JP2003067264A - Network system monitoring interval control method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the accuracy of an active/non-active check, which is performed by a monitor system, for devices in a network system. SOLUTION: A monitor manager (MGR) in a center monitor system 4 issues an active/non-active check request to a monitor agent (AGT) 2 in each site monitor device 1 at a time when a first time passes from the latest active/non- active check monitor time and then waits a response from the AGT 2. If not receiving the response from the AGT 2 within a predetermined time, the monitor manager re-issues an active/non-active check request to the agent at a time when a second time, which may be variably set, passes.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of controlling a monitoring interval of a network system, and more particularly, to a monitoring interval control method for dynamically improving the life / death monitoring confirmation accuracy by dynamically changing the life / dead monitoring interval of a monitored object. Regarding

[0002]

2. Description of the Related Art With the recent widespread use of corporate network infrastructure, servers such as large-scale computer systems, workstations, and personal computers are connected to each other on the basis of a network system to perform various tasks. Along with this, the network system becomes the basis of the information infrastructure, and the network system is available 24 hours a day, 3 days a week.
It is necessary to monitor continuously for 65 days. For monitoring the network system, the monitoring device uses SNMP (Simple N
etwork Management Protocol
It is general to monitor devices such as servers and routers in the network system based on l). Regarding SNMP, RFC (Request For Comm) relating to SNMP internetwork management.
ent) 1157.

A monitoring device based on SNMP is provided with fault information MIB (Media Information) added to a fault notification interrupt (trap) from a network system.
display the contents of the (Action Base) on the monitoring screen,
Further, it is stored in the file device as failure history information. Further, as a method of confirming the life or death of a server, a router, or the like to be monitored, as described in the above RFC, PING (P
A method of issuing an "acket Internet Group" from a monitoring server and confirming its response is generally known. Therefore, by applying the network system management technology described above, it is possible to combine and monitor the network system and the server system.

Further, Japanese Patent Laid-Open No. 11-96043 discloses a center monitoring system in a distributed system. That is, in response to failure or automatic recovery information from a large number of distributed systems, a failure message is displayed on the monitoring console screen, and a delete command is executed to delete and display the failure message. The number of messages displayed when monitoring the situation on the console screen of the monitoring center is reduced so that the operator can easily understand the situation.

Therefore, the network monitoring staff looks at the performance data such as the failure information and the CPU utilization rate of the server displayed on the monitoring screen of the monitoring device having the above monitoring function,
Judges faults in network systems and server systems and monitors performance. In particular, the above-mentioned PING method is a powerful method for easily grasping the life and death of the monitored object, and is a generally well-known method.

[0006]

The above-mentioned PING confirmation of whether the monitoring target is alive or not is sufficiently effective when the load of the monitoring target device has a relatively large margin and the load fluctuation is small.
However, when the load on the monitored device is relatively large and the load fluctuation is large, the monitored object may become unresponsive to the life-and-death confirmation command from the monitoring center. Although the monitored object is operating, there is a high possibility that it will be erroneously recognized as a dead state, that is, a down state. In particular, when outsourcing the monitoring work of the network system to an external company, that is, outsourcing, such a problem is likely to occur. Since network system outsourcing companies usually monitor the network system of multiple customers via the network from the monitoring device of the monitoring center, the load environment in which the monitored object is placed varies depending on the system on the customer side. Problems may occur.

An object of the present invention is to address the above-mentioned problems and to improve the accuracy of confirming whether a monitored object is alive or dead.

[0008]

SUMMARY OF THE INVENTION The present invention is a monitoring method executed by a monitoring manager provided in an environment of a monitoring side, which is performed by a monitoring manager for confirming whether an agent in an environment to be monitored in a network system is alive or dead. When a first time interval elapses from the life-and-death confirmation monitoring time of, a step of issuing a life-and-death confirmation request to the agent, a step of waiting for a response from the agent, and a variable setting when there is no response from the agent within a predetermined time And reissuing a life-and-death confirmation request to the agent when the second time interval has elapsed, the monitoring interval control method of the network system.

[0009]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a block diagram of the operation monitoring system of this embodiment. The monitoring system includes an information processing system such as a server center 100 to be monitored and a customer site 200, and a monitoring center 3 that monitors these information processing systems.
00 information processing system.

The server center 100 and the customer site 200
In this system, monitored devices such as a workstation (WS) 101, a server 102, a firewall (F / W) 104, and a router 105 are connected to a site monitoring device 1 via a local area network (LAN) 103. Has become. The site monitoring device 1 mounts a monitoring agent (AGT) 2 as a program. In this embodiment, the monitoring agent 2 is installed in the independent site monitoring device 1, but in general, the monitoring agent 2 may be installed in another device such as the server 102.

The system of the monitoring center 300 has a center monitoring device 4. The center monitoring device 4 is connected to the information processing system of the server center 100 and the customer site 200 via a network such as a LAN or a wide area network (WAN). The center monitoring device 4 mounts a monitoring manager (MGR) 5 as a program, and stores a monitoring management table 10 and an agent management table 11 on its memory.

Monitoring agent 2 of site monitoring device 1
Monitors the information processing devices such as the workstation 101, the server 102, the firewall 104, and the router 105, and network devices such as life and death confirmation,
Information on the monitoring result is collected and recorded in the log information file 3. The monitoring manager 5 of the center monitoring device 4 refers to the monitoring management table 10 and the agent management table 11 to send a life / death confirmation command to each site monitoring device 1. When the monitoring agent 2 of each site monitoring device 1 receives this life-and-death confirmation command, it performs life-and-death confirmation of each information processing device or network device to be monitored, and sends the result to the monitoring manager 5. The monitoring manager 5 stores the live / dead confirmation result collected from each site monitoring device 1 in the monitoring information database (DB) 6.

The monitoring agent 2 monitors the state of a device group to be monitored based on the operating rules of SNMP. When the monitoring agent 2 detects the occurrence of a failure in any of the devices, it issues a failure notification interrupt (trap) to the monitoring manager 5 and sends failure information to MIB (Media In).
It is sent to the monitoring manager 5 as a formation base). Regarding the trap and MIB,
RFC 11 for SNMP internetwork management
57 and RFC792.

FIG. 2 is a diagram showing the data structure of the monitoring management table 10 and the agent management table 11 managed by the monitoring manager 5. The monitoring interval (Tn) of the monitoring management table 10 is a standard monitoring time interval for the agent, and the unit is arbitrary such as ms, seconds, minutes. The number of manager retries (R) is the number of retries of the command when there is no response from the monitoring destination agent to the life confirmation command. The number of agents (N) is the number of agents to be monitored, and is the number in the agent management table 11. The agent pointer (Ai) stores a pointer to the agent management table 11 provided for each agent.

The agent type (TPE) of each agent management table 11 is set to a value of 0 or 1. When TPE is '0', the monitoring interval (Tn) of the monitoring management table 10 is used as the monitoring time interval,
When TPE is “1”, the constant monitoring interval value (TA) or the variable monitoring interval (T) of the agent management table 11 is used.
The value of V) is used. Monitoring interval type (TM) is 0
Alternatively, a value of 1 is set. When retrying life-or-death monitoring, when TM is "0", TP is set as the monitoring time interval.
The value of Tn or TA is used according to the value of E, and when TM is '1', the value of TV is used. When TPE is '0', TM must be '0'. Range (TR) sets the range of random numbers generated when the TV is used. The response waiting time (TW) is a time period during which the monitoring manager 5 waits for a response after issuing the life confirmation instruction to the monitoring agent 2. The final confirmation time (TL) is
The time of the last confirmation of life and death is stored. The initial value of TL is 0.

When the monitoring agent 2 and the monitored object are mounted on a device having a relatively large load and a small load fluctuation, it is sufficient to use Tn or TA as the monitoring time interval. Further, when the monitoring agent 2 and the monitoring target are installed in a device with a small load and large load fluctuation, the case where the monitoring agent 2 or the site monitoring device 1 is erroneously recognized as a dead state by using a TV Can be reduced.

FIG. 3 is a flow chart showing the flow of the monitoring process of the monitoring manager 5. The monitoring manager 5
For the monitoring agent 2 to be monitored, the agent type (TP) of the agent management table 11 pointed to by the agent pointer (Ai) of the monitoring management table 10.
The value of E) is determined (step 31). If TPE is "0", the monitoring interval (Tn) is moved to the monitoring interval (Tc) of the work area of the memory (step 32). T
If PE is '1', the constant monitoring interval (TA) is moved to Tc (step 33).

Next, the monitoring manager 5 determines whether it is the life confirmation time (step 34). If the final confirmation time (TL) is 0, it is the life and death confirmation time. If TL is not 0, it is the life-and-death confirmation time if the current time ≧ TL + Tc. Even when "1" is selected as the monitoring interval type (TM), that is, when the variable monitoring interval (TV) is selected, the monitoring interval does not need to be variable for the first life-or-death confirmation command, and therefore, it is constant in this embodiment. The monitoring interval (TA) is applied. If it is not the life confirmation time, go to step 40. If it is the life-and-death confirmation time, a life-and-death confirmation command is issued to the monitoring agent 2 (step 35). Next, wait for a response for the response waiting time (TW) (step 3).
6).

If there is a response within the waiting time (step 37)
(YES), the monitoring manager 5 stores the current time at the final confirmation time (TL) (step 38), and stores the live / dead confirmation result of the monitoring target in the monitoring information DB 6 according to the response (step 39). For example, the IP address of the monitored device and M
Store the IB. Next, the monitoring agent 2 in the next order is selected (step 40) and the process returns to step 31.

If there is no response within the waiting time (step 3
7 NO), the monitoring manager 5 determines whether or not the number of retries on the work area has reached the number of manager retries (R) (step 41). If the number of retries (R) has not been reached, it is determined whether the monitoring interval type (TM) is "0" (step 42). If TM is '0', wait for the value of the monitoring interval (Tc) (step 4
3), 1 is added to the number of retries on the work area of the memory, and the process returns to step 35. If TM is "1", the value of the variable monitoring interval (TV) is calculated according to the processing routine shown in FIG. 4 (step 44), and the value of the TV is waited for (step 45). 1 is added to and the process returns to step 35. When the number of retries on the work area reaches the number of manager retries (R) (step 41)
YES), the fact that the target object is in the failure state is stored in the monitoring information DB 6, an error is displayed on the display screen (step 46), and the process goes to step 40.

FIG. 4 is a diagram showing a processing procedure for calculating uniform random numbers as an example of calculating the value of the variable monitoring interval (TV). The monitoring manager 5 is the lower 8 of the current time timer value.
The bit is acquired and its value is placed in the work area WK (step 51). Next, a random number value in the range of 0 to (TR-1) is calculated (step 52). Specifically, the value of WK is multiplied by a prime number to obtain WK1, and the value of WK1 is divided by TR to obtain T as the quotient.
The product of R is subtracted from WK1 to obtain WK2. WK
2 is the modulo of WK1 modulo TR. Next WK
Set the value of 2 to the variable monitoring interval (TV) (step 5)
3).

Although the processing operation centering on the processing of the monitoring manager 5 has been described above, the monitoring agent 2 plays the role of an agent for the monitoring manager 5, while monitoring the device to be monitored as an agent. Operates to play the role of manager 5. The processing procedures of the monitoring management table 10, the agent management table 11, and the monitoring manager 5 described above can be similarly applied to the monitoring manager 5 portion of the monitoring agent 2.

As described above, according to the present embodiment, when the load of the monitoring agent 2 and the monitoring target varies, the life and death confirmation time for the monitoring agent 2 and the monitoring target is used by using the variable monitoring interval (TV). Can be shifted from the time when the load increases, so that the non-response state of the monitoring agent 2 can be reduced, and false recognition of life and death confirmation can be reduced, and the accuracy of life and death confirmation monitoring of the monitoring target can be improved. It is possible to

[0025]

As described above, according to the present invention, the monitoring manager uses the variable monitoring time interval at the time of retry of the life confirmation command so as to confirm the life and death so as to avoid the time point when the monitored object has a high load. It is possible to issue a command, and it is possible to improve the accuracy of confirming whether the monitored object is alive or dead.

[Brief description of drawings]

FIG. 1 is a configuration diagram of an operation monitoring system according to an embodiment.

FIG. 2 is a diagram showing a data configuration of a monitoring management table 10 and an agent management table 11 of the embodiment.

FIG. 3 is a flowchart showing a flow of monitoring processing of a monitoring manager 5 according to the embodiment.

FIG. 4 shows an example of calculating the value of a variable monitoring interval (TV).
It is a figure which shows the process sequence which calculates a uniform random number.

[Explanation of symbols]

1: Site monitoring device, 2: Monitoring agent, 4: Center monitoring device, 5: Monitoring manager, 10: Monitoring management table, 11: Agent management table

Continued front page    (72) Inventor Katsunobu Okuda             890 Kashimada, Sachi-ku, Kawasaki City, Kanagawa Stock             Company Information Service Division, Hitachi, Ltd. (72) Inventor Kazunari Takenoshita             890 Kashimada, Sachi-ku, Kawasaki City, Kanagawa Stock             Company Information Service Division, Hitachi, Ltd. F-term (reference) 5B089 GA11 GA21 GA31 GB02 JA35                       JB14 KA12 KA13 KB04 MC06                 5K030 HC01 HC14 HD03 HD06 JA10                       LA08 MA01

Claims (5)

[Claims] 1. A monitoring method executed by a monitoring manager provided in a monitoring environment, wherein the monitoring manager confirms whether an agent in a monitored environment in a network system is alive or dead. A step of issuing a life-and-death confirmation request to the agent when a time interval of 1 has elapsed, a step of waiting for a response from the agent, and a variably set second time when there is no response from the agent within a predetermined time. Re-issuing a life-and-death confirmation request to the agent when the time interval of is passed, the monitoring interval control method of the network system. 2. The method according to claim 1, wherein the first time interval is a preset constant time. 3. The method according to claim 1, wherein the second time interval is a calculated random value. 4. The first time interval is applied instead of the second time interval for some of the plurality of agents monitored by the monitoring manager. Method for controlling monitoring interval of network system. 5. The monitoring interval control of a network system according to claim 1, wherein said agent executes each of said steps by said monitoring manager when confirming the life or death of an agent to be monitored further lower. Method.