一致性（Consistency and WAIT）
WAIT does not make Redis a strongly consistent store: while synchronous replication is part of a replicated state machine, it is not the only thing needed. However in the context of Sentinel or Redis Cluster failover,
WAIT improves the real world data safety.
Specifically if a given write is transferred to one or more slaves, it is more likely (but not guaranteed) that if the master fails, we’ll be able to promote, during a failover, a slave that received the write: both Sentinel and Redis Cluster will do a best-effort attempt to promote the best slave among the set of available slaves.
However this is just a best-effort attempt so it is possible to still lose a write synchronously replicated to multiple slaves.
Since the introduction of partial resynchronization with slaves (PSYNC feature) Redis slaves asynchronously ping their master with the offset they already processed in the replication stream. This is used in multiple ways:
- Detect timed out slaves.
- Perform a partial resynchronization after a disconnection.
In the specific case of the implementation of
WAIT, Redis remembers, for each client, the replication offset of the produced replication stream when a given
write command was executed in the context of a given client. When
called Redis checks if the specified number of slaves already acknowledged
this offset or a greater one.
@integer-reply: The command returns the number of slaves reached by all the writes performed in the context of the current connection.
> SET foo bar OK > WAIT 1 0 (integer) 1 > WAIT 2 1000 (integer) 1
In the following example the first call to
WAIT does not use a timeout and asks for the write to reach 1 slave. It returns with success. In the second attempt instead we put a timeout, and ask for the replication of the write to two slaves. Since there is a single slave available, after one second
WAIT unblocks and returns 1, the number of slaves reached.