Article
Hadoop查看作业状态Rest接口
hadoop yarn提供了web端查看任务状态,同时可以通过rest的方式获取任务的相关信息。rest接口和网页端的每个界面一一对应。

上面的5个图的链接为:
http://hadoop-master1:8088/cluster/apps/RUNNING
http://hadoop-master1:8088/cluster/app/application_1417676507722_1846
http://hadoop-master1:8088/proxy/application_1417676507722_1846/
http://hadoop-master1:8088/proxy/application_1417676507722_1846/mapreduce/job/job_1417676507722_1846
http://hadoop-master1:19888/jobhistory/job/job_1417676507722_1846/mapreduce/job/job_1417676507722_1846
# 查看正在运行的任务
curl http://hadoop-master1:8088/ws/v1/cluster/apps?states=RUNNING
...
curl http://hadoop-master1:8088/proxy/application_1417676507722_1867/ws/v1/mapreduce/info
...
curl http://hadoop-master1:8088/proxy/application_1417676507722_1867/ws/v1/mapreduce/jobs
...
curl http://hadoop-master1:8088/proxy/application_1417676507722_1867/ws/v1/mapreduce/jobs/job_1417676507722_1867
...
curl http://hadoop-master1:8088/proxy/application_1417676507722_1867/ws/v1/mapreduce/jobs/job_1417676507722_1867/counters
...
curl http://hadoop-master1:8088/proxy/application_1417676507722_1867/ws/v1/mapreduce/jobs/job_1417676507722_1867/conf
如果上面的任务是已经完成的,获取对应的信息时返回的值是空的。
curl http://hadoop-master1:8088/proxy/application_1417676507722_1867/ws/v1/mapreduce/jobs/job_1417676507722_1867/counters
# 查看执行完成的任务
curl http://hadoop-master1:19888/ws/v1/history
curl http://hadoop-master1:19888/ws/v1/history/info
...
curl http://hadoop-master1:19888/ws/v1/history/mapreduce/jobs?startedTimeBegin=$(date +%s -d '-1 hour')000
...
curl http://hadoop-master1:19888/ws/v1/history/mapreduce/jobs/job_1417676507722_1867
curl http://hadoop-master1:19888/ws/v1/history/mapreduce/jobs/job_1417676507722_1867/counters
curl http://hadoop-master1:19888/ws/v1/history/mapreduce/jobs/job_1417676507722_1867/conf
curl -H "Accept: application/xml" "http://hadoop-master1:8088/ws/v1/cluster/apps?states=FINISHED&limit=1" | xmllint --format -
后面的参数和运行任务一致,只是提供服务不同。
# xml转csv

$ curl -H "Accept: application/xml" "http://hadoop-master1:8088/ws/v1/cluster/apps?startedTimeBegin=$(date +%s -d '-1 hour')000" 2>/dev/null | xsltproc yarn.xslt - | sort -r
application_1417676507722_1973,AccessLogOnlyHiveJob,RUNNING,UNDEFINED,1417942144941,0,19416
application_1417676507722_1972,InfoSecurityLogJob,FINISHED,SUCCEEDED,1417942084278,1417942098184,13906
application_1417676507722_1971,InfoSecurityLogJob,FINISHED,SUCCEEDED,1417941603456,1417941617773,14317
application_1417676507722_1970,AccessLogOnlyHiveJob,FINISHED,SUCCEEDED,1417941581080,1417942142287,561207
application_1417676507722_1969,InfoSecurityLogJob,FINISHED,SUCCEEDED,1417941422664,1417941436456,13792
# 参考
- Hadoop YARN - Introduction to the web services REST API’s.
- ResourceManager REST API’s.
- MapReduce Application Master REST API’s.
- History Server REST API’s.
- Hadoop YARN中web服务的REST API介绍
–END
Related
Related posts
-
杀鸡焉用牛刀:DuckDB 正取代部分 Spark 场景
2026-02-16
-
基于对象存储的 Spark 数据读写实战:从末尾追加到任意更新
2025-10-28
-
认真的博客
2021-12-08
-
视频自动翻译
2018-08-25