Hello @Hadoop_Admin    Thanks for using Cloudera Community. To reiterate, your Team enabled Replication from ClusterA to ClusterB & seeing Data Loss. By Data Loss, your Team means the Record Count on Source & Target isn't matching. This is observed for Large Table with ~2TB Size.   Kindly confirm the Process being used for Customer to compare the Record Count. Is VerifyRep being utilised for the concerned purpose.   Next, HBase Replication is supposed to be Asynchronous i.e. some Lags are expected, if the Source Table is being loaded. Confirm if the Command [status 'replication'] is reporting any Replication Lag.   Next, We need to establish if the RowCount Difference is Static or Dynamic during a period of No-Load on Source Table (If feasible). If Source Table has 100 Rows & Target Table has 90 Rows & remains so, We can assume 10 Rows are the Difference. If Target Table shows 91>92>93... Rows, We can assume Replication is catching up.   Finally, Any Audit Record showing any Delete Ops on the Target Table.    - Smarak ... View more

Hello @Manoj690    Thanks for contacting Cloudera Community. While taking a Full Backup, you are facing IOException while waiting on the Lock.    Kindly share the Output of Command "hbase backup history" along with "list_locks" from HBase Shell. The requested details would confirm the status of any running Backup & Locks placed on the Tables. Additionally, Share the HBase Version wherein you are using the required Backup Command.    - Smarak ... View more

>From CML, you could access Greenplum via JDBC. ... View more

Thanks @smdas for clarification.   Best Regards, Vinod ... View more

Hello @tuk    If the Post by Pabitra assisted you, Kindly mark the Post as Solution. If you utilised any other approach, Kindly share the details in the post as well.   Thanks, Smarak ... View more

backup command running with hbase super user hbase backup create full hdfs://hostname:port/backup -t table_name restore command hbase restore hdfs://hostname:port/backup -t table_name we did this on same cluster ... View more

Hello @Manoj690    RegionServer is a Service & your team can add the RegionServer Service interactively using via Ambari (HDP) or Cloudera Manager (CDH or CDP).   - Smarak ... View more

Hello @lenu    If you have Replication enabled, WALs are likely to be persisted until the WALs are replicated. If you aren't using HBase Replication, Ensure there are no Peers (via "list_peers") & "hbase.replication" Property is false. If the oldWALs aren't removed, Enable TRACE Logging for the HBase Master Service, which would print the CleanerChore Thread removing or skipping any entries.    - Smarak   ... View more