cshannon commented on issue #608:
URL: https://github.com/apache/accumulo/issues/608#issuecomment-1407425748

   That is an interesting idea for the logic for the goodTransition method, 
later this week I can look into that more and test reversing it and see what 
happens.
   
   I ran some tests yesterday and today trying to break things and so far the 
iterator has detected all the errors and reset properly. I ran tests with rapid 
GC happening in a loop concurrently while the metadata table was being split 
frequently (due to it being configured to a very small size) and also with in 
memory caching turned off. This produced a lot of messages in the logs where 
the iterator detected problems do to the splits, etc but it always reset itself 
properly. Each time the correct rfiles were GC'd and nothing that was removed 
that shouldn't.
   
   I also looked at exception handling and bubbling up from a server side 
iterator to the client and that appears to work. There's actually an 
integration test for this 
[ServerSideErrorIT](https://github.com/apache/accumulo/blob/ab8e2f518887a2505cad9dbc02e82197babfc1bd/test/src/main/java/org/apache/accumulo/test/functional/ServerSideErrorIT.java)
   
   I'll continue looking at this more next week. My goal is to try and find a 
flaw somewhere with a test case because if we can reproduce an error with 
testing this will be a lot easier to track down.
   


-- 
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.

To unsubscribe, e-mail: [email protected]

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

Reply via email to