I remember running into something like this recently. Tried to offline
mapreduce, got an error that the table couldn't be cloned. Then, I tried
to take the table offline and run the job, and got an error that the
table was offline.
I don't think I tracked it down the whole way, but there might be a bug
here.
On 12/3/13, 1:34 PM, David Medinets wrote:
A few weeks ago there was a message or two about cloning a table then
taking it offline before running a map-reduce job using
AccumuloInputFormat. I'm trying to do this but have run into an issue.
Here is the gist of my code:
String table = "AA";
String tableClone = "AA_clone";
getTableOperations().clone(table, tableClone, true, new HashMap<String,
String>(), new TreeSet<String>());
getTableOperations().offline(tableClone);
### the rest is normal Hadoop Job setup using tableClone has the input
table.
When I run the job, I get this message:
...AA_clone is offline
...at ... InputFormatBase.getSplits(InputFormatBase.java:1366)
...<normal java stack>
Did I missing something in my job setup?
