Distributed system data access methods

Abstract

A distributed system data access method is implemented by a server device and multiple node devices. Each node device includes a file system and a unified mount. Each unified mount is used to integrate and manage its own file system and a network file system mounted from each of the other node devices. The method includes: (A) the server device generating and transmitting a query request containing a read path of data to be queried to a first node device; (B) the first node device determining, based on its file system, whether the data to be queried is stored in the first node device; (C) when it is determined that the data is stored in the first node device, the first node device obtains and transmits the data to be queried to the server device through the file system.

Claims

1. A distributed system data access method, implemented by a server device and a plurality of node devices that are signal-connected to each other and all signal-connected to the server device, each node device including a file system for managing multiple data stored therein and a unified mount, the unified mount of each node device for integrating and managing the file system corresponding to its own node device and a network file system mounted from each of the other node devices, each network file system being the file system of each of the other node devices, the distributed system data access method including the following steps: (A) generating and transmitting a query request including a read path of data to be queried to a first node device, the first node device being one of the node devices; (B) after receiving the query request, determining whether the data to be queried is stored in the first node device according to the file system of the first node device; and (C) When it is determined that the data to be queried is stored in the first node device, the first node device obtains and transmits the data to the server device through its own file system.

2. The distributed system data access method as described in claim 1, wherein, Following step (B), the following steps are also included: (D) When it is determined that the data to be queried is not stored in the first node device, the first node device queries and obtains a second node device that stores the data to be queried through its own joint mount, the second node device being one of the node devices other than the first node device; and (E) The first node device obtains and transmits the data to be queried to the server device through the network file system corresponding to the second node device in its own joint mount.

3. The distributed system data access method as described in claim 1, wherein, The distributed system data access method further includes the following steps: (F) generating and transmitting a write command containing data to be written and a write path to a third node device, wherein the third node device is one of the node devices; (G) upon receiving the write command, the third node device determines whether it is a node device to be written as indicated by the write path, wherein the node device to be written is one of the node devices; (H) when the third node device is determined to be the node device to be written, the third node device stores the data to be written through its own file system; and (I) when the third node device is determined not to be the node device to be written, the third node device stores the data to be written in the node device to be written through its own network file system corresponding to the node device to be written in its own joint mount.

4. The distributed system data access method as described in claim 1, wherein each node device is assigned a storage number and stores a cold data transfer time interval, and each storage number indicates an access frequency of the data stored by the corresponding node device, wherein, The distributed system data access method further includes the following steps: (J) For each node device, a cold data transfer time point is calculated by the node device based on a current time point and the cold data transfer time interval stored in the node device; (K) For each node device, the node device determines, through its own file system and the cold data transfer time point, whether there is at least one cold data to be transferred, wherein the last access time point of each cold data to be transferred is earlier than the cold data transfer time point; (L) For each node device, the node device copies the at least one cold data to be transferred to a fourth node device through its own joint mount, wherein the fourth node device is one of the node devices, and the access frequency indicated by the storage number of the fourth node device is less than the access frequency indicated by the storage number of its own node device; and (M) For each node device, the node device deletes the stored at least one cold data to be transferred through its own file system.

5. The distributed system data access method as described in claim 1, wherein each node device is assigned a storage number and stores a hot data transfer time interval, and each storage number indicates an access frequency of the data stored by the corresponding node device, wherein, The distributed system data access method further includes the following steps: (N) For each node device, a hot data transfer time point is calculated by the node device based on a current time point and the hot data transfer time interval stored in the node device; (O) For each node device, the node device determines, through its own file system and the hot data transfer time point, whether there is at least one hot data to be transferred, wherein the last access time point of each hot data to be transferred is later than the hot data transfer time point; (P) For each node device, the node device copies the at least one hot data to be transferred to a fifth node device through its own joint mounting, wherein the fifth node device is one of the node devices, and the access frequency indicated by the storage number of the fifth node device is greater than the access frequency indicated by the storage number of its own node device; and (Q) For each node device, the node device deletes the stored at least one hot data to be transferred through its own file system.

6. The distributed system data access method as described in claim 1, wherein, The distributed system data access method further includes the following steps: (R) For each node device, the node device periodically detects a device connection status between itself and each of the other node devices, each device connection status including a normal device connection status and a device disconnection status, wherein the normal device connection status indicates that the node device can normally connect to the corresponding other node device, and the device disconnection status indicates that the node device cannot connect to the corresponding other node device.

7. The distributed system data access method as described in claim 6, wherein each node device further includes a node storage module storing the data of the node device and managed by the file system, wherein, Following step (R), the following step is further included: (S) For each node device, by means of the node device, periodically detecting the state of its own node storage module and the storage module corresponding to each of the other node storage modules, each storage module state including one of a storage module disconnection state, a storage module inaccessible state, and a storage module normally accessible state, wherein the storage module disconnection state indicates that the node device itself and the corresponding of the other node devices are in the disconnected state, so that the node device itself cannot access the corresponding of the other node storage modules. The storage module inaccessible state indicates that the node device cannot access its own storage module, or that the node device is in a normal connection state with the corresponding node device among the other node devices, but the node device cannot access the storage module of the corresponding node device among the other node devices. The storage module normal access state indicates that the node device can normally access its own storage module, or that the node device can normally access the storage module of the corresponding node device among the other node devices.

8. The distributed system data access method as described in claim 7, wherein, Following step (S), the following steps are further included: (T) For each node device, the node device periodically detects a mounting status of its own joint mount, the mounting status including one of a normal mounting status, a waiting-for-repair mounting status, and an abnormal mounting status, wherein the normal mounting status indicates that all network file systems in the joint mount of the node device are mounted normally, and the waiting-for-repair mounting status indicates that the node device detects that at least one first unmounted network file system exists in its joint mount that has not yet been mounted, and the node storage module managed by each first unmounted network file system is in a storage module disconnection state or a storage module inaccessible state, and each first unmounted network file system is one of the network file systems in the joint mount of the node device. The abnormal mounting state indicates that the node device has detected at least one unmounted second network file system in its joint mounting, and the node storage module managed by each unmounted second network file system is in normal access status. Each unmounted second network file system is one of the network file systems in the joint mounting of the node device. Alternatively, it indicates that the node device has detected at least one mounted inaccessible network file system in its joint mounting, and the node storage module managed by each inaccessible network file system is in a disconnected or inaccessible state. Each inaccessible network file system is one of the network file systems in the joint mounting of the node device.

9. The distributed system data access method as described in claim 8, wherein, Following step (T), the following steps are also included: (U) For each node device, when the joint mount state of the node device is in the abnormal mount state, the node device mounts at least one second unmounted network file system to the joint mount and unmounts at least one inaccessible network file system from the joint mount, so that the joint mount state of the node device is in the normal mount state or the mount state awaiting repair.

Images