keith-turner opened a new pull request, #3875:
URL: https://github.com/apache/accumulo/pull/3875

   Added a new FATE repo that acquires operations ids on all of the tablets in 
the merge range.  This causes the tablet group watcher to unload tablets.  This 
new repo waits for all tablets to not have a location.
   
   Modified the existing merge and delete rows FATE repos to use ample and 
conditional mutations.  Using ample simplified the code.
   
   Modified the delete rows FATE repo to only delete or fence files, but not 
merge tablets.  Had the code call the merge code to delete tablets. This code 
is working (its deleting the expected data) but its producing different splits 
than before.  The code is much much simpler now that delete rows is not merging 
tablets, need to figure out why the splits differ.
   
   Removed all merge code from tablet group watcher and related classes that 
was related to unassigning merging tablets.  Setting the operation ids takes 
care of this now.
   
   Modified the final FATE repo to delete operation ids.
   
   MergeIT and DeleteRowsIT are passing with this change, however had to 
comment out the checks of expected splits in DeleteRowsIT because of the issue 
mentioned earlier. Still looking into this issue.


-- 
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