index out of range_es 7.2 生产集群 index 无数据写入故障定位

故障分析

     zabbix 告警 es index 写入异常 ，登录 kibana 查看无新数据写入。所有的 index 都没有生成的新的数据。不同的 index 是分布在不同位置的logstash 主机向 kafka/zookeeper 集群拉取对应的 topic。zabbix 对 es 集群进行监控没有发现告警，例行检查 es 集群状态无异常。怀疑是kafka/zookeeper  集群出现问题，测试消费数据无异常，最后检查 logstash 发现异常日志 定位故障。

Zabbix 事件告警

es 集群 index 数据写入异常

检查 es 集群状态，无异常

GET _cluster/health?pretty

{  "cluster_name" : "elk-cluster",  "status" : "green",  "timed_out" : false,  "number_of_nodes" : 4,  "number_of_data_nodes" : 4,  "active_primary_shards" : 2035,  "active_shards" : 4070,  "relocating_shards" : 0,  "initializing_shards" : 0,  "unassigned_shards" : 0,  "delayed_unassigned_shards" : 0,  "number_of_pending_tasks" : 0,  "number_of_in_flight_fetch" : 0,  "task_max_waiting_in_queue_millis" : 0,  "active_shards_percent_as_number" : 100.0}

kafka/zookeeper

查看 topic

./bin/kafka-topics.sh --list --zookeeper 192.168.99.232:2181,192.168.99.233:2181,192.168.99.221:2181 

消费数据测试

./bin/kafka-console-consumer.sh --bootstrap-server 192.168.99.233:9092,192.168.99.232:9092,192.168.99.221:9092 --topic networklogs --from-beginning

查看 logstsh 日志报错提示如下

Could not index event to Elasticsearch. {:status=>400, :action=>["index", {:_id=>nil, :_index=>"apps_wms-2020.07.16", :routing=>nil, :_type=>"_doc"}, #<:event:0x682b3a66>], :response=>{"index"=>{"_index"=>"apps_wms-2020.07.16", "_type"=>"_doc", "_id"=>nil, "status"=>400, "error"=>{"type"=>"illegal_argument_exception", "reason"=>"Validation Failed: 1: this action would add [2] total shards, but this cluster currently has [4000]/[4000] maximum shards open;"}}}}

es  集群默认有分片限制最大 4000 ，导致 logstash 写入 es 的 index 数据异常

调整 es 集群默认分片限制 index写入恢复正常

PUT /_cluster/settings{  "transient": {    "cluster": {      "max_shards_per_node":10000    }  }}