You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@accumulo.apache.org by Roshan Punnoose <ro...@gmail.com> on 2013/10/22 16:03:25 UTC

Copy table between clusters

What is the best way to copy a table between two separate Accumulo
clusters. My initial thought was to export the table, use the Hadoop DistCp
to copy the files over to the new cluster, and then import the table.

Re: Copy table between clusters

Posted by Roshan Punnoose <ro...@gmail.com>.
That looks exactly like the problem I'm seeing. Great that it is fixed! I
ended up just writing a DistCpTable MR job similar to the Hadoop DistCp job
for now. I'd rather just move the tablet files over, but for now this
works. When we upgrade to 1.6.0 I'll switch to using export/import table.
Thanks!


On Tue, Oct 22, 2013 at 3:56 PM, Keith Turner <ke...@deenlo.com> wrote:

> Do you know if you are running into ACCUMULO-1558?  I fixed that recently.
>  If for some reason import table can not find a file listed in the import
> file, it just silently ignores the problem.
>
>
>
>
> On Tue, Oct 22, 2013 at 2:25 PM, Roshan Punnoose <ro...@gmail.com>
> wrote:
>
> > Whenever I try to import the table on the other cluster, the table comes
> > up; however, the tablet files in HDFS are empty. The tablet directories
> > exist, but no "rf" files are in there. I distcp'd all the locations
> > specified in the distcp.txt file, and then ran the importtable on the
> > second cluster.
> >
> >
> > On Tue, Oct 22, 2013 at 10:40 AM, Eric Newton <er...@gmail.com>
> > wrote:
> >
> > > Clone the table, offline the clone, export the clone, distcp, and
> > > import the table.  Optionally, delete the clone when you are done with
> > > the distcp.
> > >
> > > -Eric
> > >
> > >
> > > On Tue, Oct 22, 2013 at 10:03 AM, Roshan Punnoose <ro...@gmail.com>
> > > wrote:
> > > > What is the best way to copy a table between two separate Accumulo
> > > > clusters. My initial thought was to export the table, use the Hadoop
> > > DistCp
> > > > to copy the files over to the new cluster, and then import the table.
> > >
> >
>

Re: Copy table between clusters

Posted by Keith Turner <ke...@deenlo.com>.
Do you know if you are running into ACCUMULO-1558?  I fixed that recently.
 If for some reason import table can not find a file listed in the import
file, it just silently ignores the problem.




On Tue, Oct 22, 2013 at 2:25 PM, Roshan Punnoose <ro...@gmail.com> wrote:

> Whenever I try to import the table on the other cluster, the table comes
> up; however, the tablet files in HDFS are empty. The tablet directories
> exist, but no "rf" files are in there. I distcp'd all the locations
> specified in the distcp.txt file, and then ran the importtable on the
> second cluster.
>
>
> On Tue, Oct 22, 2013 at 10:40 AM, Eric Newton <er...@gmail.com>
> wrote:
>
> > Clone the table, offline the clone, export the clone, distcp, and
> > import the table.  Optionally, delete the clone when you are done with
> > the distcp.
> >
> > -Eric
> >
> >
> > On Tue, Oct 22, 2013 at 10:03 AM, Roshan Punnoose <ro...@gmail.com>
> > wrote:
> > > What is the best way to copy a table between two separate Accumulo
> > > clusters. My initial thought was to export the table, use the Hadoop
> > DistCp
> > > to copy the files over to the new cluster, and then import the table.
> >
>

Re: Copy table between clusters

Posted by Roshan Punnoose <ro...@gmail.com>.
Whenever I try to import the table on the other cluster, the table comes
up; however, the tablet files in HDFS are empty. The tablet directories
exist, but no "rf" files are in there. I distcp'd all the locations
specified in the distcp.txt file, and then ran the importtable on the
second cluster.


On Tue, Oct 22, 2013 at 10:40 AM, Eric Newton <er...@gmail.com> wrote:

> Clone the table, offline the clone, export the clone, distcp, and
> import the table.  Optionally, delete the clone when you are done with
> the distcp.
>
> -Eric
>
>
> On Tue, Oct 22, 2013 at 10:03 AM, Roshan Punnoose <ro...@gmail.com>
> wrote:
> > What is the best way to copy a table between two separate Accumulo
> > clusters. My initial thought was to export the table, use the Hadoop
> DistCp
> > to copy the files over to the new cluster, and then import the table.
>

Re: Copy table between clusters

Posted by Eric Newton <er...@gmail.com>.
Clone the table, offline the clone, export the clone, distcp, and
import the table.  Optionally, delete the clone when you are done with
the distcp.

-Eric


On Tue, Oct 22, 2013 at 10:03 AM, Roshan Punnoose <ro...@gmail.com> wrote:
> What is the best way to copy a table between two separate Accumulo
> clusters. My initial thought was to export the table, use the Hadoop DistCp
> to copy the files over to the new cluster, and then import the table.