httm VS Rdiff-backup

My projects and many projects include a THIRD-PARTY-LICENSES.html file when I distribute binaries. See: https://github.com/kimono-koans/httm/blob/master/third_party/LICENSES_THIRD_PARTY.html

The only real advantage I can think of with nested ZFS is that the files in the KVM would obviously be individual inodes in the nested ZFS filesystem and datasets, in addition to the one inode per virtual volume on the hypervisor (or zvol). This would allow for granular file management on the kvm and the use of tools like https://github.com/kimono-koans/httm which is like a command line time machine for ZFS.

> It's worth noting that copy_file_range is used by a lot of things.

Yes, but the trigger feature, block cloning, only landed in the latest 2.2 release. If you immediately hopped on 2.2, and used a system with lots copy_file_range and FICLONE use, yes, you may have a problem (like, as you note, on Gentoo, where this problem surfaced).

Most people were just hopping on the bandwagon. My distro ships 2.1.5, so I have a 6 month wait until this feature lands, so I was just building copy_file_range support into my ZFS apps, right before news of this bug hit.[0]

> There are other things required to trigger the bug that are a lot less common though.

Exactly. My guess is the incidence of this will exceedingly rare for the common user/small NAS user/etc. I've run a corruption detector[0], and what I've found mostly indicates false positives. Some are build artifact fingerprints, which I don't care about, and which were deleted with the next build. The ones with an extant file on another system, I confirmed were a diff match with the origin using `rsync -rincv` and whats on snapshots with `httm --map-aliases`. So far no positive matches.

[0]: https://github.com/kimono-koans/httm

cargo install --git https://github.com/kimono-koans/httm --branch clones strace -f -o stderr.txt -e ioctl -- httm -r -R ~/.zshenv

httm prints the size, date and corresponding locations of available unique versions (deduplicated by modify time and size) of files residing on snapshots, but can also be used interactively to select and restore files, even snapshot mounts by file! httm might change the way you use snapshots (because ZFS/BTRFS/NILFS2 aren't designed for finding for unique file versions) or the Time Machine concept (because httm is very fast!).

All btrfs snapshot tools can have different layouts. It's mostly a nightmare for any one tool to support. Although its not the tool you're looking for, FYI AFAIK httm supports all/most btrfs layouts, but it took more work than necessary to get there.

ZFS only option which requires super user privileges.

[0]: https://github.com/kimono-koans/httm/blob/master/httm.1

httm - Interactive, file-level Time Machine-like tool for ZFS/btrfs/nilfs2

For starters it has a tendency to paint itself into a corner on ENOSPC situations. You won't even be able to perform a restore if a backup was started but unfinished because it ran out of space. There's this process of "regressing" the repo [0] which must occur before you can do practically anything after an interrupted/failed backup. What this actually must do is undo the partial forward progress, by performing what's effectively a restore of the files that got pushed into the future relative to the rest of the repository, which requires more space. Unless you have/can create free space to do these things, it can become wedged... and if it's a dedicated backup system where you've intentionally filled disks up with restore points, you can find yourself having to throw out backups just to make things functional again - even ability to restore is affected.

That's the most obvious glaring problem, beyond that it's just kind of garbage in terms of the amount of space and time it requires to perform restores. Especially restores of files having many reverse-differential increments leading back to the desired restore point. It can require 2X the file's size in spare space to assemble the desired version, while it iteratively reconstructs all the intermediate versions in arriving at the desired version. Unless someone fixed this since I last had to deal with it, which is possible.

Source: Ages ago I worked for a startup[1] that shipped a backup appliance originally implemented by contractors using rdiff-backup. Writing a replacement that didn't suck but was compatible with rdiff-backup's repos consumed several years of my life...

There are far better options in 2024.

[0] https://github.com/rdiff-backup/rdiff-backup/blob/master/src...

[1] https://www.crunchbase.com/organization/axcient

and that should install the latest version, rdiff-backup-2.2.4-2.el8.x86_64.rpm. This is all described in the rdiff-backup README file.

Rdiff-backup - close to what you do currently but at least provides versioning. Based on rsync

as in just a copy of your files? This I would barely consider a backup, more of just a mirror from a point in time. What're you missing by doing this? versions of files, deduplication, and encryption (last one being very important for the best kind of backups, which should be off-site). Just because it's not files doesn't mean it's proprietary. Proprietary would mean secret and undocumented. There are many great options. Borg is my favorite but Kopia is probably better if you use windows, urbackup is an option if you want centralized management of backups and rdiff-backup is if you want something kinda what you have currently but adding versioning but lacks deduplication and encryption.

If you're comfortable with the cli and you want to have your backup in a plain file format with some incremental backups, there's rdiffbackup. It uses rsync under the hood and has worked quite well for me.

Rdiff Backup - Reverse differential backups that uses rsync, linking, and can tunnel via ssh. You get a full current backup with increments available to restore any version of the file with minimal storage space used.

borg is great. we've been using it for the past 3 years to archive hundreds of file-level backups of servers, database dumps and VM images. average size of each borg repo is few GB but there are few outliers up to few hundreds of GB.

borg replaced https://rdiff-backup.net/ for us and gave:

Robocopy is great if you don't have access to rsync. If rsync via WSL2 for instance is an option, I'd personally go with rdiffbackup.