-
-
Notifications
You must be signed in to change notification settings - Fork 742
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
evaluate Lars's vision for backups #326
Comments
That one is really important. I backup regularly on an external HDD and it might not be connected every time when the cron job fires. |
@rumpelsepp i'd appreciate if you would create an issue detailing your use case, then. :) |
Quite some of that is just out-of-scope for borgbackup (or any non-feature-creep backup tool). I've use obnam for (very) short, but neither performance (about a magnitude slower than attic back then) nor its crypto usage (forking a gpg once per file to do symmetric crypto) nor the contribution workflow (send a patch to a mailing list) impressed me. That's how I found attic. :) So, while I like to look at alternative approaches, I am a bit selective when adopting them. |
The summary here isn't a description of obnam, or isn't necessarily related to obnam, other than the original blog post happens to have been written by obnam's author. I don't know if all those features are out of scope for borg: 4 of the 7 missing features already have issues in our own issue queue. Furthermore, while some things like "automatic configuration of external drive" may not belong in the borg codebase itself, it could be part of a debian package that would ship i think it's still useful to evaluate those ideas to provide a good and smooth user experience. of course our priority is to make the backend solid, but i like to think of this as a long term project. :) |
just as a note: I'ld appreciate if all topics that we consider for borg are taken to separate issues and discussed there and removed from the list, so we can close this issue when we have decided. |
My use case is, that I backup on a external hdd and quite frequently I have problems with stale lockfiles. When these files are present, borg hangs forever when I retrigger the backup; that sucks… :/ |
@rumpelsepp this is completely unrelated to THIS issue. See #136, #157, #210 and #285. |
Yeah, you are right... Sorry for that. |
@rumpelsepp i still believe your use case would be useful to describe:
|
IMHO a tool for HD config/triggering should be distinct from the actual backup tool Separation of concerns is critical |
I'll close this issue as too unfocussed soon - if you'ld like to keep something of it, create individual, focussed and non-vague tickets. |
I do this with udev and a systemd unit. If you always use the same physical port (same USB port, or you have a SATA disk frame) you can create a udev rule that starts a systemd unit (or just runs a script).
The long "/devices/pci0000:00/0000:00:11.0/ata1/host0/target0:0:0/0:0:0:0/block/*" string can be constructed from /sys, but it's easier to just use The script can then just mount the drive based on a similar hardware path ( |
If you can't identify backup drives by physical path, you could still automate it by having udev match against the UUID (or label) of the file system. |
btw, we have an "extra" repo now for stuff that is related somehow, but should not go into borg repo. |
Maybe deployment.rst might be reorganized into a section with, well, basically "guides" to explain common setups? |
I'ld rather keep such stuff separate. There are many different setups, even many common ones. |
Quick note for people from the future: I've done this some time ago. https://borgbackup.readthedocs.io/en/latest/deployment.html |
i keep on coming back to this article from Obnam's author, Lars Wirzenius. He makes a very strong points that backups should be way more transparent for our users and while i understand this may not be our priority at first, it would be an interesting end goal for borg.
The gist of it:
Stuff we already have:
CACHEDIR.TAG
standard)thoughts?
The text was updated successfully, but these errors were encountered: