Thanks Jonathan.
Filed: https://issues.apache.org/jira/browse/CASSANDRA-2283
We'll start the scrub during our normal compaction cycle and update
this thread and the bug with the results.
-erik
On Mon, Mar 7, 2011 at 11:27 AM, Jonathan Ellis wrote:
> It sounds like it doesn't realize the data i
It sounds like it doesn't realize the data it's streaming over is
older-version data. Can you create a ticket?
In the meantime nodetool scrub (on the existing nodes) will rewrite
the data in the new format which should workaround the problem.
On Mon, Mar 7, 2011 at 1:23 PM, Erik Onnen wrote:
>
During a recent upgrade of our cassandra ring from 0.6.8 to 0.7.3 and
prior to a drain on the 0.6.8 nodes, we lost a node for reasons
unrelated to cassandra. We decided to push forward with the drain on
the remaining healthy nodes. The upgrade completed successfully for
the remaining nodes and the