Practice
Learn
← Quit
0 / 0 correct
Distributed Systems
Hard
1 of 40
Why is automatic failover in asynchronous primary-replica setups risky?
A
The promoted replica may lack recent acknowledged writes, which are lost, and the old primary may come back thinking it's still leader
B
Replicas can't become primaries
C
Failover is always safe
D
It makes reads slower