What's new in RocksDB2.7

Public API changes

• Removed PrefixHashRepFactory memtable factory. Please use NewHashSkipListRepFactory instead.
• Renamed StackableDB::GetRawDB() to StackableDB::GetBaseDB(). The call returns the underlying DB* for a given StackableDB. Before, GetRawDB() returned the raw DB* by stripping away all StackableDB layers.
• Support multi-threaded EnableFileDeletions() and DisableFileDeletions() - b60c14f
• Added DB::GetOptions() - 3ce3658
• Added DB::GetDbIdentity() - 1880268](https://github.com/facebook/rocksdb/commit/18802689b8081c4eaa9b477cf8fe1abddc887a3e)

New features

• Tailing iterator - wiki page
• Implemented BackupableDB - wiki page
• Cache index and filter block in block cache - turned off by default
• Put with SliceParts - Variant of Put() that gathers output like writev(2) -- 8a46ecd

Performance improvements

• Huge benchmark performance improvements by multiple efforts. For example, increase in readonly QPS from about 530k in 2.6 release to 1.1 million in 2.7 [1]
• Speeding up a way RocksDB deleted obsolete files - no longer listing the whole directory under a lock -- decrease in p99
• Use raw pointer instead of shared pointer for statistics: 5b825d -- huge increase in performance -- shared pointers are slow
• Optimized locking for get -- 1fdb3f -- 1.5x QPS increase for some workloads
• Cache speedup - e8d40c3
• Implemented autovector, which allocates first N elements on stack. Most of vectors in RocksDB are small. Also, we never want to allocate heap objects while holding a mutex. -- c01676e4
• Lots of efforts to move malloc, memcpy and IO outside of locks

[1] You can see the exact parameters in rocksdb/build_tools/regression_build_test.sh