Details
Description
If an exception, e.g. SQLException occurs during a wite-behind operation, the exception is lost and appears nowhere in the logs.
The logs basically say:
[2017-06-28 10:48:52] SEVERE: Unable to update underlying store: CacheJdbcPojoStore []
[2017-06-28 10:48:52] WARNING: Failed to update store (value will be lost as current buffer size is greater than 'cacheCriticalSize' or node has been stopped before store was repaired) [key=test.WriteBehindIssue$FooKey [idHash=524852907, hash=-1388553726, id=1], val=test.WriteBehindIssue$Foo [idHash=323347362, hash=574522785, value1=abcd, value2=1234], op=PUT]
If we do the same with write-through mode, it fails by throwing an exception with the SQLException as cause, deep in the trace.
Would it be possible the write-behind operation logs the SQLException, and other exception?
You can find a test case showing that here: https://github.com/bgaraude/IgniteTest