-
Type: Bug
-
Resolution: Duplicate
-
Priority: Major - P3
-
None
-
Affects Version/s: 6.0.15
-
Component/s: None
-
None
-
Storage Execution
-
ALL
-
Execution Team 2024-06-24, Execution Team 2024-08-19, Execution Team 2024-09-02, Execution Team 2024-09-16
In REP-4527, we are hitting a fatal assertion:
[j7:cl1:n1] | 2024-05-15T21:37:17.404+00:00 I STORAGE 20276 [BackgroundSync] "openCatalog: reopening all databases" [j7:cl1:n1] | 2024-05-15T21:37:17.409+00:00 I STORAGE 20277 [BackgroundSync] "openCatalog: updating cached oplog pointer" [j7:cl1:n1] | 2024-05-15T21:37:17.422+00:00 F ASSERT 23095 [BackgroundSync] "Fatal assertion","attr":{"msgid":31179,"error":"UnsupportedFormat: Index: {name: x_1, ns: test.transactions_write_conflicts_unique_indexes} has incompatible format version: 8","file":"src/mongo/db/storage/wiredtiger/wiredtiger_index.cpp","line":742}
I believe it is caused by this command, which occurs right before a stepdown cycle:
[j7:cl1:n1] | 2024-05-15T21:37:12.232+00:00 I COMMAND 5324200 [conn359] "CMD: collMod","attr":{"cmdObj":{"collMod":"transactions_write_conflicts_unique_indexes","index":{"name":"x_1","forceNonUnique":true}}} [ClusterToClusterReplication:job7] 21:37:13.541Z Waiting for thread0 to report state RUNNING (after REVERSING) [CyclicStepdown:job7] 21:37:13.755Z The last stepdown for this replicator process was 13.5 seconds ago. [CyclicStepdown:job7] 21:37:13.755Z Starting stepdown of all primaries
- is fixed by
-
SERVER-68186 Reconcile index table metadata during startup if there is a mismatch caused by collMod
- Closed
- is related to
-
SERVER-68216 Investigate atomicity issue with wiredtiger alter
- Open
-
SERVER-63578 Convert a unique index to a non-unique index via the collMod command
- Closed