-
Type: Improvement
-
Resolution: Unresolved
-
Priority: Major - P3
-
None
-
Affects Version/s: None
-
Component/s: None
-
Replication
-
200
In BF-33620 there is a race between updateTerm and getTerm. The updateTerm thread was holding the replication coordinator mutex but the getTerm thread released the mutex before calling getTerm. There is another getTerm call earlier in the function that was holding the mutex. Maybe it would be good to have a guardrail for fields like term in topology coordinator so that we are always accessing it while holding the mutex?
- is related to
-
SERVER-91374 Fix data race between getTerm and updateTerm in _doReplSetReconfig
- Closed