elasticsearch 源码简单分析之【tmp目录】

环境

服务器：centos6
elasticsearch： 6.0.0
工具：CRT
操作系统：win7

疑惑

[yutao@master01 elasticsearch-6.0.0]$ bin/elasticsearch-plugin install file:///home/yutao/download/x-pack-6.0.0.zip 

-> Downloading file:///home/yutao/download/x-pack-6.0.0.zip
Exception in thread "main" java.nio.file.AccessDeniedException: /tmp/2089611009639660923.zip
        at sun.nio.fs.UnixException.translateToIOException(UnixException.java:84)
        at sun.nio.fs.UnixException.rethrowAsIOException(UnixException.java:102)
        at sun.nio.fs.UnixException.rethrowAsIOException(UnixException.java:107)
        at sun.nio.fs.UnixFileSystemProvider.newByteChannel(UnixFileSystemProvider.java:214)
        at java.nio.file.Files.newByteChannel(Files.java:361)
        at java.nio.file.Files.createFile(Files.java:632)
        at java.nio.file.TempFileHelper.create(TempFileHelper.java:138)
        at java.nio.file.TempFileHelper.createTempFile(TempFileHelper.java:161)
        at java.nio.file.Files.createTempFile(Files.java:852)
        at org.elasticsearch.plugins.InstallPluginCommand.downloadZip(InstallPluginCommand.java:322)
        at org.elasticsearch.plugins.InstallPluginCommand.download(InstallPluginCommand.java:245)
        at org.elasticsearch.plugins.InstallPluginCommand.execute(InstallPluginCommand.java:213)
        at org.elasticsearch.plugins.InstallPluginCommand.execute(InstallPluginCommand.java:204)
        at org.elasticsearch.cli.EnvironmentAwareCommand.execute(EnvironmentAwareCommand.java:69)
        at org.elasticsearch.cli.Command.mainWithoutErrorHandling(Command.java:134)
        at org.elasticsearch.cli.MultiCommand.execute(MultiCommand.java:69)
        at org.elasticsearch.cli.Command.mainWithoutErrorHandling(Command.java:134)
        at org.elasticsearch.cli.Command.main(Command.java:90)
        at org.elasticsearch.plugins.PluginCli.main(PluginCli.java:47)

这个错误，在上一篇博客中也提到过，就是因为我使用yutao用户安装时，没有目录tmp的权限。而我想这个目录它为什么要用到？我能不能通过配置文件来进行配置？

查看源码

带着这样的疑惑，我开始查看源码：

// 从这句看出，其调用的是org.elasticsearch.plugins包下的PluginCli类的main方法
org.elasticsearch.plugins.PluginCli.main(PluginCli.java:47)

// 和上面同理 
org.elasticsearch.plugins.InstallPluginCommand.download(InstallPluginCommand.java:245)

其InstallPluginCommand类路径elasticsearch-6.0.0.zip\elasticsearch-6.0.0\lib\plugin-cli-6.0.0.jar\org\elasticsearch\plugins\InstallPluginCommand.class

    protected void execute(final Terminal terminal, final OptionSet options, final Environment env) throws Exception {
        final String pluginId = (String)this.arguments.value(options);
        final boolean isBatch = options.has((OptionSpec)this.batchOption) || System.console() == null;
        this.execute(terminal, pluginId, isBatch, env);
    }

    void execute(final Terminal terminal, final String pluginId, final boolean isBatch, final Environment env) throws Exception {
        if (pluginId == null) {
            throw new UserException(64, "plugin id is required");
        }
        final Path pluginZip = this.download(terminal, pluginId, env.tmpFile());
        final Path extractedZip = this.unzip(pluginZip, env.pluginsFile());
        this.install(terminal, isBatch, extractedZip, env);
    }

再看到其中的download方法，其有个参数env.tmpFile()；而这个方法是这个Environment类中的。

路径：elasticsearch-6.0.0.jar\org\elasticsearch\env\Environment.class

public Environment(final Settings settings, final Path configPath) {
       // 重点就是这句
        this.tmpFile = PathUtils.get(System.getProperty("java.io.tmpdir"), new String[0]);
        if (Environment.PATH_HOME_SETTING.exists(settings)) {
            final Path homeFile = PathUtils.get((String)Environment.PATH_HOME_SETTING.get(settings), new String[0]).normalize();
            if (configPath != null) {
                this.configFile = configPath.normalize();
            }
            else {
                this.configFile = homeFile.resolve("config");
            }
            this.pluginsFile = homeFile.resolve("plugins");
            final List<String> dataPaths = (List<String>)Environment.PATH_DATA_SETTING.get(settings);
            final ClusterName clusterName = (ClusterName)ClusterName.CLUSTER_NAME_SETTING.get(settings);
            if (!dataPaths.isEmpty()) {
                this.dataFiles = new Path[dataPaths.size()];
                this.dataWithClusterFiles = new Path[dataPaths.size()];
                for (int i = 0; i < dataPaths.size(); ++i) {
                    this.dataFiles[i] = PathUtils.get((String)dataPaths.get(i), new String[0]);
                    this.dataWithClusterFiles[i] = this.dataFiles[i].resolve(clusterName.value());
                }
            }
            else {
                this.dataFiles = new Path[] { homeFile.resolve("data") };
                this.dataWithClusterFiles = new Path[] { homeFile.resolve("data").resolve(clusterName.value()) };
            }
            if (Environment.PATH_SHARED_DATA_SETTING.exists(settings)) {
                this.sharedDataFile = PathUtils.get((String)Environment.PATH_SHARED_DATA_SETTING.get(settings), new String[0]).normalize();
            }
            else {
                this.sharedDataFile = null;
            }
            final List<String> repoPaths = (List<String>)Environment.PATH_REPO_SETTING.get(settings);
            if (!repoPaths.isEmpty()) {
                this.repoFiles = new Path[repoPaths.size()];
                for (int j = 0; j < repoPaths.size(); ++j) {
                    this.repoFiles[j] = PathUtils.get((String)repoPaths.get(j), new String[0]);
                }
            }
            else {
                this.repoFiles = new Path[0];
            }
            if (Environment.PATH_LOGS_SETTING.exists(settings)) {
                this.logsFile = PathUtils.get((String)Environment.PATH_LOGS_SETTING.get(settings), new String[0]).normalize();
            }
            else {
                this.logsFile = homeFile.resolve("logs");
            }
            if (Environment.PIDFILE_SETTING.exists(settings)) {
                this.pidFile = PathUtils.get((String)Environment.PIDFILE_SETTING.get(settings), new String[0]).normalize();
            }
            else {
                this.pidFile = null;
            }
            this.binFile = homeFile.resolve("bin");
            this.libFile = homeFile.resolve("lib");
            this.modulesFile = homeFile.resolve("modules");
            final Settings.Builder finalSettings = Settings.builder().put(settings);
            finalSettings.put(new Object[] { Environment.PATH_HOME_SETTING.getKey(), homeFile });
            if (Environment.PATH_DATA_SETTING.exists(settings)) {
                finalSettings.putArray(Environment.PATH_DATA_SETTING.getKey(), (List)dataPaths);
            }
            finalSettings.put(new Object[] { Environment.PATH_LOGS_SETTING.getKey(), this.logsFile });
            this.settings = finalSettings.build();
            return;
        }
        throw new IllegalStateException(Environment.PATH_HOME_SETTING.getKey() + " is not configured");
    }

我们重点看看这句：

this.tmpFile = PathUtils.get(System.getProperty("java.io.tmpdir"), new String[0]);

很明显，通过jdk提供的方法，获取到了，当前操作系统的临时文件tmp。

看到这里也就说明，不能通过配置文件的方式来修改，毕竟源码写死啦！

不过我又对上面代码中的第二个参数new String[0],产生了兴趣，虽然弄到了凌晨1点，但是还是坚持多看了几眼。

为什么要传一个字符串数组呢？还是一个大小为0的！

从源码角度来看，这个参数是选参，可传可不传；

PathUtils类的源码：

public static Path get(final String first, final String... more) {
        return PathUtils.DEFAULT.getPath(first, more);
    }

而其中getPath又是jdk里的方法：

// 类 FileSystem 
public abstract Path getPath(String first, String... more);

在往下就是java底层如何去获取文件的路径啦！我就不分析啦！

那它需要这目录干什么呢？
从源码中看出，其下载的x-pack.jar，视乎会临时存放在这个目录，进行压缩成zip文件。