Knut Anders Hatlen wrote:
Kathey Marsden <[email protected]> writes:
I see the following failures with 10.6.0.0 alpha - (891102)**
**with IBM 1.5 and 1.6.
Any ideas?
There were 10 errors:
1)
testLargeMultiByteCharLob(org.apache.derbyTesting.functionTests.tests.memory.MultiByteClobTest)java.sql.SQLException:
Log Record has been sent to the stream, but it cannot be applied to the store
(Object null). This may cause recovery problems also.
A long shot, but... Could this be another manifestation of DERBY-4463?
It's the same JVMs, and I guess a stray interrupt could cause such
failures if it happened to interrupt the writing of a log record?
I am looking back at the previous records and see it happened on the
seventh and the third of December.
http://people.apache.org/~myrnavl/derby_test_results/main/windows/testSummary-888230.html
http://people.apache.org/~myrnavl/derby_test_results/main/windows/testSummary-887040.html
I did not see it in the three weeks prior. I think the offending test
of DERBY-151 which triggered DERBY-4463 was checked in Dec 3, so I
think yours is a good theory.
Since this new manifestation seems intermittent. I think we should:
1) File a separate bug for this issue and link it to DERBY-4463 and
DERBY-151.
2) Disable the Derby151Test for the IBM vm's for now as Dag suggested in
DERBY-4463.
3) Focus fix efforts on DERBY-4463 as it seems more reliably reproducible.
Thoughts?
Kathey
