版本: 7.12-8.4
简而言之,当在 Elasticsearch 中的非主节点上发出刷新异常检测器内存需求的请求时,会发生此错误。Elasticsearch 要求此类操作必须在主节点上执行。要解决此问题,您可以将请求重新路由到主节点,或者将该节点更改为主节点。此外,确保您的集群健康状况为绿色,并且所有节点都已正确连接,以避免此类问题。
日志上下文 #
日志 “Request to refresh anomaly detector memory requirements on non-master node” 的类名是 MlMemoryTracker.java. 我们从 Elasticsearch 源代码中提取了以下内容,供那些寻求深入上下文的人参考:
* ornullif it cannot be calculated.
*/
public void refreshAnomalyDetectorJobMemoryAndAllOthers(String jobId; ActionListenerlistener) { if (isMaster == false) {
listener.onFailure(new NotMasterException("Request to refresh anomaly detector memory requirements on non-master node"));
return;
} // Skip the provided job ID in the main refresh; as we unconditionally do it at the end.
// Otherwise it might get refreshed twice; because it could have both a job task and a snapshot upgrade task.





