-
On each Standby:
sv stop cluster -
On the Master:
sv stop cluster
-
On each Standby:
sv stop etcd -
On the Master:
sv stop etcd
-
On the Master and each Standby:
Open the etcd config file:
nano /etc/etcd/etcd.configModify the two existing lines to your new values (in milliseconds):
ETCD_HEARTBEAT_INTERVAL=4000 ETCD_ELECTION_TIMEOUT=20000Add a new line:
ETCD_SNAPSHOT_COUNT=5000
-
On the Master:
sv start etcd -
On each Standby:
sv start etcd -
Verify cluster health on the Master:
If etcd is not healthy before restarting the DAP cluster service, then the master will failover immediately upon starting the cluster service.
$ etcdctl cluster-health ... cluster is healthy
-
On the Master:
sv start cluster -
On each Standby:
sv start cluster