Kerberos 相关操作
[root@mater ~]# kadmin.local
Authenticating as principal root/admin@TEST.COM with password.
kadmin.local: list_requests # 查看命令帮助
kadmin.local: list_principals # 查看用户
kadmin.local: addprinc kuser01 # 创建用户
kadmin.local: addprinc -pw kuser01 kuser01 # 带密码创建用户
kadmin.local: delprinc kuser01 # 删除用户
kadmin.local: delprinc -force kuser01 # 强制删除用户
kadmin.local: xst -k user.keytab user@BLUE.COM # 生成keytab文件
用户登录: kinit -kt /root/keytab/user.keytab user
1.故障现象
在TDH管理页面检查yarn服务发现有节点的nodemanager 服务状态异常
在yarn服务节点,发现有Unhealthy Nodes,正常的active nodes数目不对
[root@slzhd-stream03 yarn2]# kinit -kt /etc/yarn2/yarn.keytab yarn/slzhd-stream03@TDH
[root@slzhd-stream03 yarn2]# yarn node -list -all
Total Nodes:6
Node-Id Node-State Node-Http-Address Number-of-Running-Containers
slzhd-stream04:49381 RUNNING slzhd-stream04:8042 1
slzhd-nn01:43510 RUNNING slzhd-nn01:8042 1
slzhd-stream03:60255 UNHEALTHY slzhd-stream03:8042 0
slzhd-nn02:52523 RUNNING slzhd-nn02:8042 2
slzhd-stream06:46429 RUNNING slzhd-stream06:8042 1
slzhd-stream05:50346 RUNNING slzhd-stream05:8042 1
[root@slzhd-stream03 yarn2]#
查看节点监控状态
[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-Shs7O9OE-1689812527705)(/media/202307/2023-07-20_081046_4119080.8724264091443784.png)]
2.问题分析
查看resourcemanager,nodemanager的日志均未发现问题
在yarn ui 查看Node节点信息,查看Nodemanager信息,确认 slzhd-stream03 节点信息
[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-sY4P4EPG-1689812527707)(/media/202307/2023-07-19_173501_1437230.22773761327714037.png)]
发现该节点的健康检查日志的路径与实际不符
3.解决方法
检查 yarn 日志相关配置
- yarn.nodemanager.log-dirs
[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-RIEsf7dF-1689812527708)(/media/202307/2023-07-20_081324_0138240.9054526401901273.png)] - yarn.nodemanager.local-dirs
[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-rwKs3pTm-1689812527711)(/media/202307/2023-07-20_081346_4010850.40760285045153777.png)]
4.解决过程
step 1.创建yarn日志文件夹
step 2.查看节点状态
[root@slzhd-stream03 logs]kinit -kt /etc/yarn2/yarn.keytab yarn/slzhd-stream03@TDH
[root@slzhd-stream03 logs]# yarn node -list -all
Total Nodes:7
Node-Id Node-State Node-Http-Address Number-of-Running-Containers
slzhd-stream04:49381 RUNNING slzhd-stream04:8042 1
slzhd-nn01:43510 RUNNING slzhd-nn01:8042 1
slzhd-stream03:57844 RUNNING slzhd-stream03:8042 1
slzhd-nn02:52523 RUNNING slzhd-nn02:8042 2
slzhd-stream06:46429 RUNNING slzhd-stream06:8042 1
slzhd-stream05:50346 RUNNING slzhd-stream05:8042 1
slzhd-stream03:57975 LOST slzhd-stream03:8042 0
[root@slzhd-stream03 logs]#
5.参考文档
重启任务
step 1.首先要先 kill 掉所有的任务,因为当前客户的所有节点已经是 unhealthy 状态了,可以采用如下办法 kill 所有的任务:
for app in `yarn application -list | awk '$6 == "ACCEPTED" { print $1 }'`; do yarn application -kill "$app"; done
step 2.重启所有节点
step 3.重新提交任务