Hi Jignesh, for online service hbase cluster (it means there is data written during transforming), I prefer to use ExportSnapshot to copy data from cluster to cluster, because of the less performance impact, and also you can use *-bandwidth* parameter to control network costs. You can also make snapshot for each hbase table individually. After that you can use CopyTable to copy data to new cluster which is written during transforming period. But if your cluster is offline service, I think Export is also fine.
About the usage of these tools for 0.98.7, you can check this link: *https://devdoc.net/bigdata/hbase-0.98.7-hadoop1/book/ops_mgt.html#tools <https://devdoc.net/bigdata/hbase-0.98.7-hadoop1/book/ops_mgt.html#tools>* Jignesh Patel <[email protected]> 于2023年5月10日周三 01:09写道： > So which one is better approach exportsnapshot or export each table > individually? > > On Tue, May 9, 2023 at 8:54 AM Jignesh Patel <[email protected]> > wrote: > > > don't know the size of the data asI don't know the command to check. > > > > But can we follow this blog to export and then import > > > > > https://blog.clairvoyantsoft.com/hbase-incremental-table-backup-and-disaster-recovery-using-aws-s3-storage-aa2bc1b40744 > > > > On Thu, May 4, 2023 at 11:57 AM Davide Vergari <[email protected] > > > > wrote: > > > >> If hbase tables you can create a snapshot for each table then export > with > >> the ExportSnapshot mapreduce job (should be already available on > 0.98.x). > >> For data that are not in hbase you can use distcp > >> > >> Il giorno gio 4 mag 2023 alle ore 17:13 <[email protected]> ha scritto: > >> > >> > Jignesh, how much data? Is the data currently in hbase format? > >> > Very kindly, Sean > >> > > >> > > >> > > On 05/04/2023 11:03 AM Jignesh Patel <[email protected]> > wrote: > >> > > > >> > > > >> > > We are in the process of having hadoop os, however we are using a > very > >> > old > >> > > version of hadoop. > >> > > Hadoop 2.6 > >> > > and HBase 0.98.7. > >> > > > >> > > So how do we export and import the data from the cluster with the > old > >> OS > >> > to > >> > > the new OS. We are trying to use the same hadoop/hbase version. > >> > > > >> > > -Jignesh > >> > > >> > > >
