-
Notifications
You must be signed in to change notification settings - Fork 234
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
incorrectly transferred (fingerprint mismatch in file contents) #27
Comments
The file contents are also used as part of the fingerprint, though you can disable this behavior (with some loss in safety -- see the manual) by setting both the fastcheck and fastercheckUNSAFE options to true. You might try running md5 on the local and acd-mounted copies of the file and see if it agrees with diff. Otherwise, running unison with -debug=verbose (on just a single file -- the output will be large) and looking for where it's making fingerprints might give a hint. |
Thanks for your ideas. I have checked the file contents using diff and md5sum, they are identical. The debug doesn't seem to give much away about what is being fingerprinted or when. Interestingly the code suggests that the exception should indicate at least whether it was contents or metadata that differ, but that doesn't seem to be shown in the debug trace. I have tried /c/amazon/Pictures/nas9=true and the debug is quite different, but fails because of a mismatched fingerprint. I believe this is purely based on size? The files at both ends are the same size, and the four references in the debug all match the expected size. Any further ideas? |
the best shot for you is to override os.reasonForFingerprintMismatch to output as debug also the content of both fpdata. Diff after the function was done may not show everything. |
Thanks for the hints. For anyone else's benefit after some debugging I found the cause of the problem was on the acd_cli side, as by default it runs commands asynchronously. This can be changed using the --single-threaded option, after which unison and acd_cli work well together |
Thanks for the follow-up! |
I'm trying to sync a directory mounted on acd_cli (https://github.com/yadayada/acd_cli) but every file fails with
All the files are being transferred to the acd mount (rather than from it). I believe that the inode, mtime and perms are used to fingerprint? (the file itself is identical, checked with diff)
I am running with perms=0 so hope that means it is only inode and mtime. These seem to both work and be sensible values on the mount.
Any ideas on what might be going wrong, or even how I might figure out why it is failing?
I'm running v2.48.3
The text was updated successfully, but these errors were encountered: