Details
-
Improvement
-
Status: Resolved
-
Major
-
Resolution: Duplicate
-
None
-
None
-
None
-
None
Description
For user initiated compactions, I would like to be able to limit the number of threads that run compactions on each tablet server. My use case is that for Fluo tables I need to periodically compact ranges of a table to garbage collect Fluo transactional data. However, I would like these compactions to have minimal impact.
Being able to do the following would be nice, where the -m options is the max number of compactions to run on any tablet server.
compact -t myTable -b p: -e p:~ -m 1
Thinking the best implementation would be to push this information to the tablet server and let it manage, rather than having the master try to coordinate.
Attachments
Issue Links
- is related to
-
ACCUMULO-4187 Rate limiting of major compactions
- Resolved