-
Type: Bug
-
Resolution: Done
-
Priority: Major - P3
-
Affects Version/s: None
-
Component/s: None
-
None
In WT-2534 we spent a lot of time adding diagnostics to track down a bug found when porting to a system with a weaker memory model than x86.
It would be helpful to have the following simple test for lost updates, which I believe would have failed on the test system:
create a table with one record, value_format=q, initial value zero In N threads: count = 0 while count < 1000: begin a transaction with isolation=snapshot read $value update with $value+1, rollback and continue on failure commit increment count Join all threads, verify that the value is 1000 * N.