JP2008171217A - Computer system for estimating preventive maintenance of disk drive in disk sub-system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To collect error information from a disk drive composing a logic unit within a disk sub-system which cannot be recognized from a computer system and estimate a preventive maintenance replacement timing based on the collected error information. <P>SOLUTION: Error information collecting and analyzing function is added to a device driver for controlling a host bus adaptor mounted on the computer system to collect error information of the disk drive composing the logic unit LU within the disk sub-system, and the preventive maintenance replacement timing estimated based on the collected error information is provided to a system manager. The error information collecting and analyzing function is added to the device driver, and the preventive replacement timing estimated based on the collected error information is recorded as a log within a service processor SVP of a computer system. Accordingly, a computer system uninfluenced by various operating systems or versions of such operating systems can be provided. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

The present invention relates to a system for collecting error information of a disk drive constituting a logical unit in a disk subsystem connected to a host bus adapter controlled by a device driver,
In addition, the present invention relates to a computer system that estimates the preventive maintenance time of a corresponding disk drive using the collected error information.

The disk drive has a function of automatically assigning replacement blocks to defective data blocks detected at the time of data reading or writing. Further, error information such as the number of times of retry of read or write, the number of times of execution of replacement block allocation without successful retry, etc. is recorded. This automatic change is handled by the disk drive and is not informed to the computer system. In addition, a disk drive having a long operation time and a long period of time has a high possibility that the above-described read or write retry or automatic change will occur, and the access performance may be lowered.

The disk drive that can be physically recognized by the computer system has an error information collection command (in the case of a SCSI command, READ DEFECTDATA) for this disk drive.
, LOG SENSE, etc.) can be collected. Furthermore, preventive maintenance replacement of disk drives is possible by analyzing the collected information.

However, in a disk drive that constitutes a logical unit (LU) in the disk subsystem, if the disk controller in the disk subsystem cannot be recognized as an I / O device by the computer system, only the logical unit is detected from the computer system. Is I / O
It is recognized as a device and it is impossible to collect information about physical drives.

On the other hand, Patent Document 1 discloses a mechanism for providing information leading to a failure by collecting trace information in a computer system.

In Patent Document 2, a management computer equipped with a service processor board that performs fault monitoring and power control is connected to a service processor of a computer system that performs management via a network so that fault monitoring and power control of the computer system can be performed remotely. The mechanism to do is public.

Furthermore, Patent Document 3 discloses a mechanism for counting the frequency of use of consumables such as disk drives, liquid crystal displays, and batteries in the system, and transmitting the deterioration state of the consumables to the management host based on the count result. ing.

JP 2006-202076 A Japanese Patent Laid-Open No. 9-50386 JP 2001-325381 A

In recent years, computer systems have played an indispensable role in core business,
It is not allowed to stop for a long time due to a failure. Therefore, before a failure occurs, it is necessary to provide the system administrator with information on components that can cause the failure. Furthermore, a program for collecting faults needs to consider compatibility with an operating system running on a managed system, and also needs to consider the version of this operating system.

Many of the parts where failures occur are consumables such as disk drives that are mechanically operating in the system. If the maintenance replacement period is set centrally based on the operating time of these consumables, many disk drives In the disk subsystem to be installed, a situation occurs in which the disk drives in the disk subsystem are simultaneously maintained and replaced. Therefore, it is necessary to collect error information of individual disk drives mounted on the disk subsystem and determine the deterioration status.

However, in a disk drive that configures a logical unit (LU) in the disk subsystem, the disk controller in the disk subsystem is connected to the I / O from the computer system.
If it cannot be recognized as a device, only the logical unit is recognized as an I / O device from the computer system, and it is impossible to collect information about the physical disk drive.

  In order to solve the above problems, the following problems are listed.

The first problem is to collect error information from the disk drives constituting the logical units in the disk subsystem that cannot be recognized from the computer system, and to estimate the preventive maintenance replacement time based on the collected error information.

The second problem is to automatically provide the system administrator with error information of the disk drives that make up the logical units in the collected disk subsystem and the preventive maintenance replacement time estimated based on this error information. is there.

The third problem is that the system is independent of various operating systems or versions of the operating system on the computer system.

  The present invention aims to solve these problems.

The present invention provides a device driver that controls a host bus adapter mounted on a computer system.
By collecting error information and adding an analysis function, error information of the disk drives that make up the logical unit (LU) in the disk subsystem is collected, and preventive maintenance replacement estimated based on the collected error information Providing the time to the system administrator is a problem solving means.

Furthermore, instead of collecting and analyzing disk drive error information with dedicated application software, it collects error information in the device driver that controls the host bus adapter,
By adding an analysis function and recording the preventive replacement time estimated based on the collected error information as a hardware event log in the service processor (SVP) of the computer system, various operating systems or versions of this operating system are recorded. It is an object of the present invention to provide a computer system that is not affected by problems.

According to the present invention, it becomes possible to collect error information of a disk drive constituting a logical unit in a disk subsystem connected to a computer system on the computer system, and based on the error information of the corresponding disk drive. It is possible to provide the system administrator with the estimated preventive maintenance time.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 shows a disk controller 10
3 and the disk drives 107, 108, and 109 are disk subsystems 102 that can be recognized as I / O devices, and FIG. 2 is a disk subsystem 202 that can recognize the disk controller 203 as a network device. From the disk subsystem 102 of FIG. 1, the disk drive 107 mounted on the disk subsystem 102,
Means for collecting the error information 112, 113, 114 of 108, 109 is shown in FIG.
A means for collecting the error information 212, 213, 214 of the disk drives 212, 213, 214 will be described in the second embodiment.

FIG. 1 is a diagram in which a host bus adapter 101 mounted on a computer system 100 and a disk subsystem 102 are connected via a SAN or IP-SAN. The computer system 100 includes a disk controller 103 and a disk in the disk subsystem 102. The drives 112, 113, and 114 can be recognized as I / O devices. New error information collection and analysis unit 1 in the device driver 104 that controls the host bus adapter 101
05, and this error information collection and analysis unit 105 makes the disk drive 1 constituting the logical units (LU0, LU1) 111, 110 under the disk controller 103.
The error information 112, 113, 114 of 07, 108, 109 is collected.

Error information 112, 113 of the disk drives 107, 108, 109 shown in FIG.
The operation of collecting error information in the embodiment of the present invention will be described based on the flowchart of the process of collecting 114.

In FIG. 3, first, error information collection newly provided in the device driver 104 of FIG.
The analysis unit 105 inputs an error information collection command to the disk drive (i) under the disk controller 103 (S301). Where i is a variable and the initial value is 0 (
S300), i = 0, 1, 2,... The error information collection command corresponds to READ DEFECTDATA, LOG SENSE, etc., taking a SCSI command as an example.

Next, the device driver 104 receives a logical unit (L
U0, LU1) (111, 110) is received error information about the disk drive (i) (S302).

The error information collection and analysis unit 105 performs the above processes S301 and S302 as a logical unit (LU
0, LU1) 111, 110 is determined for the number of disk drives that have been performed A times (S303). If it is determined that the disk drives have been performed A times, the error information collection and analysis unit 105 receives the received disk drive. (I) 107, 108, 109 error information 112, 113, 11
4 is used to estimate the preventive maintenance time (S305). Here, A is a constant, and is a known value determined when the disk subsystem 102 is constructed. In the configuration of FIG. 1, A = 3.

Thereafter, the error information collection / analysis unit 105 starts a process of displaying the preventive maintenance time estimated based on the error information 112, 113, 114 collected from the disk drives 107, 108, 109 as a hardware event log (S306). .

FIG. 2 shows the connection between the host bus adapter 201 mounted on the computer system 200 and the disk subsystem 202 that cannot recognize the disk drives 207, 208, and 209 as I / O devices via SAN or IP-SAN. FIG. From the computer system 200, only logical unit units (LU0, LU1) 211, 210 are I / O.
Recognized as O device, disk drive 207 via SAN or IP-SAN
It is impossible to collect error information 212, 213, and 214 of 208 and 209. Therefore, the disk controller 203 is controlled from the management NIC port 217 of the disk controller 203, and error information 212, 213, 214 of the disk drives 207, 208, 209 constituting the logical units (LU0, LU1) 211, 210 are controlled. Collect.

Error information 212, 213 of the disk drives 207, 208, 209 shown in FIG.
The operation of collecting error information in the embodiment of the present invention will be described based on the flowchart of the process of collecting 214.

In FIG. 4, first, the error information collection and analysis unit 205 newly provided in the device driver 204 controls the disk controller 203 connected to the NIC port 216 of the host bus adapter 201 via the network. Error information of the drive (i) is requested (S401). Here, i is a variable and the initial value is 0 (S300), and changes to i = 0, 1, 2,.

Next, the disk controller 203 includes logical units (LU0, LU1) 211, 210.
Is collected, and the device driver 204 receives the error information of the disk drive (i) from the disk controller 203 (S402).

The error information collection / analysis unit 205 performs the above processing S401 and S402 as a logical unit (LU
(0, LU1) 211, 210 is determined for the number of disk drives that have been performed A times (S403). If it is determined that the A times have been performed, error information collection and analysis unit 205 receives the received disk drive. (I) Error information 212, 213, 21 of 207, 208, 209
4 is used to estimate the preventive maintenance time (S405). Here, A is a constant, and is a known value determined when the disk subsystem 202 is constructed. In the configuration of FIG. 2, A = 3.

Thereafter, the error information collection / analysis unit 205 starts a process of displaying the preventive maintenance time estimated based on the error information 212, 213, 214 collected from the disk drives 207, 208, 209 as a hardware event log (S406). .

The host bus adapter 201 in FIG. 2 is a LAN adapter having only NIC 216 mounted.
A host bus adapter in which the FC or iSCSI controller 219 is mounted on the same board may be used. In FIG. 2, the host bus adapter 201 has the NIC 216 and the FC or iSCSI controller 219 having different functions mounted on the same board, but the NIC 216 and the FC or iSCSI controller 219 on the host bus adapter 201 may be combined into one. I do not care.
(Preventive maintenance time estimation method)
The error information collection and analysis units 105 and 205 of the device drivers 104 and 204 estimate the preventive maintenance time based on the error information (i) collected from the disk drive (i) as follows. Here, i is a variable, and i = 0, 1, 2 in FIGS.

For example, the error information of the disk drive (i) constituting the logical unit is collected at a time H hour M determined once a day, and the collected value is as follows. Here, H is a constant from 0 to 23, M is a constant from 0 to 59, H and M are both set by the system administrator,
It can be changed.

September n, 2006 Automatic change frequency: _An (i)
Number of read errors: R _n (i)
Number of write errors: W _n (i) (a)
September n + 1, 2006 Automatic change count: A _{n + 1} (i)
Number of read errors: R _{n + 1} (i)
Number of write errors: W _{n + 1} (i) (b)
From (a) and (b),
An + 1 (i) −An (i) = A (i)
Rn + 1 (i) −Rn (i) = R (i)
Wn + 1 (i) −Wn (i) = W (i) (c)
Define (c).

The error information collection / analysis unit 105, 205 of the device driver performs A (i), R in (c).
When (i) and W (i) exceed preset A, R, and W, the hardware log 115 or 215 is generated. Here, A, R, and W are constants.

With this hardware log, it is possible to provide the system administrator with a time for preventive replacement of the disk drive, and to prompt preventive maintenance replacement of the disk drive.
(Providing means for preventive maintenance time)
A method of recording and displaying error information collected from the disk drive and preventive maintenance replacement time of the disk drive estimated based on the error information will be described.

FIG. 5 shows the error information of the disk drive collected by the device driver error information collection / analysis unit 505 and the preventive maintenance replacement time of the disk drive estimated based on this error information as a hardware event log to the system administrator. Means. Based on the flowchart of the process for recording and displaying the hardware event log shown in FIG. 6, the process for recording and displaying the hardware event log in the embodiment of the present invention will be described.

In FIG. 5, first, the error information collection / analysis unit 505 of the device driver 504 has hardware information indicating the error information 512, 513, 514 collected from the disk drives 507, 508, 509 and the preventive replacement time estimated based on this error information. The log 515 is stored in the management area of the device driver 504 (S600). In the example of FIG. 5, when the management area of the device driver 504 is physically shown, it becomes a disk drive 524 mounted on the computer system 500. This disk drive 524 is a recording medium connected outside the computer system 500. It doesn't matter. Furthermore, since the log stored in the management area of the device driver 504 can be referred to by any user operating the operating system, the contents to be stored are limited or the stored process itself is suppressed. it can.

Next, error information collection / analysis unit 505 of device driver 504 is connected to host bus adapter 5.
The physical disk number i, automatic alternation count A (i), read error count R (i), and write error count W (i) are set in the service processor report register register 521 in 01 (S
601) and the check bit 520 is set (S602). Here, the service processor reporting register register 521 and the check bit 520 of the host bus adapter 501 can be scanned in and out by the service processor 522 of the computer system 500 and have no influence on the operation of the host bus adapter 501. The register must not be given. The check pit 520 is a register that can report an event such as a failure occurring in the host adapter 501 to the service processor 522.
It must be a register that can periodically monitor the value set by the service processor. In FIG. 5, the service processor report register 521 of the host bus adapter 501
And check bit 520 for FC or iSCSI controller 519 or NIC 516
However, the registers and check bits in the FC or iSCSI controller 519 or NIC 516 may be used.

The service processor 522 periodically monitors the presence or absence of all check bits in the computer system 500 including the host bus adapter 501 (S603), and found that the check bit 520 of the host bus adapter 501 is set. The service processor 522 reads the physical disk number i, automatic alternation count A (i), read error count R (i), and write error count W (i) stored in the report register register 521 of the host adapter ( In step S604, the check bit 520 and the report register 521 of the corresponding adapter are reset (S605).

The service processor 522 generates a hardware event log based on the physical disk number i read from the host bus adapter 501, the automatic alternation count A (i), the read error count R (i), and the write error count W (i) ( In step S606, the hardware event log is stored in the log area in the service processor, and the hardware log information is displayed on the console 523 connected to the service processor 522 (S607).

By storing the hardware event log in the log area in the service processor 522 and displaying it on the console 523 connected to the service processor 522, it is possible to provide the system administrator with a preventive replacement time for the disk drive. It becomes possible to promote preventive maintenance replacement of the disk drive.

1 is a first embodiment in which a computer system to which the present invention is applied and a disk subsystem are connected (Example 1). This is a second embodiment in which a computer system to which the present invention is applied and a disk subsystem are connected (Example 2). 10 is a flowchart of processing for collecting disk drive error information (Example 1); 10 is a flowchart of processing for collecting disk drive error information (second embodiment); It is a figure which shows the means to provide the error information and preventive maintenance replacement time of a disk drive to a system administrator as a hardware event log. It is a flowchart of the process which records and displays a hardware event log.

Explanation of symbols

100, 200, 500: Computer system, 101, 201, 501: Host bus adapter, 102, 202, 502 ... Disk subsystem, 103, 203, 503 ... Disk controller, 104, 204, 504 ... Device driver, 105, 205 , 50
5 ... Error information collection / analysis unit, 106, 206, 506 ... FC or iSCSI controller control unit, 107, 207, 507 ... Disk drive 0, 108, 208, 508 ...
Disk drive 1, 109, 209, 509... Disk drive 2, 110, 210,
510 ... LU1, 111, 211, 511 ... LU0, 112, 113, 114, 212,
213, 214, 512, 513, 514 ... error information, 115, 215, 515 ... hardware log, 216, 217, 516, 517 ... NIC, 218, 518 ... NIC controller, 219, 519 ... FC or iSCSI controller, 520: Check bit, 5
21: Register register for service processor report, 522: Service processor (SV
P), 523, console, 524, disk drive.

Claims (4)

A means for issuing a command for collecting error information to a disk drive constituting a logical unit in a disk subsystem connected to the host bus adapter, and collecting error information from the disk drive periodically;
Request the disk controller in the disk subsystem connected to the host bus adapter for error information of the disk drive that configures the logical unit in the disk subsystem, and periodically collect disk drive error information from the disk controller. Means to
A computer system comprising a device driver for estimating a preventive maintenance replacement time of a disk drive based on collected error information. Means for comparing the disk drive error information periodically collected from the disk drives that make up the logical units in the disk subsystem connected to the host bus adapter, and estimating the preventive replacement time of the disk drives;
2. The computer system according to claim 1, further comprising storage means for storing the error information collected from the disk drive and the information on the estimated preventive replacement time in the management area of the device driver. 2. The computer according to claim 1, wherein the error information of the disk drive periodically collected from the disk drive constituting the logical unit in the disk subsystem connected to the host bus adapter and the estimated preventive replacement time are displayed on the monitor screen. system. A means for issuing a command for collecting error information to a disk drive constituting a logical unit in a disk subsystem connected to the host bus adapter, and collecting error information from the disk drive periodically;
Request the disk controller in the disk subsystem connected to the host bus adapter for error information of the disk drive that configures the logical unit in the disk subsystem, and periodically collect disk drive error information from the disk controller. Means to
A computer system comprising storage means for storing collected error information and estimated preventive replacement time information in a service processor of the computer system.

Images