dlmarion commented on PR #4787:
URL: https://github.com/apache/accumulo/pull/4787#issuecomment-2269330995

   > Hmm thinking about this some more. Wondering if it would be better if we 
had finite number of retries for both stores instead of infinite. The 
probability of collision is very low.
   > 
   > I'm not sure that both stores are functioning the same under these new 
changes. The MFS will throw an error if any other KeeperException is seen other 
than NodeExistsException. The UFS will retry forever if an UNKNOWN status is 
received. So if there is something wrong and we keep receiving UNKNOWN, it will 
keep retrying.
   
   So, I first went down the path of creating an IT to determine what error a 
user would receive when something like `TableOperations.create` failed when the 
number of retries in the UserFateStore would be exceeded. When I realized that 
MetaFateStore tried infinitely, I looked at how we could test for a duplicate 
FateId in the ConditionalMutationWriter. I didn't see an easy way to do that 
since the FateId is in the `row`.  If we could do that, then we could set a 
status for duplicate FateId, and continue to retry, and then fail on REJECTED 
or UNKNOWN. 


-- 
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