HazelcastAPI ClusterHeartbeatManager Sınıfı

Giriş

Şu satırı dahil ederiz

import com.hazelcast.internal.cluster.impl.ClusterHeartbeatManager;

Kodu şöyle. Hata mesajı şöyle

Resetting heartbeat timestamps because of huge system clock jump! Clock-Jump: 174692 ms, Heartbeat-Timeout: 60000 ms]] 

Bir açıklama şöyle

Hazelcast maintains an internal clock to sync the nodes (a clock vector the size of the cluster nodes). When a JVM on one of the nodes will pause longer than 10 seconds, the internal clock for that node will register the time drift and will log the following message:System clock apparently jumped ...

The following is a full example of the message:

"INFO 2019-09-09 05:45:47,408 [cached1] com.hazelcast.cluster.impl.ClusterHeartbeatManager: [10.153.201.11]:5701 [1253850] [3.6.2] System clock apparently jumped from 2019-09-09T05:45:32.370 to 2019-09-09T05:45:47.408 since last heartbeat (+10038 ms)"

These drifts would cause the nodes to have a time difference between them. The time difference accumulates each time a long pause in JVM (> 10 seconds) is registered. When the time difference starts growing more than 20 seconds you might start observing first OperationTimeoutException messages in the logs.