JPH04257035A - Fault information processing system under virtual computer system - Google Patents

Abstract

PURPOSE:To operate the early confirmation and deal of an abnormal surrounding device, in a fault information processing system under a virtual computer system. CONSTITUTION:At the time of the occurrence of the fault of a surrounding device 4, a surrounding device abnormality informing system under an operating system(OS) 20 of each guest virtual computer (guest VM) 2 receives both logging information detected by the OS 20 under each guest VM 2 and the logging information detected by a host system 1, by an interface between the OS 20 and the host system 1, and analyzes the logging information by each virtual equipment number. Then, the error code information of the virtual equipment number is converted into that of a real equipment number by each degree of importance, based on an equipment number management table equipped with virtual equipment number/real equipment number corresponding information preliminarily received from the host system 1, and it is automatically transferred to a maintenance center 3.

Description

[Detailed description of the invention]

0001

[Industrial Application Field] Recently, there has been an increasing demand for virtual computer systems in order to migrate electronic computer systems and to make effective use of hardware.

[0002] In the virtual computer system, a plurality of guest virtual machines (hereinafter referred to as guest VMs) operate under the operating system (OS) of the host system.
), it operates like an independent computer, and when a failure occurs in a peripheral device running under the guest VM, the failure information is displayed in the virtual machine number assigned to the peripheral device. , each recorded in a separate logging file.

[0003] In such a virtual computer system, as the scale of the system increases recently, the amount of logging information described above becomes enormous, and all of the failure information is recorded using virtual machine numbers, so it is difficult to manually The investigation method has problems such as overlooking serious failures and the need to match the machine number with the actual equipment, so there is a need for an effective failure information processing method that is suitable for increasing the scale and diversification of systems. It has come to be said that

0004

2. Description of the Related Art FIG. 10 is a diagram showing a conventional fault information processing method under a virtual computer system.

Conventionally, after a failure occurs or during regular maintenance,
Or, at the end of the business, the logging information ■ under the operating system (OS) of each guest VM 2 and the logging information ■ of the host system 1 are transferred to the guest VM 2.
The error information editing system 25 under the operating system (OS) of M2 edits the information, and transfers the edited list to the maintenance center 3 for output.

[0006]

[Problem to be Solved by the Invention] At the maintenance center 3, a maintenance engineer looks at the output edit list and
Although fault investigation and preventive maintenance were carried out, in this conventional method, the huge amount of logging information edit list that comes with the increase in the scale of the virtual computer system is manually investigated, so there are differences in the skills of maintenance engineers, etc. There was a problem in which serious failures were overlooked due to oversight.

[0007] Furthermore, for the operating system (OS) under each guest VM 2, the machine number of the peripheral device is
Since it is expressed as a virtual address, in the logging information that is recorded when a failure in the peripheral device is detected, all machine numbers of the peripheral device in question are output as the above virtual address (virtual machine number). Therefore, it was necessary to match the actual machine number.

[0008] Also, logging information (■) detected by the host system 1 of the virtual computer system, and logging information (■) detected by the operating systems (OS) under the plurality of guest VMs 2 constituting the virtual computer system.
There was a problem in that the files had to be edited individually.

That is, in the conventional failure information processing method under a virtual computer system, (1) there is an oversight when a maintenance engineer checks a large amount of logging information editing results;

(2) Increase in scale of virtual computer system,
Shortage of maintenance engineers due to diversification. (3) In a virtual computer system, logging information for peripheral devices is written using the virtual machine number, resulting in a misjudgment of the faulty device.

(4) There is a lot of waste in analysis work man-hours and resources (resources used for analysis in each virtual computer system, such as paper for output lists). (5) Comprehensive and accurate analysis and understanding depends on the skills of maintenance engineers.

There were problems such as [0012]. SUMMARY OF THE INVENTION In view of the above-mentioned conventional drawbacks, it is an object of the present invention to provide a fault information processing method under a virtual computer system that can quickly identify and deal with abnormal peripheral devices.

[0013]

[Means for Solving the Problems] Figures 1, 2, and 3 are diagrams showing the principle structure of the present invention, and (a) shows how the peripheral device abnormality reporting system under the operating system of each guest VM edits failure information. (b1) and (b2) show cases in which the server VM collectively manages and reports to the maintenance center. The above problem is solved by a failure information processing method under a virtual machine system configured as follows.

(1) In the virtual computer system, when a failure occurs in the peripheral device 4, the logging information detected by the operating system (OS) 20 under each guest virtual machine (guest VM) 2, and
The logging information detected by the host system 1 is sent to the operating system (OS) 20 of each guest virtual machine (guest VM) 2 by the peripheral device error reporting system 21 at the time of occurrence of the failure. , and host system 1
The above host system 1
Based on the device management table with the virtual machine number/real machine number correspondence information received from the virtual machine number→
Convert it to the error code information of the actual machine number and send it to the maintenance center 3
Configure automatic forwarding to {See Figure 1 (a)}
(2) In the virtual computer system, logging information detected by each guest virtual machine (guest VM) 2 and the underlying operating system (OS) 20
The logging information detected by the host system (1) is sent to a specific guest virtual machine (guest VM) 2a
A device management system is provided in which the specific guest virtual machine (guest VM) 2a is equipped with the virtual machine number/actual machine number correspondence information that has been received from the host system 1 in advance. Based on table ■, it is configured to analyze the failure information for each virtual machine number, convert it into error code information from virtual machine number to real machine number, edit it, and output it. {See Figure 2 (b1), Figure 3 (b2)}

0
015]

[Operation] That is, according to the present invention, first, each of the plurality of guest VMs constituting the virtual computer system is
Connection information (VM-specific information) of each guest VM is transmitted from the host system, and each guest VM generates a device management table (2) based on the VM-specific information received from the host system.

[0016] In the device management table (2), the guest VM
The subordinate device, the time when the device was connected,
Correspondence information between the real machine number and virtual machine number of each device is described. Also, among each guest VM, a specific guest VM receives the logging information detected by the host system and the logging information detected by its own operating system, and other guest VMs receive the logging information detected by their own operating system. Only information ■ is stored in the system error logging file by each peripheral device abnormality reporting system ■
For example, the failure information corresponding to the latest virtual machine number is about 1
Extract 0 items and expand them on main memory.

At this time, when a new failure occurs, in the case of the present invention, the failure information is stored in the system error logging file (2), and is also transferred to the peripheral device abnormality reporting system, and the information is expanded as described above. The latest fault information is incorporated, and the oldest fault information is excluded, so that the latest fault information is always expanded.

The present peripheral device abnormality notification system also reads out the tally file (2) that tallies the temporary errors up to the previous time, and expands it on the main memory. The tally file ■ has the following information:
For example, in the case of a disk device (DASD), a cumulative counter of temporary failures is stored for each cylinder/head.

[0019] Also, this peripheral device abnormality reporting system uses a threshold value (number of errors) for determining whether or not to transfer the information of each temporary error mentioned above to the maintenance center and the maintenance base in charge from the threshold value file (2). is read and expanded onto main memory.

[0020] The peripheral device abnormality reporting system analyzes the latest failure information stored in the main memory, classifies it by importance, and immediately encodes it into a code in the case of a permanent fixed failure. In the case of a temporary error, the data is accumulated as aggregated data, and compared with the threshold data, and if the threshold is exceeded, it is coded and transmitted to the maintenance center or the like. At this time, the virtual machine number in the fault information is converted into a real machine number and transferred based on the device management table (2) developed on the main memory.

Next, the operation of another fault information processing method of the present invention will be explained. This method collects logging information detected by the operating system under multiple guest VMs that make up the virtual machine system, and
A specific guest VM among the plurality of guest VMs stores the logging information detected by the host system on the server V.
M, which collectively collects and manages the fault information corresponding to each virtual machine number, edits it into fault information corresponding to the real machine number, and outputs it to an output device.

Specifically, in response to a request from the server VM, the batch collection program of each guest VM reads failure information from its own system error logging file (■),
After converting it into a data format common to the system and storing it in a temporary file, it is converted into a communication data format and then
Transfer to M (server system).

[0023] Each fault information transferred to the server VM is referenced to the device management table (1), which is generated in advance based on the VM-specific information transferred from the host system, and the fault information is Replace the virtual machine number with the real machine number and store it in the cumulative file, for example.

[0024] The failure information stored in the cumulative file is
The log editing program executed by the server system edits the information into the format of, for example, device number, error event, and error occurrence date information, and then transfers and outputs it to a maintenance center or the like as an analysis editing list.

That is, the output format of the failure information in this case is as follows:
Converting the virtual machine number in the fault information to the real device machine number and outputting it is different from when each guest VM mentioned above edits the logging information etc. that it manages and outputs it individually. However, in the case of this method, a specific server VM
A place where you can collect all the fault information logged by each guest VM (this includes the logging information detected by the host system) and edit and output it as a fault information list corresponding to the actual machine number. are different.

[0026] In this way, in the conventional method, maintenance personnel look at the analysis and editing list for each system corresponding to the virtual machine number edited for each system, that is, each guest VM, and perform a comprehensive analysis using their own skills. In contrast, in the fault information processing method according to the present invention, it is necessary to understand the fault situation immediately and proceed with maintenance work.However, in the fault information processing method according to the present invention, the fault is classified into fixed faults and temporary faults based on the actual machine number, not the virtual machine number. This makes it possible to accurately identify the device where the failure occurred, and if the failure information is collectively managed and output by a specific server VM, the failure information is consolidated, so it is possible to accurately identify the device where the failure has occurred. Get accurate and comprehensive information with the analysis compilation list. Therefore, the present invention is useful for early detection and early treatment of suspect parts, and has the effect of making it possible to provide high-quality maintenance services.

[0027]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Examples of the present invention will be described below in detail with reference to the drawings. The aforementioned FIGS. 1, 2, and 3 are diagrams showing the principle configuration of the present invention, and FIGS. 4 and 5 are diagrams schematically showing an embodiment of the present invention, and (a) shows VM-specific information. (b)
shows an example of developing failure information for each machine number, (c) shows an example of developing a total file, (d) shows an example of developing a threshold file, and (e) shows an example of the format of code information. , FIGS. 6 to 8 are diagrams schematically showing other embodiments of the present invention, in which (a) the fault data is transferred to the server V
(b) shows an example of the configuration of the device management table, (c) shows the operation flow, and FIG. 9 is a diagram showing the virtual machine number → real machine number conversion method of the present invention. It is.

In the present invention, while the virtual computer system is operating, peripheral devices such as magnetic disk devices are temporarily disabled.
Or, upon a permanent failure, each guest VM
The logging information detected by the operating system (OS) under 2 and the logging information detected by the host system are analyzed and classified by the peripheral device error reporting system 21 under each guest VM 2, and maintained based on the actual machine number. Means for automatically notifying the center 3, or one of the guest VMs 2 configuring the virtual computer system,
Functions as server VM 2a, and each of the above guest VMs
The present invention is implemented by means of collecting the logging information in step 2 all at once, analyzing the accumulated information, editing it into error information corresponding to the actual machine number, and outputting it to the maintenance center 3 as an analysis and editing list. It is a necessary means. Note that the same reference numerals indicate the same objects throughout the figures.

Hereinafter, while referring to FIGS. 1 to 3, FIGS.
A fault information processing method under the virtual computer system of the present invention will be explained with reference to FIG. 5, FIGS. 6 to 8, and FIG. 9. First, the peripheral device abnormality reporting system shown in Figure 1
21 is automatically started after each operating system is initial program loaded (IPL), and continues to operate until the operating system (hereinafter referred to as OS) is stopped.

(1) Peripheral device abnormality reporting system 2
1 is from the host system 1, for example, as shown in FIG. 4(a).
V received based on the VM-specific information reception format shown in
Based on the M-specific information (■), the current real/virtual machine number correspondence information is read and expanded into the main memory in the form of the device management table (■) shown in FIG. 7(b). Further, this table is updated every time the correspondence between real and virtual machine numbers is changed.

(2) For each device, read out the necessary number of information from the system error logging file ■, for example, about 10 items from the latest logging information, and store it in the main memory in the format shown in Figure 4(b). Let's expand it.

[0032] The system error logging file (2) stores all past error information regarding the peripheral devices of the virtual computer system. The peripheral device abnormality reporting system 21 of the present invention reads out the latest information, for example, 10 pieces of information, and expands it on the main memory as described above.

(3) The contents of the previous temporary error summary file ■ are developed on the main memory in the format shown in FIG. 4(c). In the case of a file storage device (DASD), for example, each cylinder/head has an error counter, and each time a temporary error is detected, it is counted on the expanded counter.

(4) In addition, the threshold file ■ is shown in FIG.
) is expanded in the main memory in the format shown in . This is a threshold for the number of temporary errors, and only when the number of temporary errors in the aggregated data expanded on the main memory exceeds the threshold, the fault information is encoded. Transferred to the maintenance center and the responsible maintenance base 3. At this time, the counter is cleared to "0".

(5) When the OS detects new failure information of peripheral devices or statistical information (consumables information, seek count information, read data byte information for DASD, etc.), the above system error logging file is At the same time, it is also transferred to the peripheral device abnormality notification system 21 and reflected in the failure information for each machine number expanded in the main memory (see FIG. 4(b)). Specifically, as described above, when the new error information is added, the oldest failure information is deleted.

(6) The log information detected by the host system 1 of the virtual machine system and stored in the file device as the host log ■ is also stored in a specific guest VM 2 through the interface between the guest VM 2 and the host.
The information is passed to the peripheral device abnormality reporting system 21.

(7) Peripheral device abnormality reporting system 2
1 classifies the logging information developed in the format shown in Figure 4(b) above on the main memory by importance level,
In the case of a permanent error, code it immediately. In addition, temporary errors are accumulated as aggregated data and compared with the threshold data stored in the main memory (see Figure 5 (d)), and only when the threshold is exceeded is the temporary error detected. Errors are coded. The format of the encoded data is shown in FIG. 5(e).

[0038] Here, the device classification is, for example, DASD/
The situation recognition code indicates the severity of the error, for example, whether it is a fixed failure or a temporary failure.The notification target device identification code indicates the actual device number, and indicates the abnormal state (overview/ details) indicates the failure information logged in the system error logging file (2), and device type details indicates the type of device.

[0039] Taking the above DASD as an example,
Based on the logging information that occurs, errors are classified into correctable errors (temporary errors) and uncorrectable errors (permanent errors). code and transfer it to maintenance center 3, etc. In the case of correctable errors, the total is calculated based on the location (cylinder/head) where the error occurred. When the total value exceeds the threshold value, it is encoded and transmitted. At this time, set the corresponding counter of the above aggregated data to “0”.
clear. Also, past events may be included depending on the degree of disability. Also, at this time, the actual device machine number is set in the code based on the device management table {see FIG. 7(b)} expanded in advance on the main memory.

(8) The coded information is
7, immediately to the maintenance center and the responsible maintenance base.
3 will be notified. {See FIG. 1(a)} (9) Based on the coded information, the suspected location and the severity of the suspected problem are displayed on the display 3a on the center side.

(10) When the operation of the peripheral device abnormality reporting system 21 is stopped, the above-mentioned tally data developed on the main memory is recorded in the tally file (2). In the above embodiment, in a virtual computer system composed of a plurality of guest VMs, a peripheral device abnormality reporting system 21 that operates under the OS of each guest VM collects logging information detected by the OS and the host system. ■，
■ Read out, classify by importance, code immediately in the case of a fixed failure, aggregate it in the case of a temporary failure, and if the total value exceeds a predetermined threshold. It is unique in that it is coded and notified to a maintenance center etc. for the first time.

Next, another embodiment of the present invention will be described with reference to FIGS. 6 to 8. In this embodiment, for example, it is assumed that three guest VMs (systems) 2 are operating under the virtual computer system.

First, the host system (meaning a virtual computer system) 1 has the number of operating systems (guest V
M number), system name information, and correspondence information between the actual machine numbers of devices (peripheral devices) built into the system and logical machine numbers (virtual machine numbers) uniquely defined for each system. Created and managed as information ■.

The host system constantly monitors the operating system management information (2), and creates a device management table {see FIG. 7(b)} each time information is added or changed.

The created device management table ■ is recorded as management information in the error information storage file (host log) ■ of the host system 1, and is used as reference information when accumulating log information together with the above operating system management information. used.

[0046] When a hardware error occurs in each subordinate device while each guest VM 2 is operating, error information is created in the guest VM to which the device is connected and stored in the system error logging file (LOGREC). be done.

In the case of the present invention, a log batch collection program 22 is prepared in each guest VM 2, and when a certain guest VM 2 stands as a server VM and is recognized by the host system 1, the guest VM 2 becomes a server VM (server system) 2a, and the log batch collection program 22 is started. {Figure 8 (c
) Processing step 200} The server VM 2a
When the log batch collection program 22 is started, it establishes a data communication path between the server VM 2a and each guest VM 2, as shown in FIG. Bulk log collection program 22
Start. {Processing step 201 in FIG. 8(c)
203 Reference} The log batch collection program 22 of each guest VM 2 started by the server VM 2a is
Own system error logging file (LOGREC)
) Convert the data in ■ into a data format common to each system and store it in a temporary file.

[0048] After that, the data stored in each temporary file is converted into the communication data format, and the data stored in the above-mentioned server V
M 2a , data transfer is performed via the data communication path. {Processing steps 208 to 210 in FIG. 8(c)
, 206} On the other hand, the log batch collection program 22 of the server VM 2a collects the data in its own system error logging file (LOGREC), and
Transfer the data in the host system 1 error information storage file (host log) ■. {See processing steps 204 and 205 in FIG. 8(c)} The server VM 2
The data transferred to the log batch collection program 22 of a is determined by referring to the operating system management information ■ received from the host system 1 and the device management table ■, and identifying the devices in the error information of each guest VM 2. Replace the machine number information (logical machine number) with the actual machine number and store it in the cumulative file. {See processing step 207 in FIG. 8(c)) The data stored in the cumulative file is stored in the server V.
Log editing program provided in M2a 23
Analyzed and edited by.

At this time, as shown in FIG. 3 (b2), an analysis and editing list is output in the form of, for example, actual device machine number, error type, and date of occurrence, and is sent to the maintenance center.
Transferred to 3rd class.

[0050] The maintenance engineer (CE) of the maintenance center 3 accurately grasps the operating status of the hardware and the failure status using the actual machine number based on the output analysis and editing list.
Can perform maintenance work.

Next, referring to FIG. 9, a means for converting error information corresponding to a virtual machine number recorded under the guest VM 2 into error information corresponding to a real machine number will be explained. As mentioned above, the host system 1 stores information such as the number of systems (guest VMs) in operation and the names of the systems (guest VMs) as operating system management information ■, and the system (guest VM).
The VM-specific information (2) contains correspondence information between the actual machine number of the device incorporated in the guest VM (guest VM) and the logical machine number (virtual machine number) uniquely defined for each system.

[0052] The host system 1 always monitors the operating system (guest VM) and the device machine number information connected to that system (guest VM), and updates the information every time there is a change or addition of information. In host system 1,
Store in the error information storage file (host log) described above.

[0053] The host system 1 is a certain guest VM.
2 is running as a server VM, for example, when the power is turned on or an instruction from the user is triggered.
The VM-specific information ■ stored in the above error information storage file (host log) ■ on the server VM 2a
In addition, if there is a new change or addition, it is stored in the error information storage file (host log), and the change or addition information is directly transferred to the server VM 2a.
Transfer to.

The server VM 2a creates the device management table {see FIG. 7(b)} based on the VM-specific information transferred from the host system 1. In server VM 2a, log editing program 2
3. When creating the analysis edit list using step 3, convert the virtual machine number in the log information transferred from each guest VM 2 to a real device machine number by referring to the device management table Create a list and forward it to Maintenance Center 3.

The conversion process from the virtual machine number to the real machine number is performed using the fault information processing method explained in FIGS. , Temporary errors are transferred from the host system 1 in each guest VM 2 in advance, even when the number of error occurrences is counted, and when the threshold is exceeded, the code is coded and transferred to the maintenance center 3. This is performed based on the virtual/actual machine number correspondence table (2) created based on the VM-specific information (2).

[0056]

[Effects of the Invention] As explained above in detail, the present invention provides a failure information processing method under a virtual computer system, when a peripheral device such as a magnetic disk device temporarily or , each guest virtual machine (guest VM
) The logging information detected by the underlying OS and the logging information detected by the host system are sent to each guest VM.
The subordinate peripheral device abnormality reporting system analyzes and classifies the information.
Automatically notify the maintenance center with the actual machine number. Or, one of the guest VMs configuring the virtual computer system functions as a server VM, and the logging information detected by each of the above guest VMs and the host system is collectively collected and accumulated. is analyzed, edited into error information corresponding to the actual machine number, and output to the maintenance center as an analysis edit list, so it is possible to identify suspect parts due to failures in peripheral devices such as magnetic disk devices in virtual computer systems. It is useful for early extraction, early treatment, and early detection of consumable parts of the peripheral device. Additionally, in the past, maintenance staff had to use their skills to comprehensively grasp the situation and proceed with maintenance work by looking at the analysis and editing list corresponding to virtual machine numbers for each system, which was edited for each system. On the other hand, according to the present invention, the information is consolidated, and some analysis and editing lists are accurate and accurate.
Comprehensive information is now easily available. Furthermore, since the actual machine number is notified rather than the virtual machine number, it is possible to accurately identify the device in which the failure has occurred. Furthermore, the above functions have the effect of reducing maintenance work time and strengthening comprehensive judgment, thereby making it possible to provide higher quality maintenance services.

[Brief explanation of the drawing]

[Figure 1] Principle configuration diagram of the present invention (Part 1)

[Figure 2] Principle configuration diagram of the present invention (Part 2)

[Figure 3] Principle configuration diagram of the present invention (
Part 3)

[Fig. 4] A diagram showing one embodiment of the present invention (Part 1)
)

[Fig. 5] Diagram showing one embodiment of the present invention (Part 2)

[Fig. 6] Diagram showing another embodiment of the present invention (Part 1)

[Figure 7]
Diagram showing another embodiment of the present invention (Part 2)

[Fig. 8] Diagram showing another embodiment of the present invention (Part 3)

[Figure 9] Diagram showing the virtual machine number → real machine number conversion method of the present invention

[Figure 10] Diagram showing a failure handling method under a conventional virtual computer system

[Explanation of symbols]

1 Host system 2 Guest virtual machine (guest VM, system)
2a Server VM (server system) 20 Operating system (OS) 21 Peripheral device error reporting system 25 Error information editing system ■ Host system error information storage file or logging information ■ System error logging file or logging information ■ Total file
■ Threshold file ■ Operating system management information ■
VM-specific information ■ Device management table or virtual/actual machine number correspondence table

Claims (2)

[Claims] Claim 1: In a virtual computer system, a peripheral device (
4) When a failure occurs in each guest virtual machine (
Guest VM) (2) The underlying operating system (
Logging information (■) detected by OS) (20) and logging information (■) detected by host system (1)
■) The peripheral device error reporting system (21) under the operating system (OS) (20) of each guest virtual machine (guest VM) (2) ), as well as,
It receives it through the interface with the host system (1), analyzes the logging information for each virtual machine number, and
Device management table with virtual machine number/actual machine number correspondence information received from the above host system (1) (■)
A failure information processing method under a virtual computer system characterized by converting error code information from virtual machine number to real machine number based on the degree of importance and automatically transmitting the error code information to a maintenance center (3). [Claim 2] Logging information detected by each guest virtual machine (guest VM) (2) under an operating system (OS) (20) in a virtual computer system.
(■), and the logging information (■) detected by the host system (1) is collected by a specific guest virtual machine (guest VM) (2a), and a means is provided for collective management. Virtual machine (guest VM) (
2a) analyzes the failure information for each virtual machine number based on the device management table (■) with the virtual machine number/real machine number correspondence information received from the host system (1) above in advance, and A fault information processing method under a virtual computer system characterized by converting and editing error code information from machine number to actual machine number, and outputting it as an analysis and editing list.