JP2008217202A - Disk array device and firmware update method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a disk array device and an update method for firmware, capable of updating firmware of a hard disk without dropping a processing speed. <P>SOLUTION: This disk array device has: a RAID controller constructing a RAID system, and controlling operation of a plurality of hard disks; and a management agent updating the firmware to an updated hard disk among the plurality of hard disks. The management agent separates the updated hard disk from the RAID system to update the firmware, stores a logic address on the RAID system performed with writing when the writing into the RAID system is performed while the updated hard disk is separated, restores data of a portion corresponding to the stored logic address out of the updated hard disk by the RAID controller after the update, and restores the updated hard disk to the RAID system. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a disk array device, and more particularly to a disk array device in which a RAID system is constructed and a method for updating the firmware.

  There is known a disk array device in which data is distributed and stored in a plurality of hard disks. An example of such a disk array device is one that manages a plurality of hard disks by constructing a RAID system with a RAID controller.

  Each of the plurality of hard disks normally stores firmware for controlling various operations of the hard disk. This firmware may be updated for the purpose of homogenizing performance among a plurality of hard disks or improving performance.

  Such firmware update may be performed via a network. In this case, it is performed by a firmware update program (hereinafter referred to as a management agent) installed in the disk array device. The operation in this case will be described. A management agent accesses a computer (hereinafter referred to as a management PC) that provides the latest version of firmware via a network. The management agent acquires the latest firmware version information from the management PC and compares it with the firmware version of the hard disk in the disk array device. If the firmware version of the hard disk is old, the latest firmware is downloaded from the management PC. After downloading, the firmware of the hard disk in the disk array device is updated. During the update, the writing and reading processes for all hard disks are interrupted. In addition, communication via the network is interrupted, and the firmware is updated locally. During the firmware update, data cannot be written to or read from the disk array device, which is inconvenient for the user.

  Patent Document 1 describes a method for updating firmware and drivers in a disk array device via a network.

  Japanese Patent Application Laid-Open No. 2004-228561 discloses a technique that enables firmware to be updated without interrupting the processing of the disk array device. In the update method described in Patent Document 2, when updating the firmware, first, data recorded on the hard disk A to be updated is copied to another hard disk B. Next, when copying is completed, the I / O processing for the hard disk A is stopped, and the hard disk B is substituted for the I / O processing for the hard disk A. Further, storage management of the data change history caused by the I / O processing is started during the proxy mode. Then, the firmware of the hard disk A is updated. When the update of the hard disk A is completed, the contents of the hard disk A are set up in the latest update state based on the data change history and the data recorded in the hard disk A or B. Further, the I / O process for the hard disk A that has been substituted for the hard disk B is returned to the hard disk A directly, and the process ends.

JP 2002-278906 A JP 2002-318666 A

  If the technique of Patent Document 2 is used, the firmware can be updated without interrupting the processing of the disk array device. However, (1) copy the data of the hard disk to be updated to another hard disk, (2) compare the contents of the hard disk that was delegated in setting up the data of the hard disk to be updated to the latest update state, etc. It is expected that the processing speed will be slow during the firmware update.

  Accordingly, it is an object of the present invention to provide a disk array device and a firmware update method that can update hard disk firmware without reducing the processing speed.

  [Means for Solving the Problems] will be described below using the numbers and symbols used in [Best Mode for Carrying Out the Invention]. These numbers and symbols are added in parentheses in order to clarify the correspondence between the description of [Claims] and [Best Mode for Carrying Out the Invention]. However, these numbers and symbols should not be used for the interpretation of the technical scope of the invention described in [Claims].

  A disk array device (100) according to the present invention accesses a RAID controller (14) that logically controls a plurality of hard disks (31 to 35) by constructing a RAID system having redundancy, and accesses the RAID controller (14). A management agent (61) which is connected in a possible manner and updates the firmware of the updated hard disk (31) among the plurality of hard disks (31 to 35). When updating the firmware of the updated hard disk (31), the management agent (61) logically separates the updated hard disk (31) from the RAID system and updates the firmware. The management agent (61) stores a logical address to be written on the RAID system when writing to the RAID system is performed while the updated hard disk (31) is disconnected. After the update is completed, the management agent (61) accesses the RAID controller (14), restores the data of the sector corresponding to the stored logical address in the updated hard disk (31), and updates the updated hard disk (31). Is restored to the RAID system.

  According to the above-described configuration, only the hard disk (31) to be updated is disconnected from the RAID system, so that writing and reading processing to the RAID system can be performed even during the update. If there is a write to the RAID system during the update, the logical address to be written is stored. After the update is completed, only the sector data corresponding to the stored logical address among the sectors of the updated hard disk (31) is restored. At this time, since it is not necessary to restore all the data recorded on the hard disk to be updated, the processing required for restoring the data can be minimized. Therefore, it is possible to update the hard disk firmware without reducing the processing speed.

  In the disk array device (100), when the RAID controller (14) writes to the RAID system while the update hard disk (31) is disconnected, the recovery flag is added to the sector information of the sector to be written. Is preferably written. At this time, when restoring the data of the updated hard disk (31), the data of the sector to which the recovery flag is not written is generated based on the data of the sector to which the recovery flag is written. Restore data.

  In the disk array device (100) described above, the RAID controller (14) preferably controls the operation of the plurality of hard disks (31 to 35) by constructing a RAID system having parity.

  The firmware update method according to the present invention is a firmware update method for updating hard disk firmware of a RAID system. This firmware update method includes a step of disconnecting an updated hard disk to be updated from the RAID system (step S40), a firmware update step of updating the firmware of the disconnected updated hard disk (step S50), and writing during the update. The step of storing the logical address on the RAID system (step S52), the step of rebooting the hard disk to be updated after the firmware update step (S50) (step S70), A restoration step (step S90) for restoring the portion of data corresponding to the stored logical address, a step of returning the updated hard disk with the restored data to the RAID system (step S100), Comprising.

  The firmware update method preferably further includes a step (step S51) of writing a recovery flag in the sector information of the sector that was written during the update. At this time, in the restoration step (step S90), the data of the updated hard disk is restored by restoring the data of the sector where the recovery flag is not written based on the data of the sector where the recovery flag is written.

  According to the present invention, there is provided a disk array apparatus and a firmware updating method capable of updating the hard disk firmware without reducing the processing speed.

  Embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a configuration of a network system 100 including a disk array device 11 according to the present embodiment. The network system 100 has one management PC 1 and a plurality of disk array devices (11 to 14). Each disk array device (11 to 14) is connected to the management PC 1 via the network 2. The management PC 1 is installed with manager software 41 for communicating with each disk array device (11 to 14) regarding management of the disk array device.

  FIG. 2 is a schematic block diagram functionally depicting the configuration of each disk array device 11. Each disk array device 10 includes an interface 12 interposed between the network 2, an input / output device 13 for enabling data input / output with a user, a RAID controller 14, a memory 15, a CPU 16, and a storage. These units 17 are connected to each other via a general-purpose bus 11. The storage unit 17 stores an operating system 17 and a management agent 61 as a firmware update program. A plurality of hard disks (31 to 35) are connected under the RAID controller 14.

  The operating system 17 and the management agent 61 are expanded in the memory when the disk array device 11 is started up and executed by the CPU to realize its functions. The management agent 61 implements a function related to hard disk firmware update, and operates on the operating system 51. Specifically, it periodically communicates with the manager 41 of the management PC 1 via the network 2, downloads the latest firmware from the manager 41 as necessary, and updates the firmware of the hard disk (31 to 35). This function is realized.

  The RAID controller 14 constructs a RAID system and manages a plurality of hard disks (31 to 35). At this time, the RAID controller 14 logically constructs a RAID system having redundancy, and controls writing and reading operations to the plurality of hard disks 30. The RAID system referred to in this specification does not include RAID 0 that does not have redundancy.

  FIG. 3 is a diagram for explaining the RAID system. As shown in FIG. 3, the RAID controller 14 divides each of the plurality of hard disks (31 to 35) into a plurality of areas (hereinafter referred to as sectors), and manages them by assigning logical addresses. Here, the logical address is composed of a plurality of sectors included in separate hard disks, and is composed of three sectors in the example shown in FIG.

  For example, when data is written by the user via the input / output device 13, the operating system 51 notifies the RAID controller 14 of data in units of logical addresses. The RAID controller 14 divides the notified data into two according to the access from the operating system 51, and stores the divided data in two of the three sectors in the logical address. Then, parity data is generated from the data divided into two, and this parity data is stored in the remaining one sector. In the example shown in FIG. 3, when writing to the logical address 201, the data divided into two is stored in the sector 201-1 of the hard disk 31 and the sector 201-2 of the hard disk 32 one by one. Parity data is stored in 33 sectors 201-P. Similarly, for the logical address 202, the data divided into 202-1 and 202-2 are stored one by one, and the parity data is stored in 202-P. Further, the RAID controller 14 writes sector information in the header portion of each sector when storing data as necessary.

  As described above, by storing parity data in one sector in a logical address, the data content is protected even if one hard disk is disconnected from the RAID system.

  Next, an operation when updating the firmware of the hard disk 31 in the disk array device 10 described above will be described. FIG. 4 is a flowchart showing the firmware update method according to the present embodiment.

  It is assumed that the manager 41 operating on the management PC 1 issues a firmware update instruction to the management agent 61 operating on the disk array device 10 via the network 2. Upon receiving the instruction, the management agent 61 acquires the target name and firmware version of the hard disk to be updated from the manager 41. Then, it is confirmed whether or not a hard disk with the target name exists under the RAID controller 14 and the firmware version of each hard disk is confirmed (step S10).

  When the hard disk 31 with the target name exists and the firmware version is old, the management agent 61 downloads the latest firmware from the manager 41 (step S20). After that, consistency check (consistency check) of the hard disks 31 to 35 under the RAID controller 14 is performed (step S30). After the consistency check, the management agent 61 accesses the RAID controller 14, logically disconnects the update target hard disk 31 from the RAID system, and puts the RAID system into a degenerated state (step S40). After entering the degenerated state, the management agent 61 updates the firmware of the separated hard disk 31 to be updated to the downloaded firmware (step S50).

  Here, it is assumed that the data writing process is performed from the operating system 51 to the RAID controller 14 while the update target hard disk 31 is disconnected from the RAID system. In this case, the RAID controller 14 writes data to each hard disk (32 to 35) other than the hard disk 31 while the RAID system is in a degenerated state. At this time, the RAID controller 14 writes data to each sector of each hard disk (32 to 35) to be written while writing a recovery flag for later recovery as sector information (step S51). Further, the management agent 61 manages the history of writing, and stores the logical address on the RAID system where writing is performed (or performed) in the memory 15 (step S52).

  After completing the firmware update, the management agent 61 confirms whether the firmware update has been completed normally (step S60). In the case of normal termination, the management agent 61 restarts the update target hard disk (step S70). Further, the management agent 61 confirms whether or not the new firmware is normally activated (step S80).

  When starting normally in S80, the management agent 61 accesses the RAID controller 14 and restores the data stored in the hard disk 31 to be updated. At this time, based on the logical address stored in the memory 15, only the sector corresponding to the logical address is restored. Specifically, the management agent 61 transmits a recovery command designating the stored logical address to the RAID controller 14. When the RAID controller 14 receives the recovery command, the RAID controller 14 reads the data of the sector corresponding to the designated logical address among the plurality of sectors of each hard disk. At this time, the sector of the updated hard disk 31 is also read. Then, another data without the recovery flag is generated from the two data with the recovery flag among the data of the three sectors in the logical address, and written to the sector without the recovery flag. At the same time, the recovery flag is cleared from the sector information of the two sectors having the recovery flag. If there is a recovery flag in all three sectors, only the three recovery flags are cleared to indicate that there is no hard disk sector to be updated at the logical address. It does not change. With such an operation, data is restored only in the sector of the logical address that was written during the detachment among the sectors of the hard disk to be updated (step S90).

  Next, the management agent 61 causes the RAID controller 14 to restore the update-target hard disk 31 that has been disconnected to the RAID system (step S100). Further, the consistency check is performed again (step S110), and the process is terminated. If there are a plurality of update targets on the hard disk in the disk array device 11, the processes in steps S10 to S110 are repeated.

  The above operation will be described more specifically. In FIG. 3, it is assumed that the firmware of the hard disk 31 is updated and writing processing is performed on the logical addresses 301 to 306 while the hard disk 31 is disconnected.

  In this case, in step S40, the hard disk 31 is disconnected from the RAID system. In step S51, data is written to the hard disks (32 to 35) other than the hard disk 31. At this time, as shown in FIG. 3, a recovery flag (“R” in the figure) is written in the sector information of each sector to be written. Further, in step S52, the management agent 61 stores the logical addresses 301 to 306.

  In step S90, when the data is restored, the management agent 61 transmits a recovery command to the stored logical addresses 301 to 306 to the RAID controller 14. The RAID controller 14 reads the data of the sectors 301-1, 301-2, and 301-P from the hard disks 31 to 33 corresponding to the logical address 301. At this time, sector information is also read. Then, data of the sector 301-1 having no recovery flag is generated from the sectors 301-2 and 302-P having the recovery flag “R”, and the data is written to the sector 301-1. After writing, the recovery flag “R” is cleared from the sector information of sectors 301-2 and 301-P. Similarly, the data in the sectors 302-P, 304-2, and 306-1 in the hard disk 31 are updated. For logical addresses 303 and 305, since the recovery flag “R” exists in all sectors, only the recovery flag “R” is cleared, and the data stored in the sector is not changed.

  By operating as described above, of the data stored in the hard disk 31, the sector (301-1, 302-P, 304-2, 306-1) corresponding to the logical address written during the detachment is stored. Only the data is restored.

  In step S60, if the firmware update does not end normally, the management agent 61 restarts the updated hard disk (step S120), and confirms whether the old version is normally started (step S120). Step S130). If the old version is started normally, the data contents of the updated hard disk are written back to only the logical address that was written during detachment, similar to S90 (step S140). Further, the management agent 61 returns the update target hard disk to the RAID system (step S150), and after confirming the consistency (step S160), notifies the manager 41 that the firmware update has failed (step S170).

  In S80, if the new version of firmware cannot be started normally by reboot after updating the firmware of the hard disk to be updated, or if the old version of firmware cannot be started normally in step S130, The management agent 61 holds the degraded state in which the hard disk to be updated is disconnected, and notifies the manager 41 that the firmware update has failed (step S200).

  By operating as described above, the RAID system is constructed in a degenerate state even while the hard disk to be updated is disconnected from the RAID system, so that write and read processing to the RAID system can be performed even during firmware update. It can be carried out.

  Here, when restoring the data written during detachment, since the logical address to be written is stored, only the data of the sector corresponding to the logical address to be written in the disconnected hard disk is stored. Can be restored. Since it is not necessary to perform restoration processing for sectors other than the logical address to be written, processing required for data restoration can be minimized. That is, as in the conventional example, when the hard disk to be updated is disconnected, the data recorded on the hard disk to be updated is copied and transferred to another device (such as a hard disk), or the data on the hard disk to be updated is restored. There is no need to compare the contents with the data recorded on the hard disk that has been substituted, and the processing speed can be increased.

  Similarly, the disk array devices 12 to 14 connected to the management PC 1 via the network 2 can also update the firmware without stopping the writing and reading processing to the RAID system.

  Further, since there is no need to interrupt the writing / reading processing to / from the disk array device, the firmware can be updated while being connected to another computer via the network 2. Therefore, it is not necessary to update the firmware locally for each disk array device, and the update can be performed via the network.

  In this embodiment, the RAID system constructed by the RAID controller has been described with respect to the type in which the logical address is composed of three sectors of different hard disks, and parity data is stored in one of these sectors. Other types of RAID systems may be used if they are not RAID 0 that does not have sex.

1 is a schematic configuration diagram of a network system. It is a schematic block diagram which shows the structure of a disk array apparatus. It is a conceptual diagram for demonstrating a RAID system. It is a flowchart which shows the firmware update method.

Explanation of symbols

1 management PC
2 Network 10 Disk array device 11 General-purpose bus 12 Interface 13 Input / output device 14 RAID controller 15 Memory 16 CPU
17 Storage Unit 41 Manager 31 Hard Disk 32 Hard Disk 33 Hard Disk 34 Hard Disk 35 Hard Disk 51 Operating System 61 Management Agent 100 Network System

Claims (6)

A RAID controller that logically controls a plurality of hard disks by constructing a RAID system having redundancy;
A management agent connected to the RAID controller so as to be accessible, and performing a firmware update on an updated hard disk of the plurality of hard disks;
Comprising
When updating the firmware of the updated hard disk, the management agent logically disconnects the updated hard disk from the RAID system and updates the firmware.
When writing to the RAID system is performed while the hard disk to be updated is disconnected, the management agent stores a logical address to be written on the RAID system, and accesses the RAID controller after the update is completed. A disk array device that restores data of a sector corresponding to the stored logical address among the hard disks to be updated and returns the hard disk to be updated to the RAID system. The disk array device according to claim 1,
The RAID controller is
When writing to the RAID system is performed while the updated hard disk is disconnected, a recovery flag is written in the sector information of the sector to be written,
When restoring the data on the hard disk to be updated, the data on the hard disk to be updated is restored by generating the data on the sector on which the recovery flag is not written based on the data on the sector on which the recovery flag is written. Disk array device. The disk array device according to claim 1 or 2,
The RAID controller is a disk array device that controls a plurality of hard disks by constructing a RAID system having parity. A firmware update method for updating firmware of a hard disk in a RAID system,
Disconnecting the hard disk to be updated, from which the firmware is updated, from the RAID system;
A firmware update step for updating the firmware of the updated hard disk that has been disconnected; and
Storing a logical address to be written when writing to the RAID system during detachment of the updated hard disk;
Restarting the updated hard disk after completion of the firmware update step;
A restoring step of restoring data of a sector corresponding to the stored logical address in the updated hard disk based on the stored logical address;
Returning the data-restored hard disk to be restored to the RAID system;
A firmware update method comprising: The firmware update method according to claim 4,
Furthermore,
Writing a recovery flag to sector information of a sector written while the hard disk to be updated is disconnected;
Comprising
A firmware updating method for restoring data of the hard disk to be updated by restoring data of a sector in which no recovery flag is written based on the data of the sector in which the recovery flag is written in the restoration step.   A firmware update program for realizing the firmware update method according to claim 4 or 5 by a computer.

Images