Cluster: re-order failover operations to make it safer.

We need to:

1) Increment the configEpoch.
2) Save it to disk and fsync the file.
3) Broadcast the PONG with the new configuration.

If other nodes will receive the updated configuration we need to be sure
to restart with this new config in the event of a crash.
This commit is contained in:
antirez 2013-09-30 10:13:58 +02:00
parent b187517719
commit 62b1591439

View File

@ -1693,16 +1693,16 @@ void clusterHandleSlaveFailover(void) {
}
}
/* 3) Pong all the other nodes so that they can update the state
* accordingly and detect that we switched to master role. */
clusterBroadcastPong();
/* 4) Update my configEpoch to the epoch of the election. */
/* 3) Update my configEpoch to the epoch of the election. */
server.cluster->myself->configEpoch = server.cluster->failover_auth_epoch;
/* 5) Update state and save config. */
/* 4) Update state and save config. */
clusterUpdateState();
clusterSaveConfigOrDie();
/* 5) Pong all the other nodes so that they can update the state
* accordingly and detect that we switched to master role. */
clusterBroadcastPong();
}
}