Details
-
Bug
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Fixed
-
25.10
-
None
-
MXS-SPRINT-279
-
Can result in hang or crash
-
Description
A query against Exasol hangs MaxScale when the result set contains both a very large VARCHAR and a TIMESTAMP. The large VARCHAR disables batching, so rows are fetched column by column with SQLGetData, and the TIMESTAMP then truncates because the driver returns more characters than it declared. The retry loop in mxq::ODBC::get_normal_result() resizes the buffer to exactly the indicator value, leaving no room for the terminating NUL, so the driver truncates again and reports the same indicator. Nothing changes between iterations and the loop spins forever inside a driver call that holds a driver-internal mutex.
Attachments
Issue Links
- is duplicated by
-
MXS-7587 query not get translated and not make it to EXAsol instance
-
- Closed
-
- resolves
-
MCDEV-5387 Loading...