Thanks, I did now.

However it failed miserably.

It completed successfully on server A as destination and as source, however
only after I created the table with all the correlating column families
(specified by "--new.name=new_table_name"). Without that step being done
manually it failed as well.

When running:

hbase org.apache.hadoop.hbase.mapreduce.CopyTable
--peer.adr=serverB:2181:/hbase table_name

It failed, regardless whether I created the table with the same name and
column families, or a new table and specified 'new.name'.

Would very much appreciate any additional feedback.

Thanks!
Tom
On Wed, Sep 7, 2011 at 6:03 PM, Doug Meil <[email protected]>wrote:

>
> Hi there-
>
> Have you tried this?
>
> http://hbase.apache.org/book.html#copytable
>
> It's the Java invocation of the copy-table function (without the ruby)
>
>
>
> On 9/7/11 8:29 PM, "Tom Goren" <[email protected]> wrote:
>
> >So I have read http://blog.sematext.com/2011/03/11/hbase-backup-options/
> >
> >The built-in solution, namely the 'Export' and 'Import' classes are
> >relatively straight forward to use, however, the exports include the data
> >alone, and nothing in regards to the table description (column families,
> >versioning etc.).
> >
> >Manually creating the table seems wrong to me, and parsing the output of
> >'describe table_name' seems like something that someone must have done
> >before me.
> >
> >There is this:
> >http://svn.apache.org/repos/asf/hbase/trunk/bin/copy_table.rb
> >But surely there is something nicer?
> >
> >Is there an all encompassing solution that I am unaware of?
> >
> >What is the basic correlating option to:
> >
> >mysqldump db_name table_name > file
> >mysql db_name < file
> >
> >Or, if this is too Utopian in an hbase world, what is the next best thing?
> >
> >Thanks.
> >Tom
>
>

Reply via email to