Command Line Upgrade
Also available as:
PDF
loading table of contents...

Configure and Start Apache HBase

[Note]Note

The su commands in this section use "hbase" to represent the HBASE Service user. If you are using another name for your HBASE Service user, you need to substitute your HBASE Service user name for "hbase" in each of the su commands.

The hbase.bucketcache.percentage.in.combinedcache is removed in HDP 2.6.0. This simplifies the configuration of block cache. BucketCache configurations from HDP 2.5 needs to be recalculated to attain identical memory allotments in HDP 2.6.0. The L1 LruBlockCache is whatever hfile.block.cache.size is set to and the L2 BucketCache is whatever hbase.bucketcache.size is set to.

  1. Replace your configuration after upgrading. Replace the Apache HBase template configuration in /etc/hbase/conf.

  2. Start services. From root, assuming that $HBASE_USER=hbase:

    su - hbase -c "/usr/hdp/current/hbase-master/bin/hbase-daemon.sh start master; sleep 25"

    su - hbase -c "/usr/hdp/current/hbase-regionserver/bin/hbase-daemon.sh start regionserver"

  3. Check processes.

    ps -ef | grep -i hmaster

    ps -ef | grep -i hregion

  4. If bulk load support for backup/restore is required, follow these steps:

    1. Register org.apache.hadoop.hbase.backup.BackupHFileCleaner through hbase.master.hfilecleaner.plugins.

      org.apache.hadoop.hbase.backup.BackupHFileCleaner is responsible for keeping bulk loaded hfiles so that incremental backup can pick them up.

    2. Register org.apache.hadoop.hbase.backup.BackupObserver through hbase.coprocessor.region.classes.

      org.apache.hadoop.hbase.backup.BackupObserver is notified when bulk load completes and writes records into hbase:backup table.