hadoop 客户的的使用

${HADOOP_HOME}/bin/hadoop job
Usage: JobClient <command> <args>
        [-submit <job-file>]
        [-status <job-id>]
        [-counter <job-id> <group-name> <counter-name>]
        [-kill <job-id>]
        [-abort <job-id>]
        [-suspend <job-id> [hours]]
        [-recover <job-id> [-force] [-jobconf name=value] [-file local-path] [-cacheArchive]]
        [-set-priority <job-id> <priority>]. Valid values for priorities are: VERY_HIGH HIGH NORMAL LOW VERY_LOW
        [-set-map-capacity <job-id> <map-capacity>]
        [-set-reduce-capacity <job-id> <reduce-capacity>]
        [-set-map-over-capacity <job-id> <true/false>]
        [-set-reduce-over-capacity <job-id> <true/false>]
        [-events <job-id> <from-event-#> <#-of-events>]
        [-history <jobOutputDir>]
        [-list [all]]
        [-kill-task <task-id>]
        [-fail-task <task-id>]
        [-input-add <job-id> <input>]
        [-input-done <job-id>]

 

• -kill <job-id> kill一个job，job的最终状态是KILLED
• -kill-task <task-id> kill一个task attempt，task attempt的最终状态是KILLED，对应的task会重新启动一个task attempt计算，kill不会导致task失败
• -fail-task <task-id> fail一个task attempt，task attempt的最终状态是FAILED，如果task attempt fail超过一定次数（默认4次），对应task会失败
• -set-priority <job-id> 设置job的优先级
• -status <job-id> 获取job的状态
• -list [all] 获取作业列表，没有参数表示获取运行的作业列表，参数all表示获取所有作业列表
• -suspend <job-id> [hours], -recover <job-id> 在断点重启中介绍