keith-turner opened a new pull request, #5675:
URL: https://github.com/apache/accumulo/pull/5675

   When attempting to merge #5588 to main, found a bug with the change. The bug 
was that the code did not consider transitive compactions when searching past 
currently running compactions.  This bug would not cause any errors, it was 
just suboptimal. After these changes no compactions that would transitively 
include any smaller files will be planned.
   
   For a tablet that has 6 1M files compacting, 5 7M files and 4 40M files the 
code without this fix would find the 4 40M files to compact. However if we wait 
for the compaction of the 5 7M files (which is waiting on the compaction of the 
6 1M files), then the 4 40M files would be compacted with the 41M file that 
would be produced. These changes would wait for that making the three 
compactions run sequentially so they can include each others output.
   
   A second test added in this change gives examples cases of compactions that 
can be found when there are running compaction.


-- 
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