JP2000059385A - Method for managing plural systems at time of overlapping ip addresses - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide the method for managing plural systems by which plural systems are managed by a single manager even when IP addresses are overlapped. SOLUTION: In this method for managing plural systems 15, 16 and 17 through one monitor manager 19, each system is provided with a proxy receiving server 30 for mediating the communication between a server machine 18 belonging to that system and the monitor manager and a system ID capable of specifying the system, to which the server machine to become the transmission destination belongs, is added to communication data in the case of communication from the monitor manager to the server machine. The proxy receiving server deletes the system ID from the communication data in the case of communication from the monitor manager to the server machine but adds the system ID capable of specifying the system, to which the server machine as a transmission source belongs, to the communication data in the case of communication from the server machine to the monitor manager.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a distributed system operation management method.

[0002]

2. Description of the Related Art First, TCP / IP and SNMP as the background of the present invention
Will be described. Communication in distributed systems on the Internet is based on TCP / IP (Transmission Control Protocol).
/ Internet Protocol). In TCP / IP, each machine in the system is identified using an IP address, so a unique IP address must be used on the Internet. IP
If the address is used repeatedly, it becomes impossible to specify the machine and communication cannot be performed. Table 1 shows the IP address system.

[Table 1]

[0003] In addition, although an IP address is basically impossible to connect to the Internet, an address that can be freely used by a user is defined. This address is called a private IP address. Table 2 shows this private IP address system.

[Table 2]

Next, SNMP (Simple Network Management P)
rotocol) will be described. When managing the operation of machines in a distributed system, it is common to take a centralized management system from a monitoring manager (hereinafter referred to as a manager). In order to realize this management system, a monitoring application called a monitoring agent (hereinafter referred to as an agent) is resident in each of the distributed server machines to perform monitoring. Normally, the agent employs an event-driven system, and notifies the manager of information only when the agent detects a failure. The operator of the manager who has learned the failure communicates with the agent to obtain detailed information. This method has the advantage of reducing resources between managers and agents, and is widely used.

[0005] SNMP is a protocol on TCP / IP designed to notify a failure or the like of a device on an SNMP network, and is widely used in fields such as failure monitoring. Communication by SNMP can be classified into two types: TRAP in which communication is started from the agent shown in FIG. 1A, and GET and the like in which communication is started from the manager shown in FIG.

In Version 1 of SNMP, TRAP which is a command when communication is started from an agent, and GET and GET NEX which are commands when communication is started from a manager.
A total of five types of commands, T, SET, and GET RESPONSE, which is a reply from the agent to the communication from the manager, are defined. The contents of these five types of commands are described below. TRAP Communication generated by the agent in the event of a failure GET Acquisition of MIB value GET NEXT Acquisition of MIB value (continued) SET MIB value set ) Means SNMP
Is an internal database that represents the status of the device.

[0007] TRAP, which is a command when communication is started from the agent, transmits information to the manager when an emergency failure occurs in the device. In communication started by the manager, the value of the corresponding MIB is returned to the manager in response to a request from the manager. In addition,
By changing the value of this MIB from the manager side,
The state of the corresponding function of the device can also be changed.

FIG. 2 shows the data structure of SNMP. SNMP is
It is composed of version 1, community name 2, and PDU3. When TRAP which is a command when communication is started from the agent is used, the PDU 3
Type 4, company ID 5, agent address 6, general
A TRAP number 7, an extended TRAP number 8, an occurrence time 9, and related management information 10 are entered. When commands GET, GET NEXT, SET, and GET RESPONSE are used when communication is started from the manager, the PDU 3 includes a PDU type 4, a request identification 11, an error status 12, and an error position number 1.
3. Management information 14 and the like are entered. In the following description of the drawings, the same components are denoted by the same reference numerals, and description thereof will be omitted.

By the way, in a system that does not assume a connection to the Internet at the time of system construction, it takes time and effort to obtain an IP address that can be used on the Internet. They often use addresses.

FIG. 3 shows an example in which a plurality of independent systems which do not need to communicate with the outside at the time of system construction are constructed. The A system 15, B system 16, and C system 17 in the figure are independent systems. These systems include one or more server machines 18. At this time, a system in which the IP address system of the included server machine 18 is completely the same can be considered from the viewpoint of ease of system design and trouble in construction. In FIG. 3, the B system 16 and the C system 17 have exactly the same IP address system.

Further, when performing a system monitoring operation,
It is conceivable that monitoring target systems are added in accordance with customer requests. However, when monitoring a plurality of customer systems, non-unique and duplicate IP addresses may be used.

[0012]

In the case where connection to another system or integrated monitoring is required after the system is constructed, as shown in FIG. Since the used IP addresses are duplicated, the monitoring manager 19 cannot uniquely identify the machine 18 and cannot perform communication using TCP / IP.

To enable communication by TCP / IP, FIG.
As shown in (1), the proxy function of communication, that is, the use of PROXY 20 is common. However, the PROXY 20 is merely a technique for securing a general communication path, and since SNMP does not hold information for specifying an agent in the inside thereof, there are the following two problems. . First, when a failure notification is generated from the agent, for example, the agent having the IP address 192.168.10.5 of the B system 16 in FIG. 5 cannot be distinguished from the agent having the IP address 192.168.10.5 of the C system 17, On the manager 19 side,
It cannot be determined from which agent the failure notification was issued. Second, in the communication from the manager 19 side,
The agent in the B system cannot be distinguished from the agent in the C system, and the PROXY 20 cannot distribute the communication.

There are two solutions to the above problems in the prior art. In the first method, as shown in FIG. 6, for a customer system having a duplicate IP address, monitoring is not performed by a single manager, but individual managers 21 and 22 are prepared, and a network interferes with each other. This is a method that enables monitoring by not allowing it to be performed. However, when monitoring a large number of customer systems at the same time, in the worst case, it is necessary to prepare the same number of managers as the system to be monitored, which leads to inefficient monitoring of both systems and personnel. Become.

In the second method, as shown in FIG. 7, a conversion table for an address system that does not cause duplication is prepared, and when communication occurs,
This is a method of converting the source and destination IP addresses using this table. In this way, monitoring by a single manager can be possible. However, in this method, it is necessary to convert all overlapping IP addresses to new IP addresses, so it is necessary to prepare as many new IP addresses as the overlapping IP addresses. Then, when the number of monitored systems increases, there is a possibility that the IP address becomes insufficient and the system cannot respond.

The present invention has been made in order to solve the above-mentioned problem, and enables monitoring by a single manager, and enables the use of a duplicate IP address even if the IP address used is duplicated.
It proposes a method that does not require conversion to a new IP address.

[0017]

According to a first aspect of the present invention, there is provided a method for managing a system including a plurality of at least one server machine by a single monitoring manager, wherein the system is included in a certain system. If the IP address of the server machine to be used and the IP address of the server machine included in another system are duplicated, a proxy that mediates communication between each server machine belonging to this system and the monitoring manager is provided to each system. A receiving server is provided, and the monitoring manager, when communicating from the monitoring manager to the server machine, adds, to the communication data, a system ID capable of specifying a system to which the destination server machine belongs, and the proxy receiving server When communication from the monitoring manager to the server machine, Deletes the ID and adds a system ID that can identify the system to which the originating server machine belongs to the communication data when communicating from the server machine to the monitoring manager. This is a system management method.

According to a second aspect of the present invention, the communication uses SNMP as a protocol, and the monitoring manager communicates with the server machine from the monitoring manager when the communication is performed from the monitoring manager to the server machine.
To the protocol, a system ID that can specify a system to which the server machine as a transmission destination belongs is added, and the proxy receiving server deletes the system ID from the protocol when communicating from the monitoring manager to the server machine. 2. The communication method according to claim 1, wherein, when communication is performed from the server machine to the monitoring manager, a system ID capable of specifying a system to which the server machine that is the transmission source belongs is added to the protocol. Is a multiple system management method.

According to a third aspect of the present invention, the proxy receiving server has a unique IP address, and a communication from the monitoring manager to the server machine is provided between the monitoring manager and the proxy receiving server. 3. In this case, a communication distribution server for converting the system ID into an IP address of a proxy reception server provided in a system corresponding to the system ID is provided. Is a method for managing multiple systems when IP addresses overlap.

According to a fourth aspect of the present invention, there is provided a method of managing a system including a plurality of at least one server machine by one monitoring manager, wherein the system includes an IP address of a server machine included in a certain system. And when the IP address of a server machine included in another system is duplicated, when the server machine notifies the monitoring manager of a failure, the server machine that is the notification source includes the server notification in the failure notification data. This is a method for managing a plurality of systems at the time of IP address duplication, characterized by adding a system name capable of specifying a system to which a machine belongs.

According to a fifth aspect of the present invention, the communication uses SNMP as a protocol, and when a failure notification is performed from a server machine to a monitoring manager, the server machine which is a notification source transmits the failure to the server machine according to the protocol. 5. The method according to claim 4, wherein a system name that can specify a system to which the IP address belongs is added.

[0022]

First, a first embodiment of the present invention, that is, a configuration in which a failure is notified from a server machine 18 to a manager 19 will be described with reference to FIG.
A system 15, B system 16, C system 1 shown
Server machine 1 belonging to a distributed system such as 7
In order to centrally manage the server 8 by a single manager 19, a monitoring application called an agent is resident in each server machine 18 to perform monitoring. The agent is the server machine 1 where this agent resides
When the failure of No. 8 is detected, the failure information is notified to the manager 19. Hereinafter, the server machine 18 where the agent resides is simply referred to as the agent 18.

The agents 18 belonging to each system include:
There is a unique IP address within the system, but there may be duplicate IP addresses outside the system. In the present embodiment, the B system and the C system have exactly the same IP address system.

The A system 15, the B system 16, and the C system 17 have agents 1 belonging to these systems.
A proxy reception function 30 that mediates communication between the manager 8 and the manager 19 is provided. Manager 19
Are connected to the proxy reception function 30 of each system via the communication destination selection manager 31. Further, a database 35 in which a virtual-real space conversion table is stored is connected to the communication destination selection manager 31.

Next, the operation of this embodiment will be described. Communication between the manager 19 and the agent 18 is performed by a protocol called SNMP. When the agent 18 detects a failure, the agent 18 notifies the manager 19 of the failure information. At this time, a command called TRAP is used in SNMP. For example, when a certain agent 18 of the C system 17 detects a failure, this information is notified as TRAP communication to the communication destination selection manager 31 via the proxy reception function 30 of the C system 17.

At the time of TRAP transmission, the agent 18
The system information 23, the host name 24, the large item 25, and the medium item 2 are included in the related management information 10 in the SNMP data structure shown in FIG.
6, hierarchical fault information such as a small item 27, a state 28, and a value 29 is entered. FIG. 8 shows, as an example, a warning when the capacity of the hard disk becomes 70% in the host (1) of the trading company A. At this time, system name 23 = A-SYOUSYA, host name 24 = HOST (1), major item 25 = H
D, middle item 26 = DISK (1), small item 27 = USAGE, state 28 =
WARNING, value 29 = 70. Table 3 shows an example of hierarchized fault information.

[Table 3]

Upon receiving the hierarchical fault information, the communication destination selection manager 31 recognizes the source agent 18 from the system name 23 and the host name 24 included in this information, and furthermore, the large item 25 The fault is identified from the item 26, the small item 27, the state 28, and the value 29. As a subsequent operation, the manager 19 changes the symbol of the fault information, and the communication destination selection manager 31 determines which symbol to change with reference to the database 35.

Based on this determination, the manager 19 changes the symbol of the corresponding fault information shown in FIG. 10 and displays the fault information. By the above method, Manager 1
The operator of No. 9 can identify the source agent even when the IP address of the agent 18 is duplicated.

Next, referring to FIGS. 12 and 14, a second embodiment of the present invention, that is, a request communication from the manager 19 to the agent 18 and a response to the request from the agent 18 to the manager 19 will be described. The configuration will be described.

The operator of the manager 19 who has learned the above-mentioned trouble, in order to obtain detailed information,
Request communication is performed from 9 to the agent 18, and the agent 18 returns detailed information to the manager 19. FIG. 12 shows the communication from the manager 19 to the agent 18, and FIG. 14 shows the reply from the agent 18 to the manager 19, but the same configuration is used.

The manager 19 has a communication distribution function 32
Is connected to a proxy reception function 30 provided in each system. Further, a system ID database 36 is connected to the communication distribution function 32. The proxy receiving function 30 has a unique IP address that does not overlap with each other.

In the SNMP of this embodiment, GET, GET NE
Commands such as XT, SET, and GET RESPONSE are used.
As shown in FIG. 11, an agent IP address 33 and a system ID 34 are added to the SNMP data structure. The agent IP address 33 is an IP address assigned to each agent 18. System ID 3
4 is a unique ID number assigned to each of the monitoring target systems 15, 16, and 17.

In the database 36 of the system ID,
A table for converting the system ID 34 into an IP address of the corresponding proxy reception function 30 as shown in Table 4 is stored.

[Table 4] The configuration other than the above is the same as that of the first embodiment.

Next, the operation of request communication (GET or the like) from the manager 19 to the agent 18 will be described with reference to the flowcharts of FIGS. Note that S1 to S8 in the following text indicate the steps in FIG.

When communication from the manager 19 occurs (S1), an extended SNMP packet to which an agent IP address 33 and a system ID 34 are added is created as transmission data (S2). When this packet is sent to the communication sorting function 32, the communication sorting function 32 refers to the conversion table stored in the system ID database 36, and sets the system ID 34 to the IP of the proxy receiving function 30.
The address is converted (S3). Then, it is confirmed whether or not this IP address is registered (S4), and if not registered, the communication failure is notified to the manager 19 (S5). If registered, the packet is transmitted to the proxy reception function 30 having the converted IP address (S6).

Proxy reception function 3 having the corresponding IP address
0 receives the extended SNMP packet, deletes the agent IP address 33 and the system ID 34 from the packet (S7), and transmits the packet to the destination agent 18 as a normal SNMP packet. Then, the agent 18 receives the normal SNMP packet (S8). Therefore, in this method, since the agent 18 only needs to have a function of receiving a normal SNMP packet, there is no need to make any special change to the existing agent.

Next, a reply (GET) from the agent 18 to the manager 19 in response to the request from the manager 19 will be described with reference to the flowcharts of FIGS.
RESPONSE) will be described. Note that S9 to S16 shown in the following text indicate the steps in FIG.

For example, when a communication from an agent belonging to the C system 17 occurs (S9), a normal SNMP packet transmitted from this agent is
Is received by the proxy receiving function 30 of the C system 17 to which the. The proxy reception function 30 adds
Agent IP address 33 of source agent 18
And the system ID 34, and sends the packet to the communication distribution function 32 as an extended SNMP packet (S10).

The communication sorting function 32 that has received the extended SNMP packet sets the system ID 34 added to the packet to
It is checked whether it is correct registered in the system ID database 36, and if it is not registered, the communication failure is notified to the manager 19 (S11 to S11).
14). If it is correct, the packet is sent to the manager 19 (S15). And Manager 1
9 receives the extended SNMP packet (S16). Therefore, the system ID 34 is added to the reply from the agent 18 in the protocol.
Can uniquely identify from which agent 18 the reply is.

[0040]

According to the present invention, a plurality of systems having overlapping IP addresses can be managed by a single manager. Therefore, efficient centralized management can be performed. Also, when performing system monitoring operations, it is not necessary to set up multiple managers even if duplicated IP addresses are used by adding monitored systems according to customer requests, etc. Can be used and the cost does not increase. Furthermore, at this time, there is no need to make any special changes to the already installed agents, such as changing the IP address or supporting extended SNMP communication, and the user recognizes the mechanism of the present invention. You don't have to.

[Brief description of the drawings]

FIG. 1 is a diagram showing communication by SNMP.

FIG. 2 is a diagram showing a conventional SNMP protocol.

FIG. 3 is a diagram showing a system in which IP addresses are duplicated.

FIG. 4 is a diagram showing an example in which communication becomes impossible due to duplication of IP addresses.

FIG. 5 is a diagram showing an example in which communication is enabled by a conventional PROXY function.

FIG. 6 is a diagram showing a method of using a plurality of managers according to the related art.

FIG. 7 is a diagram showing a method based on address conversion according to the related art.

FIG. 8 is a PDU of SNMP TRAP according to the first embodiment of the present invention.
The figure which shows the example of mapping of information with respect to.

FIG. 9 is a diagram showing a notification of a failure from an agent to a manager according to the first embodiment of the present invention.

FIG. 10 is a diagram showing a state change of the symbol of the failure information according to the first embodiment of the present invention.

FIG. 11 shows an extended SN according to the second embodiment of the present invention.
The figure which shows MP protocol.

FIG. 12 is a diagram showing request communication from a manager to an agent according to the second embodiment of the present invention.

FIG. 13 is a flowchart of request communication from a manager to an agent.

FIG. 14 is a diagram showing a reply from an agent to a manager in response to a request from the manager according to the second embodiment of the present invention.

FIG. 15 is a flowchart of a response from the agent to the manager in response to the request of the manager.

[Explanation of symbols]

1 Version 2 Community name 3 PDU 4 PDU type 5 Company ID 6 Agent address 7 General TRAP number 8 Extended TRAP number 9 Occurrence time 10 Related management information 11 Request identification 12 Error status 13 Error location number 14 Management information 15 A system 16 B system 17 C system 18 Server machine (agent) 19 Manager 20 PROXY 21 Manager 22 Manager 23 System name 24 Host name 25 Large item 26 Medium item 27 Small item 28 Status 29 Value 30 Proxy reception function 31 Communication destination selection manager 32 Communication distribution Function 33 Agent IP address 34 System ID 35 Database 36 System ID
Database

Claims (5)

[Claims] 1. A method for managing a system including a plurality of at least one server machine by one monitoring manager, wherein the system includes an IP address of a server machine included in a certain system and an IP address of a server machine included in another system. When the IP address of the server machine to be used is duplicated, each system is provided with a proxy receiving server that mediates communication between each server machine belonging to this system and the monitoring manager, and the monitoring manager includes the monitoring manager System that can identify the system to which the destination server machine belongs in the communication data when communicating from the server to the server machine
ID, the proxy receiving server deletes the system ID from the communication data at the time of communication from the monitoring manager to the server machine, and transmits the communication data at the time of communication from the server machine to the monitoring manager. And a system ID for identifying a system to which the server machine as a transmission source belongs is added to the system. 2. The system according to claim 1, wherein said communication uses SNMP as a protocol, and said monitoring manager can specify a system to which a destination server machine belongs in said protocol when said monitoring manager communicates with said server machine. ID, the proxy receiving server removes the system ID from the protocol when the communication is performed from the monitoring manager to the server machine, and performs the communication with the protocol when the communication is performed from the server machine to the monitoring manager. 2. The method according to claim 1, further comprising adding a system ID capable of specifying a system to which the server machine as the transmission source belongs. 3. The proxy receiving server has a unique IP address, and the system ID is provided between the monitoring manager and the proxy receiving server when communication from the monitoring manager to the server machine is performed. 3. A communication server according to claim 1 or 2, further comprising: a communication distribution server for converting an IP address into an IP address of a proxy reception server provided in a system corresponding to the system ID. System management method. 4. A method of managing a system including a plurality of at least one server machine by one monitoring manager, wherein the system includes an IP address of a server machine included in one system and an IP address of a server machine included in another system. When the server machine sends a failure notification to the monitoring manager when the IP address of the server machine is duplicated, the server machine that is the notification source can specify the system to which the server machine belongs in the failure notification data. A method for managing multiple systems at the time of duplication of an IP address, characterized by adding a system name. 5. The communication uses SNMP as a protocol, and when a failure notification is performed from a server machine to a monitoring manager, a server machine that is a notification source can specify a system to which the server machine belongs according to the protocol. 5. The method according to claim 4, wherein a name is added.