-
Type:
Bug
-
Resolution: Unresolved
-
Priority:
Minor - P4
-
None
-
Affects Version/s: None
-
Component/s: None
-
None
-
Query Execution
-
ALL
-
17
Problem 1:
We can execute a very large query successfully, even though its explain fails with BSONObjectTooLarge
Problem 2:
A very large query might succeed or fail sporadically because of the following LOGV2_WARNING call in find_cmd.cpp requiring the explain output (which exceeds 16MB):
try { numResults = batchedExecute(batchSize, exec.get(), firstBatch, docUnitsReturned); } catch (DBException& exception) { firstBatch.abandon(); auto&& explainer = exec->getPlanExplainer(); auto&& [stats, _] = explainer.getWinningPlanStats(ExplainOptions::Verbosity::kExecStats); LOGV2_WARNING(23798, "Plan executor error during find command", "error"_attr = exception.toStatus(), "stats"_attr = redact(stats), "cmd"_attr = cmdObj); exception.addContext(str::stream() << "Executor error during find command: " << nss.toStringForErrorMsg()); throw; }
It is unlikely that customers will be running so large queries (therefore priority = minor), but it is still an inconsistency.
Possible solutions could be:
- allow explain output exceed 16MB
- fail large queries early and consistently if their explain output is above 16MB