Hi Everyone! I recently encountered a failing nvme drive in my main PC and saved 99% of it using the amazing GNU ddrescue tool. (https://www.gnu.org/software/ddrescue/) Sadly there were one or two files in my NTFS partition that were not saveable according to ddrutility, another amazing tool that with NTFS partitions can tell you exactly which files are bad. (https://sourceforge.net/p/ddrutility/wiki/Home/)
After getting worried by this experience, I also used the tool to recover two older external HDDs that had been sitting on a self and they also had bad sectors but running the tool overnight with a lot of retries, I was able to save 99% of sectors with no NTFS file system errors.
Anyway, my question is that I'm now stuck with these raw disk images that are the size of the entire drives, as shown in the photo. However the NTFS partition that spans the drives is only 40% full in the case of the 4TiB partition with the rest being for the most part almost all 0s.
So, I would imagine these files would compress really well with even gzip, but the more reading I have done the more lost I am on how I can ever safely delete these raw DD images.
It seems like the safest thing to do would be to copy these files to a ZFS partition and run regular scrubs to ensure they are not corrupt, but that would waste 60% of the storage space tracking the 60% of the empty NTFS partition on the drive.
The creator of lzip has a very long article(https://www.nongnu.org/lzip/safety//_of//_the//_lzip//_format.html) on why it is the best tool for error detection and recovery for large data stream compression, and I believe the recommendation of the lzip creator would be for me to TAR all the drive images I have in the photo together and then use lzip to compress the TAR.
However when researching additional LZMA compression technologies that competed with lzip I ran into the concept of the PAR file as an alternative to a TAR for stream compression to get even more redundancy and error detection/recovery.
Some folks pointed out though, that for this type of long term storage, its much better to use a file archive compression tool rather than a stream compression tool and I was recommended WIN-RAR which actually incorporates PAR files into the archives. I don't mind paying for that software if its the best tool for this job, but the settings page for the PAR portion is quite complex... I don't really know how much redundancy I want or need, it would be nice if I new the probability of a bit flip occurring and what it might do if undetected, but I don't know what that would be given this setup.
So I guess my question is, does anyone have experience archiving large sparse compressible DD images in a safe way, I'm leaning towards just copying them to my ZFS Zraid pool and taking the space hit to ensure ensure the ZFS checksums will keep the files safe, but if someone more knowledgeable about lzip/WIN-RAR can recommend settings that have worked well maybe I can delete the originals and store the archives instead.
https://preview.redd.it/cp5v854woc9e1.png?width=1638&format=png&auto=webp&s=4c0398d9e1153524ac5afcbfc0ba0ec709a5e08c