[转]HBase hbck——检察HBase集群的一致性

 Hbase提供了hbck命令来检查各种不一致问题。hbck的名字仿效了HDFS的fsck命令，后者是一个用于检查HDFS中不一致问题的工具。下面这段非常易懂的介绍出自于hbck的源程序。

       检查数据在Master及RegionServer的内存中状态与数据在HDFS中的状态之间的一致性。

       HBase的hbck不仅能够检查不一致问题，而且还能够修复不一致问题。

       在生产环境中，应当经常运行hbck，以便及早发现不一致问题并更容易地解决问题。

一、问题

       首先，在HBase上创建一张表usertable。

       然后，在HDFS上直接删除usertable表的目录/hbase/usertable来删除表的数据（错误的做法）。

二、解决方法

       1.hbase hbck

              Status：OK，表示没有发现不一致问题。

              Status：INCONSISTENT，表示有不一致问题。

       2.hbase hbck -fixAssignments

              Try to fix region assignments.  Replaces the old -fix

              [zk: 192.168.147.15:2181(CONNECTED) 3] ls /hbase/unassigned

       3.hbase hbck -fixMeta

              Try to fix meta problems.  This assumes HDFS region info is good.

              hbase(main):007:0>  scan '.META.'

       4.hbase(main):015:0> create 'usertable','f1'

              ERROR: Table already exists: usertable!

              

              [zk: 192.168.147.15:2181(CONNECTED) 4] ls /hbase/table

              [usertable]

              [zk: 192.168.147.15:2181(CONNECTED) 5] get /hbase/table/usertable

              ENABLING

              [zk: 192.168.147.15:2181(CONNECTED) 6] rmr /hbase/table/usertable

              [zk: 192.168.147.15:2181(CONNECTED) 7] get /hbase/table/usertable

              Node does not exist: /hbase/table/usertable

              

              此时，create 'usertable','f1'，依然提示同样的错误。

              查看ZKTable源码：

/**
 * Cache of what we found in zookeeper so we don't have to go to zk ensemble
 * for every query.  Synchronize access rather than use concurrent Map because
 * synchronization needs to span query of zk.
 */
private final Map<String, TableState> cache =
  new HashMap<String, TableState>();

 this.cache.put(tableName, state);

              把ZK中的数据缓存在本地内存中，以加速访问，减少访问时间。

              为了让本地内存中的数据失效，必须重启HBase集群。

              

              重启HBase集群。

              此时，create 'usertable','f1'，没有提示错误，表创建成功。

三、hbck

[bigdata@tbe192168147015 ~]$ hbase hbck -h
Usage: fsck [opts] {only tables}
 where [opts] are:
   -help Display help options (this)
   -details Display full report of all regions.
   -timelag {timeInSeconds}  Process only regions that  have not experienced any metadata updates in the last  {{timeInSeconds} seconds.
   -sleepBeforeRerun {timeInSeconds} Sleep this many seconds before checking if the fix worked if run with -fix
   -summary Print only summary of the tables and status.
   -metaonly Only check the state of ROOT and META tables.
  Repair options: (expert features, use with caution!)
   -fix              Try to fix region assignments.  This is for backwards compatiblity
   -fixAssignments   Try to fix region assignments.  Replaces the old -fix
   -fixMeta          Try to fix meta problems.  This assumes HDFS region info is good.
   -fixHdfsHoles     Try to fix region holes in hdfs.
   -fixHdfsOrphans   Try to fix region dirs with no .regioninfo file in hdfs
   -fixHdfsOverlaps  Try to fix region overlaps in hdfs.
   -fixVersionFile   Try to fix missing hbase.version file in hdfs.
   -maxMerge <n>     When fixing region overlaps, allow at most <n> regions to merge. (n=5 by default)
   -sidelineBigOverlaps  When fixing region overlaps, allow to sideline big overlaps
   -maxOverlapsToSideline <n>  When fixing region overlaps, allow at most <n> regions to sideline per group. (n=2 by default)

   -repair           Shortcut for -fixAssignments -fixMeta -fixHdfsHoles -fixHdfsOrphans -fixHdfsOverlaps -fixVersionFile -sidelineBigOverlaps
   -repairHoles      Shortcut for -fixAssignments -fixMeta -fixHdfsHoles -fixHdfsOrphans