvax!ittvax!dcdwest!greg