Split-brain
Two parts of a system both believe they hold the truth and accept changes independently, so their data drifts apart.
In clusters it happens when a network split leaves two nodes each thinking it is the primary; both accept writes, and merging them later loses data. Quorums and fencing prevent it there. The same thing happens at the architecture level when a new system is built next to an old one and both accept edits: every record now has two versions, and someone has to decide which is right, often by hand and for years.
Related terms
- Event sourcing: Storing every change as an event in an append-only log and rebuilding current state by replaying it, instead of storing only the latest state.
- Eventual consistency: A guarantee that copies of data will agree once updates stop arriving, but may briefly show different values in the meantime.
- Strangler fig pattern: Replacing a legacy system piece by piece, routing one feature at a time to new code until the old system can be switched off.
