Details
-
Improvement
-
Resolution: Unresolved
-
Major
-
2.5.0
-
Security Level: Public
Description
for the recent issues we have seen we need a tool that cam compare metadata (specifically revids) for a given replication definition in XDCR. To scale to large data sizes, being able to do this per vbucket or per doc range would be great but we can do without these. for clarity, here is a high level desc.
Ideal case:
xdcr_compare cluster1_connectioninfo cluster1_bucketname cluster2connectioninfo cluster2_bucketname [vbucketid] [keyrange]
should return a line per docid for each row where cluster1 metadata and clustermetadata for the given key differ.
docID - cluster1_metadata cluster2_metadata
simplification: the tool is expected to return false positives in a moving system but we will tackle that by rerunning the tool multiple times.