On 5/25/23 15:09, Jerry Feldman wrote:
If you use rsnapshot or back in time or the many other rsync based backup/snapshot tools onto a Unix/Linux file system, you effectively have an incremental backup system. They use hard links to link all the duplicates
I have a home-brewed script I run. I think I based it on rsnapshot, though I might be remembering wrong.
Yes, the link-dest feature of rsync is very cool. Something people forget about backups is how to go about doing a restore, and how would they even know whether the restore will even work, until it is too late?
The cool thing about these link-dest trees is the "restore" step can sometimes be as simple as mounting the backup disk and a cd into the backup directory. The other cool thing is that doing a backup basically exercises everything needed to know whether a restore *would* work.
A downside is backups can take a while. Even if there is little new data to backup, the entire directory tree still needs to be copied. Another downside to what I am doing is there is no modern snapshot magic happening here, so version skew over the backup time could be serious. On my boring laptop this is not a practical problem.
I am looking forward to seeing how much faster even the cheap M.2 SSD will be when compared to a spinning disk. And looking forward to how much smaller it will be to use as the backup I carry when traveling.
-kb, the Kent who, when he next sets up a laptop, will have to finally get slightly modern and use btrfs so he can finally do snapshots, except for the fact that snapshots are so powerful that he will have to think through his whole backup approach about then.
_______________________________________________ Discuss mailing list [email protected] http://lists.blu.org/mailman/listinfo/discuss
