RAID apparatus, module therefor, disk incorporation appropriateness judgment method and program

Abstract

A disk incorporation process unit 54 shares information (i.e., a common table) managed by a disk statistics unit 53 and judges whether or not to permit an incorporation of an installed disk by referring to the common table in the event of a discretionary disk having been isolated followed by the aforementioned disk being installed.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

FIGS. 1A and 1B each is a diagram exemplifying a conventional hot swap;

FIG. 2 is a diagram of a common configuration of a RAID apparatus;

FIG. 3 is diagram of a hardware configuration of a centralized module (CM) shown in FIG. 2;

FIG. 4 is a functional block diagram of the CM shown in FIG. 2;

FIG. 5 is a diagram exemplifying a structure of a common table;

FIG. 6 is a flow chart of a disk incorporation process unit according to a first embodiment;

FIG. 7A is a diagram exemplifying an FC system error; FIG. 7B is a diagram showing a specific example of a “disk isolation factor”; and

FIG. 8 is a process flow chart of a disk incorporation process unit according to a second embodiment.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

The following is a description of the preferred embodiment of the present invention by referring to the accompanying drawings.

FIG. 2 is a diagram of a common configuration of a RAID apparatus.

The shown RAID apparatus 1 comprises two Centralized Modules (CM) 10 (i.e., 10a and 10b), a FRT 3, Backend Routers (BRT) 4 and 5, and Drive Enclosures (DE) 6 and 7.

The CM 10 manages/controls various access and error recovery processes within the RAID apparatus 1. The BRTs 4 and 5 are positioned between the CMs 10 and DEs 6 and 7, and perform the roles of switches connecting the CM 10 and each DE (i.e., disk groups). There are two paths for a host 2 to access a discretionary DE by way of the CM 10. These two access paths are respectively equipped with the BRTs 4 and 5. Therefore, even if either of the access paths becomes unusable by a certain cause (e.g., a BRT failure, et cetera), an access is enabled by using the other access path.

Here, the CM 10a is connected to both of the systems of the BRT 4 and BRT 5, and likewise is the CM 10b connected. Note that a later described reincorporation appropriateness judgment process, et cetera, is carried out by the CM 10a and CM 10b individually. The FRT 3 is disposed for relay-controlling communications between the CMs 10a and 10b.

The DE 6 includes port bypass circuits (PBCs) 6a and 6b, and a disk group 6c. The DE 7 likewise includes PBCs 7a and 7b, and a disk group 7c. The PBC is hardware having the function of making a certain abnormal disk bypassed from a loop (i.e., the function of isolating the disk) in order to prevent the disk from becoming a “dam” in the loop when an abnormality occurrs to the disk in an FC transmission path formed by the loop. The PBC notifies the CM 10 of the isolated disk.

Each port of the BRT 4 is connected to the PBC 6a and PBC 7a, while each port of the BRT 5 is connected to the PBC 6b and PBC 7b, and each of the CM 10s accesses the disk groups 6c and 7c by way of the BRT 4 or BRT 5 and PBC.

Each of the CMs 10 is connected to the hosts 2 (i.e., 2a and 2b) by way of a random telecommunication line.

Each of the CMs 10 is also connected with an FST 20 on an as required basis (e.g., for maintenance and repair works). The FST 20 is a specific maintenance-use personal computer (PC) An operator (e.g., a maintenance technician, et cetera) operates the FST 20 on an as required basis to instruct the CM 10 for isolating a discretionary disk.

FIG. 3 shows a diagram of a hardware configuration of the above noted CM 10.

The CM 10 shown in FIG. 3 comprises individual DIs 31, individual direct memory accesses (DMAs) 32, two Central Processing units (CPUs) 33 and 34, a Memory Controller Hub (MCH) 35, memory 36 and individual channel adaptors (CAs) 37.

The DIs 31 are FC controllers connecting to respective BRTs. The DMA 32 is a telecommunication line connecting to the FRT 3. The MCH 35 is a circuit connecting the so-called host side bus, such as external buses of the CPUs 33 and 34, to a peripheral interconnect (PCI) bus for enabling intercommunications. The CAs 37 are adaptors for connecting to the respective hosts.

Later described processes of various flow charts shown in FIGS. 6 and 8, and functions of various function units shown in FIG. 4, are accomplished by the CPU 33 or CPU 34 reading an application program stored in memory 36 and executing it. A later described common table 60, et cetera, are also stored in the memory 36.

FIG. 4 is a functional block diagram of the CM 10.

The CM 10 comprises a monitor unit 51, a configuration management unit 52, a disk statistics unit 53 and a disk incorporation unit 54. Among these, the functions of the monitor unit 51, configuration management unit 52 and disk statistics unit 53 may be approximately the same as a conventional technique (whereas the difference lies in the aspect of reflecting data respectively detected and managed by the units to the common table 60). The characteristic of the CM according to the present embodiment lies in the disk incorporation unit 54. Although a functional unit for judging an appropriateness of disk incorporation has conventionally existed, there has been the above noted problem because its judgment only used to utilize a Disk WWN as described above.

Having received a notification from the PBC as described above in the event of isolating according to a judgment of the PBC, the monitor unit 51 sets it to a PBC cause 63 of the later described common table 60. The configuration management unit 52 judges whether or not the respective disks are in recovery (i.e., in a rebuild/copyback state) and sets the judgment result to a later described recovery in-progress 64 of the common table 60.

Meanwhile, information of an error occurring in each disk is integrated in the disk statistics unit 53. That is, the disk statistics unit 53 is a module disposed for performing the processes of counting up a point corresponding to an error phenomenon for every occurrence of the error for each disk equipped in the RAID apparatus 1, and isolating a disk of which the count-up value exceeds a threshold value.

And in the case of isolating a disk, the disk statistics unit 53 according to the present embodiment sets a cause for isolating (noted as “isolation cause” hereinafter) the disk to an isolation cause 61 of the common table 60. There are two kinds of isolation causes, i.e., a device system error and an FC system error. The difference between the device system error and FC system error is that the former is a hardware-wise abnormality and the latter is an error in a view point of the FC loop. The disk statistics unit 53 further sets a disk isolation factor, as detailed information of the isolation cause, to a factor 65 of the common table 60. The disk isolation factor includes for example an isolation based on a disk statistics, an isolation due to a forced degeneration, an isolation due to “disk not ready”, et cetera.

The disk incorporation unit 54, sharing information managed by the disk statistics unit 53, judges whether or not to permit a reincorporation of the isolated disk by referring to the common table 60. Note that the disk incorporation unit 54 first judges by a Disk WWN as in the conventional technique. Therefore, if a Disk WWN of a disk installed after isolating a discretionary disk is different from the registered Disk WWNs (i.e., in the case of installing a new disk such as a maintenance-use disk, et cetera), it of course permits the incorporation. Contrarily, if a Disk WWN of a disk installed after isolating a discretionary disk is the same as a registered Disk WWN (i.e., in the case of carrying out a hot swap by using the above noted same disk), an incorporation has conventionally not been permitted without an exception, whereas the present method may sometimes permit an incorporation by making a judgment as shown in the following.

<A Judgment Method for Whether or not to Permit a Reincorporation of an Isolated Disk Once from an Apparatus>

(1) Basically done is to permit a reincorporation and carry out an incorporation process only if all of the following conditions 1 through 4 are satisfied. However, all these conditions are not necessarily satisfied, although a possibility of a problem occurrence as a result of reincorporating the isolated disk is considered to be extremely low if all the conditions are satisfied:

Condition 1: the isolation cause is not a device system error (i.e., not a hardware-wise failure of a disk)

Condition 2: in the case of an FC system error (i.e., a disk transmission path error), a judgment is made as to whether or not to permit an incorporation according to a category of the FC error. That is, a reincorporation is not permitted if either of the following conditions is not satisfied:

• • not recovery in progress (i.e., in a rebuild/copyback state) (that is, a recovery failure disk due to an FC system error (i.e., a disk of rebuild/copyback in progress) is not incorporated in order to prevent a delay of the rebuild/copyback process)
  • not an isolation according to a statistical count-up of points of the apparent disk cause

Condition 3: not an isolation according to a PBC judgment (i.e., a reincorporation of a disk autonomously isolated by the PBC is not permitted)

Condition 4: the above noted “disk isolation factor” is a factor of an incorporation target (i.e., the “disk isolation factor” is referred to and if it is a factor of a incorporation target, a reincorporation is permitted and it is carried out)

(2) In the case of carrying out a reincorporation, the disk incorporation unit 54 monitors the disk statistics unit 53 for a certain time period after the incorporation and, if there is a point being added to other disks, isolates the aforementioned disk by determining that the incorporated disk is the cause. In other words, it-monitors a statistics of the FC transmission path in which a disk is incorporated for a certain time period after a reincorporation and, if a point is added to the transmission path, isolates the aforementioned disk as a suspect disk. Incidentally, the above noted “other disks” is defined as all disks existing in the same loop as the incorporated disk for example.

FIG. 5 is a diagram exemplifying a structure of the above noted common table.

The common table 60 shown in FIG. 5 is furnished with storage areas for storing various kinds of information such as the above noted isolation cause through factor for each disk, and the stored data are cleared at the time of a disk replacement.

The common table 60 shown in the diagram stores an isolation cause 61, a reincorporation 62, a PBC cause 63, a recovery in-progress 64 and a factor 65, all for each disk.

The information other than the factor 65, i.e., the isolation cause 61, reincorporation 62, PBC cause 63 and recovery in-progress 64, each is one bit flag information for example.

The isolation cause 61 is set by a judgment of the disk statistics unit 53 as to whether the present disk has been isolated by a device system error (i.e., a hardware-wise breakdown) or an FC system error (i.e., an abnormality in the transmission path). An example setup is “1” for a device system error and “0” for an FC system error.

The reincorporation 62 is set, for example, at “1” by the disk incorporation unit 54 in the case of the present disk having been reincorporated. It is cleared to “0” when a certain time period elapses following being set at “1”.

The PBC cause 63 is set, for example, at “1” by the monitor unit 51 according to a notification from the PBC in the case of carrying out an isolation of the present disk based on a PBC judgment.

The recovery in-progress 64 is set, for example, at “1” by the configuration management unit 52 in the case of a rebuild/copyback being in operation prior to a reincorporation relating to the present disk (that is, the present disk is in the process of recovery).

In the factor 65, an eventual isolation cause (e.g., an error code such as a later described “0x0028”, et cetera) is set by a judgment of the disk statistics unit 53. That is, the above noted “disk isolation factor” is set.

Incidentally, a Disk WWN of each of the currently equipped disks is also stored while it is not specifically shown herein.

FIG. 6 is a flow chart of the disk incorporation process unit 54. This process is according to the first embodiment.

Having detected that a discretionary disk has once been isolated followed by being connected, the PBC, et cetera, for example, reads a Disk WWN of the disk (called as “target disk” hereinafter) and notifies the disk incorporation unit 54 of the Disk WWN (step S11) (simply noted as “S11” hereinafter). The disk incorporation unit 54 carries out the processes of the S12 and thereafter.

That is, it first compares the notified Disk WWN with the stored Disk WWNs (S12) and, if they are not identical, that is, if a disk different from the isolated disk is installed for example (“no” for S13), carries out a normal incorporation process (S14). Contrarily, if the Disk WWNs are identical, that is, if the isolated disk is reinstalled (“yes” for S13), it carries out the processes of the S15 and thereafter.

It carries out the processes of the S15 and thereafter by referring to various kinds of information stored in the common table 60 relating to the reinstalled target disk.

That is, as it is possible to know whether the isolation cause of the target disk is a device system error (i.e., a hardware-wise breakdown of the disk itself) or FC system error (i.e., an abnormality in the transmission path) by referring to the isolation cause 61 to begin with; if the cause is the device system error (“yes” for S16), the disk incorporation unit 54 cancels an incorporation process for the target disk (i.e., a reincorporation is not permitted) (S21).

Comparably, if the isolation cause of the target disk is an FC system error (i.e., an abnormality in the transmission path) (“yes” for S17), if the state of the target disk is “in recovery” (i.e., the recovery in-progress 64 is “1” for example) (“yes” for S18), it cancels an incorporation process for the target disk (i.e., a reincorporation is not permitted) (S21)

Then, if the target disk has been isolated according to a PBC judgment (i.e., the PBC cause 63 is “1” for example) (“yes” for S19), or if the “disk isolation factor” (refer to the factor 65) is not an “incorporation target factor” (“no” for S20), it also cancels an incorporation process for the target disk (i.e., a reincorporation is not permitted) (S21).

Incidentally, the disk isolation factor is described later by showing a specific example. Note that the case of the judgment of the S19 being “no” (i.e., not a case of an isolation according to a PBC judgment) includes, for example, an event of an operator (i.e., a maintenance technician, et cetera) operating the FST 20 to instruct the CM 10 for the isolation of the target disk or that of isolating it according to a CM 10 judgment.

With an exception of judging as a cancellation of the reincorporation process (i.e., the reincorporation being not permitted), it permits and carries out the incorporation process for the present target disk (S22).

Then, having completed the incorporation of the target disk, the disk incorporation unit 54 starts a timer which shifts to a time-out in a predetermined length of time (S23). And it monitors the disk statistics unit 53 (i.e., monitors a situation of the above noted counting of points by the disk statistics unit 53) until the timer shits to a time-out, judges whether or not the count-up points added to other disks exceed a preset second threshold value and, if the points exceed the threshold value (“yes” for S24), carries out an isolation process for the incorporated disk (S26). Comparably, if the timer shifts to a time-out before the count-up points to other disks exceed the threshold value (“no” for S24), it does nothing (S25). Note that the second threshold value used in the above described step S24 is one different from the threshold value (which is called a “first threshold value”) for judging whether the above described isolation is to be carried out (that is, the second threshold value is smaller than the first threshold value).

As described above, the process according to the first embodiment permits the incorporation if all the conditions shown in the above described steps S16 through S20 are satisfied even in the case of the same disk having been reinstalled. In other words, an incorporation of the same disk is permitted if the cause of a disk failure or a situation at the event thereof is considered to create no problem associated with the reinstallation of the same disk. However, the configuration is such that a monitor is performed for a certain period of time after the reincorporation because the incorporated disk may adversely affect other disks, and the disk is isolated again if there is a problem.

FIG. 7A exemplifies an FC system error.

The “0x0028”, “0x100b”, et cetera, are error codes of the FC system errors of which the meanings and fault causes are shown in a list thereof as shown in FIG. 7A.

The error code “0x0028” means that “the disk was not existent in an FC loop although it existed in the configuration information”, and the error code “0x1083” means that “a disk was not existent in an FC loop”. These two errors are examples of the above noted “FC system error (i.e., an error on a disk transmission path) but is apparently a disk cause error”.

Note that FIG. 7A also shows an example of the FC system error of which the failure cause is a transmission path, for reference. That is, the error code “0x0002” means that “a DMA error detected during a data transfer”, the error code “0x0015” means that “a data under-run” detected”, and the error code “0x10b” means that “a driver time-out detected”.

FIG. 7B shows a specific example of the above noted “disk isolation factor”. The factor of which the “reincorporation appropriateness” is “appropriate” shown in FIG. 7B is the above noted “incorporation target factor”. That is, examples shown in the chart, i.e., “isolation due to a disk statistics”, “isolation due to a forced degeneration”, “isolation due to a preventive maintenance” and “disk not ready”, are the above noted “incorporation target factor”. The individual factors other than these factors do not constitute the above noted “incorporation target factor” in the examples shown in the chart, and therefore an incorporation is not permitted even though other conditions are satisfied.

That is, the respective factors, i.e., “Write & Verify Error”, “SMART notification from a disk”, “disk isolation from a RAID recovery”, “isolation due to detecting a Disk Event” and “isolation due to DE Off/On” in the examples shown in the chart do not constitute the above noted “incorporation target factor”.

The next is a description of a second embodiment.

FIG. 8 shows a process flow chart of a disk incorporation process unit according to the second embodiment.

The second embodiment premises an execution of a certain process immediately after the judgment of the S13 shown in FIG. 6 being “yes”. That is, premising the execution of the process of “referring to the reincorporation 62 of the common table 60 and, if it is “1” (meaning that the present disk is reincorporated), “the incorporation is canceled” immediately instead of shifting to the S15”. Moreover, if the judgment of the above noted S24 is “no”, the disk incorporation unit 54 sets “1” to the reincorporation 62 instead of doing nothing (S31).

Then it starts a timer (called as “monitor timer” hereinafter) different from the timer of the above noted S23 (S32). A set time of the monitor timer is basically longer than that of the timer of the S23.

Then, in the case of the reincorporated disk having been isolated again before the monitor timer shifts to a time-out (“yes” for S33), it carries out the process of FIG. 6; however, the judgment becomes that “the incorporation is canceled” by the above noted added process since the reincorporation 62 remains as being set as “1” in the S31. That is, the judgment of “the incorporation is canceled” is forced, in lieu of applying the judgment logic of FIG. 6 (S35).

Contrarily, if the monitor timer shifts to time-out without the reincorporated disk being isolated again (“no” for S33), it clears the reincorporation 62 to “0” (S34). In this case, the judgment logic of FIG. 6 is applied in lieu of being judged as “the incorporation is canceled” forcibly, even if the reincorporated disk is isolated again thereafter.

The RAID apparatus, module therefor, etcetera, according the present invention, relating to a RAID apparatus managing a how swap by using a Disk WWN, is contrived to permit the incorporation of the same disk if a predefined condition is satisfied even in the case of a hot swap being performed by the aforementioned disk, thus solving the above described problem of taking an extraneous work and increasing a cost in the event of replacing with a new disk.

Claims

1. A module within a Redundant Array of Inexpensive Disks (RAID) apparatus including a RAID group constituted by a plurality of disks, comprising: a first storage unit for registering an identifier name of each of the disks;a second storage unit for storing a cause for isolating each of the disks; anda disk incorporation process unit for judging whether or not a predefined series of conditions are satisfied by referring to the second storage unit and carrying out an incorporation process for an installed disk if the conditions are satisfied even in the case of an identifier name registered in the first storage unit being identical with that of the installed disk when detecting the facts of a discretionary one of the disks having been isolated and a discretionary disk having been installed.

2. The module according to claim 1, further comprising a disk statistics unit for carrying out the processes of counting up a point corresponding to a phenomenon of an error for each occurrence of the error by each of said disks, and isolating a disk of which the count-up result of points exceeds a preset first threshold value, whereinsaid disk incorporation process unit monitors a count-up situation by the disk statistics unit for a certain period of time following the carry-out of an incorporation process for said installed disk, and isolates the aforementioned disk if a count-up result of points for a disk other than the installed disk exceeds a preset second threshold value.

3. The module according to claim 1, wherein said predefined series of conditions includes at least a condition of said cause for isolating said disk being not a cause of the disk itself hardware-wise.

4. The module according to claim 3, wherein said predefined series of conditions are further added by a condition of a detail factor being a “factor of an incorporation target”.

5. The module according to claim 3, wherein said second storage unit further stores information, for each of said disks, indicating whether or not an isolation has been performed according to a judgment of a port bypass circuit (PBC), andsaid predefined series of conditions are further added by a condition of said isolated disk having not been isolated according to a judgment of the PBC.

6. The module according to claim 3, wherein said second storage unit further stores information indicating whether or not a state of each of the said disks is in recovery, andsaid predefined series of conditions are further added by a condition of a state of said isolated disk being not in recovery.

7. The module according to claim 1, wherein said disk incorporation process unit does not permit a reincorporation of said installed disk regardless of said conditions being satisfied, if the disk is isolated within a predefined period of time following a carry-out of an incorporation process for the disk.

8. A Redundant Array of Inexpensive Disks (RAID) apparatus, including: a RAID group constituted by a plurality of disks; anda module for collecting, and managing, contents of an error occurring in each of the disks, and also carrying out an incorporation process for a discretionary disk, whereinthe module comprises:a first storage unit for registering an identifier name of each of the disks;a second storage unit for storing a cause for isolating each of the disks; anda disk incorporation process unit for judging whether or not a predefined series of conditions are satisfied by referring to the second storage unit and carrying out an incorporation process for an installed disk if the conditions are satisfied even in the case of an identifier name registered in the first storage unit being identical with that of the installed disk when detecting the facts of a discretionary one of the disks having been isolated and a discretionary disk having been installed.

9. A disk incorporation appropriateness judgment method used for a controller module within a Redundant Array of Inexpensive Disks (RAID) apparatus comprising a RAID group constituted by a plurality of disks, carrying out an incorporation process for an installed disk if a predefined condition is satisfied even in the case of a stored identifier name of each of the disks being identical with that of the installed disk when detecting the facts of a discretionary one of the disks having been isolated and a discretionary disk having been installed.

10. A program for making a computer used for a Redundant Array of Inexpensive Disks (RAID) apparatus comprising a RAID group constituted by a plurality of disks, wherein the program makes the computer execute the function of carrying outan incorporation process for an installed disk if a predefined condition is satisfied even in the case of a stored identifier name of any of the disks being identical with that of the installed disk when detecting the facts of a discretionary one of the disks having been isolated and a discretionary disk having been installed.