Hello,
just want to know if somebody else is aware of such an effect.
Observed both on Windows 10 and 11, and on different drives respectively PCs.
Also for a quite long time meanwhile (years), meaning several FreeFileSync versions.
Using FreeFileSync on folders with >60,000 files (mostly pictures and videos) and ca. 300-500 GB big, organized in multiple sub-folders with mostly >1,000 files.
Usually one as master on the local drive and the others on internal and external backup drives.
Scanning and comparing such a folder files content takes a while (ca. 30-40 minutes) and FreeFileSync nearly always finds one or more often more files (up to 20) which are indicated as different / changed.
If I scan just the sub-folders one after another, (ca. 1000 files) this issue never happens.
Also comparing files marked as different with a binary compare tool, the tool shows no difference.
Is it just a statistical thing? Kind of shit happens?
Regards
File content compare fails sporadically (marks files as different) if many files scanned
- Posts: 3
- Joined: 6 Jan 2026
- Posts: 1222
- Joined: 8 May 2006
Shouldn't happen.
What type of scan are you running?
Are you sure you are in fact (externally) binary comparing the same pairs of files (that FFS is comparing)?
What type of scan are you running?
Are you sure you are in fact (externally) binary comparing the same pairs of files (that FFS is comparing)?
- Posts: 3
- Joined: 6 Jan 2026
I'm was running "Compare" "File content" from time to time with the intention to see if the files are really the same and unfortunately observed this issue.
Because it takes hours to verify, I've mostly just merged / overwritten the files.
I've tried it on 2 SSDs recently because these are faster than USB drives or HDDs.
"Just" copied ca. 500 GB of files from one drive to another and run compare multiple times.
FreeFileSync reported always different amount of files which supposed to be different (rarely none).
Selected just the folders with such files and compared them only. No issues.
Also took an extra tool for binary compare. No issues.
I do not really trust the FreeFileSync anymore, but it is one of the few tools which does such a compare, so I brake the amount of files to check to subfolders to be sure.
It may be something systematical, like too long file names or something broken in the files, but no idea what.
Because it takes hours to verify, I've mostly just merged / overwritten the files.
I've tried it on 2 SSDs recently because these are faster than USB drives or HDDs.
"Just" copied ca. 500 GB of files from one drive to another and run compare multiple times.
FreeFileSync reported always different amount of files which supposed to be different (rarely none).
Selected just the folders with such files and compared them only. No issues.
Also took an extra tool for binary compare. No issues.
I do not really trust the FreeFileSync anymore, but it is one of the few tools which does such a compare, so I brake the amount of files to check to subfolders to be sure.
It may be something systematical, like too long file names or something broken in the files, but no idea what.
- Posts: 1222
- Joined: 8 May 2006
> FreeFileSync reported always different amount of files which supposed to be different (rarely none).
Odd (IMO).
So I ran a scan, Content compare (which I don't typically do), ~40 GB of data (so smaller then your dataset).
20th Century. Same size, day/month .happen. to be reversed, target being newer then "source".
That's a bit odd?
And I would of, kind of, expected my regular Date/Time compare (with an Update sync) to have at least pointed that out (but it does not - maybe that's "normal"?).
940.LEX. Again target is newer, by ~1 hour, so maybe DST is playing in? That the size is the same is not unexpected. The files are not the same, though unsure what the differences are (& again I was unaware of this).
BGEFIG6, source is newer, so my normal Time/Size compare will update the target.
D13124.0, is odd (to me), target newer, but smaller.
REYNOLDS, ditto.
So with my brief exploration, not really seeing anything amiss - at least not on the actual Content compare end.
.
Odd (IMO).
So I ran a scan, Content compare (which I don't typically do), ~40 GB of data (so smaller then your dataset).
20th Century. Same size, day/month .happen. to be reversed, target being newer then "source".
That's a bit odd?
And I would of, kind of, expected my regular Date/Time compare (with an Update sync) to have at least pointed that out (but it does not - maybe that's "normal"?).
940.LEX. Again target is newer, by ~1 hour, so maybe DST is playing in? That the size is the same is not unexpected. The files are not the same, though unsure what the differences are (& again I was unaware of this).
BGEFIG6, source is newer, so my normal Time/Size compare will update the target.
D13124.0, is odd (to me), target newer, but smaller.
REYNOLDS, ditto.
So with my brief exploration, not really seeing anything amiss - at least not on the actual Content compare end.
.
- FFS - Content Compare.png (31.26 KiB) Viewed 93 times
-
- Site Admin
- Posts: 7523
- Joined: 9 Dec 2007
It's unlikely to be a FFS bug. I suggest checking your RAM (but it could also be other components). The fact that differences are only detected when the system is under high load seems to indicate a hardware issue.
- Posts: 3
- Joined: 6 Jan 2026
I've checked it (semi-systematically) with 2 different PCs but unfortunately with same drives, same effect.
Could be the drives in this case.
What else could be is the the operating system or which tool else ever is accessing or checking files for which reasons ever.
Just wanted to know, if someone is experiencing same issues.
I know that it is very difficult to debug.
Could be the drives in this case.
What else could be is the the operating system or which tool else ever is accessing or checking files for which reasons ever.
Just wanted to know, if someone is experiencing same issues.
I know that it is very difficult to debug.
- Posts: 4908
- Joined: 11 Jun 2019
It's also not crazy that a NAS touches files for some reason. Some offer scanning for malware, some have indexing, some can change metadata for organizing/sorting, etc. As far as locally connected devices, differences can still be caused by other software that may have the same purposes.
And of course, hardware issues, typically RAM for something like this. You can run memtest to check memory. Failing means bad memory, but passing doesn't mean there isn't an issue. The memory controller inside the CPU could also be a source of instability.
If this is happening with the same drives on different PCs, is it finding the same files are different every comparison? If so, could be failing in the spot those files are located, and enough heat buildup after a long scan causes a read error vs scanning the parent directory directly.
And of course, hardware issues, typically RAM for something like this. You can run memtest to check memory. Failing means bad memory, but passing doesn't mean there isn't an issue. The memory controller inside the CPU could also be a source of instability.
If this is happening with the same drives on different PCs, is it finding the same files are different every comparison? If so, could be failing in the spot those files are located, and enough heat buildup after a long scan causes a read error vs scanning the parent directory directly.
