KR100449490B1 - Management method of san-based distributed shared raid system - Google Patents

Abstract

Disclosed are a method of operating and managing a storage area network (SAN) based distributed shared redundant array of inexpensive disks (RAID) systems.

The present invention relates to a method for providing a web-based integrated management function by synchronizing a managed object to a server sharing a synchronized distributed shared RAID controller system, a web server and a diagnostic management program to be burdened by the distributed shared RAID controller server It is installed on the network and the private network using an Ethernet switch between the server and the RAID controller.

In addition, the present invention provides a browser control module, a distributed server control module, and a RAID controller diagnostic management module in a server to independently update a managed object maintained by the server by broadcasting a query response between the server and the RAID controller. It aims to provide a web-based integrated management function by synchronizing the information and notifying the connected browsers of the result.

Therefore, the present invention can reduce the load of the web server and the diagnostic management program to be handled by the distributed shared RAID controller, and can be directly connected to the Internet to improve security vulnerabilities. In addition, the present invention can connect to any server and run the diagnostic management program on the web by synchronizing the management objects maintained between each server.

Description

MANAGEMENT METHOD OF SAN-BASED DISTRIBUTED SHARED RAID SYSTEM}

The present invention relates to a technology for operating a RAID system, and more particularly, to operating a SAN-based distributed shared RAID system suitable for providing a synchronized management object and a web-based integrated management function through a server sharing a synchronized distributed shared RAID controller system. And a management method.

In a general RAID system, the master server directly manages and distributes managed objects. Therefore, the master server directly connects a managed system to the Internet network and manages it independently.

The present invention has been made to solve the above-described problems of the prior art, by providing an Ethernet path for maintaining the same management object to a plurality of servers sharing a distributed shared RAID controller system synchronized using a private network of Ethernet web The purpose is to provide a method of operating and managing a SAN-based distributed shared RAID system to reduce the load on the web server and the diagnostic management program and increase the security reliability by managing and controlling the system through any server.

Another object of the present invention is to provide a web-based integrated management function by blocking access from an external network to a managed system and synchronizing managed objects to a server sharing a synchronized distributed shared RAID controller system.

According to a preferred embodiment of the present invention for achieving the above object, SAN-based distributed sharing for connecting the server and the distributed shared RAID controller to the private network, and running a diagnostic management program on the web to the other interface of the server through a web service 1. A method of operating and managing a RAID system, comprising: a first step of determining a current state of a server when a management object is downloaded using node information obtained through a heartbeat message; A second step of terminating the download process if the current state of the server is IN_SERVICE as a result of the determination of the first step; As a result of the judgment in the first step, when the managed object is required to be downloaded, the server state is changed to the service switching state (GOING_IN_SERVICE) and a standby message is sent to all servers and controllers, thereby changing the state of all the nodes in service to the standby state. A third step of allowing standby nodes to process only restricted messages to eliminate inconsistencies while downloading managed objects; Determining whether a server in service exists; A fifth step of downloading all managed objects from the server in the service state if the server in service exists as a result of the determination in the fourth step; A sixth step of downloading the management object from the RAID controller if the server in the service state does not exist as a result of the determination in the fourth step; Determining whether a managed object download is successful, and if the managed object download is successful, standby the server state by changing the state of the server to a service state and sending a Resume message to the RAID controller and all servers. A seventh step of changing from a state to a service state; A method of operating and managing a SAN-based distributed shared RAID system, which includes an eighth step of notifying all connected browsers of the change.

1 is a configuration block diagram of a SAN-based distributed shared RAID system for performing the method according to the present invention;

2 is an operation management module of the SAN-based distributed shared RAID system of FIG.

3 is an exemplary diagram of a state management heartbeat message of a node (server, RAID controller) of FIG. 1;

4 is a flowchart of an initial managed object synchronization process according to the present invention;

5 is a system configuration diagram for performing managed object synchronization in accordance with the present invention.

<Explanation of symbols for the main parts of the drawings>

100: web based diagnostic management browser 102: internet network

104: server 106: fiber channel switch

107: Ethernet Switch 108: Distributed Shared RAID Controller

110: FC-JBOD

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

The present invention comprises a synchronized distributed shared RAID controller, a managed object synchronized in a server sharing the same, and a web-based integrated management system using APPLET.

1 is a block diagram of a SAN-based distributed shared RAID system for performing the method according to the present invention, which includes a web-based diagnostic management browser 100, the Internet 102, a server 104, a fiber channel switch 106, and an Ethernet. Switch 107, distributed shared RAID controller 108, and FC-JBOD 110.

As shown in FIG. 1, the web-based diagnostic management browser 100 is connected to the Internet 102 and connects to an arbitrary server 104 to download and execute a diagnostic management program implemented by APPLET on the web. do.

The server 104 has two Ethernet interfaces, one connected to the Internet 102 to perform web services, and the other connected to a private network to communicate with the distributed shared RAID controller 108.

The distributed shared RAID controller 108 is connected to the Ethernet switch 107 and included in the private network for communication with the server 104.

FIG. 2 illustrates an operation management module of the SAN-based distributed shared RAID system of FIG. 1. The server 202 according to the present embodiment includes a web server 21, a diagnostic management program 22, and a browser control module 23. And a distributed server control module 24 and a RAID controller diagnostic management module 25.

As shown, the web browser 200 can be executed on the web by connecting to the web server 21 of the server 202 and downloading the diagnostic management program 22 to APPLET. In this case, when the server 202 is logged in, a TCP connection is established with the browser control module 23 of the server 202.

The web server 21 and the diagnostic management program 202 are loaded on all the servers so that the operator can access them via the Internet.

The browser control module 23 manages a list of the diagnostic management programs 22 logged in to the server 202, performs a query from the diagnostic management program 22, and gives a response.

The distributed server control module 24 transmits and stores the TCP connection and the managed object between the servers 202 during the synchronization process of the initial managed object, as described in FIG. 4 to be described later.

The RAID controller diagnostic management module 25 is responsible for managing the state of the server 202 and the distributed shared RAID controller nodes 204 and transmitting and receiving RAID control commands through a UDP connection to an interface of a private network.

The diagnostic and management module 26 in the distributed shared RAID controller 204 broadcasts the response of all queries by the representative among the distributed shared RAID controllers, and each controller periodically broadcasts its status.

3 is an exemplary diagram of a state management heartbeat message of a node (server, RAID controller) of FIG. 1.

As shown, 'TYPE' of the heartbeat message for status management of each node is divided into a server and a distributed shared RAID controller, and 'STATUS' is divided into GOING_IN_SERVICE, IN_SERVICE, STANDBY, ENTERING_SHUTDOWN, and SHUTDOWN.

Here, the IN_SERVICE and STANDBY states indicate a state in which service of each node is available, and the remaining states indicate a state in which the service is stopped or ready.

In addition, the heartbeat message is broadcast periodically, and if the node does not arrive for a predetermined time by periodically checking the heartbeat message at each node, the node is considered to be 'SHUTDOWN'.

4 is a flowchart of an initial management object synchronization process according to the present invention, which illustrates a process of synchronizing a management object with a server and a RAID control period when the system is initially operated.

First, when the management object is downloaded using the node information obtained through the heartbeat message (S400), the process proceeds to step S402 to determine the current state.

If it is determined in step S402 that the state is IN_SERVICE, the process proceeds to step S424 to end the download process.

If it is necessary to download a managed object, change its status to GOING_IN_SERVICE and send a standby message to all servers and controllers to change the status of all IN_SERVICE nodes to STANDBY so that STANDBY nodes can handle only restricted messages. Inconsistency is removed while downloading the management object (S404) (S406).

In the next step S408, in order to reduce the load of the RAID controller due to downloading all managed objects from the RAID controller, it is first determined whether a server having a status of IN_SERVICE exists.

As a result of the determination in step S408, if there is a server in the IN_SERVICE state, the process proceeds to step S410 and step S412 to download all management objects from the old server.

As a result of the determination in step S408, if there is no server in the IN_SERVIC state, the flow advances to step S414 to download from the representative RAID controller.

On the other hand, if the management object has been downloaded successfully, proceed to step S418 and step S420 to change its status to IN_SERVICE and send a resume message to all servers and RAID controllers, and STANDBY status to IN_SERVICE. Change the status to allow the service to proceed normally.

Then, in step S422, these changes are notified to all connected browsers.

5 is a configuration diagram of a system for performing management object synchronization according to the present invention, and illustrates a synchronization structure of management information reflected during operation of the system.

As shown in FIG. 5, all query messages are transmitted from the diagnostic management program to the server, and the server broadcasts the query to the private network so that all RAID controllers receive the query.

The response to the query is broadcast by the representative RAID controller. All servers analyze the response to update the managed object and pass the information to all browsers connected to it. Update in real time.

As described above, in the present invention, when a management target is locally limited to manage nodes of a distributed structure, a dedicated network is configured through Ethernet to indirectly manage a RAID controller and JBODs through a server through a server. By providing this, security vulnerabilities can be improved.

In addition, since a web server and a diagnostic management program are not directly provided in the RAID controller, the load of the RAID controller can be reduced.

In addition, with the exception of the first server running when multiple servers are running, the remaining servers download initial managed objects from other servers, which significantly reduces the load on the RAID controller and keeps logs and alarms constant. By allocating storage space for the information to be managed on the server, it eases the concern about the local storage of the RAID controller.

As mentioned above, although this invention was demonstrated concretely based on the Example, this invention is not limited to this Example, Of course, various changes are possible within the range which does not deviate from the summary of the claim mentioned later.

Claims (4)

A storage area network (SAN) -based distributed shared RAID system that connects servers to distributed shared RAID (Redundant Array of Inexpensive Disks) controllers in a private network and runs diagnostic management programs on the Web through different web interfaces. In the operation and management method, A first step of determining a current state of the server when a management object is downloaded using node information obtained through a heartbeat message; A second step of terminating a download process if the current state of the server is IN_SERVICE as a result of the determination of the first step; As a result of the determination of the first step, when the management object needs to be downloaded, the state of the server is changed to a service switching state (GOING_IN_SERVICE) and a standby message is transmitted to all servers and controllers, thereby bringing the state of all nodes in service to the standby state. Altering, causing the standby nodes to process only restricted messages to eliminate inconsistencies while downloading managed objects; Determining whether a server in service exists; A fifth step of downloading all managed objects from the server in the service state if the server in service exists as a result of the determination in the fourth step; A sixth step of downloading a management object from the RAID controller when there is no server in service state as a result of the determination in the fourth step; It is determined whether the download of the management object is successful, and when the download of the management object is successful, the state of the server is changed to a service state and a resume message is transmitted to the RAID controller and all servers. A seventh step of changing a state of the server from a standby state to a service state; And an eighth step of notifying all connected browsers of the change. The method of claim 1, The SAN-based distributed shared RAID system includes a web server, a diagnostic management program, a browser control module, a distributed server control module, and a RAID controller diagnostic management module. The method of claim 1, The method, Method of operating and managing a SAN-based distributed shared RAID system, characterized in that for transmitting and receiving query / response messages between the browser, server and distributed shared RAID controller. The method of claim 2, The web server, diagnostic management program, browser control module, distributed server control module, RAID controller diagnostic management module is provided in the server operation and management method of the SAN-based distributed shared RAID system.