Question

我正在尝试使用uni cast在Windows机器上设置ES群集。我想我已经完成了所有必需的配置更改，但我的ES节点仍然没有形成集群。有人请告诉我我缺少的内容。请在下面找到我的elasticseach.yml配置

   =======Noed 8=======
    cluster.name: elasticsearch
    node.name: NODE8
    node.data: true
    network.host: "10.249.167.8"
    network.publish_host: "10.249.167.8"
    network.bind: "10.249.167.8"
    discovery.zen.ping.multicast.enabled: false
    discovery.zen.ping.unicast.hosts: ["10.249.167.9", "10.249.167.10", "10.249.167.8"]

    transport.tcp.port: 9300

=======Node9 Config========
cluster.name: elasticsearch
node.name: NODE9
node.data: true
network.host: "10.249.167.9"
network.publish_host: "10.249.167.9"
network.bind: "10.249.167.9"
discovery.zen.ping.multicast.enabled: false

discovery.zen.ping.unicast.hosts: ["10.249.167.9", "10.249.167.10", "10.249.167.8"]

transport.tcp.port: 9300

I can query both ES node individually, but they dont form cluster

Node 8 Get : http://10.249.167.8:9200/_cat/nodes?h=ip,port,heapPercent,name
10.249.167.8 9300 2 Cecilia Reyes

Node 9 Get : http://10.249.167.9:9200/_cat/nodes?h=ip,port,heapPercent,name
10.249.167.9 9300 9 Victorius

Following are the startup logs, any help would be appreciated a ton, I am stuck on this for a while now:(
[2016-02-13 01:08:06,395][WARN ][bootstrap ] unable to install syscall filter: syscall filtering not supported for OS: 'Windows Server 2012 R2'
[2016-02-13 01:08:06,645][INFO ][node ] [NODE8] version[2.1.1], pid[7628], build[40e2c53/2015-12-15T13:05:55Z]
[2016-02-13 01:08:06,645][INFO ][node ] [NODE8] initializing ...
[2016-02-13 01:08:07,020][INFO ][plugins ] [NODE8] loaded [cloud-azure], sites []
[2016-02-13 01:08:07,051][INFO ][env ] [NODE8] using [1] data paths, mounts [[(C:)]], net usable_space [94.6gb], net total_space [126.6gb], spins? [unknown], types [NTFS]
[2016-02-13 01:08:09,170][INFO ][node ] [NODE8] initialized
[2016-02-13 01:08:09,170][INFO ][node ] [NODE8] starting ...
[2016-02-13 01:08:09,357][INFO ][transport ] [NODE8] publish_address {10.249.167.8:9300}, bound_addresses {10.249.167.8:9300}
[2016-02-13 01:08:09,373][INFO ][discovery ] [NODE8] elasticsearch/i42Qv-qNSJaSoLRCt2e5tg
[2016-02-13 01:08:13,936][INFO ][cluster.service ] [NODE8] new_master {NODE8}{i42Qv-qNSJaSoLRCt2e5tg}{10.249.167.8}{10.249.167.8:9300}, reason: zen-disco-join(elected_as_master, [0] joins received)
[2016-02-13 01:08:13,983][INFO ][http ] [NODE8] publish_address {10.249.167.8:9200}, bound_addresses {10.249.167.8:9200}
[2016-02-13 01:08:13,983][INFO ][node ] [NODE8] started
[2016-02-13 01:08:16,715][INFO ][gateway ] [NODE8] recovered [1] indices into cluster_state

Node 9 Log========================================================================

[2016-02-13 01:08:44,988][WARN ][bootstrap ] unable to install syscall filter: syscall filtering not supported for OS: 'Windows Server 2012 R2'
[2016-02-13 01:08:45,237][INFO ][node ] [NODE9] version[2.1.1], pid[6468], build[40e2c53/2015-12-15T13:05:55Z]
[2016-02-13 01:08:45,237][INFO ][node ] [NODE9] initializing ...
[2016-02-13 01:08:45,601][INFO ][plugins ] [NODE9] loaded [cloud-azure], sites []
[2016-02-13 01:08:45,625][INFO ][env ] [NODE9] using [1] data paths, mounts [[(C:)]], net usable_space [113.6gb], net total_space [126.6gb], spins? [unknown], types [NTFS]
[2016-02-13 01:08:47,554][INFO ][node ] [NODE9] initialized
[2016-02-13 01:08:47,554][INFO ][node ] [NODE9] starting ...
[2016-02-13 01:08:47,753][INFO ][transport ] [NODE9] publish_address {10.249.167.9:9300}, bound_addresses {10.249.167.9:9300}
[2016-02-13 01:08:47,763][INFO ][discovery ] [NODE9] elasticsearch/ys7WjfT3QR2DqwLFr-m6Ew
[2016-02-13 01:08:52,292][INFO ][cluster.service ] [NODE9] new_master {NODE9}{ys7WjfT3QR2DqwLFr-m6Ew}{10.249.167.9}{10.249.167.9:9300}, reason: zen-disco-join(elected_as_master, [0] joins received)
[2016-02-13 01:08:52,342][INFO ][http ] [NODE9] publish_address {10.249.167.9:9200}, bound_addresses {10.249.167.9:9200}
[2016-02-13 01:08:52,342][INFO ][node ] [NODE9] started
[2016-02-13 01:08:53,649][INFO ][gateway ] [NODE9] recovered [0] indices into cluster_state

Answer 1

问题看起来像裂脑问题。

这意味着节点被分开并且自己充当主节点。

从您的日志中可以清楚地看到您的节点已逐字地创建了两个不同的群集。

避免分裂大脑，可能的方法是 提及最小主节点

discovery.zen.minimum_master_nodes

可以使用以下计算

来计算

minimum master node = (N/2)+1

其中N是节点数

例如，如果群集中有3个节点，则可以设置为

discovery.zen.minimum_master_nodes: 2

弹性搜索节点未形成群集

1 个答案: