-
Notifications
You must be signed in to change notification settings - Fork 342
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
Support for advanced filesystem features (setuid/setgid/xattrr/hardlinks/sockets, etc) #544
Comments
Yes, there's a whole bunch of things currently not captured at the filesystem level, including setuid/gid, hardlinks, mount points, sockets, xattr, ACLs, etc. Currently only files (simple, without alternate streams) and symlinks are supported, and only simple filemode + owner/group are supported. I'm going to repurpose this issue to track adding the missing pieces and close the others as duplicates. |
Sounds fine to me, thanks. Wasn't sure if you preferred individual issues or not. |
Hm, see #543. When running |
Ideally this will include - under Windows - support for:
IMO, for the latter, the required behaviour should be to just capture the symlink/junction point/hard link itself, and not it's target. |
If you write this feature kopia will be good tool to backup whole VM, now i can do this with xfs_dump tool. |
@jkowalski |
Symlinks and hardlinks are correctly backed up. Both can be restored but hardlinks are restored as individual files and not as hardlinks. |
thnx @jkowalski , just would like to clarify how does it work( restore) if there are mutliply hard links to the same file?
|
|
It's mostly some additional restore work, architecture is fine. |
Sockets: until full support is implemented, would you reword this message:
to indicate that (1) the file in question is a socket, and (2) it has not been backed up? Also in the interim, would you provide a non-default option to ignore sockets, in order to reduce noise? |
@jkowalski |
No it does not, will reopen. |
One extra comment: it'd be worthwhile to add this limitation to the documentation, to make sure people are aware of this limitation, instead of discovering it at the test - or worse, at restore time. |
I'm currently snapshotting a disk full of backups made with the excellent RaspiBackup. It has a folder structure of incremental backups using hardlinks. It is taking a very long time to do the first snapshot! Would it be quicker if I snapshot the latest backup first? Then add the whole disk? |
In case it's useful for discussion or when writing documentation, this forum page on Duplicati has some good explanations of the difficulties around hardlinks. |
Hi Jarek, you say supporting hardlinks is "mostly some additional restore work" - Kopia backups with no hardlinks take ~11 seconds for 2.5 TB of files, but a much smaller directory of hardlinks takes ~9 hours. Nothing seems to bring this time down. Is this time expected to come down with an update? |
Still needed for macOS users. Thanks! |
I don't think the fix from @PhracturedBlue is merged into kopia yet? Still a wip maybe? |
Something screwy going on in github. I don't think I ever touched this code, it somehow got attributed to be due to an unrelated PR. The fix in question seems to be this one: #2597 which was merged, though I don't think it entirely fixes the issue (based on description). |
Hi all, I've found myself in need for support for xattr (on Linux) and was thinking along the following lines, put here to see if this is something I should undertake. Restic is also in Go and matches license wise and does support xattr. Can I, should I ?, lift and modify that code and prepare a PR here? |
this has become: #3643 |
Version: 0.6.1
Followup to #543
After the
kopia restore
I assume aside from the SUID bit, SGID and sticky bit are also not (yet) properly supported.
The text was updated successfully, but these errors were encountered: