Details
-
Task
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Won't Do
-
None
Description
MCOL-5587 - discovery place. The reproduction there can be used here, one iteration and 10K-1M records will suffice.
Running heaptrack on the mariadb test -e "select * from cptest; shows about three stack traces that have peak consumption that grows linearly with the number of rows returned. This is due to logic that results are stored for later consumption if command line option "-q" is not given.
I think that some memory threshold on the results' storage would be nice.