StevenLuMT commented on code in PR #3378:
URL: https://github.com/apache/bookkeeper/pull/3378#discussion_r926911566
##########
bookkeeper-server/src/main/java/org/apache/bookkeeper/bookie/storage/ldb/SingleDirectoryDbLedgerStorage.java:
##########
@@ -401,24 +401,7 @@ public long addEntry(ByteBuf entry) throws IOException,
BookieException {
log.debug("Add entry. {}@{}, lac = {}", ledgerId, entryId, lac);
}
- // First we try to do an optimistic locking to get access to the
current write cache.
- // This is based on the fact that the write cache is only being
rotated (swapped) every 1 minute. During the
- // rest of the time, we can have multiple thread using the optimistic
lock here without interfering.
- long stamp = writeCacheRotationLock.tryOptimisticRead();
- boolean inserted = false;
-
- inserted = writeCache.put(ledgerId, entryId, entry);
- if (!writeCacheRotationLock.validate(stamp)) {
- // The write cache was rotated while we were inserting. We need to
acquire the proper read lock and repeat
- // the operation because we might have inserted in a write cache
that was already being flushed and cleared,
- // without being sure about this last entry being flushed or not.
- stamp = writeCacheRotationLock.readLock();
- try {
- inserted = writeCache.put(ledgerId, entryId, entry);
- } finally {
- writeCacheRotationLock.unlockRead(stamp);
- }
- }
+ boolean inserted = writeCache.put(ledgerId, entryId, entry);
Review Comment:
@eolivelli @lordcheng10
I think this lock is necessary
when swapWriteCache cost long time, during this time,
writeCacheRotationLock.writeLock takes effect
we need to care this action, block writeCache's put action, prevent more
write pressure on writecache
--
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]
