背景:
每个节点都有以下网络配置:
<network>
<port auto-increment="false">5701</port>
<outbound-ports>
<ports>10000-10005</ports>
</outbound-ports>
<join>
<multicast enabled="false" />
<tcp-ip enabled="true">
<member-list>
<member>{assume ip of node 1}</member>
<member>{assume ip of node 2}</member>
<member>{assume ip of node 3}</member>
</member-list>
</tcp-ip>
</join>
因此,每个节点有6个端口用于出站通信,根据我目前对hazelcast的理解,这对于3个节点就足够了。但是,在每个节点上执行 netstat 时,我注意到每个节点上的跟踪类似(假设{n}为某些非零数字)
@node1
tcp {n} 0 {ip of node 1}:5701 {ip of node 2}:35576 ESTABLISHED
tcp {n} 0 {ip of node 1}:5701 {ip of node 3}:10001 ESTABLISHED
tcp 0 {n} {ip of node 1}:54941 {ip of node 2}:5701 ESTABLISHED
tcp 0 {n} {ip of node 1}:10000 {ip of node 3}:5701 ESTABLISHED
@node2
tcp {n} 0 {ip of node 2}:5701 {ip of node 1}:54941 ESTABLISHED
tcp {n} 0 {ip of node 2}:5701 {ip of node 3}:10000 ESTABLISHED
tcp 0 {n} {ip of node 2}:10003 {ip of node 1}:5701 ESTABLISHED
tcp 0 {n} {ip of node 2}:35576 {ip of node 3}:5701 ESTABLISHED
@node3
tcp {n} 0 {ip of node 3}:5701 {ip of node 1}:10000 ESTABLISHED
tcp {n} 0 {ip of node 3}:5701 {ip of node 2}:10003 ESTABLISHED
tcp 0 {n} {ip of node 3}:10001 {ip of node 1}:5701 ESTABLISHED
tcp 0 {n} {ip of node 3}:10000 {ip of node 2}:5701 ESTABLISHED
查询:
注意:我对hazelcast很新,所以如果我错过了提供的话,请询问具体细节。
答案 0 :(得分:0)
是否可以在每个节点上分享lsof -a -i -n -P -p ${HAZELCAST_PID}
和ifconfig
的结果?
据我所知(参见SocketConnector),当属性hazelcast.socket.client.bind
为true
(默认值)时,Hazelcast将尝试绑定到提供的特定端口许多重试的<outbound-ports>
元素,然后失败,异常与您看到的超出范围的客户端端口不一致。