* You want diff backup
* You want the files to be directly accessible
* You are storing large files
* You want to do this on Linux
A solution that werks on Loonix and isn't some gentooman's github is ZFS snapshots + send/recv. Have all your drives formatted as ZFS and use send/recv to push snapshots to your backup.
lolwut?
If I knew how to do that I wouldn't need to ask Jow Forums for assistance. I don't even know what tools would be necessary for my use case. I will take nudges in the right direction if you have them.
Unfortunately zfs is not an option. My hard drives are already formatted as ntfs and they'll sadly have to stay like that for the rare occasion I might need to access them on windows. Also, on GNU I'll be using btrfs, but zfs. But thanks.
Does anyone else have some ideas?
I don't know, man. How big are the files that we're talking about here?
Nothing extreme like the 2 TB this user uses their program with, but some can be upwards of 20 GB which already would be very overkill to move back and forth every week on a USB 3 connection. Not to mention the unnecessary writes to the HDD. That's why I need to send only the changes to the files, which apparently is called delta.
rsync --inplace --no-whole-file
Thanks. After researching for a bit it does seem like rsync would probably be the logical choice. Unless anyone else has a better solution.
This might be a solution if you don't expect any speed improvements. It will write less to the hdd but it will also take much longer than a regular copy, just saying it. Rsync has to read the section from the hdd before it can compare their contents. This will lead to increased copy times.
cron and rsync - google it, it'll take about 2 min to setup if you've got half a brain.
Why would reading from the external hard drive take longer than writing to it?