一、集群环境

java-1.8.0-openjdk-1.8.0.181-7.b13.el7
hadoop-2.5.2
spark-2.3.3
hbase-1.3.1
hbase-2.1.0
zookeeper-3.5.5-bin
janusgraph-0.2.0-hadoop2-gremlin
mysql-5.7.27
hive-2.1.1

这两天我配置了mysql和hive，本文记录遇到的问题。

二、MySQL

使用了arm架构下的mysql.tar.gz离线安装。

参考文章：ARM架构部署mysql-5.7.27
文章内容：

cd /usr/local将部署包：mysql-5.7.27-aarch64.tar.gz 上传到 /usr/local 下tar xvf mysql-5.7.27-aarch64.tar.gzmv /usr/local/mysql-5.7.27-aarch64 /usr/local/mysqlmkdir -p /usr/local/mysql/logsln -sf /usr/local/mysql/my.cnf /etc/my.cnfcp -rf /usr/local/mysql/extra/lib* /usr/lib64/mv /usr/lib64/libstdc++.so.6 /usr/lib64/libstdc++.so.6.oldln -s /usr/lib64/libstdc++.so.6.0.24 /usr/lib64/libstdc++.so.6groupadd mysqluseradd -g mysql mysqlchown -R mysql:mysql /usr/local/mysqlcp -rf /usr/local/mysql/support-files/mysql.server /etc/init.d/mysqldchmod +x /etc/init.d/mysqldsystemctl enable mysqldvim /etc/profileexport MYSQL_HOME=/usr/local/mysqlexport PATH=$PATH:$MYSQL_HOME/binsource /etc/profilemysqld --initialize-insecure --user=mysql --basedir=/usr/local/mysql --datadir=/usr/local/mysql/datasystemctl start mysqldsystemctl status mysqld移动文件 mv /usr/local/mysql-5.7.27-aarch64 /usr/local/mysql创建logs目录 mkdir -p /usr/local/mysql/logsln -sf a b 建立软连接，b指向a:ln -sf /usr/local/mysql/my.cnf /etc/my.cnfcp是linux里的拷贝命令-r 是用于目录拷贝时的递归操作-f 是强制覆盖:cp -rf /usr/local/mysql/extra/lib* /usr/lib64/创建mysql组:ln -s /usr/lib64/libstdc++.so.6.0.24 /usr/lib64/libstdc++.so.6创建mysql用户添加到mysql组:groupadd mysql && useradd -g mysql mysql将/usr/loca/mysql目录包含所有的子目录和文件，所有者改变为root,所属组改变为mysql:chown -R mysql:mysql /usr/local/mysql设置开机启动:cp -rf /usr/local/mysql/support-files/mysql.server /etc/init.d/mysqldchmod +x /etc/init.d/mysqldsystemctl enable mysqld添加环境变量:vim /etc/profileexport MYSQL_HOME=/usr/local/mysqlexport PATH=PATH:PATH:PATH:MYSQL_HOME/binsource /etc/profile初始化mysql:mysqld --initialize-insecure --user=mysql --basedir=/usr/local/mysql --datadir=/usr/local/mysql/data开启mysql:systemctl start mysqld查看状态:systemctl status mysqld

2.1 MySQL初始化失败

重点检查my.cnf文件，所有目录的创建，权限，初始化命令参数。
关于my.cnf的详细介绍参考：MySQL 配置文件 my.cnf / my.ini 逐行解析
原文内容：

MySQL 配置文件详解
文件位置: Windows、Linux、Mac 有细微区别，Windows 配置文件是 .ini，Mac/linux 是 .cnf[Windows]
MySQL\MySQL Server 5.7\my.ini[Linux / Mac]
/etc/my.cnf
/etc/mysql/my.cnf 
当然我们也可以使用命令来查看 MySQL 默认配置文件位置mysql --help|grep 'cnf'[client]
客户端设置。当前为客户端默认参数port = 3306
默认连接端口为 3306socket = /tmp/mysql.sock
本地连接的 socket 套接字default_character_set = utf8
设置字符集，通常使用 uft8 即可[mysqld_safe]
mysqld_safe 是服务器端工具，用于启动 mysqld，也是 mysqld 的守护进程。当 mysql 被 kill 时，mysqld_safe 负责重启启动它。open_files_limit = 8192
此为 MySQL 打开的文件描述符限制，它是 MySQL 中的一个全局变量且不可动态修改。它控制着 mysqld 进程能使用的最大文件描述符数量。默认最小值为 1024需要注意的是这个变量的值并不一定是你在这里设置的值，mysqld 会在系统允许的情况下尽量取最大值。当 open_files_limit 没有被配置时，比较 max_connections*5 和 ulimit -n 的值，取最大值当 open_file_limit 被配置时，比较 open_files_limit 和 max_connections*5 的值，取最大值user = mysql
用户名log-error  = error.log
错误 log 记录文件[mysqld]
服务端基本配置port = 3306
mysqld 服务端监听端口socket = /tmp/mysql.sock
MySQL 客户端程序和服务器之间的本地通讯指定一个套接字文件max_allowed_packet  = 16M
允许最大接收数据包的大小，防止服务器发送过大的数据包。当发出长查询或 mysqld 返回较大结果时，mysqld 才会分配内存，所以增大这个值风险不大，默认 16M，也可以根据需求改大，但太大会有溢出风险。取较小值是一种安全措施，避免偶然出现但大数据包导致内存溢出。default_storage_engine = InnoDB
创建数据表时，默认使用的存储引擎。这个变量还可以通过 –default-table-type 进行设置max_connections  = 512
最大连接数，当前服务器允许多少并发连接。默认为 100，一般设置为小于 1000 即可。太高会导致内存占用过多，MySQL 服务器会卡死。作为参考，小型站设置 100 - 300max_user_connections = 50
用户最大的连接数，默认值为 50 一般使用默认即可。thread_cache_size = 64
线程缓存，用于缓存空闲的线程。这个数表示可重新使用保存在缓存中的线程数，当对方断开连接时，如果缓存还有空间，那么客户端的线程就会被放到缓存中，以便提高系统性能。我们可根据物理内存来对这个值进行设置，对应规则 1G 为 8；2G 为 16；3G 为 32；4G 为 64 等。Query Cache
query_cache_type = 1
设置为 0 时，则禁用查询缓存（尽管仍分配query_cache_size个字节的缓冲区）。
设置为 1 时 ，除非指定SQL_NO_CACHE，否则所有SELECT查询都将被缓存。
设置为 2 时，则仅缓存带有SQL CACHE子句的查询。
请注意，如果在禁用查询缓存的情况下启动服务器，则无法在运行时启用服务器。query_cache_size = 64M
缓存select语句和结果集大小的参数。查询缓存会存储一个select查询的文本与被传送到客户端的相应结果。如果之后接收到一个相同的查询，服务器会从查询缓存中检索结果，而不是再次分析和执行这个同样的查询。如果你的环境中写操作很少，读操作频繁，那么打开query_cache_type=1，会对性能有明显提升。如果写操作频繁，则应该关闭它（query_cache_type=0）。Session variables  sort_buffer_size = 2M
MySQL 执行排序时，使用的缓存大小。增大这个缓存，提高 group by，order by 的执行速度。tmp_table_size = 32M
HEAP 临时数据表的最大长度，超过这个长度的临时数据表 MySQL 可根据需求自动将基于内存的 HEAP 临时表改为基于硬盘的 MyISAM 表。我们可通过调整 tmp_table_size 的参数达到提高连接查询速度的效果。read_buffer_size  = 128k
MySQL 读入缓存的大小。如果对表对顺序请求比较频繁对话，可通过增加该变量值以提高性能。read_rnd_buffer_size = 256k
用于表的随机读取，读取时每个线程分配的缓存区大小。默认为 256k ，一般在 128 - 256k之间。在做 order by 排序操作时，会用到 read_rnd_buffer_size 空间来暂做缓冲空间。join_buffer_size  = 128k
程序中经常会出现一些两表或多表 Join （联表查询）的操作。为了减少参与 Join 连表的读取次数以提高性能，需要用到 Join Buffer 来协助 Join 完成操作。当 Join Buffer 太小时，MySQL 不会将它写入磁盘文件。和 sort_buffer_size 一样，此参数的内存分配也是每个连接独享。table_definition_cache = 400
限制不使用文件描述符存储在缓存中的表定义的数量。table_open_cache   = 400
限制为所有线程在内存中打开的表数量。MySQL 错误日志设置
log_error = error.log log_warnings = 2
log_warnings 为0， 表示不记录告警信息。
log_warnings 为1， 表示告警信息写入错误日志。
log_warnings 大于1， 表示各类告警信息，例如有关网络故障的信息和重新连接信息写入错误日志。
慢查询记录
slow_query_log_file = slow.log slow_query_log  = 0 log_queries_not_using_indexes  = 1 long_query_time = 0.5 min_examined_row_limit = 100
slow_query_log ：全局开启慢查询功能。
slow_query_log_file ：指定慢查询日志存储文件的地址和文件名。
log_queries_not_using_indexes：无论是否超时，未被索引的记录也会记录下来。
long_query_time：慢查询阈值（秒），SQL 执行超过这个阈值将被记录在日志中。
min_examined_row_limit：慢查询仅记录扫描行数大于此参数的 SQL。

2.2 MySQL启动报错

报错内容：
Job for mysqld.service failed because the control process exited with error code. See “systemctl status mysqld.service” and “journalctl -xe” for details.
解决思路：
mysqld.pid目录权限问题，请把我们组群mysql:mysql给到权限，这个组群是我们安装mysql时创建的。
参考文章：
[1]:关于Job for mysqld.service failed because the control process exited with error code报错解决办法
[2]:启动mysql报错Job for mysqld.service failed because the control process exited with error code.

2.3 启动时报不能打开日志错

报错内容：
ERROR Could not open file ‘***/log/mysql/error.log‘ for error logging: Permission denied
**错误原因：**日志文件夹的权限问题,请重点用chmod检查权限是否够组群用户使用。
参考文档：
[1]:centos系统中MySQL无法启动的问题
[2]:Docker中mysql启动错误Could not open file ‘/var/log/mysqld.log‘ for error logging: Permission denied

2.4 mysql启动时pid报错

报错内容：
Starting MySQL... ERROR The server quit without updating PID file
参考文档：
启动mysql服务时一直提示ERROR The server quit without updating PID file
该文章分析了启动mysql的几个服务脚本源码，非常详细，介绍了用户权限对mysql初始化的影响以及pid文件在此的作用，包括mycnf的配置目录描述。再次重启时可以使用mysql.service或者mysqld_safe来启动mysql。

二、Hive

2.1 Hive修改core-site.xml文件后刷新权限

core-site.xml中的这两个配置名：
hadoop.proxyuser.root.hosts
hadoop.proxyuser.root.groups
要分发集群并且刷新权限，有的文档说yarn rmadmin也需要刷新超级用户组配置。
hdfs dfsadmin -refreshSuperUserGroupsConfiguration

2.2 Hive启动元数据时报错

根本原因来自于hive-site.xml文件中配置的路径信息。
hive.metastore.uris
hive.metastore.warehouse.dir
hive.exec.scratchdir
hdfs仓库的路径和元数据仓库拼写前是否带有mycluster，cluster，或者不写集群名，甚至是集群名和高可用的配置名拼写错。
还有的host名称没有配置正确的映射。
有较多的文档可以参考，具体需要结合集群是否有高可用配置。

2.3 Hive初始化MySQL报错

2.3.1 报错信息

执行初始化代码：schematool -dbType mysql -initSchema

Underlying cause: java.sql.SQLException : null,  message from server: "Host 'hadoop01' is not allowed to connect to this MySQL server"
SQL Error code: 1130

2.3.2 错误原因

MySQL数据权限问题。

2.3.3 参考文档

给权限然后flush刷新，看到root有%就可以了。
连接Mysql服务器提示：1130-Host XXX is not allowed to connect to this MySQL server的处理方法

2.4 mr shuffle不存在

2.4.1 报错信息：**org.apache.hadoop.yarn.exceptions.InvalidAuxServiceException: The auxService:mapreduce_shuffle does not exist

以下报错日志是我在我的个人集群中更改配置后复现的报错结果，其中运行了mapreduce官方案例的wordcount和pi：

[root@hadoop11 data]# hadoop jar /opt/installs/hadoop3.1.4/share/hadoop/mapreduce/hadoop-mapreduce-examples-3.1.4.jar wordcount /wc.txt /out3
2023-10-09 15:42:20,213 INFO mapreduce.JobResourceUploader: Disabling Erasure Coding for path: /tmp/hadoop-yarn/staging/root/.staging/job_1696837267350_0001
2023-10-09 15:42:20,536 INFO input.FileInputFormat: Total input files to process : 1
2023-10-09 15:42:20,684 INFO mapreduce.JobSubmitter: number of splits:1
2023-10-09 15:42:20,955 INFO mapreduce.JobSubmitter: Submitting tokens for job: job_1696837267350_0001
2023-10-09 15:42:20,959 INFO mapreduce.JobSubmitter: Executing with tokens: []
2023-10-09 15:42:21,212 INFO conf.Configuration: resource-types.xml not found
2023-10-09 15:42:21,213 INFO resource.ResourceUtils: Unable to find 'resource-types.xml'.
2023-10-09 15:42:21,620 INFO impl.YarnClientImpl: Submitted application application_1696837267350_0001
2023-10-09 15:42:21,743 INFO mapreduce.Job: The url to track the job: http://hadoop13:8088/proxy/application_1696837267350_0001/
2023-10-09 15:42:21,745 INFO mapreduce.Job: Running job: job_1696837267350_0001
2023-10-09 15:42:30,000 INFO mapreduce.Job: Job job_1696837267350_0001 running in uber mode : false
2023-10-09 15:42:30,002 INFO mapreduce.Job:  map 0% reduce 0%
2023-10-09 15:42:32,053 INFO mapreduce.Job: Task Id : attempt_1696837267350_0001_m_000000_0, Status : FAILED
Container launch failed for container_e50_1696837267350_0001_01_000002 : org.apache.hadoop.yarn.exceptions.InvalidAuxServiceException: The auxService:mapreduce_shuffle does not existat sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)at sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:62)at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)at java.lang.reflect.Constructor.newInstance(Constructor.java:423)at org.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.instantiateExceptionImpl(SerializedExceptionPBImpl.java:171)at org.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.instantiateException(SerializedExceptionPBImpl.java:182)at org.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.deSerialize(SerializedExceptionPBImpl.java:106)at org.apache.hadoop.mapreduce.v2.app.launcher.ContainerLauncherImpl$Container.launch(ContainerLauncherImpl.java:163)at org.apache.hadoop.mapreduce.v2.app.launcher.ContainerLauncherImpl$EventProcessor.run(ContainerLauncherImpl.java:394)at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)at java.lang.Thread.run(Thread.java:748)2023-10-09 15:42:33,087 INFO mapreduce.Job: Task Id : attempt_1696837267350_0001_m_000000_1, Status : FAILED
Container launch failed for container_e50_1696837267350_0001_01_000003 : org.apache.hadoop.yarn.exceptions.InvalidAuxServiceException: The auxService:mapreduce_shuffle does not existat sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)at sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:62)at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)at java.lang.reflect.Constructor.newInstance(Constructor.java:423)at org.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.instantiateExceptionImpl(SerializedExceptionPBImpl.java:171)at org.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.instantiateException(SerializedExceptionPBImpl.java:182)at org.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.deSerialize(SerializedExceptionPBImpl.java:106)at org.apache.hadoop.mapreduce.v2.app.launcher.ContainerLauncherImpl$Container.launch(ContainerLauncherImpl.java:163)at org.apache.hadoop.mapreduce.v2.app.launcher.ContainerLauncherImpl$EventProcessor.run(ContainerLauncherImpl.java:394)at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)at java.lang.Thread.run(Thread.java:748)2023-10-09 15:42:35,113 INFO mapreduce.Job: Task Id : attempt_1696837267350_0001_m_000000_2, Status : FAILED
Container launch failed for container_e50_1696837267350_0001_01_000004 : org.apache.hadoop.yarn.exceptions.InvalidAuxServiceException: The auxService:mapreduce_shuffle does not existat sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)at sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:62)at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)at java.lang.reflect.Constructor.newInstance(Constructor.java:423)at org.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.instantiateExceptionImpl(SerializedExceptionPBImpl.java:171)at org.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.instantiateException(SerializedExceptionPBImpl.java:182)at org.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.deSerialize(SerializedExceptionPBImpl.java:106)at org.apache.hadoop.mapreduce.v2.app.launcher.ContainerLauncherImpl$Container.launch(ContainerLauncherImpl.java:163)at org.apache.hadoop.mapreduce.v2.app.launcher.ContainerLauncherImpl$EventProcessor.run(ContainerLauncherImpl.java:394)at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)at java.lang.Thread.run(Thread.java:748)2023-10-09 15:42:38,147 INFO mapreduce.Job:  map 100% reduce 100%
2023-10-09 15:42:39,167 INFO mapreduce.Job: Job job_1696837267350_0001 failed with state FAILED due to: Task failed task_1696837267350_0001_m_000000
Job failed as tasks failed. failedMaps:1 failedReduces:0 killedMaps:0 killedReduces: 02023-10-09 15:42:39,245 INFO mapreduce.Job: Counters: 10Job CountersFailed map tasks=4Killed reduce tasks=1Launched map tasks=4Other local map tasks=3Data-local map tasks=1Total time spent by all maps in occupied slots (ms)=5Total time spent by all reduces in occupied slots (ms)=0Total time spent by all map tasks (ms)=5Total vcore-milliseconds taken by all map tasks=5Total megabyte-milliseconds taken by all map tasks=5120
[root@hadoop11 data]#

[root@hadoop11 data]# hadoop jar /opt/installs/hadoop3.1.4/share/hadoop/mapreduce/hadoop-mapreduce-examples-3.1.4.jar pi 1            10
Number of Maps  = 1
Samples per Map = 10
Wrote input for Map #0
Starting Job
2023-10-09 15:45:28,177 INFO mapreduce.JobResourceUploader: Disabling Erasure Coding for path: /tmp/hadoop-yarn/staging/roo           t/.staging/job_1696837267350_0002
2023-10-09 15:45:28,335 INFO input.FileInputFormat: Total input files to process : 1
2023-10-09 15:45:28,441 INFO mapreduce.JobSubmitter: number of splits:1
2023-10-09 15:45:28,629 INFO mapreduce.JobSubmitter: Submitting tokens for job: job_1696837267350_0002
2023-10-09 15:45:28,631 INFO mapreduce.JobSubmitter: Executing with tokens: []
2023-10-09 15:45:28,856 INFO conf.Configuration: resource-types.xml not found
2023-10-09 15:45:28,857 INFO resource.ResourceUtils: Unable to find 'resource-types.xml'.
2023-10-09 15:45:28,939 INFO impl.YarnClientImpl: Submitted application application_1696837267350_0002
2023-10-09 15:45:29,009 INFO mapreduce.Job: The url to track the job: http://hadoop13:8088/proxy/application_1696837267350_           0002/
2023-10-09 15:45:29,011 INFO mapreduce.Job: Running job: job_1696837267350_0002
2023-10-09 15:45:36,147 INFO mapreduce.Job: Job job_1696837267350_0002 running in uber mode : false
2023-10-09 15:45:36,149 INFO mapreduce.Job:  map 0% reduce 0%
2023-10-09 15:45:37,190 INFO mapreduce.Job: Task Id : attempt_1696837267350_0002_m_000000_0, Status : FAILED
Container launch failed for container_e50_1696837267350_0002_01_000002 : org.apache.hadoop.yarn.exceptions.InvalidAuxServic           eException: The auxService:mapreduce_shuffle does not existat sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)at sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:62)at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)at java.lang.reflect.Constructor.newInstance(Constructor.java:423)at org.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.instantiateExceptionImpl(SerializedExceptio           nPBImpl.java:171)at org.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.instantiateException(SerializedExceptionPBI           mpl.java:182)at org.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.deSerialize(SerializedExceptionPBImpl.java:           106)at org.apache.hadoop.mapreduce.v2.app.launcher.ContainerLauncherImpl$Container.launch(ContainerLauncherImpl.java:16           3)at org.apache.hadoop.mapreduce.v2.app.launcher.ContainerLauncherImpl$EventProcessor.run(ContainerLauncherImpl.java:           394)at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)at java.lang.Thread.run(Thread.java:748)2023-10-09 15:45:39,235 INFO mapreduce.Job: Task Id : attempt_1696837267350_0002_m_000000_1, Status : FAILED
Container launch failed for container_e50_1696837267350_0002_01_000003 : org.apache.hadoop.yarn.exceptions.InvalidAuxServic           eException: The auxService:mapreduce_shuffle does not existat sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)at sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:62)at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)at java.lang.reflect.Constructor.newInstance(Constructor.java:423)at org.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.instantiateExceptionImpl(SerializedExceptio           nPBImpl.java:171)at org.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.instantiateException(SerializedExceptionPBI           mpl.java:182)at org.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.deSerialize(SerializedExceptionPBImpl.java:           106)at org.apache.hadoop.mapreduce.v2.app.launcher.ContainerLauncherImpl$Container.launch(ContainerLauncherImpl.java:16           3)at org.apache.hadoop.mapreduce.v2.app.launcher.ContainerLauncherImpl$EventProcessor.run(ContainerLauncherImpl.java:           394)at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)at java.lang.Thread.run(Thread.java:748)2023-10-09 15:45:41,260 INFO mapreduce.Job: Task Id : attempt_1696837267350_0002_m_000000_2, Status : FAILED
Container launch failed for container_e50_1696837267350_0002_01_000004 : org.apache.hadoop.yarn.exceptions.InvalidAuxServic           eException: The auxService:mapreduce_shuffle does not existat sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)at sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:62)at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)at java.lang.reflect.Constructor.newInstance(Constructor.java:423)at org.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.instantiateExceptionImpl(SerializedExceptio           nPBImpl.java:171)at org.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.instantiateException(SerializedExceptionPBI           mpl.java:182)at org.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.deSerialize(SerializedExceptionPBImpl.java:           106)at org.apache.hadoop.mapreduce.v2.app.launcher.ContainerLauncherImpl$Container.launch(ContainerLauncherImpl.java:16           3)at org.apache.hadoop.mapreduce.v2.app.launcher.ContainerLauncherImpl$EventProcessor.run(ContainerLauncherImpl.java:           394)at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)at java.lang.Thread.run(Thread.java:748)2023-10-09 15:45:44,293 INFO mapreduce.Job:  map 100% reduce 100%
2023-10-09 15:45:44,307 INFO mapreduce.Job: Job job_1696837267350_0002 failed with state FAILED due to: Task failed task_16           96837267350_0002_m_000000
Job failed as tasks failed. failedMaps:1 failedReduces:0 killedMaps:0 killedReduces: 02023-10-09 15:45:44,387 INFO mapreduce.Job: Counters: 10Job CountersFailed map tasks=4Killed reduce tasks=1Launched map tasks=4Other local map tasks=3Data-local map tasks=1Total time spent by all maps in occupied slots (ms)=5Total time spent by all reduces in occupied slots (ms)=0Total time spent by all map tasks (ms)=5Total vcore-milliseconds taken by all map tasks=5Total megabyte-milliseconds taken by all map tasks=5120
Job job_1696837267350_0002 failed!
[root@hadoop11 data]#

2.4.2 查看yarn任务：

2.4.3 问题描述：

最开始发现这个mr shuffle的报错是在beeline中执行走mr的查询代码时，发现程序不走mr，经过测试发现是hadoop中yarn的mr配置原因，因为我们服务器之前装hadoop的哥们只用spark，在服务器的yarn中只有spark的成功任务，从没跑过mr。问题已经很明确，yarn的环境配置缺少mapreduce_shuffle。

网上有比较多的方案，参考文档¹这一篇的配置方法提到了spark和mr的shuffle。但是经过配置后仍然报mr shuffle不存在的错。

于是检查yarn配置，如下图：

只搜索到了一个mr_shuffle,而且不是我在yarn-site.xml中添加的。
原因再次定位到yarn配置文件没有生效。

而后经过检查发现，我的集群在重启yarn时，stop-yarn.sh报no resourcemanager to stop,由于集群数量太多，一直没有仔细看yarn的关机提示，导致我在分发yarn配置后，其实资源管理器并没有重启。

于是解决掉yarn无法重启的问题，原因见参考文档²。

再次去8088检查yarn配置，这次查到了两个mr shuffle配置，其中一个正是我添加的。

no resourcemanager to stop

2.4.4 参考文档

[1] AWS EMR S3DistCp: The auxService:mapreduce_shuffle does not exist
[2] no resourcemanager to stop

---

1. 1 ↩︎

2. 2 ↩︎