> On Oct. 21, 2016, 10:25 p.m., Bruce Schuchardt wrote:
> > is it possible to unit test this change?  
> > PdxClientServerDUnitTest.testNonPersistentServerRestartAutoSerializer ran 
> > into the problem but seems to be marked Flaky.  Can you modify that test to 
> > always hit the problem and remove the Flaky label?

Possibly we can do that by introducing some testHook and call from 
PdxRegistoryRecoveryListener and then just wait "before clearing the pdx 
Registry".


- Hitesh


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/53092/#review153605
-----------------------------------------------------------


On Oct. 21, 2016, 6:52 p.m., Hitesh Khamesra wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/53092/
> -----------------------------------------------------------
> 
> (Updated Oct. 21, 2016, 6:52 p.m.)
> 
> 
> Review request for geode, Bruce Schuchardt, Udo Kohlmeyer, and Dan Smith.
> 
> 
> Repository: geode
> 
> 
> Description
> -------
> 
> We clear pdx registry in client when we re-connect to cluster. But during 
> that time other thread may end up using old registry while other thread is 
> clearing this. Thus we need to synchronize that. In current code it happens 
> through listener events and I modified that notification to synchronize this.
> 
> 
> Diffs
> -----
> 
>   
> geode-core/src/main/java/org/apache/geode/cache/client/internal/EndpointManagerImpl.java
>  3f3d725 
> 
> Diff: https://reviews.apache.org/r/53092/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Hitesh Khamesra
> 
>

Reply via email to