virajjasani commented on a change in pull request #939: HBASE-23349 : Config based Scanner reset after compaction if low refCount is preventing archival of compacted away store files URL: https://github.com/apache/hbase/pull/939#discussion_r364571746
########## File path: hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/StoreScanner.java ########## @@ -157,6 +158,15 @@ private final ReentrantLock flushLock = new ReentrantLock(); // lock for closing. private final ReentrantLock closeLock = new ReentrantLock(); + // scanner reset lock - ensure scanners are reopened only on new store files(non compacted) + private final ReentrantLock compactionLock = new ReentrantLock(); Review comment: The reason why I thought of this is because `flushLock` in updateReaders also uses `closeLock` internally and try to clear and close memStoreScanners if required. However, in case of Compaction, this would not be the case. Now that I think of this again, I think it's good to use same lock because the ultimate goal is Scanner reset which should follow same lock for flush and compaction even though an extra lock is being used within flushLock. ---------------------------------------------------------------- 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] With regards, Apache Git Services
