KR102571197B1 - Method for maintaining cache coherency of cluster file system - Google Patents

Abstract

A method for maintaining cache coherency between a first server and a second server sharing one storage device in a cluster file system through a network is provided. The method for maintaining cache coherency includes a first server generating first data and storing the first data in a storage device, and then a file for the first data stored in a storage device in a cache memory of a second server based on a command input to a second server. update information

Description

Method for maintaining cache coherency of cluster file system}

According to the present invention, in a cluster file system in which a plurality of servers are connected to a plurality of storage devices through a network, when at least two servers share a storage device, cache coherency can be ensured by rereading the storage device rather than accessing the network. It is about how to maintain cache coherency.

A cluster file system refers to a configuration in which a plurality of servers or computers are connected through a network and operated as a single system. In such a cluster file system, each of a plurality of servers communicates with another server to share data stored in the storage device of the other server.

1 is a diagram showing a conventional cluster file system.

Referring to FIG. 1, a conventional cluster file system 10 includes a plurality of servers 11 and 12 and a plurality of storage devices 13 and 14 mounted and connected to each of them.

Specifically, the cluster file system 10 includes the first server 11 and the second server 12, the first storage device 13 is mounted and connected to the first server 11, and the second storage device 13 The device 14 is mounted and connected to the second server 12 .

In such a conventional cluster file system 10, a processor (not shown) of the first server 11 executes a daemon process to connect a communication channel with the second server 12, and The consistency of the file system is maintained through communication between the second servers 12 .

Here, the consistency of the file system between the first server 11 and the second server 12 is maintained by the first server 11 and the second storage device 14 for sharing data stored in the first server 11. 2 This means maintaining consistency of each cache memory of the server 12.

Accordingly, in the conventional cluster file system 10, after the consistency of the file system between the first server 11 and the second server 12 is maintained, the first server 11 transfers the file system through the second server 12. By accessing the second storage device 14, data stored in the second storage device 14 is shared.

As such, when another server shares data stored in a storage device connected to one server in the conventional cluster file system 10, it is necessary to maintain the consistency of the file system through communication between the two servers. There is a problem in that processor resources of the processor are wasted and performance is degraded due to increased network traffic.

Korean Patent Registration No. 10-1638727 (2016.07.20.)

According to the present invention, in a cluster file system in which a plurality of servers are connected to a plurality of storage devices through a network, when at least two servers share a storage device, cache coherency can be ensured by rereading the storage device rather than accessing the network. It is intended to provide a method for maintaining cache coherency.

A method for maintaining cache coherency in a cluster file system according to an embodiment of the present invention is provided in a first server and a second server sharing one storage device over a network, wherein the first server generates first data and storing in a storage device; and updating file information about the first data stored in the storage device in a cache memory of the second server based on a command input to the second server.

The updating may include: retrieving file information for the first data from the cache memory of the second server; accessing, by the second server, the storage device through the network based on a search result, and acquiring the file information of the first data from the storage device; and copying the file information to a buffer page of the cache memory based on the obtained file information.

The updating may further include generating an inode object for the first data in an inode cache of the cache memory based on the file information copied to the buffer page and updating the inode cache. include

The updating may further include generating a dentry object for the first data in a dentry cache of the cache memory based on the updated inode cache.

The second server may further include, when the file information for the first data is retrieved from the cache memory, executing the command for the first data based on the retrieved file information.

In addition, if the file information for the first data is not retrieved from the cache memory and the file information is not retrieved from the storage device, the second server outputs an error message of a command for the first data. Include more steps.

The second server may further include executing the command for the first data based on the updated file information after the first file information for the first data is updated in the cache memory. do.

The method for maintaining cache coherency according to the present embodiment may include requesting, by the middleware of the second server, a file path for the first data from the cache memory based on the command; Retrieving file information including a requested file path among a plurality of file information previously stored in the cache memory; and updating the file information for the first data in the cache memory based on a search result.

A method for maintaining cache coherence according to the present invention, when at least one of a plurality of servers sharing a storage device receives a command for data stored in a storage device, by updating file information for the corresponding data in its own cache memory. , it is possible to prevent waste of processor resources and excessive network traffic to maintain cache coherency.

1 is a diagram showing a conventional cluster file system.
2 is a diagram showing a cluster file system according to an embodiment of the present invention.
3 is a block diagram showing the internal configuration of a cluster file system according to an embodiment of the present invention.
4 is a flowchart of a method for maintaining cache coherency according to an embodiment of the present invention.
FIG. 5 is a flowchart of the cache memory update method of FIG. 4 .

Advantages and features of the present invention, and methods of achieving them, will become clear with reference to the detailed description of the following embodiments taken in conjunction with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in various different forms, and only these embodiments make the disclosure of the present invention complete, and common knowledge in the art to which the present invention belongs. It is provided to completely inform the person who has the scope of the invention, and the present invention is only defined by the scope of the claims.

In describing the embodiments of the present invention, if it is determined that a detailed description of a known function or configuration may unnecessarily obscure the subject matter of the present invention, the detailed description will be omitted. In addition, terms to be described later are terms defined in consideration of functions in the embodiment of the present invention, which may vary according to the intention or custom of a user or operator. Therefore, the definition should be made based on the contents throughout this specification.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

2 is a diagram showing a cluster file system according to an embodiment of the present invention, and FIG. 3 is a block diagram showing the internal configuration of the cluster file system according to an embodiment of the present invention.

2 and 3, the cluster file system 100 of this embodiment includes a plurality of servers, for example, a first server 110-1 to an n-th server 110-n and a network 130 with them. It may include a plurality of storage devices connected through , for example, a first storage device 120-1 to an m-th storage device 120-m.

Here, each of the first server 110-1 to n-th server 110-n may be a computer device. Each of the first storage device 120-1 to the m-th storage device 120-m stores a solid state drive (SSD) using a semiconductor such as a hard disk drive or flash memory. may be a device. Also, n and m may be natural numbers having the same or different values.

The network 130 connecting each of the first server 110-1 to n-th server 110-n and each of the first storage device 120-1 to m-th storage device 120-m is wired/wireless. It may be a communication-type network, and communication may be performed using a predetermined protocol, for example, an NVMe over Fabrics (NVMe-of) protocol.

Accordingly, in the cluster file system 100 of this embodiment, each of the first server 110-1 to the n-th server 110-n connects to the first storage device 120-1 to the m-th storage device 120-1 through the network 130. It is connected to at least one of the devices 120-m and can use it as its own virtual local disk.

On the other hand, as shown in FIG. 3, at least two servers among a plurality of servers, for example, a first server 110-1 and a second server 110-2 are connected to one storage device, for example, a first storage device ( 120-1) can be shared.

Accordingly, each of the first server 110-1 and the second server 110-2 accesses the first storage device 120-1 through the network 130 to store the first storage device 120-1. Stored data may be read and opened, or new data may be created and stored in the first storage device 120-1.

As shown in FIG. 3, each of the first server 110-1 and the second server 110-2 may include a processor and a cache memory. For example, the first server 110-1 includes the first processor 111-1 and the first cache memory 113-1, and the second server 110-2 includes the second processor 111-2. and a second cache memory 113-2.

The first storage device 120-1 includes a data block 121 in which a plurality of data is stored and an inode table in which file information for each of the plurality of data stored in the data block 121 is stored ( 123) may be included.

The first processor 111-1 of the first server 110-1 may control the overall operation of the first server 110-1 based on a command input to the first server 110-1. . Similarly, the second processor 111-2 of the second server 110-2 controls the overall operation of the second server 110-2 based on a command input to the second server 110-2. can

The first cache memory 113-1 may include a first dentry cache 111-1, a first inode cache 115-1, and a first buffer page 116-1. The second cache memory 113-2 has substantially the same structure as the first cache memory 113-1, that is, the second dentry cache 114-2, the second inode cache 115-2 and the second cache memory 115-2. A buffer page 116-2 may be included.

Each of the first cache memory 113-1 and the second cache memory 113-2 may store file information for each of a plurality of data stored in the first storage device 120-1. The file information may include metadata information such as data type, name, size, access information, modification information, and directory information indicating the location of the data.

Accordingly, when a command for data stored in the first storage device 120-1, for example, a data reading command is input to the first server 110-1 from the user, the first server 110-1 executes a 1 Processor 111-1 may process an input command based on file information of corresponding data stored in the first cache memory 113-1. Of course, the second processor 111-2 of the second server 110-2 also processes the command based on the file information of the data stored in the second cache memory 113-2 in response to the input data read command. can do.

In this way, the first server 110-1 and the second server 110-2 share the first storage device 120-1 through the network 130, and the first cache memory 113-1 of each server 1) and the second cache memory 113-2 may store file information about a plurality of data of the first storage device 120-1.

Accordingly, each of the first processor 111-1 of the first server 110-1 and the second processor 111-2 of the second server 110-2 is connected to the first storage device according to the inputted data read command. Instead of accessing 120-1, each cache memory, that is, accessing each of the first cache memory 113-1 and the second cache memory 113-2 may be accessed to process a command for corresponding data. . Accordingly, the first server 110-1 and the second server 110-2 can increase the operating speed according to command processing.

Meanwhile, each of the first cache memory 113-1 and the second cache memory 113-2 must maintain consistency with each other. In other words, the same file information, that is, file information for each of a plurality of data stored in the first storage device 120-1 is stored in each of the first cache memory 113-1 and the second cache memory 113-2. It should be. Accordingly, when the same command for one piece of data is input, the first server 110-1 and the second server 110-2 can process it identically.

In order to maintain the consistency of the first cache memory 113-1 and the second cache memory 113-2, each cache memory properly updates file information of data stored in the first storage device 120-1. is needed

To this end, when one of the first server 110-1 and the second server 110-2 creates new data in the first storage device 120-1, the one server transfers data to the other server. Consistency between the two cache memories can be maintained by notifying the first storage device 120-1 that new data is generated so that the cache memory of the other server is updated.

However, in this method of maintaining consistency of cache memory, communication between a plurality of servers is performed whenever new data is generated in the first storage device 120-1, so heavy traffic may be generated in the network 130.

Accordingly, in the cluster file system 100 of the present embodiment, a cache coherency maintenance method capable of maintaining coherence between cache memories of each server while preventing traffic from being generated in the network 130 can be performed.

4 is a flowchart of a method for maintaining cache coherency according to an embodiment of the present invention.

Hereinafter, as shown in FIG. 3 for convenience of description, the first server 110-1 and the second server 110-2 connect the first storage device 120-1 through the network 130 A method for maintaining cache coherency between the first server 110-1 and the second server 110-2 of the shared cluster file system 100 will be described in detail.

Referring to FIGS. 3 and 4 , the first processor 111-1 of the first server 110-1 may generate new data, for example, first data, according to an input command. The first processor 111-1 stores file information about the first data in the first cache memory 113-1, and transfers the first data to the first storage device 120-1 through the network 130. It can be transmitted and stored as (S10).

Next, a command for the first data, for example, a command for reading the first data may be input to the second server 110-2 from the user (S20).

Accordingly, the second processor 111-2 of the second server 110-2 searches for file information about the first data in the second cache memory 113-2 based on the input command, and displays the search result. Based on this, file information of the first data may be updated in the second cache memory 113-2 (S30).

FIG. 5 is a flowchart of the cache memory update method of FIG. 4 .

Referring to FIGS. 4 and 5 , the second processor 111-2 may search for file information of the first data in the second cache memory 113-2 based on the input command (S110).

When file information of the first data is searched in the second cache memory 113-2 (Y), the second processor 111-2 may execute and process a command for the first data based on the searched file information. Yes (S40).

On the other hand, if the file information of the first data is not retrieved from the second cache memory 113-2 (N), the second processor 111-2 sends the second server 110-2 through the network 130. and the first storage device 120-1 may be connected.

Subsequently, the second processor 111-2 may obtain file information about the first data from the inode table 123 of the first storage device 120-1.

The second processor 111-2 copies the acquired file information of the first data to the second buffer page 116-2 of the second cache memory 113-2 and stores the second buffer page 116-2. can be updated (S120)

Subsequently, the second processor 111-2 stores the second inode cache 115 of the second cache memory 113-2 based on the file information of the first data updated in the second buffer page 116-2. In -2), the second inode cache 115-2 may be updated by creating an inode object for the first data (S130).

Next, the second processor 111-2 stores the first data in the second dentry cache 114-2 of the second cache memory 113-2 based on the updated second inode cache 115-2. A dentry for , for example, directory information for a directory in which the first data is stored may be generated (S140).

Then, when the second cache memory 113-2 of the second server 110-2 is updated based on the file information of the first data stored in the first storage device 120-1, the second processor 111-2 2) may be processed by executing a previously input command based on the file information of the first file stored in the updated second cache memory 113-2 (S40).

As described above, in this embodiment, the second server 110-2 is the user, not the time when the first data is generated by the first server 110-1 and stored in the first storage device 120-1. A plurality of servers, that is, the first server 110-1 and the second server are updated by updating the file information for the first data in the second cache memory 113-2 at the time when a command for the first data is input from Cache coherency between (110-2) can be maintained.

Accordingly, the present invention can reduce traffic generated in the network 130 by not requiring communication between a plurality of servers whenever new data is generated in a storage device to maintain cache coherency.

Combinations of each block of the block diagram and each step of the flowchart of the present invention described above may be performed by computer program instructions. Since these computer program instructions may be loaded into an encoding processor of a general-purpose computer, special-purpose computer, or other programmable data processing equipment, the instructions executed by the encoding processor of the computer or other programmable data processing equipment are each block or block diagram of the block diagram. Each step in the flow chart creates means for performing the functions described. These computer program instructions may also be stored in a computer usable or computer readable memory that can be directed to a computer or other programmable data processing equipment to implement functionality in a particular way, such that the computer usable or computer readable memory The instructions stored in are also capable of producing an article of manufacture containing instruction means that perform the function described in each block of the block diagram or each step of the flowchart. The computer program instructions can also be loaded on a computer or other programmable data processing equipment, so that a series of operational steps are performed on the computer or other programmable data processing equipment to create a computer-executed process to generate computer or other programmable data processing equipment. It is also possible that the instructions performing the processing equipment provide steps for executing the functions described in each block of the block diagram and each step of the flow chart.

Additionally, each block or each step may represent a module, segment or portion of code that includes one or more executable instructions for executing specified logical function(s). It should also be noted that in some alternative embodiments it is possible for the functions recited in blocks or steps to occur out of order. For example, two blocks or steps shown in succession may in fact be performed substantially concurrently, or the blocks or steps may sometimes be performed in reverse order depending on their function.

The above description is merely an example of the technical idea of the present invention, and various modifications and variations can be made to those skilled in the art without departing from the essential qualities of the present invention. Therefore, the embodiments disclosed in the present invention are not intended to limit the technical idea of the present invention, but to explain, and the scope of the technical idea of the present invention is not limited by these embodiments. The protection scope of the present invention should be construed according to the claims below, and all technical ideas within the scope equivalent thereto should be construed as being included in the scope of the present invention.

100: cluster filesystem
110: server
120: storage device
130: network

Claims (10)

A method for maintaining cache coherency between a first server and a second server sharing one storage device over a network in a cluster file system,
generating, by the first server, first data and storing the first data in the storage device;
retrieving file information for the first data from the cache memory of the second server based on a command for the first data input to the second server;
accessing, by the second server, the storage device through the network when the file information for the first data is not retrieved from the cache memory; and
The second server obtains the file information for the first data from the inode table of the storage device, copies the acquired file information to a buffer page of the cache memory, and stores the file information for the first data. A method for maintaining cache coherency of a cluster file system comprising the step of updating.
delete According to claim 1,
The updating step is
and updating the inode cache by creating an inode object for the first data in the inode cache of the cache memory based on the file information copied to the buffer page. maintenance method.
According to claim 3,
The updating step is
and generating a dentry object for the first data in a dentry cache of the cache memory based on the updated inode cache.
According to claim 1,
When the file information for the first data is retrieved from the cache memory,
and executing, by the second server, the command for the first data based on the searched file information.
According to claim 1,
If the file information for the first data is not searched in the cache memory and the file information is not searched in the storage device,
and outputting, by the second server, an error message of a command for the first data.
According to claim 1,
After the file information for the first data is updated in the cache memory,
and executing, by the second server, the command for the first data based on the updated file information.
According to claim 1,
requesting, by the middleware of the second server, a file path for the first data from the cache memory based on the command;
Retrieving file information including a requested file path among a plurality of file information previously stored in the cache memory; and
and updating the file information for the first data in the cache memory based on a search result.
A computer-readable recording medium storing a computer program for maintaining cache coherency between a first server and a second server sharing one storage device over a network in a cluster file system,
The computer program is executed by a processor,
generating, by the first server, first data and storing the first data in the storage device;
retrieving file information for the first data from the cache memory of the second server based on a command for the first data input to the second server;
accessing, by the second server, the storage device through the network when the file information for the first data is not retrieved from the cache memory; and
The second server obtains the file information for the first data from the inode table of the storage device, copies the acquired file information to a buffer page of the cache memory, and stores the file information for the first data. A computer-readable recording medium including instructions for the processor to perform the step of updating.
A computer program for maintaining cache coherency between a first server and a second server sharing one storage device over a network in a cluster file system,
The computer program is executed by a processor,
generating, by the first server, first data and storing the first data in the storage device;
retrieving file information for the first data from the cache memory of the second server based on a command for the first data input to the second server;
accessing, by the second server, the storage device through the network when the file information for the first data is not retrieved from the cache memory; and
The second server obtains the file information for the first data from the inode table of the storage device, copies the acquired file information to a buffer page of the cache memory, and stores the file information for the first data. A computer program comprising instructions for the processor to perform the step of updating.