Uploaded image for project: 'HBase'
  1. HBase
  2. HBASE-27537

Optimize the compaction triggering logic if table has set TTL

Add voteVotersWatch issueWatchersCreate sub-taskLinkCloneUpdate Comment AuthorReplace String in CommentUpdate Comment VisibilityDelete Comments
    XMLWordPrintableJSON

Details

    • Bug
    • Status: Open
    • Critical
    • Resolution: Unresolved
    • 2.4.15
    • None
    • regionserver
    • None

    Description

      Although table has a TTL set in the cluster,some regions do not perform periodic major compactions,and the expired TTL data is still stored in these large files. Not trigger the reason is that the periodic check of the thread according to number of files in the region  If > hbase.hstore.compaction.min then major compaction is not triggered, Instead, minor compaction is triggered first. But minor compaction which excludes large files (by the hbase.hstore.compaction.max.size), when the larger files reach a certain number, the expired data under the larger files cannot be deleted normally.

       

       

       

      Attachments

        Activity

          This comment will be Viewable by All Users Viewable by All Users
          Cancel

          People

            sujun mahuacai
            sujun mahuacai

            Dates

              Created:
              Updated:

              Slack

                Issue deployment