We are observing Intermittent empty results from Iceberg table queries Environment: StarRocks 4.0.6, 5 FE nodes, REST Iceberg catalog Symptom: The same query against the same table intermittently returns 0 rows. Re-establishing the MySQL connection sometimes fixes it (presumably landing on a different FE). The table has not been updated — same snapshot is used in both cases. Root cause evidence from query profiles: - Both queries use snapshot 5731107429248102568 with identical predicates - Failing query: resultDataFiles=0, skippedDataFiles=414, totalFileSizeInBytes=0 - Working query: resultDataFiles=4, skippedDataFiles=9345, totalFileSizeInBytes=1.2GB - Both scanned 2 data manifests, skipped 1 — but the failing query only enumerated 414 files from those manifests vs 9,349 on the working query - ICEBERG.hasMoreOutput iterations: 120 (failing) vs 131 (working) - REFRESH EXTERNAL TABLE did not resolve the issue Is anyone else experiencing this on 4.0.6? Is it possible that the Iceberg manifest cache is incomplete/corrupt on one front-end? Anything we should look for specifically?