Elasticsearch Rest 客户端仍然给出 IOException：打开的文件太多

Question

这是对上一篇文章中提供给我的解决方案的跟进：

How to Properly Close Raw RestClient When Using Elastic Search 5.5.0 for Optimal Performance?

同样的错误信息又回来了！

2017-09-29 18:50:22.497 ERROR 11099 --- [8080-Acceptor-0] org.apache.tomcat.util.net.NioEndpoint   : Socket accept failed

java.io.IOException: Too many open files
    at sun.nio.ch.ServerSocketChannelImpl.accept0(Native Method) ~[na:1.8.0_141]
    at sun.nio.ch.ServerSocketChannelImpl.accept(ServerSocketChannelImpl.java:422) ~[na:1.8.0_141]
    at sun.nio.ch.ServerSocketChannelImpl.accept(ServerSocketChannelImpl.java:250) ~[na:1.8.0_141]
    at org.apache.tomcat.util.net.NioEndpoint$Acceptor.run(NioEndpoint.java:453) ~[tomcat-embed-core-8.5.15.jar!/:8.5.15]
    at java.lang.Thread.run(Thread.java:748) [na:1.8.0_141]

2017-09-29 18:50:23.885  INFO 11099 --- [Thread-3] ationConfigEmbeddedWebApplicationContext : Closing org.springframework.boot.context.embedded.AnnotationConfigEmbeddedWebApplicationContext@5387f9e0: startup date [Wed Sep 27 03:14:35 UTC 2017]; root of context hierarchy
2017-09-29 18:50:23.890  INFO 11099 --- [Thread-3] o.s.c.support.DefaultLifecycleProcessor  : Stopping beans in phase 2147483647
2017-09-29 18:50:23.891  WARN 11099 --- [Thread-3] o.s.c.support.DefaultLifecycleProcessor  : Failed to stop bean 'documentationPluginsBootstrapper'

    ... 7 common frames omitted

2017-09-29 18:50:53.891  WARN 11099 --- [Thread-3] o.s.c.support.DefaultLifecycleProcessor  : Failed to shut down 1 bean with phase value 2147483647 within timeout of 30000: [documentationPluginsBootstrapper]
2017-09-29 18:50:53.891  INFO 11099 --- [Thread-3] o.s.j.e.a.AnnotationMBeanExporter        : Unregistering JMX-exposed beans on shutdown
2017-09-29 18:50:53.894  INFO 11099 --- [Thread-3] com.app.controller.SearchController  : Closing the ES REST client

我尝试使用上一篇文章中的解决方案。

ElasticsearchConfig:

@Configuration
public class ElasticsearchConfig {

@Value("${elasticsearch.host}")
private String host;

@Value("${elasticsearch.port}")
private int port;

@Bean
public RestClient restClient() {
    return RestClient.builder(new HttpHost(host, port))
    .setRequestConfigCallback(new RestClientBuilder.RequestConfigCallback() {
        @Override
        public RequestConfig.Builder customizeRequestConfig(RequestConfig.Builder requestConfigBuilder) {
            return requestConfigBuilder.setConnectTimeout(5000).setSocketTimeout(60000);
        }
    }).setMaxRetryTimeoutMillis(60000).build();
}

搜索控制器：

@RestController
@RequestMapping("/api/v1")
public class SearchController {

    @Autowired
    private RestClient restClient;

    @RequestMapping(value = "/search", method = RequestMethod.GET, produces="application/json" )
    public ResponseEntity<Object> getSearchQueryResults(@RequestParam(value = "criteria") String criteria) throws IOException {

        // Setup HTTP Headers
        HttpHeaders headers = new HttpHeaders();
        headers.add("Content-Type", "application/json");

        // Setup query and send and return ResponseEntity...

        Response response = this.restClient.performRequest(...);
    }

    @PreDestroy
    public void cleanup() {
        try {
            logger.info("Closing the ES REST client");
            this.restClient.close();
        } 
        catch (IOException ioe) {
            logger.error("Problem occurred when closing the ES REST client", ioe);
        }
    }
}    

pom.xml：

<parent>
    <groupId>org.springframework.boot</groupId>
    <artifactId>spring-boot-starter-parent</artifactId>
    <version>1.5.4.RELEASE</version>
</parent>

<dependencies>
    <!-- Spring -->
    <dependency>
        <groupId>org.springframework.boot</groupId>
        <artifactId>spring-boot-starter-web</artifactId>
    </dependency>

    <dependency>
        <groupId>org.springframework.boot</groupId>
        <artifactId>spring-boot-starter-test</artifactId>
        <scope>test</scope>
    </dependency>

    <!-- Elasticsearch -->
    <dependency>
        <groupId>org.elasticsearch</groupId>
        <artifactId>elasticsearch</artifactId>
        <version>5.5.0</version>
    </dependency>

    <dependency>
        <groupId>org.elasticsearch.client</groupId>
        <artifactId>transport</artifactId>
        <version>5.5.0</version>
    </dependency>

    <!-- Apache Commons -->
    <dependency>
        <groupId>org.apache.commons</groupId>
        <artifactId>commons-lang3</artifactId>
        <version>3.6</version>
    </dependency>

    <!-- Log4j -->
    <dependency>
        <groupId>log4j</groupId>
        <artifactId>log4j</artifactId>
        <version>1.2.17</version>
    </dependency>
</dependencies>

这让我觉得 RestClient 从来没有明确地关闭连接，首先......

这令人惊讶，因为我的基于 Elasticsearch Spring Boot 的微服务在两个不同的 AWS EC-2 服务器上进行了负载平衡。

日志文件报告的异常发生了 2000 次，最后 preDestroy() 才关闭客户端。查看 StackTrace 末尾记录的 @PreDestroy() 清理方法中的信息。

是否需要在 SearchController 中显式放置 finally 子句并显式关闭 RestClient 连接？

此 IOException 不会再次发生非常重要，因为此搜索微服务依赖于许多不同的移动客户端（iOS 和 Android）。

需要它具有容错性和可扩展性......或者，至少不要破坏。

这是在日志文件底部的唯一原因：

2017-09-29 18:50:53.894  INFO 11099 --- [Thread-3] com.app.controller.SearchController : Closing the ES REST client

是因为我这样做了：

kill -3 jvm_pid

我是否应该保留 @PreDestory cleanup() 方法但更改 SearchController.getSearchResults() 方法的内容以反映如下内容：

@RequestMapping(value = "/search", method = RequestMethod.GET, produces="application/json" )
public ResponseEntity<Object> getSearchQueryResults(@RequestParam(value = "criteria") String criteria) throws IOException {
    // Setup HTTP Headers
    HttpHeaders headers = new HttpHeaders();
    headers.add("Content-Type", "application/json");

    // Setup query and send and return ResponseEntity...
    Response response = null;

    try {
        // Submit Query and Obtain Response
        response = this.restClient.performRequest("POST", endPoint, Collections.singletonMap("pretty", "true"), entity);
    } 
    catch(IOException ioe) {
        logger.error("Exception when performing POST request " + ioe);
    }
    finally {
        this.restClient.close();
    }
    // return response as EsResponse();
}

这样，RestClient 连接总是关闭...

如果有人可以帮助我，将不胜感激。

Answer 1

在我看来，您做错的事情很少，但我会直接解决。

我不会写完整的解决方案（事实上，我没有执行或测试任何东西），但重要的是理解它。此外，最好将所有与数据访问相关的内容移至另一层。 无论如何，这只是一个例子，所以设计并不完美。

第 1 步：导入正确的库。

与您的示例几乎相同。 我更新了示例以使用版本 5.6.2 中推荐的最后一个客户端库

<project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
  xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
  <modelVersion>4.0.0</modelVersion>

  <groupId>com.acervera</groupId>
  <artifactId>elastic-example</artifactId>
  <version>1.0-SNAPSHOT</version>
  <packaging>jar</packaging>

  <name>elastic-example</name>
  <url>http://maven.apache.org</url>

  <properties>
    <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
    <es.version>5.6.2</es.version>
  </properties>

  <parent>
    <groupId>org.springframework.boot</groupId>
    <artifactId>spring-boot-starter-parent</artifactId>
    <version>1.5.4.RELEASE</version>
  </parent>

  <dependencies>
    <!-- Spring -->
    <dependency>
      <groupId>org.springframework.boot</groupId>
      <artifactId>spring-boot-starter-web</artifactId>
    </dependency>

    <dependency>
      <groupId>org.springframework.boot</groupId>
      <artifactId>spring-boot-starter-test</artifactId>
      <scope>test</scope>
    </dependency>

    <!-- Elasticsearch -->
    <dependency>
      <groupId>org.elasticsearch</groupId>
      <artifactId>elasticsearch</artifactId>
      <version>${es.version}</version>
    </dependency>

<dependency>
  <groupId>org.elasticsearch.client</groupId>
  <artifactId>elasticsearch-rest-high-level-client</artifactId>
  <version>${es.version}</version>
</dependency>

    <!-- Log4j -->
    <dependency>
      <groupId>log4j</groupId>
      <artifactId>log4j</artifactId>
      <version>1.2.17</version>
    </dependency>
  </dependencies>
</project>

第 2 步：在 bean 工厂中创建和关闭客户端。

在 bean factory 中，创建和销毁它。您可以重复使用同一个客户端。

import org.apache.http.HttpHost;
import org.elasticsearch.client.RestClient;
import org.elasticsearch.client.RestHighLevelClient;
import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.Configuration;

import javax.annotation.PostConstruct;
import javax.annotation.PreDestroy;
import java.io.IOException;

@Configuration
public class ElasticsearchConfig {


    // Here all init stuff with @Value(....)


    RestClient lowLevelRestClient;
    RestHighLevelClient client;

    @PostConstruct
    public void init() {
        lowLevelRestClient = RestClient.builder(new HttpHost("host", 9200, "http")).build();
        client = new RestHighLevelClient(lowLevelRestClient);
    }

    @PreDestroy
    public void destroy() throws IOException {
        lowLevelRestClient.close();
    }

    @Bean
    public RestHighLevelClient getClient() {
        return client;
    }

}

第 3 步：使用 Java 传输客户端执行查询。

使用 Java 传输客户端执行查询。

import org.elasticsearch.action.search.SearchRequest;
import org.elasticsearch.action.search.SearchResponse;
import org.elasticsearch.client.RestHighLevelClient;
import org.elasticsearch.index.query.QueryBuilders;
import org.elasticsearch.search.SearchHit;
import org.elasticsearch.search.builder.SearchSourceBuilder;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.http.ResponseEntity;
import org.springframework.web.bind.annotation.RequestMapping;
import org.springframework.web.bind.annotation.RequestMethod;
import org.springframework.web.bind.annotation.RequestParam;
import org.springframework.web.bind.annotation.RestController;

import java.io.IOException;

@RestController
@RequestMapping("/api/v1")
public class SearchController {

    @Autowired
    private RestHighLevelClient client;

    @RequestMapping(value = "/search", method = RequestMethod.GET, produces="application/json" )
    public ResponseEntity<Tweet> getSearchQueryResults(@RequestParam(value = "criteria") String criteria) throws IOException {

        // This is only one example. Of course, this logic make non sense and you are going to put it in a DAO
        // layer with more logical stuff
        SearchRequest searchRequest = new SearchRequest();
        SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
        searchSourceBuilder.query(QueryBuilders.matchAllQuery());
        SearchResponse searchResponse = client.search(searchRequest);

        if(searchResponse.getHits().totalHits > 0) {
            SearchHit searchHit = searchResponse.getHits().iterator().next();

            // Deserialize to Java. The best option is to use response.getSource() and Jackson
            // This is other option.
            Tweet tweet = new Tweet();
            tweet.setId(searchHit.getField("id").getValue().toString());
            tweet.setTittle(searchHit.getField("tittle").getValue().toString());
            return ResponseEntity.ok(tweet);
        } else {
            return ResponseEntity.notFound().build();
        }

    }

}

另外，使用 bean 来构建响应。

public class Tweet {

    private String id;

    private String tittle;

    public String getTittle() {
        return tittle;
    }

    public void setTittle(String tittle) {
        this.tittle = tittle;
    }

    public String getId() {
        return id;
    }

    public void setId(String id) {
        this.id = id;
    }

    // Here rest of bean stuff (equal, hash, etc) or Lombok
}

步骤：4

享受 Elasticsearch！

备注：Java REST Client [5.6] » Java High Level REST Client

PS。有必要重构这个例子。只是为了了解方式。

Answer 2

您确定不会在每个 HTTP 请求上都启动与 elasticsearch 服务器的新（线程池）连接吗？即，排队

        Response response = this.restClient.performRequest(...);

在单个 HTTP 请求后仔细检查 elasticsearch 服务器上的日志。您应该尝试在没有 @Autowired 注释的情况下实现单例模式，看看问题是否仍然存在。

Answer 3

从您的堆栈跟踪来看，由于too many open files 错误，嵌入式 tomcat（您的应用程序容器）似乎不再能够接受新连接。从您的代码来看，elasticsearch rest client 似乎没有问题。

由于您在为您的搜索请求提供服务时重复使用 RestClient 的单个实例，因此与 ES 集群的打开连接可能不超过 30 个 (org.elasticsearch.client.RestClientBuilder.DEFAULT_MAX_CONN_TOTAL)。因此，RestClient 不太可能导致问题。

其他潜在的根本原因可能是您的服务使用者与您的 (tomcat) 服务器保持连接打开的时间更长，或者他们没有正确关闭连接。

我是否需要明确地将 finally 子句放在 SearchController 并显式关闭 RestClient 连接？

没有。你不应该。在关闭您的服务时应关闭休息客户端（在 @PreDestroy 方法中，因为您已经正确执行了）。