Description
Related to: MB-21574
Our customers are going to attempt to use XDCR to replicate from <4.6 clusters into >= 4.6 clusters and so we need to be prepared for that.
Two issues at hand:
1 - Given that we will not support non-LWW buckets being used in an XDCR stream with LWW buckets, we should do our best in the product to prevent that from happening. Either a warning or an actual prevention would be appropriate. Instead of the "CAS drift" message, I could see us showing an alert that says "looks like you're trying to do something we don't support" so that the user can self-remedy
2 - We will have to continue to support non-LWW buckets on both ends of the stream between <4.6 and >=4.6. Given that we are still tracking the CAS drift exceptions even for non-LWW buckets, we should ensure that there are no unforeseen consequences of leaving XDCR running for a long time and potentially receiving an infinite number of these exceptions in the stats, buffers, etc.
Attachments
Issue Links
For Gerrit Dashboard: MB-21599 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
70015,3 | MB-21599: Add new *_with_meta options | watson | memcached | Status: MERGED | +2 | +1 |
70016,12 | MB-21599: Prevent LWW XDCR from non LWW cluster | watson | ep-engine | Status: MERGED | +2 | +1 |
70247,1 | Merge remote-tracking branch 'couchbase/watson' | master | memcached | Status: MERGED | +2 | +1 |
70250,1 | Merge remote-tracking branch 'couchbase/watson' | master | ep-engine | Status: MERGED | +2 | +1 |