官术网_书友最值得收藏!

Jobs

Jobs is the default tab of Spark UI. It shows the status of all the applications executed within a SparkContext. It can be accessed at http://localhost:4040/jobs/.

It consists of three sections:

  • Active Jobs: This section is for the jobs that are currently running
  • Completed Jobs: This section is for the jobs that successfully completed
  • Failed Jobs: This section is for the jobs that were failed

It is shown in the following screenshot:

Sections of Spark UI are created lazily and can be visible if required. For example, the Active Jobs section will only be visible if there is a job that is currently running. Similarly, the Failed Jobs and Completed Jobs sections are visible only if there is a job that failed or successfully completed.

The Jobs tab section of Spark UI is rendered using the org.apache.spark.ui.jobs.JobsTab class that uses org.apache.spark.ui.jobs.JobProgressListener to get the statistics of the job.

After executing all the jobs mentioned in the Spark REPL also known as CLI section, Spark UI will look as follows:

Also, if you expand the Event Timeline section, you can see the time at which SparkContext started (that is, driver was initiated) and the jobs were executed along with their status:

Also, by clicking on any of the jobs, you can see the details of the job, that is, the Event Timeline of the job and the DAG of the transformations and stages executed during the execution of the job, as follows:

主站蜘蛛池模板: 怀来县| 中江县| 金堂县| 曲周县| 塘沽区| 长治市| 龙口市| 朝阳市| 山西省| 高唐县| 东至县| 鹿泉市| 昆明市| 乐平市| 利辛县| 察雅县| 灵宝市| 海门市| 嘉义市| 平江县| 盖州市| 万宁市| 舒兰市| 苗栗市| 南京市| 杂多县| 邻水| 木兰县| 玉林市| 灵寿县| 古田县| 溆浦县| 湟中县| 五大连池市| 巴塘县| 长治市| 沙河市| 伊金霍洛旗| 昌黎县| 华宁县| 克拉玛依市|