Error with sync.ffs_db

[Fubax]

When the target hard drive has cluster size bigger than the sync.ffs_db file itself, the FFS fills the remaining empty space with 0x00, which then produces CRC error.

Cannot read database file "path\to\sync.ffs_db".
File content is corrupted. (invalid checksum)

I'm using Btrfs formatted drive with 4kb cluster size.

[Zenju]

The cluster size should not have any effect on file size. If it does it can be considered a file system bug.

[Fubax]

I think it's a file system bug as you said, since I used WinBtrfs driver on Windows, which is buggy and causes CRC errors on parallel write operations. That's a shame, since Btrfs is very nice file system for backup storage :(

[therube]

When the target hard drive has cluster size bigger than the sync.ffs_db file itself, the FFS fills the remaining empty space with 0x00

If it does that with sync.ffs_db, wouldn't you expect it to do that with all files where size < cluster size?
Yet, it does not - is that so?

If you did something like, cal > abc123, abc123 would show the expected calendar & without any extraneous 0x00, no?

Is compression enabled (in case that may be playing in)?

On windows, small files may be stored directly in the MFT (so in that respect, "size on disk" is 0).

[Fubax]

@therube If you like to investigate more, it did that only to sync.ffs_db not to other files, but files created by some other specific programs where also corrupted. What they had in common was that they where all written to the drive in a highly parallel order, not sequentially like byte, by byte. Only dose where affected.

And the compression was disabled. "size on disk" was correctly reporting the cluster size for small files.