Major Description
The current GoXDCR DCP checkpoint interval is set to 30 minutes. When GoXDCR is forced to restart due to any disruption, it must go back to the previous checkpoint to resume progress. There is the potential that if there are enough disruptions (due to network issues, errors from memcached, etc.), XDCR could be stuck in a loop without progress as we continuously revert back to the same checkpoint.
It sounds like we could probably lower the interval to a value that reduces the possibility of this occurring.
