Backups Aren't Simple
summary
Effective data backup is not a simple act of copying files, but a complex system of managing snapshots, rotation, and redundancy. While basic copies are prevent data loss from hardware failure, they must be evolved into a sophisticated strategy to protect against ransomware, human error, and bit rot. Ultimately, the author argues that the complexity of implementing these features manually is too high, making professional open-source tools like Borg or Restic indispensable for reliable data preservation.
key points
- Simple mirroring or RAID 1 is insufficient because it does not protect against accidental deletion or ransomware, which are immediately replicated to the backup.
- To optimize storage and bandwidth, a Grandfather-Father-Son (GFS) rotation scheme combined with deduplication is used to keep frequent recent backups and fewer older ones.
- The 3-2-1 rule—three copies of data on two different media types with one offsite—provides the same level of protection against physical disasters and hardware-specific failures.
who should read it
Individuals and hobbyists who manage their own data or homelabs, as well as those who may believe a simple copy-paste backup strategy is sufficient.
discussion positive
consensus
Commenters agree that the true value of a backup system lies in its ability to restore data rather than the act of backing up itself. There is a strong preference for tools like Restic and Borg due to their deduplication and append-only logic, which helps prevent accidental data deletion. Many emphasize the 3-2-1 strategy and the necessity of regular restore exercises to ensure viability.
top insight
A user explains how using GNU tar with the --listed-incremental flag allows for a portable, simple backup system that avoids complex repository formats. By separating the small index file from the large archive, the main data can be stored in write-once, low-cost storage like AWS Glacier while still allowing for daily incremental updates.
Generated by an LLM from the article and thread. Check the sources before relying on it.