On Thu, Feb 11, 2010 at 12:32 PM, Ryan LeCompte <[email protected]> wrote:
> I see. Here's my problem, though. I already have a new derby metastoredb
> setup and configured in the NEW cluster. I don't want to blow that one away,
> since I have some tables that I don't want to get rid of.
>
> I essentially want to do this:
>
> 1) Grab tables from the old cluster and move them into the new cluster (one
> time operation)
>
> I guess I would need to somehow merge table information from the old derby
> metastoredb into the new one?
>
> I guess I could easily distcp the table directories under
> /user/hive/warehouse from the old cluster and upload them to
> /user/hive/warehouse in the new cluster.
>
> But what about Derby?
>
>
> On Thu, Feb 11, 2010 at 12:25 PM, Edward Capriolo <[email protected]>
> wrote:
>>
>> I think it is pretty simple
>> 1)distcp the warehouse
>> 2)rsync your derby DB
>> --or--
>> backup restore derby
>>
>> This assumes you are not going to edit anything while moving.
>>
>>
>> On Thu, Feb 11, 2010 at 12:01 PM, Ryan LeCompte <[email protected]>
>> wrote:
>> > Hey guys,
>> >
>> > I have another Hadoop cluster that has Hive installed with its own
>> > metastore
>> > and all. I would like to move/copy/export data from a bunch of Hive
>> > tables
>> > from a different Hadoop cluster into this one.
>> >
>> > Is this possible? What's the best way to do it? The hadoop/hive/derby
>> > versions are the same.
>> >
>> > Thanks!
>> >
>> > Ryan
>> >
>> >
>
>

I opened up https://issues.apache.org/jira/browse/HIVE-1161, I think
whipping up tools to handle simple master/slave replication between
two clusters should be fun, also might help quell some SPOF questions.
+1

Reply via email to