ES 关于 remote_cluster 的一记小坑

最近有小伙伴找到我们说 Kibana 上添加不了 Remote Cluster，填完信息点 Save 直接跳回原界面了。具体页面，就和没添加前一样。

我们和小伙伴虽然隔着网线但还是进行了深入、详细的交流，梳理出来了如下信息：

两个集群：集群 A 和集群 B ，版本都是 7.10.0 ；
1. 集群 A 没区分节点角色；
2. 集群 B 设置了独立的 master 节点、coordinator 节点和 data 节点，其中 data 节点还带 remote_cluster_client 角色；
在集群 A 的 Kibana 可以添加集群 B 为远程集群；
在集群 B 添加集群 A 就不行，Kibana 跳回之前的页面；
网络组确认已经放开策略，网络测试也正常；

翻看了 ES 和 Kibana 的日志， ES 日志中有连接失败的错误信息， Kibana 日志中无对应输出。

yaml 复制代码

[2023-09-13T11:38:41,055][INFO ][o.e.c.s.ClusterSettings  ] [710-1] updating [cluster.remote.test.mode] from [SNIFF] to [sniff]
[2023-09-13T11:38:41,055][INFO ][o.e.c.s.ClusterSettings  ] [710-1] updating [cluster.remote.test.seeds] from [[]] to [["127.0.0.1:7102"]]
[2023-09-13T11:38:41,056][INFO ][o.e.c.s.ClusterSettings  ] [710-1] updating [cluster.remote.test.mode] from [SNIFF] to [sniff]
[2023-09-13T11:38:41,056][INFO ][o.e.c.s.ClusterSettings  ] [710-1] updating [cluster.remote.test.seeds] from [[]] to [["127.0.0.1:7102"]]
[2023-09-13T11:38:41,057][INFO ][o.e.c.s.ClusterSettings  ] [710-1] updating [cluster.remote.test.mode] from [SNIFF] to [sniff]
[2023-09-13T11:38:41,057][INFO ][o.e.c.s.ClusterSettings  ] [710-1] updating [cluster.remote.test.seeds] from [[]] to [["127.0.0.1:7102"]]
[2023-09-13T11:38:41,093][INFO ][o.e.c.s.ClusterSettings  ] [710-1] updating [cluster.remote.test.mode] from [SNIFF] to [sniff]
[2023-09-13T11:38:41,094][INFO ][o.e.c.s.ClusterSettings  ] [710-1] updating [cluster.remote.test.seeds] from [[]] to [["127.0.0.1:7102"]]
[2023-09-13T11:38:41,094][INFO ][o.e.c.s.ClusterSettings  ] [710-1] updating [cluster.remote.test.mode] from [SNIFF] to [sniff]
[2023-09-13T11:38:41,094][INFO ][o.e.c.s.ClusterSettings  ] [710-1] updating [cluster.remote.test.seeds] from [[]] to [["127.0.0.1:7102"]]
[2023-09-13T11:38:41,095][INFO ][o.e.c.s.ClusterSettings  ] [710-1] updating [cluster.remote.test.mode] from [SNIFF] to [sniff]
[2023-09-13T11:38:41,096][INFO ][o.e.c.s.ClusterSettings  ] [710-1] updating [cluster.remote.test.seeds] from [[]] to [["127.0.0.1:7102"]]
[2023-09-13T11:38:51,099][WARN ][o.e.t.RemoteClusterService] [710-1] failed to connect to new remote cluster test within 10s
[2023-09-13T11:39:11,101][WARN ][o.e.t.SniffConnectionStrategy] [710-1] fetching nodes from external cluster [test] failed
org.elasticsearch.transport.ConnectTransportException: [][127.0.0.1:7102] handshake_timeout[30s]
	at org.elasticsearch.transport.TransportHandshaker.lambda$sendHandshake$1(TransportHandshaker.java:73) ~[elasticsearch-7.10.2.jar:7.10.2]
	at org.elasticsearch.common.util.concurrent.ThreadContext$ContextPreservingRunnable.run(ThreadContext.java:684) ~[elasticsearch-7.10.2.jar:7.10.2]
	at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1130) [?:?]
	at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:630) [?:?]
	at java.lang.Thread.run(Thread.java:832) [?:?]

已经测试过网络是通的，开始尝试把集群 B 的角色配置统统去掉再尝试。

嘿，添加成功了。

接下来为了找出问题，开始一步一步给每个节点再加回原来的角色。而且如果节点原来没有 remote_cluster_client 角色，这次也一并加上。
操作过程