-
Type:
Improvement
-
Resolution: Done
-
Priority:
Major - P3
-
None
-
Affects Version/s: None
-
Component/s: None
-
None
-
Server Programmability
-
Workload Resilience 2025-11-24, Workload Resilience 2025-12-08, Workload Resilience 2025-12-22
-
None
-
None
-
None
-
None
-
None
-
None
-
None
When part of the work for an operation requires running a nested operation inside the same server, we should always use DBDirectClient for that. Currently, there are other places in non-test code that call ServiceEntryPointCommon::handleRequest() directly:
- in ClusterSEPTransactionClientBehaviors::handleRequest()
- in UMCTransaction::runCommand()
- in DefaultSEPTransactionClientBehaviors::handleRequest()
- in runCommandInLocalTxn() in sharding_catalog_manager.cpp
- in commitOrAbortTransaction() in sharding_catalog_manager.cpp
- in AsyncWorkScheduler::scheduleRemoteCommand()
- related to
-
SERVER-119078 Have a clean interface for "peer" cleanup sub-operations
-
- Backlog
-
-
SERVER-119080 Improve management of operations generated by long-lived tasks.
-
- Backlog
-
-
SERVER-119082 Improve interface for reentrant "remote" command requests
-
- Backlog
-