How to compare data cluster to cluster?
Stop updates to all clusters, wait for XDR lag to go to 0 and then scan partition by partition comparing all clusters.
Depends on what you’re comparing.
If it’s checking that all the records exist in both, when you’re scanning by partition you could choose to just get the digests back (‘no bins’ scan policy), and compare that the digests in one cluster match the other. You could turn batches of digests from cluster 1 into batch exists operations against cluster 2.
Or are you talking about comparing the actual data? You might use asbackup of both to compare. I’m not sure why you would need to do this.
These suggestions will not work for typical cases where new keys are being written and old keys are being deleted (through client deletes, expiration, or eviction). In typical scenarios you will find some keys in A not in B and vice versa.
Thank you ! But how to form those keys and compare with both clusters. Iam actually comparing the data.