I have a Couchbase cluster that serves around 40K updates per second; And I am using DCP to streaming the changes to HDFS.
the question is:
if my streaming process is down for some time, for example, 1 day, there will be around 3 Billion updates, then I restart my DCP, will I still get exact the 3 Billion updates? will the updates get compacted(for example, I updated all documents 10 times with 300M documents)?