XDCR New Feature in CB 7.1

Hi All,

While going through new features of XDCR in CB 7.1 in official doc .
I did not find any more details related to XDCR.IS there any blog or other links. Please share .


Are you asking about this in the What’s New in Version 7.1?

XDCR checkpointing is now entirely persistent through topology-changes on the source cluster. This provides improved performance when failover and rebalance occur on the source cluster.

This is an internal change. Previous to 7.1, XDCR checkpoints (which exist on the source cluster) were not saved to replica vBuckets (behavior since beginning of XDCR). This meant that on a source cluster topology change (i.e. rebalance or failover), XDCR checkpoint information (which is kept for each vBucket) may not have been available for all active vBuckets (when replicas are promoted to active, etc.) – in such cases, XDCR needs to recheck if the affected documents need to be replicated to the target, which can be resource intensive. In 7.1, XDCR will push the checkpoints from active vBuckets to the replicas on a periodic basis. So, when a failover occurs, for example, once the replica vBuckets gets promoted to the active, they’ll have the XDCR checkpoints and be able to resume replication, and the previous behavior of rechecking items will be minimized. Either way (previous to or after this change), XDCR automatically handles cluster topology changes (either on the source or the target clusters). In 7.1 and above, however, XDCR behavior on source cluster topology changes should show improved performance (less resource consumption), especially on large systems with billions of documents. On smaller clusters, the change may not be as noticeable.

Thanks @hyunjuV for you reply.