1、说明
集群本来有三个节点，但是异常情况导致两个节点安装es的磁盘丢失了，之后恢复了磁盘，然后恢复集群，恢复集群我是把好的es的整个目录拷贝到刚恢复的节点上，然后修改配置文件。
好的节点的配置文件如下所示：
$ egrep -v "^#|^$" elasticsearch.yml
cluster.name: elasticsearch
node.name: "node 14.69"
bootstrap.mlockall: true
network.host: 192.168.14.69
discovery.zen.minimum_master_nodes: 1
discovery.zen.ping.timeout: 60s
discovery.zen.ping.multicast.enabled: false
discovery.zen.ping.unicast.hosts: ["192.168.14.40","192.168.14.177","192.168.14.69"]
拷贝后修改刚恢复的两个节点的配置文件，但是由于疏忽，忘记修改node.name这个参数，所以在集群起来后所有节点的node.name都是node 14.69，之后又新的数据进来后，就出现了 unassigned 的分片，在head插件的页面上就会在最上面出现一行 unassigned 的分片，集群的状态也变为red。
上面的情况是怎么出现的？下面分析下：
新建索引，默认是5个分片，1个副本，副本分片的主要目的就是为了故障转移，如果持有主分片的节点挂掉了，一个副本分片就会晋升为主分片的角色。
副本分片和主分片是不能放到一个节点上面的，当副本分片没有办法分配到其他的节点上，所以出现所有副本分片都unassigned得情况。因为配置失误，所以集群被识别为只有一个节点。
2、解决办法：
查看节点的状态
$ curl -XGET http://192.168.14.69:9200/_cluster/health\?pretty
{
"cluster_name" : "elasticsearch",
"status" : "red",
"timed_out" : false,
"number_of_nodes" : 3,
"number_of_data_nodes" : 3,
"active_primary_shards" : 56,
"active_shards" : 112,
"relocating_shards" : 0,
"initializing_shards" : 0,
"unassigned_shards" : 52,
"delayed_unassigned_shards" : 0,
"number_of_pending_tasks" : 0,
"number_of_in_flight_fetch" : 0
}
可以看到集群状态时red，未赋值的分片数是52个。
首先设置副本数为0
$ curl -XPUT "http://192.168.14.69:9200/_settings" -d'
{
"number_of_replicas" : 0
}'
再次查看节点的状态
$ curl -XGET http://192.168.14.69:9200/_cluster/health\?pretty
{
"cluster_name" : "elasticsearch",
"status" : "red",
"timed_out" : false,
"number_of_nodes" : 3,
"number_of_data_nodes" : 3,
"active_primary_shards" : 56,
"active_shards" : 56,
"relocating_shards" : 0,
"initializing_shards" : 0,
"unassigned_shards" : 26,
"delayed_unassigned_shards" : 0,
"number_of_pending_tasks" : 0,
"number_of_in_flight_fetch" : 0
}
可以看到集群未赋值的分片数是26，少了一半，说明少的是正常节点上的副本分片。
查看所有的分片信息
$ curl http://192.168.14.40:9200/_cat/shards
voicereprot 2 p UNASSIGNED
voicereprot 0 p UNASSIGNED
voicereprot 3 p STARTED 15301 5.8mb 192.168.14.40 node 14.40
voicereprot 1 p STARTED 15461 5.8mb 192.168.14.177 node 14.177
voicereprot 5 p STARTED 14540 5.3mb 192.168.14.40 node 14.40
voicereprot 4 p STARTED 16375 6.2mb 192.168.14.69 node 14.69
spipe 2 p UNASSIGNED
spipe 0 p STARTED 0 144b 192.168.14.69 node 14.69
spipe 3 p STARTED 0 144b 192.168.14.177 node 14.177
...........此处省略
查看所有的节点信息
$ curl http://192.168.14.40:9200/_nodes/process?pretty
{
"cluster_name" : "elasticsearch",
"nodes" : {
"mc3rloswRgqUJ5VkL4nxBw" : {
"name" : "node 14.177",
"transport_address" : "inet[/192.168.14.177:9300]",
"host" : "SZB-L0019761",
"ip" : "192.168.14.177",
"version" : "1.7.3",
"build" : "05d4530",
"http_address" : "inet[/192.168.14.177:9200]",
"process" : {
"refresh_interval_in_millis" : 1000,
"id" : 21695,
"max_file_descriptors" : 65536,
"mlockall" : false
}
},
"bICASdrQSe2ddNLhHw0Vyw" : {
"name" : "node 14.40",
"transport_address" : "inet[/192.168.14.40:9300]",
"host" : "DEV-L0003234",
"ip" : "192.168.14.40",
"version" : "1.7.3",
"build" : "05d4530",
"http_address" : "inet[/192.168.14.40:9200]",
"process" : {
"refresh_interval_in_millis" : 1000,
"id" : 12999,
"max_file_descriptors" : 8192,
"mlockall" : false
}
},
"kKzGBiXXTICg6f0UrT9_BA" : {
"name" : "node 14.69",
"transport_address" : "inet[/192.168.14.69:9300]",
"host" : "DEV-L0000155",
"ip" : "192.168.14.69",
"version" : "1.7.3",
"build" : "05d4530",
"http_address" : "inet[/192.168.14.69:9200]",
"process" : {
"refresh_interval_in_millis" : 1000,
"id" : 29986,
"max_file_descriptors" : 65535,
"mlockall" : false
}
}
}
}
查看节点信息，其中第一行 mc3rloswRgqUJ5VkL4nxBw、bICASdrQSe2ddNLhHw0Vyw、kKzGBiXXTICg6f0UrT9_BA 分别是节点的唯一标识。
一个一个转移未赋值的分片
$ curl -XPOST '192.168.14.40:9200/_cluster/reroute' -d '{
"commands" : [ {
"allocate" : {
"index" : "voicereprot",
"shard" : 0,
"node" : "bICASdrQSe2ddNLhHw0Vyw",
"allow_primary" : true
}
}
]
}'
之后的只需要修改四个变量， 节点IP、index（索引名）、shard（分片标记）、node（节点唯一标识）
这里只转移到14.40这一个节点上
再次再次查看节点的状态，可以发现已经转移了一个分片
$ curl http://192.168.14.40:9200/_cluster/health\?pretty
{
"cluster_name" : "elasticsearch",
"status" : "red",
"timed_out" : false,
"number_of_nodes" : 3,
"number_of_data_nodes" : 3,
"active_primary_shards" : 57,
"active_shards" : 57,
"relocating_shards" : 0,
"initializing_shards" : 0,
"unassigned_shards" : 25,
"delayed_unassigned_shards" : 0,
"number_of_pending_tasks" : 0,
"number_of_in_flight_fetch" : 0
}
陆续转移其他分片
$ curl -XPOST '192.168.14.40:9200/_cluster/reroute' -d '{
"commands" : [ {
"allocate" : {
"index" : "voicereprot",
"shard" : 2,
"node" : "bICASdrQSe2ddNLhHw0Vyw",
"allow_primary" : true
}
}
]
}'
...............此处省略
最后执行完毕可以发现所有的未赋值分片都已经转移完毕
$ curl http://192.168.14.40:9200/_cat/shards
voicereprot 4 p STARTED 16375 6.2mb 192.168.14.69 node 14.69
voicereprot 0 p STARTED 0 144b 192.168.14.177 node 14.177
voicereprot 3 p STARTED 15301 5.8mb 192.168.14.40 node 14.40
voicereprot 1 p STARTED 15461 5.8mb 192.168.14.177 node 14.177
voicereprot 5 p STARTED 14540 5.3mb 192.168.14.40 node 14.40
......此处省略
$ curl http://192.168.14.40:9200/_cluster/health\?pretty
{
"cluster_name" : "elasticsearch",
"status" : "green",
"timed_out" : false,
"number_of_nodes" : 3,
"number_of_data_nodes" : 3,
"active_primary_shards" : 82,
"active_shards" : 82,
"relocating_shards" : 0,
"initializing_shards" : 0,
"unassigned_shards" : 0,
"delayed_unassigned_shards" : 0,
"number_of_pending_tasks" : 0,
"number_of_in_flight_fetch" : 0
}
最后恢复副本为1
$ curl -XPUT "http://192.168.14.69:9200/_settings" -d'
{
"number_of_replicas" : 1
}'
可以把转移分片的过程写入到一个脚本里边，把索引名和分片标记定义为变量，把变量的值过滤到一个文档中，脚本通过循环执行对对应的索引和分片标记做处理。
参考：
https://blog.csdn.net/x4609883/article/details/79926267
https://blog.csdn.net/u013905744/article/details/81508029
https://cloud.tencent.com/info/527778d6d1b6588d0c27d9622a6fa7b6.html
https://blog.csdn.net/envinfo2012/article/details/82767674
转载于:https://blog.51cto.com/xiaoxiaozhou/2344652
扫一扫
3394
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
