Distributed shared file system based on block storage

Abstract

The invention discloses a distributed shared file system based on block storage. The system comprises a metadata module, a cluster management module, a node manager, a storage heartbeat wire, a network heartbeat wire, a file system isolating module, a distributed lock management module, a user space and kernel space interface and a standard file system interface, wherein the metadata module is responsible for the synchronization of nodes of the metadata. All the nodes share the read-write data in parallel, so that the purpose of data sharing can be achieved by quickly synchronizing the metadata. According to the invention, the multipoint simultaneous mounting is realized; the mounting mode is simple; no transfer node is required; after the nodes in the cluster are directly mounted and stored in blocks, the distributed shared file system can realize a data sharing function; no single point failure problem exists; any nodes in the cluster have no influence on a cluster system; a calculating node in each cluster can directly access the sharing storage; and the high-performance parallel read-write is realized.

Description

technical field [0001] The invention relates to a shared file system, in particular to a distributed shared file system based on block storage. Background technique [0002] Due to the emergence of cloud computing, the computing power of servers is no longer limited to a single node. Computing pools composed of cluster nodes provide virtualization, cloud desktop and other business carriers and become the mainstream. However, there are many forms of combining computing pools and data carrier storage, which are not uniform and have many problems. The birth of the distributed shared file system based on block storage solves many defects in the traditional system: First, it cannot be mounted and read and written by multiple nodes at the same time. Second, the mounting method is cumbersome. After mounting to the node (transit node) through the iscsi protocol, format the local file system, export it into a traditional shared file system such as NFS or cifs, and then mount it to m...

AI Technical Summary

Problems solved by technology

However, the combination of computing pool and data carrier storage is varied, not uniform, and there are many problems

The birth of the distributed shared file system based on block storage solves many defects in the traditional system: First, it cannot be mounted and read and written by multiple nodes at the same time

Second, the mounting method is cumbersome. After mounting to the node (transit node) through the iscsi protocol, format the local file system, export it into a traditional shared file system such as NFS or cifs, and then mount it to multiple nodes (computing nodes in the cluster) Realize the sharing function

3. There is a single point of failure problem in the transit node. If a downtime occurs, it will affect the storage access of computing nodes in the entire cluster.

Fourth, there is a performance bottleneck in the transit node. All data reads and writes must be converted by the transit node and finally written to storage. The network interface, CPU, memory and other configurations of the transit node will become bottlenecks.

Images