albertogpz commented on pull request #697:
URL: https://github.com/apache/geode-native/pull/697#issuecomment-734421299


   > This is passing on Windows now, and I have at least an example of it 
working on RHEL7 Release build. Our Debug and Coverage builds of RHEL7 appear 
to now be timing out on a test, though, which has been a problem in the past. 
Here's the tail end of the output, which looks kind of strange to me.
   > 
   > ```
   > [       OK ] 
PartitionRegionOpsTest.putPartitionedRegionWithRedundancyServerGoesDownSingleHop
 (60480 ms)
   > [----------] 1 test from PartitionRegionOpsTest (60480 ms total)
   > 
   > [----------] Global test environment tear-down
   > [==========] 1 test from 1 test suite ran. (60481 ms total)
   > [  PASSED  ] 1 test.
   > [debug 2020/11/25 21:03:47.396293 UTC 
heavy-lifter-80e74939-d075-5450-9833-2e242c016911:1522 139806642039104] 
ThinClientPoolDM::destroy...
   > [debug 2020/11/25 21:03:47.396528 UTC 
heavy-lifter-80e74939-d075-5450-9833-2e242c016911:1522 139806642039104] 
Internal fair queue size while closing is 0
   > [debug 2020/11/25 21:03:47.396541 UTC 
heavy-lifter-80e74939-d075-5450-9833-2e242c016911:1522 139806642039104] 
ConnectionQueue::close( ): queue closed 
   > [debug 2020/11/25 21:03:47.396559 UTC 
heavy-lifter-80e74939-d075-5450-9833-2e242c016911:1522 139806642039104] 
recursive lock: CacheImpl::doIfDestroyNotPending
   > 
   > 
   > 0% tests passed, 1 tests failed out of 1
   > 
   > Total Test time (real) = 650.16 sec
   > 
   > The following tests FAILED:
   >     27 - 
PartitionRegionOpsTest.putPartitionedRegionWithRedundancyServerGoesDownSingleHop
 (Timeout)
   > Errors while running CTest
   > ```
   
   Interesting... @gaussianrecurrence told me that he had seen a deadlock in 
the feature/asio branch once related to that mutex but had not been able to 
reproduce it again. We looked together at the code today and it appeared that a 
deadlock is possible after the change included by @mreddington about which I 
him asked above. We could try removing the call to 
CacheImpl::doIfDestroyNotPending in TcrEndpoint::closeConnections() to see if 
the timeout does not occur.


----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
[email protected]


Reply via email to