[GitHub] [hadoop] steveloughran commented on issue #666: HADOOP-16221 add option to fail operation on metadata write failure

[GitBox <gi...@apache.org>]

steveloughran commented on issue #666: HADOOP-16221 add option to fail operation on metadata write failure
URL: https://github.com/apache/hadoop/pull/666#issuecomment-478079151
 
 
   right, if you are getting into the failure handling, you are going to have to look closely at the invoker retry stuff and those @Retry annotations we do our best to keep up to date.
   
   This means: any change to stop swallowing DDB exceptions will need changes in the annotation values and commentary on all uses of the {{finishedWrite()}} method, and review of "is everything consistent with my changes". Yes, it's a pain, yes it's manual, but its there to help us all understand how things fail, what kind of exceptions get raised, etc. It was that or javadocs: at least with these annotations you could imagine doing some graph of invocations
   
   * there must not be retries around retries; {{WriteOperationsHelper.finalizeMultipartUpload()}} is doing retries though. It may be best to pull the finishedWrite there into its own once() call
   * I'm also not sure if we want to wrap the IOE with another IOE, not unless we really think it makes sense either in retry handling or diagnostics. Normally I'd say "don't do at all", but this time I'm open to discussion, precisely because it lets us change policies. But
   
   

----------------------------------------------------------------
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.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services

---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org