Uploaded image for project: 'HBase'
  1. HBase
  2. HBASE-26116

Parallel scheduling of getClusterstatus interface will lead to linear growth of HMaster memory

    XMLWordPrintableJSON

Details

    • Bug
    • Status: Open
    • Major
    • Resolution: Unresolved
    • 1.3.1, 2.2.3
    • None
    • mapreduce, master
    • None

    Description

      when Mapreduce job and Spark job use TableInputFormat.getSplits  will call getclusterstatus ,and when running more ten job, HMaster memory is very high，Of course, the region quantity itself is a reason

       

      Attachments

        Activity

          People

            idol ShuangQi Xia
            idol ShuangQi Xia
            Votes:
            0 Vote for this issue
            Watchers:
            4 Start watching this issue

            Dates

              Created:
              Updated: