Hey Gester,

At second glance, I see this comment regarding the retry logic in AcceptorImpl:

```
// fix for bug 36617. If BindException is thrown, retry after
// sleeping. The server may have been stopped and then           
// immediately restarted, which sometimes results in a bind exception          
```

This makes me a bit nervous about bypassing that logic in a GatewayReceiverImpl 
context, since it is just a subclass of CacheServerImpl and may be subject to 
the same issue of encountering bind exceptions on server restart.

I'd vote we remove the changes I suggested in AcceptorImpl after looking at 
this more closely.  Your other changes should reduce the chance of collision 
dramatically so I think it's unlikely a customer would encounter the 2 minute 
retry.

[ Full content available at: https://github.com/apache/geode/pull/2425 ]
This message was relayed via gitbox.apache.org for [email protected]

Reply via email to