keith-turner opened a new issue #2154:
URL: https://github.com/apache/accumulo/issues/2154


    The ample methods 
[here](https://github.com/apache/accumulo/blob/92a4dd3874b906a87166ae329c7f8d33e74a5a93/core/src/main/java/org/apache/accumulo/core/metadata/schema/Ample.java#L154),
 
[here](https://github.com/apache/accumulo/blob/92a4dd3874b906a87166ae329c7f8d33e74a5a93/core/src/main/java/org/apache/accumulo/core/metadata/schema/Ample.java#L140),
 and 
[here](https://github.com/apache/accumulo/blob/92a4dd3874b906a87166ae329c7f8d33e74a5a93/core/src/main/java/org/apache/accumulo/core/metadata/schema/TabletsMetadata.java#L307)
 all take KeyExtent.  These methods have the behavior that they may return 
tablet metadata for a different KeyExtent.  The impl looks up a tablet using 
only the table id and end row in the extent, so this could result in finding a 
tablet in the metadata table that has a different prev end row than the one 
passed to Ample.  This behavior seems confusing to me and like it could 
possibly lead to incorrect code.  Seems like it wold be better to return 
nothing in the
  case where the prev end rows do not match, which would be a simple change.  
However before this changes would need to survey all code using this methods 
and make sure the change in behavior does not break anything.  Noticed this 
while working on #2152.  If a changes is made, would also need to change the 
new method from #2152.


-- 
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:
[email protected]


Reply via email to