ElasticSearch系列：报错收集（一）

1 日志提示: UseConcMarkSweepGC was deprecated

• 具体日志如下:

OpenJDK 64-Bit Server VM warning: Option UseConcMarkSweepGC was deprecated in version 9.0 and will likely be removed in a future release.

• 解决办法: 修改jvm.options

-XX:+UseConcMarkSweepGC 改为 -XX:+UseG1GC

2 java.security.AccessControlException

• 集群日志如下:

[2021-10-21T14:03:34,533][ERROR][o.a.l.DicLibrary ] [aid-cluster-node-20] Init dic library error :java.security.AccessControlException: access denied ("java.io.FilePermission" "dic" "read"), path: dic

• 解决办法：修改 java.policy 文件

打开文件

vi /data/900x/elasticsearch/jdk/conf/security/java.policy

找到 gant ，在最后增加语句:

permission java.io.FilePermission "*", "read,write";

[2021-10-21T15:19:44,409][ERROR][o.a.l.DicLibrary ] [aid-cluster-node-01] Init dic library error :
org.ansj.exception.LibraryException: path :dic file:/data/900x/elasticsearch/dic not found or can not to read, path:
dic

Init dic library error :org.ansj.exception.LibraryException: path :dic file:/data/900x/elasticsearch/dic not found or
can not to read, path: dic

3 java 代码调用 ES 报错: Data Too Large

• 参考文档

• ElasticSearch 429 Too Many Requests circuit_breaking_exception

• ElasticSearch学习6_从[FIELDDATA]Data too large错误看FieldData配置

• 报错日志如下:

Elasticsearch exception [type=circuit_breaking_exception, reason=[parent] Data too large, data for [] would be [27568521612/25.6gb], which is larger than the limit of [27487790694/25.5gb], real usage: [27568521280/25.6gb], new bytes reserved: [332/332b], usages [request=0/0b, fielddata=0/0b, in_flight_requests=332/332b, accounting=16612791/15.8mb]]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-94mC8w4w-1665671536336)(00_images/00_image_datatoolarge.png)]
DataTooLarge

• 报错原因

ES 内存缓存数据 + 当前查询需要的缓存数据量 达到了 断路器。

• 解决办法

在 ES 配置文件 elasticsearch.yml 中加入如下参数 indices.fielddata.cache.size, 参数的大小与 indices.breaker.fielddata.limit 保持一致；

indices.fielddata.cache.size: 60%

4 调用 ES 报错: reactor I/O status STOPPED: 为真正解决

• 报错日志如下:

Request cannot be executed; I/O reactor status: STOPPED

reactor I/O status: STOPPED

• 报错原因

  • 多线程状态下，某个线程关闭了 RestHighLevelClient (此说法可能性不大)
  • 连接超时，索引数据太大，有相关聚合统计，导致 ES 处理时间过长
  • ES 自带 bug

• 解决办法

• 重启 ES 节点即可解决问题；