Device and method for controlling redundant pass

Description: TECHNICAL FIELD [Detailed description of the invention]

【0001】

[Technical field to which the invention belongs]

The present invention relates to a redundant path control device and a method for performing retry processing on a logical disk using an alternative path including an initiator when an failure occurs in an I / O access path.

【0002】

[Conventional technology]

The path redundancy driver 3'is used when a failure occurs in the components of the I / O access path, that is, HBA (Host Bus Adapter) 5,6, interface cable 31,32, controller 11', 12', etc. Also provides a means to perform retry processing with an alternative path so as not to affect the operation of the application 7 running on the host computer 1. That is, the redundant disk array device 10 and the software (path redundancy driver 3') running on the host computer 1 can realize the redundancy of the I / O path for the logical disks 13 to 15.

【0003】

In addition, arbitrary logical disks 13 to 15 used by the application 7 running on the host computer 1 may be occupied (reserved) so that they cannot be accessed by other applications (not shown) or other host computers. In general, this occupancy state is released by a release command from the initiator that previously occupied it (HBA is applicable here), or is forcibly released by a reset by HW (hardware).

【0004】

[Problems to be Solved by the Invention]

Here, for example, considering the case where the application 7 running on the host computer 1 is used while occupying the logical disk 13 via the HBA5, the interface cable 31 and the controller 11', the following problems occur. To do.

【0005】

The first problem is that the function of the path redundancy driver 3'becomes meaningless in the following cases. The path redundancy driver 3'receives the execution result of the application 7 running on the host computer 1 making an I / O request to the logical disk 13 from the driver 4, and determines normal termination or abnormal termination. At this time, it is determined that the termination is abnormal, and that the cause is a failure (failure) in the component of the path (HBA5, interface cable 31 or controller 11'). As a result, an attempt is made to retry the abnormally terminated I / O using the alternative path, that is, via the HBA6, the interface cable 32, and the controller 12'for the logical disk 13. However, since the logical disk 13 is occupied by another initiator (HBA5), it cannot be read / written, and as a result, the I / O request of the application 7 ends abnormally. Therefore, the path redundancy driver 3'cannot serve its intended purpose.

【0006】

The second problem is that when the HW forcibly resets and releases the occupied state, one or more of the processes received from the host computer 1 side due to the initialization of the controllers 11'and 12'. Since all I / O requests are discarded, extra retry processing occurs on the host computer 1 side. Also, if the bus (host interface) is reset, all the devices (single devices 20, 21) connected to the bus will also be initialized, so the burden on the host computer 1 side will be even greater (retry). Processing) is applied.

【0007】

Further, considering the case where the host computer 1 occupies an arbitrary logical disk 13 to 15 for the purpose of protecting the data integrity on the logical disks 13 to 15, the following problems occur.

【0008】

The third problem is that temporarily releasing the occupied state of the logical disk from a certain host computer allows access from a completely unintended host computer, so the data in the logical disk is changed. There is a risk of destroying it.

【0009】

[Purpose of Invention]

Therefore, an object of the present invention is to provide a redundant path control device and a method that enable access to a logical disk that has already been reserved by using an alternative path.

【0010】

[Means for solving problems]

The redundant path control device (or redundant path control method) according to the present invention performs retry processing on a logical disk using an alternative path including an initiator when a failure occurs in the I / O access path. , If the retry process ends abnormally because the logical disk is reserved by another initiator, a redundant path control means (or redundant path control step) is provided to reset the logical disk to the reserved state by the initiator of the alternative path. (Claims 1 and 5).

【0011】

It is assumed that when a retry process is executed on a logical disk using an alternative path, the retry process ends abnormally because the logical disk has already been reserved by another initiator. Then, the redundant path control means (or the redundant path control step) resets the logical disk to the reserve state by the initiator of the alternative path. This makes it possible to access the logical disk using an alternative path.

【0012】

In addition, the redundant path control means (or redundant path control step) issues a forced reserve command from the initiator of the alternative path when the retry process ends abnormally because the logical disk is reserved by another initiator. Control means (or first control step), and when a forced reserve command is issued, the second control means (or second control step) that resets the logical disk to the reserve state by the initiator of the alternative path. May be provided (claims 2, 6).

【0013】

Further, a plurality of logical disks may be assembled to form a disk array device (claims 3, 7), or the initiator may be an HBA (claims 4, 8).

【0014】

BEST MODE FOR CARRYING OUT THE INVENTION

Hereinafter, embodiments of the redundant path control device according to the present invention will be described with reference to the drawings. The redundant path control method according to the present invention is used in the redundant path control device according to the present invention. Therefore, by explaining the present embodiment, one embodiment of the redundant path control method according to the present invention will also be described.

【0015】

FIG. 1 is a block diagram showing an embodiment of a redundant path control device according to the present invention. Hereinafter, description will be given based on this drawing. The first means and the second means in the claims correspond to the path redundancy driver and the controller in the present embodiment, respectively.

【0016】

The HBAs 5 and 6 of the host computer 1 are connected to the controllers 11 and 12 of the disk array device 10 via the interface cables 31 and 32, respectively. The host computer 1 executes I / O for the logical disks 13 to 15 in the disk array device 10. Driver 4 controls HBA5 and 6 to perform I / O processing.

【0017】

The path redundancy driver 3 operates as follows. Pass the I / O received from file system 2 to driver 4. The execution result of I / O for any of the logical disks 13 to 15 in the disk array device 10 is received from the driver 4 via the HBAs 5 and 6, and normal termination or abnormal termination is determined. If it is determined that the cause of the abnormal termination is a failure (failure) in the path components (HBA5,6, interface cable 31,32, controller 11,12, etc.), the alternative path is used to determine the abnormal termination I / O. Perform retry processing.

【0018】

The controllers 11 and 12 of the disk array device 10 are connected to the logical disks 13 to 15 via the internal buses 16 and 17, respectively. Each of the logical disks 13 to 15 can be accessed from any of the controllers 11 and 12.

【0019】

For example, the path redundancy driver 3 issues a forced reserve command from the initiator 6 of the alternative path when the retry process ends abnormally because the logical disk 13 is reserved by another initiator 5. When the forced reserve command is issued, the controller 12 resets the logical disk 13 to the reserve state by the initiator 6 of the alternative path.

【0020】

FIG. 2 is a flow showing a part of the I / O command processing process received from the host computer 1, and these processes are executed by the controllers 11 and 12 of the disk array device 10. In Fig. 3, the normal termination or abnormal termination is determined based on the I / O execution result received from the driver 4, and the cause of the abnormal termination is the path component (HBA5,6, interface cable 31,32, controller 11). , 12 etc.), it is a flow showing a part of the process of retrying the abnormally terminated I / O using the alternative path, and these processes are path redundant. Implemented by driver 3. Hereinafter, the operation of the redundant path control device of the present embodiment will be described with reference to FIGS. 1 to 3.

【0021】

The data (write I / O) written to the disk array device 10 by the application 7 running on the host computer 1 is passed through the application 7, the file system 2, the path redundancy driver 3, the driver 4, the HBA 5, and the interface cable 31. It reaches controller 11 and is written to any of the specified logical disks 13 to 15.

【0022】

The data (read I / O) read from the disk array device 10 by the application 7 running on the host computer 1 is sent to the HBA 5 from any of the specified logical disks 13 to 15 via the controller 11 and the interface cable 31. It reaches application 7 via driver 4, path redundancy driver 3, and file system 2.

【0023】

In addition, the execution result of each I / O by the host computer 1 is generally judged by each layer of HBA5, driver 4, file system 2 and application 7, and some action is taken as necessary. is there. Here, the path redundancy driver 3 determines whether the I / O execution result received from the driver 4 ends normally or abnormally, and the cause of the abnormal end is the component of the path (HBA, interface cable, controller, etc.). If it is determined that there is a failure in the above, an alternative path is used to retry the I / O that ended abnormally.

【0024】

Next, the state in which the application 7 reserves the logical disk 13 will be described.

【0025】

A reserve command is a command for exclusively occupying a logical disk specified by a specific initiator (HBA). For example, when a reserve is performed by designating a logical disk 13 from the HBA 5 via the controller 11, I / O (read or write) from the HBA 5 to the logical disk 13 via the controller 11 is performed unless there is a problem such as a failure. It ends normally. However, the I / O (read or write) issued from the HBA 6 to the logical disk 13 via the controller 12 terminates abnormally with the error message "Reservation Conflict".

【0026】

As can be seen from these facts, the path redundancy driver 3 cannot make the function work well for the I / O for the logical disk 13 occupied by the reserve command.

【0027】

A method for solving this problem will be described using the flows of FIGS. 2 and 3.

【0028】

FIG. 2 is a flow showing a part of the I / O command processing process executed by the controllers 11 and 12 of the disk array device 10 and received from the host computer 1. In order to solve this problem, the disk array device 10 is provided with a forced reserve means.

【0029】

First, it is determined whether or not the command received from the host computer 1 can be executed correctly (S101). If the command cannot be executed correctly, it shifts to abnormal termination processing (S111). If it is a command that can be executed correctly, it is determined whether it is an existing command or a newly provided forced reserve command (S102). If it is a forced reserve command, the flag is cleared (S103), otherwise it shifts to the existing command decoding process (S112).

【0030】

Then, it is determined whether or not the specified logical disk is already reserved (S104). If it is not reserved, set it to the reserved state by the initiator that issued the forced reserve command (S107). If it is already reserved, it is determined whether the reserve is reserved by an initiator other than the initiator that issued the forced reserve command (S105). If the reserve is from the initiator that issued the forced reserve command, the process shifts to termination processing (S113). If the reserve is from an initiator other than the initiator that issued the forced reserve command, set the flag (S106) and set the reserve state by the initiator that issued the forced reserve command (S107).

【0031】

Then, the flag is confirmed (S108), and if the flag is not set, the process proceeds to the termination process (S110). If the flag is set, it is reserved by another initiator, so the reserve state is canceled, the used flag is cleared (S109), and the process proceeds to the end process (S110).

【0032】

FIG. 3 is a flow performed by the path redundancy driver 3 and showing a part of the retry processing process by the alternative path.

【0033】

First, the path redundancy driver 3 receives the execution result of I / O for the logical disks 1 to 15 in the disk array device 10 from the driver 4 and determines normal termination or abnormal termination. If it is determined that the cause of the abnormal termination is a failure (failure) in the path components (HBA, interface cable, controller, etc.), the alternative path is used to retry the abnormally terminated I / O (S201). ). Then, the retry result is judged (S202), and if there is no abnormality, the process proceeds to the normal termination process (S209).

【0034】

If there is an anomaly, determine whether the cause of the anomaly is due to a logical disk reserved by another initiator (S203). In the case of an abnormality caused by something else, the process proceeds to abnormal termination processing (S210). On the other hand, if the cause of the error is a logical disk reserved by another initiator, execute the forced reserve command for the same path (S204).

【0035】

Then, the result is judged (S205). If the forced reserve command does not end normally, the process proceeds to abnormal termination processing (S211). If the forced reserve command ends normally, the occupancy of the logical disk has been changed to this initiator, so the I / O that ended abnormally in step 202 is retried again with the same path (S206). Subsequently, the result of step 206 is determined (S207), and if the retry is successful, the process proceeds to the normal termination process (S208), and if the retry is unsuccessful, the process proceeds to the abnormal termination process (S212).

【0036】

As described above, an I / O path redundancy system for the reserved logical disk is realized.

【0037】

Needless to say, the present invention is not limited to the above embodiment. For example, the following embodiments can be adopted.

【0038】

The number of HBAs 5 and 6 is limited by the type of OS, driver 4, or HW such as host computer 1, and there is no limitation on path redundancy driver 3. There is no limit to the number of controllers 11 and 12 and the number of ports for connecting each controller 11 and 12 to the host computer 1 and the like via the interface cables 31 and 32. An example is shown in which the HBAs 5 and 6 and the controllers 11 and 12 are directly connected by the interface cables 31 and 32, but a hub or a switch or the like may intervene in the middle of these.

【0039】

There is no limit to the number of disk array devices 10 connected to the host computer 1. There is no limit to the number of host computers 1 connected to the disk array device 10. There is no limit to the number of logical disks 13 to 14 in the disk array device 10. The present invention is not limited to the disk array device 10.

【0040】

[Effect of the invention]

According to the redundant path control device and method according to the present invention, when the retry process ends abnormally because the logical disk has already been reserved by another initiator, the logical disk is set to the reserved state by the initiator of the alternative path. By re-doing, the logical disk can be accessed using an alternative path. In other words, in the event of a failure (failure) in a path component (HBA, interface cable, controller, etc.), even if there is an application that occupies and uses the logical disk, it is replaced so as not to affect the operation of the application. It is possible to retry with a pass.

【0041】

In addition, since the occupancy is not forcibly released by resetting by HW, it is possible to suppress the occurrence of unnecessary retry processing on the host computer side.

【0042】

Furthermore, by not releasing the occupied state of the logical disk even temporarily, it is possible to avoid the risk of destroying the data in the logical disk because it does not give an opportunity to allow access from a host computer that is not intended at all.

[Simple explanation of drawings]

[Figure 1]

It is a block diagram which shows one Embodiment of the redundant path control apparatus which concerns on this invention.

[Figure 2]

It is a flowchart which shows an example of the operation of the controller in the redundant path control device of FIG.

[Fig. 3]

It is a flowchart which shows an example of the operation of the path redundancy driver in the redundant path control device of FIG.

[Fig. 4]

It is a block diagram which shows the conventional redundant path control device.

[Explanation of symbols]

1 Host computer 2 file system 3-path redundancy driver 4 driver 5,6 HBA 7 application 30 ~ 33 interface cable 10 Disk array device 11,12 controller 13 ~ 15 Logical disk 16,17 Internal bus 20,21 Single device