Hello, A recent hard disk read failure lost two blocks of a compressed, multipart .hdd package image. This was a hardware failure of the disk and affected two blocks in the .hdd file. This was the only virtual disk for a Windows XP VM. Each block lost was 512k, both well into the expansive (35GB compressed) disk (not the start or end). Using dd I was able to copy each part of the whole .hdd package to another disk (using conv=noerror,sync, so bad blocks were padded out with nulls). Only one of the multiple compressed disk image sections was damaged. Unfortunately for my recovery, the disk is also encrypted, so low-level disk read tools like dd_rescue are not reasonably possible which might allow me to put the two broken blocks back together.. I'm stuck with going through the MacOS hw abstraction layer of PGP WDE. When I try to run the VM Parallels (Build 6.0.12106 Revision 692267; September 7, 2011; on MacOsX 10.6.8) identifies the hdd image as corrupt, and will not allow it to be attached. On boot it reports no media, and on configuration edit and save Parallels reports: I was able to use the Parallels Mounter to mount and copy all files from that virtual disk. My data is safe. I would like to know if there is a way to tell Parallel's disk compression layer to ignore a couple errors and allow the system to boot (since a chkdisk run would likely resolve the issues and I wouldn't have to do an entire VM re-install). I don't mind a few blocks data loss for compress checksum failures, but I do mind if a data corruption error isn't recoverable for the VM as a whole. Have I missed some support trick or tool that would allow me to do this?
