KR20070080363A - Integrated server system - Google Patents

Abstract

An integrated server system is provided to facilitate extension of servers by replacing the failed server with the other server if one of the clustered servers is failed. Each server(204-210), which is a blade server, is connected to a user client(200) through the network and is managed as an entity through an integrated application. A SAN(Storage Area Network) switch(212) is connected to the servers, and a storage(214) is connected to the SAN switch and stores data read/written by the servers. The integrated application provides a file sharing function for enabling the server to write the data to the storage and read the data from the storage at the same time. The integrated application keeps a cache area of each server uniform, synchronizes a cache in every update, and monitors the servers. The storage is managed by a CVM(Cluster Volume Manager) and the CVM generates a virtual volume sharing the data among the servers. The servers include at least one of a web, message, game, application, web mail, database, and backup server.

Description

Integrated server system {INTEGRATED SERVER SYSTEM}

1 is a view showing the configuration of a server system according to the prior art.

2 is a diagram illustrating a configuration of an integrated server system according to an exemplary embodiment of the present invention.

3 illustrates functional modules of a plurality of servers forming a cluster according to an exemplary embodiment of the present invention.

4 is a diagram illustrating a module configuration of an integrated application according to an exemplary embodiment of the present invention.

5A to 5C illustrate an example of a failover process according to the present invention.

The present invention relates to an integrated server system, and more particularly, to an integrated server system that clusters a plurality of servers and allows a system including a plurality of servers to operate without interruption.

The server market is reshaping as new technologies related to servers are pouring in from the hardware and software sectors. Due to the improved performance of x86 processors based on Intel or AMD processors, x86 servers used in existing web servers have been eroding the midrange server market, and have recently surpassed the high-end market.

In the past, high-ends were generally equipped with Unix or mainframes, but recently, the number of concurrent users has been increasing rapidly in Internet services, and users have large files such as video, VOD, CAD / CAM and Internet broadcasting. In the situation where the system needs to be provided, there is a problem that a conventional UNIX-based server cannot meet the needs of users.

1 is a diagram illustrating a configuration of a server system according to the prior art. Referring to FIG. 1, a conventional server system includes an Ethernet switch 102 and a plurality of servers 104, 106, 108, and 114 connected to an internal network. ).

Here, the plurality of servers may include a web server 102 for a web service, a VOD server 106 for a VOD streaming service, a file server 108 for providing a file service, and a DB server 114 for storing data. In particular, the file server 108 is connected to one or more storages 112 through a storage area network switch 110.

The server system configured as described above is connected to one or more user clients 100 through an internet network, and performs a process of providing an appropriate service according to a user's request.

As shown in FIG. 1, in the related art, in the case where a plurality of servers provide a service, each server operates independently of each other, and thus has the following problems.

First, when the number of users increases, the service speed decreases due to an increase in network load, and when the transaction increases, there is a problem that the service speed decreases due to a disk I / O speed limit, and it is difficult to implement network load balancing.

In particular, according to the conventional server system, when a failure occurs in at least one server, it takes a long time to recover in the future, there is a problem that it is necessary to purchase a separate high-availability (HA) cluster software.

In addition, according to the conventional server system, since there is no data sharing function, duplicate storage of the same data is inevitable in the case of web, VOD / Streaming, and file service. .

In addition, there is a problem in that file systems cannot be shared, server collaboration is impossible, and configuration of large file systems is impossible in a SAN environment.

Lastly, according to the related art, data coherency cannot be maintained according to data replication for data redundancy, and excessive costs are required for service monitoring, performance tuning, storage management for each server, and backup. There was a problem, and there was a problem of lack of scalability due to data growth.

In the present invention, to solve the problems of the prior art as described above, it is proposed an integrated server system that can be easily expanded server.

Another object of the present invention is to provide an integrated server system that can increase disk utilization through data sharing.

Another object of the present invention is to provide an integrated server system capable of maintaining data consistency.

Another object of the present invention is to provide an integrated server system capable of quickly replacing a failed server with another server when at least one server fails.

In order to achieve the object as described above, according to a preferred embodiment of the present invention, provided. A plurality of servers connected via a network with a user client, the server being managed as an entity through an integrated application; A SAN switch connected to the plurality of servers; And storage coupled to the SAN switch, the storage storing data that is read and written by the plurality of servers, wherein the integrated application allows the plurality of servers to simultaneously write data to the storage and read data stored in the storage at the same time. An integrated server system is provided that provides a file sharing function.

Preferably, the server is a blade server.

The integrated application is characterized in that the cache areas of the plurality of servers are kept the same and cache synchronization is performed every update.

In addition, the integrated application is characterized in that for monitoring the plurality of servers.

According to the present invention, when at least one of the plurality of servers fails, the integrated application may replace the failed server according to at least one of N: 1, N: N, and N: M.

Preferably, the storage is managed by a cluster volume manager (CVM), and the cluster volume manager can create a virtual volume that can share data among the plurality of servers.

According to the present invention, the plurality of servers include at least one of a web server, a messenger server, a game server, an application server, a web mail server, a database server, and a backup server.

The integrated server system according to the present invention further comprises a console connected to the plurality of servers, the console is characterized in that for monitoring the usage of the plurality of connected servers.

Preferably, the integrated application may distribute and manage lock information and metadata.

According to another aspect of the invention, a plurality of servers connected via a network with a user client, the server being managed as an entity through an integrated application; A SAN switch connected to the plurality of servers; And storage connected to the SAN switch and storing data read and written by the plurality of servers, wherein the storage is managed by a cluster volume manager (CVM), wherein the cluster volume manager is managed by the plurality of cluster volumes. An integrated server system is provided, which creates a virtual volume capable of sharing data between servers of a server.

According to another aspect of the present invention, a plurality of servers connected to a user client through a network, a SAN switch connected to the plurality of servers and storage connected to the SAN switch, the storage for storing data that the plurality of servers read and write A recording medium in which an application program for integrating the plurality of servers is recorded, the system comprising: a file sharing module for allowing the plurality of servers to simultaneously write data to the storage and to simultaneously read data stored in the storage; A cache consolidation module which maintains the same cache area of the plurality of servers and caches synchronization every update; An integrated monitoring module for monitoring the plurality of servers; And when a failure occurs in at least one of the plurality of servers, the integrated application includes a failover module configured to replace the failed server according to at least one of N: 1, N: N, and N: M. A computer readable recording medium is provided.

Hereinafter, with reference to the accompanying drawings will be described in detail a preferred embodiment of the integrated server system according to the present invention.

2 is a diagram illustrating a configuration of an integrated server system according to an exemplary embodiment of the present invention.

As shown in FIG. 2, in the integrated server system according to the present invention, the plurality of servers form a cluster in a symmetrical structure and are connected to one or more storages 214 through the SAN switch 212.

According to the present invention, a plurality of servers may be an Intel-based or AMD-based server, it is preferable that a blade server is used to reduce power consumption.

Here, the blade server is a server that composes a computer system by inserting an ultra-thin blade into a slot, which can solve the problem of power consumption and heat generation, and modularizes to a small size based on low-power chips, thereby reducing power consumption. There are several advantages.

In this case, an operating system (OS) necessary for providing a service may be installed in the plurality of servers, and preferably, a Linux or a Windows operating system may be installed.

The storage 214 may use any FA Redundant Array of Independent Disks (FA RAID) storage certified by the operating system, and the SAN switch may use a switch such as CISCO, Brocade, McData, MDC, Qlogic.

According to the present invention, a plurality of servers form a cluster and are managed as an entity. Looking at the characteristics of the administrative aspects in the system according to the present invention, the present invention provides an integrated fabric management. to provide.

Here, integrated fabric management refers to the management of QLogic Brocade and McData fabrics, by controlling the access of nodes in the cluster to support the removal of non-working elements within the fabric, and to replace those that do not work automatically. In this case, it means to restart the data path based on the substitution factor.

In addition, integrated fabric management may include fabric load balancing among a plurality of servers constituting the cluster.

The integrated server system according to the present invention uses a consistent device naming within the scope of the cluster. By using the same name, it is possible to ensure that each server in the cluster has access to all devices.

In other words, consistent device naming ensures portability of scripts from one server to another in the cluster, and prevents operational failures in managing the cluster when preventing device loss and when servers boot, detect and name devices in a different order. It can greatly reduce the probability of occurrence.

In addition, the integrated server system according to the present invention performs a logical unit number (LUN) discovery and management function, more specifically, allows the administrator to discover the LUN present in the shared storage, this LUN Function to map files to file systems.

In addition, the integrated server system according to the present invention provides a centralized Graphic User Interface (GUI) and Command Line Interface (CLI), and allows an administrator to configure and manage a cluster in a central management office.

In addition, the integrated server system according to the present invention is capable of coordinating, alerting and exchanging important management data with enterprise system management consoles such as IBM Tivoli, CA Unicenter, HP Openview, and Micro Operation Manager (MOM). In this case, SNMP MIB supporting GET, SET, and TRAP can be provided.

In the above, the management aspect of the integrated server system has been described. The characteristics of the file system for the integrated server system are as follows.

First, the file system is used as a journaling file system, which can restore and maintain the consistency of data existing among all nodes (servers) in the cluster.

The file system also supports linear scalability and at least 16 node support, which is made possible by distributed management of lock information and metadata.

The file system according to the present invention supports direct and asynchronous input / output (I / O), which allows applications such as backup software and database servers to access a file without the file system caching data or maintaining lock state information. Allow bufferless I / O.

According to the present invention, an application for server integration is installed in the plurality of servers so that the plurality of servers can be managed as one entity.

4 is a diagram illustrating a module configuration of an integrated application according to the present invention. The integrated application according to the present invention includes a file sharing module 400, a cache integration module 402, an integrated monitoring module 404, and a failover module ( 406 and multi-pass I / O module 408.

The file sharing module 400 allows a plurality of servers included in a cluster to simultaneously read (write / write) and write all data stored in the storage 214.

As a result, data consistency can be maintained and data duplication becomes unnecessary, thereby increasing system efficiency.

In addition, the cache consolidation module 402 keeps the cache areas of the plurality of servers the same, and synchronizes the cache with each update.

The cache consolidation module 402 can solve a performance degradation problem caused by a lack of storage resources, and if a need for retrieval of previously stored data is required, the cache area can be searched together to output the data in the cache area. It can increase.

The integrated monitoring module 404 performs a function of integrally monitoring an application, a server, a network, an operating system, and storage existing in the integrated server system, and performs a function of promptly reporting to an administrator when an error occurs.

For example, when the at least one server fails in the integrated monitoring module 404, the failover module 406 performs a function of quickly recovering the failed server according to a preset policy.

Here, the failover policy may be shown as shown in Figs. 5a to 5c.

FIG. 5A illustrates N: 1 failover and illustrates a process in which N different servers replace the function of one failed server.

5B illustrates 1: 1 failover, in which one adjacent failure server is replaced by another adjacent server, and FIG. 5C is N: M failover, in which N failure servers are replaced by N servers. It is shown.

Conventionally, when one server fails, the entire system has to be stopped, but according to the present invention, since the server replacement is preferably performed within 3 seconds by the failover function of the integrated application, it is possible to effectively cope with the occurrence of the failure. do.

The multipath I / O module 408 performs a function of supporting a plurality of data paths, so that the data paths can be multiplexed to prepare for failures quickly.

At this time, the data input and output through the multi-pass can be achieved through the HBA card,

On the other hand, according to the present invention, it is possible to provide the system scalability for the enterprise environment according to the user's needs, which can be made through the Cluster Volume Manager (CVM) that controls the storage 214.

The SAN volume manager according to the present invention may create a shareable virtual volume of data on a plurality of servers and provide N-way data striping.

This enables support for unlimited number of cluster servers, number of storage support, and online storage and server reconfiguration, overcoming the limitations of disk I / O speed by adding storage and creating multi-storage striped volumes using volume managers, especially as transactions increase. can do.

As mentioned above, a journaling file system can be used for the file system for storage, so that in case of file system damage, it can be recovered within seconds.

On the other hand, according to the present invention, up to 16 nodes can be added in an online state, which is possible because the integrated server system according to the present invention distributes and manages lock information and metadata.

As a result, no errors or bottlenecks occur in a plurality of servers in which the distributed locks and the management of meta data form a cluster.

Meanwhile, according to the present invention, the integrated application allows a plurality of servers to mount the same file system at the same time, and thus active-active configuration is possible for the plurality of servers.

That is, one server can provide a video service, a web service, a mail service, and a database service.

According to the present invention, since a plurality of servers are connected in an active-active manner, it is possible to shorten a considerable time for failover as described above. Preferably, 365 days of non-stop service can be provided, and the server responds quickly. You have time.

FIG. 3 is a diagram illustrating functional modules of a plurality of servers constituting a cluster according to an exemplary embodiment of the present invention. As shown in FIG. 3, the plurality of servers according to the present invention may include Oracle 9i RAC and IBM-DB2. Database servers such as Data WareHouse, messenger servers, web servers, game servers, application servers, web mail servers, and backup servers.

These plurality of servers are connected in one by the system fabric 302 and the integrated application described above, and may be connected to the storage 214 through the storage fabric and the SAN switch 212.

In addition, the plurality of servers are connected to the console through the management fabric 304, the console 308 performs the function of monitoring each usage of the plurality of servers.

Preferred embodiments of the present invention described above are disclosed for purposes of illustration, and those skilled in the art will be able to make various modifications, changes, and additions within the spirit and scope of the present invention. Additions should be considered to be within the scope of the following claims.

As described above, according to the present invention, the server can be easily extended through the integrated application and the cluster volume manager.

In addition, according to the present invention, there is an advantage that can increase the disk utilization through data sharing.

In addition, according to the present invention, a plurality of servers constituting the cluster has the advantage of maintaining the consistency of data by reading and writing data at the same time.

In addition, according to the present invention, when a failure occurs in at least one server according to the effective failover policy, there is an advantage that can quickly replace the failed server to another server.

Claims (12)

A plurality of servers connected via a network with a user client, the server being managed as an entity through an integrated application; A SAN switch connected to the plurality of servers; And It is connected to the SAN switch, including a storage for storing data that the plurality of servers read and write, The integrated application, the integrated server system, characterized in that the plurality of servers write data to the storage at the same time, and provides a file sharing function to read data stored in the storage at the same time. The method of claim 1, The server is an integrated server system, characterized in that the blade server. The method of claim 1, The integrated application, the integrated server system, characterized in that to maintain the same cache area of the plurality of servers, the cache synchronization for each update. The method of claim 1, The integrated application system, characterized in that for monitoring the plurality of servers. The method of claim 1, And when the failure occurs in at least one of the plurality of servers, the integrated application replaces the failed server according to at least one of N: 1, N: N and N: M. The method of claim 1, The storage is managed by a cluster volume manager (CVM), wherein the cluster volume manager creates a virtual volume capable of sharing data among the plurality of servers. The method of claim 1, The plurality of servers, at least one of a web server, a messenger server, a game server, an application server, a web mail server, a database server and a backup server. The method of claim 1, Further comprising a console connected to the plurality of servers, The console is integrated server system, characterized in that for monitoring the usage of the plurality of servers connected. The method of claim 1, The integrated application integrated server system, characterized in that for managing the distribution (Lock) and metadata. A plurality of servers connected via a network with a user client, the server being managed as an entity through an integrated application; A SAN switch connected to the plurality of servers; And It is connected to the SAN switch, including a storage for storing data that the plurality of servers read and write, The storage is managed by a cluster volume manager (CVM), wherein the cluster volume manager creates a virtual volume capable of sharing data among the plurality of servers. In a system comprising a plurality of servers connected to a user client through a network, a SAN switch connected to the plurality of servers and storage connected to the SAN switch, the storage to store data read and written by the plurality of servers, A recording medium on which an application program for integrating a server is recorded, A file sharing module for allowing the plurality of servers to simultaneously write data to the storage and to read data stored in the storage at the same time; A cache consolidation module which maintains the same cache area of the plurality of servers and caches synchronization every update; An integrated monitoring module for monitoring the plurality of servers; And In the event that at least one of the plurality of servers fails, the integrated application is a computer having a program including a failover module that replaces the failed server according to at least one of N: 1, N: N, and N: M. Readable record carrier. The method of claim 11, The storage is managed by a cluster volume manager (CVM), wherein the cluster volume manager creates a virtual volume capable of sharing data among the plurality of servers.

Images