runzhiwang commented on a change in pull request #317:
URL: https://github.com/apache/incubator-ratis/pull/317#discussion_r535198688



##########
File path: 
ratis-server/src/main/java/org/apache/ratis/server/raftlog/segmented/SegmentedRaftLog.java
##########
@@ -388,10 +388,12 @@ public TermIndex getLastEntryTermIndex() {
     try(AutoCloseableLock writeLock = writeLock()) {
       validateLogEntry(entry);
       final LogSegment currentOpenSegment = cache.getOpenSegment();
+      // The stateMachineData will be cached inside the StateMachine itself 
when enable stateMachineCaching.
+      LogEntryProto toAppendEntry = stateMachineCachingEnabled? 
ServerProtoUtils.removeStateMachineData(entry) : entry;
       if (currentOpenSegment == null) {
         cache.addOpenSegment(entry.getIndex());
         fileLogWorker.startLogSegment(entry.getIndex());
-      } else if (isSegmentFull(currentOpenSegment, entry)) {
+      } else if (isSegmentFull(currentOpenSegment, toAppendEntry)) {

Review comment:
       @szetszwo The problem is cache.appendEntry(entry) will put entry with 
StateMachineData into LogSegment#entryCache. For example, each entry occupy 8KB 
after remove StateMachineData, each StateMachineData is 80M, the max 
segment.size.max is 8MB, when `isSegmentFull` returns true, open segment has 
1000 entry,  and LogSegment#entryCache is 80M * 1000, i.e. 80GB. 
   
   ```
          if (stateMachineCachingEnabled) {
            // The stateMachineData will be cached inside the StateMachine 
itself.
   -        cache.appendEntry(ServerProtoUtils.removeStateMachineData(entry));
   +        cache.appendEntry(removedStateMachineData));
          } else {
            cache.appendEntry(entry);
          }
   ```
   
   
   




----------------------------------------------------------------
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]


Reply via email to