Issue verifying media

[jdeus]

Hello and thank you for forking!

I'm writing a utility to make multi-media WORM backups to blu-rays and I'm facing the following issue:
On PC1 (Debian 12 inside a Docker 12 docker with /dev/sr0 mounted) with Verbatim Slimline (USB) (externally powered) which is has a Pioneer drive inside :

• Creation of an ISO file (genisoimage -allow-limited-size -R -J -iso-level 4 -o media.iso file.tar)
• ECC augmentation of the ISO (dvdisaster -c -i media.iso -mRS03 -n BD --ecc-target image)
• Burning of the media (growisofs -dvd-compat -use-the-force-luke=notray -speed=1 -Z /dev/sr0=media.iso)

Verification always fails (same machine):

dvdisaster -a RS03 -r -d /dev/sr0 -i media1_pioneer.iso
dvdisaster -a RS03 --ecc-target image -i media1_pioneer.iso -t

Error correction properties:
- type             : Augmented image
- method           : RS03, 56 roots, 28.1% redundancy.
- created by       : dvdisaster-0.79.10 (unstable)
- requires         : dvdisaster-0.79
- data md5sum      : none available

Data integrity:
- medium sectors   : 11826135 total / 9162560 data
- good image/file  : all sectors present
- data md5sum      : 749fd65d5ea80ec554e24c54bf8ccca2
* Ecc block test   : 46313 good, 64 bad; 130560 bad sub blocks

If repeated, the result is the same.

On PC2 (Windows 10 with a separate reader ASUS BW-16D1HT) the exact same disk can be read without any issue :

dvdisaster -a RS03 -r -d e: -i media1_asus.iso
dvdisaster -a RS03 --ecc-target image -i media1_asus.iso -t

Error correction properties:
- type             : Augmented image
- method           : RS03, 56 roots, 28.1% redundancy.
- created by       : dvdisaster-0.79.10 (unstable)
- requires         : dvdisaster-0.79
- data md5sum      : none available

Data integrity:
- medium sectors   : 11826135 total / 9162560 data
- good image/file  : all sectors present
- data md5sum      : 749fd65d5ea80ec554e24c54bf8ccca2
- Ecc block test   : pass

In both cases I'm using the latest CLI release (win64-cli for Windows) : dvdisaster-0.79.10 pl3

I have not investigated further at the moment. Is there something obvious that can explain this situation ?

[speed47]

Hello @jdeus,

Thanks for the detailed report! The pl3 version does fix an issue under Windows where if a silently unsupported option was enabled (#82), errors would be encountered. However, you seem to observe the opposite, i.e. problems under Linux but not under Windows, which seems quite strange.

The original authors stopped supporting the Windows version some time before stopping updating the whole software entirely, hence I knew re-introducing the Windows version might indeed imply having a few bugs under this OS, but I'm not expecting the Linux version to behave like this. I consider it the "reference" version, so to speak, because that was the most recently updated version.

Now, the first thing I would do if I were you, would be to run the program with the verbose option on both OSes, and see when/whether the output differs. I would be interested in having both logs if you do that (of course, feel free to remove any sensitive information from them, if any is present, such as the image label or such).

[speed47]

So I used your exact commands (thanks for detailing them) to reproduce thf bug step by step:

• Creation of an ISO file (genisoimage -allow-limited-size -R -J -iso-level 4 -o media.iso file.tar)
• ECC augmentation of the ISO (dvdisaster -c -i media.iso -mRS03 -n BD --ecc-target image)
• Burning of the media (growisofs -dvd-compat -use-the-force-luke=notray -speed=1 -Z /dev/sr0=media.iso)

I bought a small spindle of BD-RE to do it. I first generated a file.tar full of random data to end up with a very similar number of sectors.

Unfortunately, I can't reproduce the bug:

Error correction properties:
- type             : Augmented image
- method           : RS03, 56 roots, 28.1% redundancy.
- created by       : dvdisaster-0.79.10 (unstable)
- requires         : dvdisaster-0.79
- data md5sum      : none available
Data integrity:
- medium sectors   : 11826135 total / 9162984 data
- good image/file  : all sectors present
- data md5sum      : 860e9332e8b7df6a3b051463ae6a13a3
- Ecc block test   : pass

So I'm wondering whether this could be a hardware issue on your side, such as faulty ram flipping bits, or your pioneer drive of PC1. You might want to live boot a Linux on PC2 and try to read the disc back with this hardware using the Linux version, this should work if this hypothesis is correct.