-
Type: Task
-
Resolution: Fixed
-
Priority: Major - P3
-
Affects Version/s: None
-
Component/s: None
-
None
-
Fully Compatible
-
Sharding 2019-08-26, Sharding 2019-09-09
Transactions which are alive for too long are killed by the server's session management code (owned by the sharding team). Testing shows that this scenario can produce the following unclear error message:
Exec error resulting in state DEAD :: caused by :: operation was interrupted: generic server error
The former part of the message ("Exec error resulting in state DEAD") is context produced by query execution code. This error string has already been improved in SERVER-38316 and SERVER-41996. The latter part of the error message ("operation was interrupted: generic server error") describes the underlying cause of the error and is produced by the session/transaction management code. This message should be improved to indicate the reason for interruption (e.g. "transaction exceeded transactionLifetimeLimitSeconds").
- is duplicated by
-
SERVER-42820 Improve error message when transaction is killed due to exceeding timeout
- Closed
- related to
-
DRIVERS-2535 Add UnknownTransactionCommitResult error label to TransactionExceededLifetimeLimitSeconds errors during commitTransaction
- Backlog