Implement physiological logging

Implementation of physiological logging should improve the speed of writes because of the following factors:

1. Fewer data will be logged on the WAL level.
2. The overhead of tracking of logical operations is much less than for tracking of binary operations.

Following subtasks should be impelemented:

• Implementation of doulbe write log. So pages written first into this log and only after that to the data files
• Implementation of support physiological logging for OPaginatedClusterV1
• Implementation of support physiological logging for OPaginatedClusterV2
• Implementation of support physiological logging for OPaginatedClusterV3
• Implementation of support physiological logging for OSBTreeV1
• Implementation of support physiological logging for OSBTreeV2
• Implementation of support physiological logging for CellBTreeSingleValueV1
• Implementation of support physiological logging for CellBTreeSingleValueV3
• Implementation of support physiological logging for CellBTreeMultiValueV2
• Implementation of support physiological logging for CellBTreeMultiValueV3
• Implementation of support physiological logging for OSBTreeBonsaiLocal
• Implementation of support physiological logging for OLocalHashTableV2
• Implementation of support physiological logging for OLocalHashTableV3
• In case of not high-level exception is thrown inside of component storage should be switched to read only mode.
• Replace file truncation by the algorithm of reuse of pages in hash index.
• Create restore procedure based on physiological logging.
• Run crash tests
• Run benchmarks
• Handle binary compatibility of incremental backups between 3.0 and 3.1 versions in WAL