-
Type:
Bug
-
Resolution: Unresolved
-
Priority:
Major - P3
-
None
-
Affects Version/s: None
-
Component/s: None
-
None
-
Storage Execution
-
ALL
-
-
Storage Execution 2026-06-08, Storage Execution 2026-06-22, Storage Execution 2026-07-06, Storage Execution 2026-07-20, Storage Execution 2026-08-03, Storage Execution 2026-08-17
-
None
-
None
-
None
-
None
-
None
-
None
-
None
In SERVER-125800, I added the following log to initializeMetadata():
LOGV2_WARNING( 12580001, "Replicated fast count eligible namespace found in the collection " "catalog but not tracked in the metadata store or the oplog", "nss"_attr = coll->ns().toStringForErrorMsg(), "uuid"_attr = coll->uuid());
We expect this log to not be triggered. In general, we expect that the accumulator map in initializeMetadata() contains entries for every replicated fast count eligible collection in the catalog except the oplog. On ASC, the oplog has a different UUID on every node, so we won't have the old oplog UUID in the catalog.
However, in practice, there are sometimes five collections absent from accumulator:
1 [js_test:catalog_shard] c20042| {"t":{"$date":"2026-05-08T11:02:23.081-04:00"},"s":"E", "c":"STORAGE", "id":12580001,"svc":"S", "ctx":"initandlisten","msg":"Replicated fast count eligible namespace found in the collection catalog but not tracked in the metadata store or the oplog","attr":{"nss":"admin.system.keys","uuid":{"uuid":{"$uuid":"60b7aa30-b63d-47f9-a39a-86b455440914"}}}}
2 [js_test:catalog_shard] c20042| {"t":{"$date":"2026-05-08T11:02:23.081-04:00"},"s":"E", "c":"STORAGE", "id":12580001,"svc":"S", "ctx":"initandlisten","msg":"Replicated fast count eligible namespace found in the collection catalog but not tracked in the metadata store or the oplog","attr":{"nss":"config.sampledQueries","uuid":{"uuid":{"$uuid":"44cfda13-dad8-4ef9-b79c-5fc8bf7c238f"}}}}
3 [js_test:catalog_shard] c20042| {"t":{"$date":"2026-05-08T11:02:23.081-04:00"},"s":"E", "c":"STORAGE", "id":12580001,"svc":"S", "ctx":"initandlisten","msg":"Replicated fast count eligible namespace found in the collection catalog but not tracked in the metadata store or the oplog","attr":{"nss":"config.reshardingOperations","uuid":{"uuid":{"$uuid":"a2dd6189-7481-4d34-8d86-6bcede80571a"}}}}
4 [js_test:catalog_shard] c20042| {"t":{"$date":"2026-05-08T11:02:23.081-04:00"},"s":"E", "c":"STORAGE", "id":12580001,"svc":"S", "ctx":"initandlisten","msg":"Replicated fast count eligible namespace found in the collection catalog but not tracked in the metadata store or the oplog","attr":{"nss":"config.sampledQueriesDiff","uuid":{"uuid":{"$uuid":"a3cbf51d-2721-4fad-baa6-9a3e738f4555"}}}}
5 [js_test:catalog_shard] c20042| {"t":{"$date":"2026-05-08T11:02:23.081-04:00"},"s":"E", "c":"STORAGE", "id":12580001,"svc":"S", "ctx":"initandlisten","msg":"Replicated fast count eligible namespace found in the collection catalog but not tracked in the metadata store or the oplog","attr":{"nss":"config.analyzeShardKeySplitPoints","uuid":{"uuid":{"$uuid":"e1808b48-6643-488c-b86a-fe3fddf90a77"}}}}
When an entry is not in accumulator, it means that (1) the fast count store does not contain an entry for that UUID and (2) the oplog (since the valid-as-of timestamp in the timestamp store) does not contain an entry for that UUID.
Investigate why the UUIDs for the above collections are not found during initialization. Is this expected? Should we convert the log to debug or remove it?
- is related to
-
SERVER-125800 Implement WiredTigerRecordStore size/count initialization
-
- Closed
-
-
SERVER-129848 Set beginFetchingTimestamp based on primary's fast count validAsOf timestamp
-
- Closed
-