File System Mounting via Network Metadata Caching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing methods for mounting a file system are inefficient due to the time required to read metadata from disk, which is limited by the speed of the I/O interface, leading to decreased operating system efficiency in reading and writing operations.
Innovation Solution
A method and apparatus that request and cache metadata from a node with a storage stack over a network, utilizing higher network bandwidth than the I/O bandwidth of the storage stack, allowing for faster metadata retrieval and mounting of the file system without the need for disk or magnet seek operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If metadata is read from disk during file system mounting, then the file system can be mounted, but the mounting time is prolonged due to I/O interface speed limitations
Solution Approach 1:
The patent applies preliminary action by pre-reading and caching metadata from the disk into memory before the actual file system mounting operation. This allows the metadata to be readily available when mounting is needed, eliminating the time-consuming disk I/O operation during the critical mounting phase. The metadata is prepared in advance and stored in a cache structure that can be quickly accessed and parsed.
Solution Approach 2:
The patent creates a copy of the metadata from the disk and stores it in a cache in memory. Instead of reading the original metadata from the disk during mounting, the system uses this cached copy, which can be accessed much faster. This copying approach separates the slow disk I/O operation from the fast mounting operation, resolving the time contradiction.
2Reliability
If metadata is read sequentially and parsed one by one from disk, then the file system structure can be built, but the reading and writing efficiency of the operating system decreases
Solution Approach 1:
The patent performs preliminary reading and caching of all necessary metadata into memory before the file system mounting process begins. This preliminary action collects the metadata in a structured cache format, allowing subsequent parsing and file system construction to proceed much faster without repeated disk I/O operations, thereby improving overall reading and writing efficiency.
Solution Approach 2:
The patent creates a cached copy of the metadata in memory that can be efficiently parsed and processed. This copy eliminates the need for sequential disk reading during the mounting process, allowing the system to work with in-memory data structures that provide much faster access and processing speeds for building the file system structure.
Data Source
AI summary
The present invention belongs to the field of computer, and particularly provides a method, an apparatus and a system for mounting a file system. The method comprises: requesting, from a node mounted with a file system of a storage stack at the network distal end, for metadata corresponding to the mounting when the file system of the storage stack is to be mounted; receiving metadata transmitted by the node mounted with the file system of the storage stack, and caching the metadata in a memory; and mounting the file system of the storage stack according to the metadata; wherein the network bandwidth is larger than the IO bandwidth of the storage stack. The technical solutions provided by the present invention have the advantages that the time for mounting the file system is short and the reading and writing efficiency of the operating system is high.