Don't compact bottommost level in SuggestCompactRange

Summary: Before the fix we also marked the bottommost level for compaction. This is wrong because then RocksDB has N+1 levels instead of N as before the compaction.

Test Plan: SuggestCompactRangeTest in db_test

Reviewers: yhchiang, rven

Reviewed By: rven

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D37869

db/db_impl_experimental.cc

@@ -38,7 +38,7 @@ Status DBImpl::SuggestCompactRange(ColumnFamilyHandle* column_family,
   {
     InstrumentedMutexLock l(&mutex_);
     auto vstorage = cfd->current()->storage_info();
-    for (int level = 0; level < vstorage->num_non_empty_levels(); ++level) {
+    for (int level = 0; level < vstorage->num_non_empty_levels() - 1; ++level) {
       std::vector<FileMetaData*> inputs;
       vstorage->GetOverlappingInputs(
           level, begin == nullptr ? nullptr : &start_key,

db/db_test.cc

@@ -12547,7 +12547,7 @@ TEST_F(DBTest, SuggestCompactRangeTest) {
   options.compaction_style = kCompactionStyleLevel;
   options.write_buffer_size = 100 << 10;  // 100KB
   options.level0_file_num_compaction_trigger = 2;
-  options.num_levels = 3;
+  options.num_levels = 4;
   options.max_bytes_for_level_base = 400 * 1024;
 
   Reopen(options);