Hi Everyone, 1. We are currently facing a data discrepancy issue where a UDT (User-Defined Type) column is returning different values across multiple data centers. We are running on DSE 6.9.6 on Cassandra 3.11 2. To resolve this, we have already attempted a full repair and a -pr (primary range) repair on the affected table, but the data inconsistency still persists. Interestingly, when we run a targeted repair using nodetool repair -st <start_token> -et <end_token> for a specific partition, the data is successfully repaired. 3. Upon reviewing the repair logs, we did not encounter any explicit errors. However, we noticed that certain ranges were skipped during the repair process, as indicated by the following messages in the logs.
INFO [RepairJobTask:1]2025-06-03 21:04:17, 621 SyncTask.java:131 - [repair#3b812a50-40be-11f0-b089-795611d3898f6] Endpoints /x.x.x.x and x.x.x.x have 3 range(s) out of sync for table (3 and 0 ranges skipped respectively) 4. Just wondering why repair would skip these token ranges. Please let us know your suggestions