Github user joshelser commented on the pull request:
https://github.com/apache/accumulo/pull/54#issuecomment-160762876
> Can you expand on this? I am really curious about the behavior of thrift
when a oneway throws an exception on the server side.
Best as I understand it: the oneway modifier just equates to the generated
code not calling the method to read the response for an RPC off the wire. This
has multiple implications:
1. The caller does not know if the application received the message (only
that the network connection succeeded).
2. The caller does still receive a response object (this is how void works
though, not oneway)
3. If you try to run another RPC that isn't oneway on the same connection,
it's possible that the synchronous call will get goofed up by that oneway's
void response (this might be a Thrift bug -- it's certainly a pain to work
around).
Let me test what actually happens in the client code when the oneway throws
an exception. I have a little test harness that I used to play around with this.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
