zoukankan      html  css  js  c++  java
  • 【转】HBASE Region in Transition issue on Master UI

    【From】https://community.hortonworks.com/content/supportkb/244808/hbase-region-in-transition-issue-on-master-ui.html

    Problem Description: 
    When the HBASE Region In Transition issue occurs on Master UI, check the hbck -details. There could be multiple reasons for this issue to occur. However, if there are any inconsistencies, an error like the following will be displayed in hbck -details: 
    1. ERROR: Region { meta => status_hbase_historic_2018_bckup,COL_P_11057715947,
    2. 1545968041870.5d44a1eec86e57eab8bfc85c2764c371., hdfs =>
    3. hdfs://<nameservice>/apps/hbase/data/data/default/status_hbase_historic_2018_bckup/
    4. 5d44a1eec86e57eab8bfc85c2764c371, deployed => , replicaId => 0 } is a split parent in META,
    5. in HDFS, and not deployed on any region server.
    6. This could be transient, consider to run the catalog janitor first!

    Cause: 
    When a Region Split is marked as Yes, it means that the Region is candidate for Split, and 02 Daughter Regions would be hosted. However, during the process of splitting, if a region get stuck in splitting stage (because of region server restart), the above error will be displayed in hbck -details. 
    Solution: 
    To resolve this issue, run the following command:
    1. hbase hbck -fixSplitParents
    The -fixSplitParents option makes a parent region online forcibly, so after fixing with this option, overlapped regions (the parent region and the daughter regions) will happen normally. 
    In that case where we have overlapped regions, we need to run below to fix them-
    1. hbase hbck -fixHdfsOverlaps
    Then check Hbase master UI and if still any RIT is there we have to manually assign them like below--
    1. hbase> assign "4edbb392a982b7846a91311d30e9fe21"

      

    About: 
    This article created by Hortonworks Support (Article: 000008066) on 2019-04-18 12:50 
    OS: Linux 
    Type: Cluster-Administration 
    Version: HDP
    Support ID: 000008066
  • 相关阅读:
    mapreduce 函数入门 二
    mapreduce 函数入门 一
    Flume+Kafka+Storm+Redis 大数据在线实时分析
    mapReduce 大数据离线分析
    docker 简介
    flume安装使用+根据数据源分类
    hiho 171周
    如何新建一个空的optix工程
    读 Real-Time Rendering 收获
    hiho 1590
  • 原文地址:https://www.cnblogs.com/pekkle/p/11340359.html
Copyright © 2011-2022 走看看