JP2010033379A - Virtualization system and restoration method for virtualization - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To automatically restore configuration information of a changed virtualization storage. <P>SOLUTION: This virtualization system 1 connected to a plurality of storages 30a-c, virtualizing physical capacity of the plurality of storages to logical capacity has: a first virtualization switch 11 executing the virtualization according to first configuration information for prescribing a configuration of the logical capacity; a second virtualization switch 12 redundantly configured with the first virtualization switch 11, executing the virtualization according to second configuration information for prescribing the configuration of the logical capacity during operation of the first virtualization switch 11; and a management device 20 storing third configuration information for prescribing the configuration of the logical capacity, and changing one piece of the configuration information such that the configuration information accords with the other two pieces of the configuration information when the one piece of the configuration information of the first, second, and third configuration information does not accord with the other two pieces of the configuration information. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a virtualization system and a virtualization recovery method.

  As the use of digital data increases, there is a demand for efficient storage of digital data in an external storage device (hereinafter referred to as “storage”) and efficient use of digital data stored in the storage. Many storages used as devices for storing digital data (for example, direct-attached storage (DAS)) are not flexible with respect to storage specification changes, capacity increases, and the like.

  For this reason, a technology that converts the physical capacity of multiple storages to a specified logical capacity (hereinafter referred to as “storage virtualization”) in order to make it possible to expand the capacity by adding storage or use a combination of different types of storage. Is being used. In addition, the logical capacity is physically connected to a plurality of storages, and another device (hereinafter referred to as “host”) connected to the switch by a switch (hereinafter referred to as “virtualization switch”) that executes a virtualization program. Provided). For this reason, the host cannot see the physical storage, but can see the virtual storage having the logical capacity (hereinafter referred to as “virtual storage”).

  Further, the virtualization switch is arranged in a configuration in which a plurality of virtualization switches are made redundant in an operating state in order to obtain high robustness. Each virtualization switch uses configuration information that defines the configuration of exactly the same logical capacity, so that no matter which virtualization switch the host computer (hereinafter referred to as “host”) accesses, it is the same as the host. Virtual storage can be provided.

However, if the configuration information stored in the virtualization switch is erroneously altered or deleted due to an operation error or the like, the host cannot access the data stored in the physical capacity. In this case, the virtualization switch whose configuration information has been modified returns an error to the host, so that the host can access data via another virtualization switch made redundant by multipath control.
Although the host access to the storage is maintained by such redundancy, the stoppage of the use of one virtualization switch reduces the robustness of the virtualization switch. Therefore, the administrator needs to restore correct configuration information for a virtualization switch whose configuration information has been modified.

  If there is a discrepancy in the configuration information between the two redundantly configured processing devices, the configuration information of the standby standby device is changed to match the configuration information of the active device, and the operation is performed normally. A method is known (Patent Document 1 below). However, such a conventional technique is not a technique for matching the configuration information of each processing apparatus when all the processing apparatuses are operating.

JP 2007-122627 A

  This virtualization system and virtualization recovery method automatically recovers the modified or deleted configuration information when the configuration information of the operating virtualization switch is modified or deleted in a system configuration in which all virtualization switches are operated. For the purpose.

  In order to solve the above problem, a virtualization system that is connected to a plurality of storages and virtualizes the physical capacity of the plurality of storages to a logical capacity, the first configuration information defining the configuration of the logical capacity According to the first virtualization switch, the first virtualization switch is configured redundantly with the first virtualization switch, and the logical capacity configuration is defined during operation of the first virtualization switch. A second virtualization switch that executes the virtualization according to the configuration information of 2, and third configuration information that defines the configuration of the logical capacity, and stores the first, second, and third configuration information A management device that changes any one of the configuration information so that the configuration information matches with the other two configuration information when any one of the configuration information does not match with the other two configuration information. Virtual features System is provided.

  In order to solve the above problem, the virtualization is executed according to the first configuration information that is connected to a plurality of storages and defines the configuration information for virtualizing the physical capacity of the plurality of storages to a logical capacity. The virtualization switch according to second configuration information that is redundantly configured with the first virtualization switch and that defines the configuration of the logical capacity during operation of the first virtualization switch. Virtual in a virtualization system having a second virtualization switch that executes the configuration and a management device that stores third configuration information that defines the configuration of the logical capacity and manages the first and second configuration information The method of requesting the acquisition of the first and second configuration information by the management device, and the first, second, and third configuration information by the management device. If any one of the configuration information does not match the other two configuration information, the step of changing any one of the configuration information so as to match the other two configuration information, and the other two configurations And storing the changed configuration information in the virtualization switch or management device having configuration information that does not match the information.

  The virtualization system can maintain high robustness without causing an error or an operation stop by automatically recovering the configuration information when there is a mismatch in the configuration information between the two operating virtualization switches.

Hereinafter, embodiments of the present virtualization system and this virtualization restoration method will be described with reference to the drawings.
An example of the configuration of the virtualization system 1 will be described with reference to FIG. The virtualization system 1 includes virtualization switches 11 and 12 and a management device 20.

The virtualization switches 11 and 12 are connected to the hosts 41 and 42 and the storages 30a to 30c via a network 5 indicated by a broken line using FC (Fiber Channel) or the like, and use a LAN (Local Area Network) cable or the like. A network 7 indicated by a solid line is used to connect to the management device 20 or the like.
The virtualization switches 11 and 12 have a function as a switching hub for switching data transmission / reception from the hosts 41 and 42 and the storages 30a to 30c, and perform virtualization according to configuration information (described later) that defines the configuration of the logical capacity. For example, the physical capacity of the storages 30a to 30c can be virtualized into a logical capacity.

  The virtualization switches 11 and 12 are connected to a single or a plurality of CPUs (Central Processing Units), volatile memories (SRAM, DRAM, etc.), non-volatile memories (EEPROMs, etc.), a hard disk, an input unit, and a network 7 An NIC (Network Interface Card) and an HBA (Host Bus Adapter) for connecting to the network 5 are provided.

The virtualization switches 11 and 12 can store a virtualization program and configuration information in a nonvolatile memory or a hard disk. In addition, the configuration information is preferably stored in a non-volatile memory so as to be quickly transmitted to the management device 20 in accordance with a request from the management device 20. Further, the virtualization switches 11 and 12 can realize virtualization of the storages 30a to 30c by executing a virtualization program on the CPU.
As shown in the figure, the virtualization switches 11 and 12 have two redundant configurations in order to allow operation to continue even if one system goes down. The virtualization switches 11 and 12 are both in an operating state, and by performing a heartbeat of each other's switch via the network 7, when one system goes down, the one system down is reported to the HBAs of the hosts 41 and 42. Processing can also be executed.
Further, the virtualization switches 11 and 12 can be configured to transmit and receive various data to and from the management device 20, the hosts 41 and 42, and the storages 30a to 30c via the network 5 or 7 using a known protocol. Note that the configuration information may be transmitted / received between the virtualization switches 11 and 12 and the management apparatus via the network 7 having a low transfer rate because the data amount is smaller than the input / output data of the hard disk.

  The management device 20 executes a management program for managing the virtualization switches 11 and 12, holds the configuration information of the virtualization switches 11 and 12, and changes the configuration information stored in the virtualization switches 11 and 12. can do.

In addition, the management device 20 includes a single or a plurality of CPUs, a memory, a hard disk, a display, an input unit, an HBA, and a NIC.
The management device 20 can store a management program and configuration information in a nonvolatile memory or a hard disk. Further, the management device 20 can manage the virtualization switches 11 and 12 by executing a management program on the CPU. Furthermore, the management apparatus 20 can be configured to transmit and receive various data to and from the virtualization switches 11 and 12, the hosts 41 and 42, and the storages 30a to 30c via the network 5 or 7 using a known protocol. As described above, the configuration information may be transmitted / received between the virtualization switches 11 and 12 and the management apparatus 20 via the network 7.

The storages 30a to 30c can include FC ports, RAID (Redundant Arrays of Inexpensive) controllers, and hard disks.
The storages 30a to 30c can be configured to transmit / receive input / output data to / from the virtualization switches 11 and 12 via the network 5 using a fiber channel protocol. The storages 30a to 30c can be configured to be able to transmit / receive storage setting information such as LUN (Logical Unit Number) information to / from the management apparatus 20 via the network 7 using SNMP (Simple Network Management Protocol).
As illustrated, the storage 30a, the storage 30b, and the storage 30c may include LUNs 31a to 33a, 31b and 32b, and 31c and 32c, respectively.

The hosts 41 and 42 can acquire and use various data stored in the storages 30 a to 30 c via the network 5 and the virtualization switches 11 and 12.
The hosts 41 and 42 have a single or a plurality of CPUs, memories, hard disks, HBAs, and NICs. The hosts 41 and 42 can be configured to transmit and receive input / output data to and from the virtualization switches 11 and 12 via the network 5 using a known protocol.

An example of storage virtualization processing will be described with reference to FIG.
Based on the LUN information received from the storages 30a to 30c, the virtualization switches 11 and 12 form a virtual storage pool 51 that is an area in which resources taken from the physical storage are stored in the hosts 41 and 42. The virtual storage pool 51 can further generate virtual storages 53 and 56. The virtual storages 53 and 56 are separate logical areas provided in the virtual storage pool 51. As illustrated, the virtualization switches 11 and 12 can allocate LUNs 31a to 33a and 31b and 32b, 31c and 32c to the virtual storage 53 and the virtual storage 56, respectively, according to the configuration information.

The virtual storage 53 and the virtual storage 56 have virtual targets 54 and 55, which are virtual storage FC ports, and virtual targets 57 and 58, respectively.
The host 41 is physically connected to each of the virtualization switches 11 and 12, but is connected to the virtual targets 54 and 55 as the connection to the virtual storage 53. Similarly, the host 42 is physically connected to each of the virtualization switches 11 and 12, but connected to the virtual targets 57 and 58 as a connection to the virtual storage 56.
That is, the hosts 41 and 42 do not see the physical storages 30a to 30c, but can see the virtual storages 53 and 56.

  Before the LUN is allocated to the virtual target, the hosts 41 and 42 cannot access the virtual storages 53 and 56. However, after the LUN is allocated, the hosts 41 and 42 can access the virtual storages 53 and 56.

  As described above, the virtualization switches 11 and 12 collectively combine the resources of the three storages 30a to 30c, and further generate the virtual storages 53 and 56, thereby changing the physical capacity of the plurality of storages to the plurality of logical capacities. Can be divided into

The configuration information stored in the virtualization switches 51 and 52 includes the following information, for example.
1) Virtual resource configuration information Virtual storage information: Virtualization switch information (information associated with a virtual target, used when responding to a SCSI command from a host. Different from the storage information of the storages 30a to 30c), vendor ID (including vendor information), product ID (including information such as device name), serial number (virtual storage identifier), series name, model type name, box ID (Virtual storage ID)

  Virtual volume information: Virtual volume disk information (which LUN of which LUN of the storage LUN area is used, or RAID type), volume information of the virtual volume (volume size, concatenation (different) Information indicating that the LUN is used as one virtual volume), shared information of the virtual volume (whether the volume is in a shared state (multipath control from the host is possible if sharing is set))

  Virtual target information: Virtual target (corresponding to storage FC port), virtual target WWN (World Wide Name), path ID (identifying the maintenance unit allocated to each virtual target)

  Virtual host LUN information: link information (indicating which virtual volume and which virtual target are linked to form a host LUN), number of virtual volume to be recognized by the host

  Other: Bitmap table setting (Management area setting for realizing the function to temporarily stop copying)

2) Copy session configuration information Copy session configuration information Session ID, copy status, copy source virtual volume, copy destination virtual volume can be included.

  Such configuration information is appropriately modified in accordance with a change in the configuration of the virtual storage pool 51 of FIG. 2, LUN data migration, and the like, and is held in the management device 20 and the virtualization switches 11 and 12.

An example of an automatic restoration flow of configuration information will be described with reference to FIG. First, the management device 20, the virtualization switch 11, and the virtualization switch 12 initially have the same configuration information A.
Although not shown, the hosts 41 and 42 always access the virtualization switch 11 or 12 for data access.

The operator changes the configuration information A of the virtualization switch 11 via the input unit to make the configuration information B (step S101). Note that the operation of changing the configuration information A to the configuration information B by the operator may occur due to an erroneous operation.
The management device 20 requests the virtualization switch 11 to acquire configuration information via the network 7 by polling (step S102), and the virtualization switch 11 responds to the request to obtain the configuration information B via the network 7. Transmit (step S103).

Similarly, the management device 20 requests the virtualization switch 12 to acquire configuration information via the network 7 by polling (step S104), and the virtualization switch 12 configures via the network 7 in response to the request. Information A is transmitted (step S105).
The management device 20 compares its own configuration information and the received configuration information, and when only one of the configuration information is inconsistent with the other two configuration information, the other 2 Any one piece of configuration information is changed to match one piece of configuration information (step S106). In other words, the management apparatus 20 includes the configuration information A and the virtualization switch 12 held by the management apparatus 20 in the configuration information B received from the virtualization switch 11 among the configuration information A and the received configuration information A and B. Therefore, the configuration information B received from the virtualization switch 11 is changed to the configuration information A.
The management device 20 transmits the configuration information A, which is the configuration information after the change, to the virtualization switch 11, which is the device that has transmitted the changed configuration information (step S107), and performs an automatic recovery flow of the configuration information. finish.

In step S101, the operator changes the configuration information stored in the virtualization switch 11. However, the operator transmits the configuration information changed by the operator using the management device 20 to the virtualization switch 11, so that the virtualization is performed. The configuration information of the switch 11 may be changed.
In step S101, the configuration information is changed. However, this processing may be processing for deleting the configuration information by mistake. In that case, since the configuration information of the virtualization switch 11 is different from the configuration information A of the virtualization switch 12 and the management device 20 in step S106, the management device 20 generates the configuration information A, and in step S107, the virtualization switch 11 11, the configuration information A is transmitted.

In this way, the virtualization system maintains high robustness without causing an error or an operation stop by automatically recovering the configuration information when there is a mismatch in the configuration information between the two operating virtualization switches. be able to.
In addition, since the administrator causes an abnormal state due to an operation error and is avoided from the recovery operation of the abnormal state, the virtualization system can reduce the burden on the administrator.

An example of the configuration information automatic recovery flow will be described with reference to FIG. In this example, a recovery flow of configuration information caused by a work mistake in the replacement work of the management apparatus will be described. The replacement of the management device here corresponds to a case where a new management device is replaced with an old management device and introduced into the virtualization system 1 or a case where a new management program is installed on the management device. The old management device 20a shown in the figure is a management device before replacement work (or before installation), and the new management device 20b is a management device after replacement work (or after installation). As illustrated, at the start of the flow, the old management device 20a and the virtualization switches 11 and 12 have configuration information, but the new management device 20b after the replacement work does not have configuration information.
Although not shown, the hosts 41 and 42 always access the virtualization switch 11 or 12 for data access.

The old management device 20a instructs the virtual storages 11 and 12 through the network 7 to transfer data from one of the LUNs of the storages 30a to 30c input by the operator via the input unit of the old management device 20a to another LUN. (Step S201). The old management device 20a backs up the configuration information in accordance with the configuration information backup instruction input by the operator via the input unit (step S202).
The virtual storages 11 and 12 that have received the data migration instruction start data migration according to the instruction (step S203). When the data migration is completed, the virtual storages 11 and 12 update the configuration information A to the configuration information C according to the data migration instruction (step S204), and transmit the configuration information C to the old management apparatus 20a (step S205). The old management device 20a that has received the configuration information C stops operating (step S206).

Next, the new management apparatus 20b is activated in accordance with the activation process by the operator (step S207), and the operator uses the configuration information A backed up in step S202 to read the configuration information A in the memory of the new management apparatus 20b (step S207). S208). As described above, the configuration information of the management device 20b is not immediately changed to the backed-up configuration information A, but the configuration information A read into the memory is used for comparison processing described later.
The new management device 20b requests the virtualization switches 11 and 12 to obtain configuration information (step S209), and the virtualization switches 11 and 12 send the configuration information C to the new management device 20b in response to the request. Transmit (step S210).

The new management device 20b compares the configuration information A read into the memory and the received configuration information C, and only one of the configuration information is inconsistent with the other two configuration information. Any one configuration information is changed so as to coincide with the other two configuration information (step S211). That is, the new management apparatus 20b receives the two pieces of configuration information received from the virtualization switches 11 and 12 by the configuration information A of the new management apparatus 20b among the configuration information A and the received two pieces of configuration information C. Therefore, the configuration information A of the new management apparatus 20b is changed to the configuration information C.
The management apparatus 20 displays the changed configuration information C on the output unit of the management apparatus 20, thereby notifying the operator of the result of automatic recovery due to the mismatch of the configuration information and generating a log (step S212). ), The configuration information automatic recovery flow is terminated.

In the above example, there is an operation error that is not the backup of the configuration information C reflecting the data migration in step S202 but the backup of the configuration information A before the reflection. Further, in step S208, the configuration information is restored using the configuration information A that was backed up by mistake. Even in such a case, if the configuration information does not match between the management device and the two virtualization switches that operate, the virtualization system is high without causing an error or operation stop by automatically recovering the configuration information. Robustness can be maintained. For this reason, the administrator can avoid an abnormal state due to an operation error and the recovery from the abnormal state. Therefore, the virtualization system can reduce the burden on the administrator.
Therefore, systematic automatic response is possible without restricting the migration / maintenance work at the time of configuration information backup by a work manual or the like. In addition, high robustness can be maintained even when the data migration and the management device replacement work are performed in parallel.

An example of the configuration information automatic restoration flow will be described with reference to FIG. In this example, a recovery flow of configuration information from a work mistake of the replacement work of the management apparatus will be described.
Steps S201 to S207, S209, and S210 in FIG. 5 are the same as steps S201 to S207, S209, and S210 in FIG.
However, in the flow shown in FIG. 5, the backup of the configuration information as in step S202 of FIG. 4 and the recovery process of the configuration information as in step S208 are not performed.
Although not shown, the hosts 41 and 42 always access the virtualization switch 11 or 12 for data access.

In step S301, the management apparatus 20 compares its own configuration information and the received configuration information, and only one of the configuration information does not match the other two configuration information. Any one of the configuration information is changed so as to coincide with the other two configuration information. In the case of this example, the new management apparatus 20b does not have the configuration information that it has, so the configuration information C is generated in accordance with the received two configuration information C and stored in the new management apparatus 20b.
The management device 20 displays the generated configuration information C on the output unit of the management device 20, thereby notifying the operator of the result of automatic recovery due to the mismatch of the configuration information and generating such a log. (Step S302), the configuration information automatic restoration flow is terminated.

An example of configuration information change processing performed by the management apparatus will be described with reference to FIG.
Patterns 1 and 2 shown in the figure are examples of the change process described with reference to FIG. Pattern 3 is an example of the change process described with reference to FIG. 4, and pattern 4 is an example of the change process described with reference to FIG.
Pattern 5 is a case where both the configuration information of the virtualization switches 11 and 12 have been deleted due to an operation mistake. In this case, the virtualization switch 11 is used by using the configuration information managed by the management apparatus 20. And 12 configuration information can be recovered.
In the case of patterns 6 to 9, an abnormality can be displayed on the display of the management apparatus 20 in order to perform a recovery operation by the operator.

The recovery processing when the configuration information held by the virtualization switch and the management apparatus do not match each other in the virtualization system has been described above. In the description, the event that causes the mismatch of the configuration information has been described as a work error. However, the event that causes the mismatch of the configuration information is not limited to the work error, and the management apparatus stops (the management apparatus is down, exchanged, Management software version upgrade, etc.), virtualization switch stops (switch down, replacement, etc.), network failure.
Even if there is a mismatch in configuration information due to a reason not associated with a work error, this virtualization system and this virtualization recovery method do not cause an error or stop operation by automatically recovering the configuration information. High robustness can be maintained.

The embodiment described above is as follows.
(Appendix 1)
A virtualization system connected to a plurality of storages and virtualizing the physical capacity of the plurality of storages to a logical capacity,
A first virtualization switch that executes the virtualization according to first configuration information that defines a configuration of the logical capacity;
A second virtual system configured redundantly with the first virtualization switch and executing the virtualization according to second configuration information defining a configuration of the logical capacity while the first virtualization switch is in operation; Switch
When third configuration information that defines the configuration of the logical capacity is stored, and any one of the first, second, and third configuration information does not match the other two configuration information A management device that changes any one of the configuration information so as to match the other two configuration information;
A virtualization system comprising:
(Appendix 2)
When the first or second configuration information does not exist, the management device generates the same configuration information as the other two existing configuration information, and the first or second virtualization without the configuration information The virtualization system according to appendix 1, wherein the generated configuration information is transmitted to a switch.
(Appendix 3)
The supplementary note 1 or 2, wherein when the third configuration information does not exist, the management device generates the same configuration information as the other two existing configuration information and stores the configuration information Virtualization system.
(Appendix 4)
A first virtualization switch connected to a plurality of storages and executing the virtualization according to first configuration information defining configuration information for virtualizing a physical capacity of the plurality of storages to a logical capacity; and A second virtualization switch configured redundantly with the first virtualization switch and executing the virtualization according to second configuration information defining a configuration of the logical capacity while the first virtualization switch is in operation; A virtualization restoration method in a virtualization system having a management device that stores third configuration information that defines the configuration of the logical capacity and manages the first and second configuration information,
The management device requesting acquisition of the first and second configuration information;
When any one of the first, second, and third configuration information does not match the other two configuration information, the management device matches the other two configuration information. Changing any one of the configuration information to:
Storing the changed configuration information in the virtualization switch or management device having configuration information inconsistent with the other two configuration information;
A recovery method characterized by comprising:
(Appendix 5)
When either the first or second configuration information does not exist, the management device generates the same configuration information as the other two configuration information present; and
The restoration method according to appendix 4, further comprising a step of transmitting the generated configuration information to the first or second virtualization switch having no configuration information by the management device.
(Appendix 6)
Item 6. The appendix 4 or 5, further comprising a step in which, when the third configuration information does not exist, the management device generates the same configuration information as the other two configuration information present and stores the configuration information Recovery method.

FIG. 1 is a diagram illustrating an example of a configuration of a virtualization system. FIG. 2 is a diagram illustrating an example of the configuration of the virtual storage. FIG. 3 is a diagram illustrating an example of an automatic restoration flow of configuration information. FIG. 4 is a diagram illustrating an example of an automatic restoration flow of configuration information. FIG. 5 is a diagram illustrating an example of an automatic restoration flow of configuration information. FIG. 6 is a diagram illustrating an example of changing the configuration information.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Virtualization system 11, 12 Virtualization switch 20 Management apparatus 30a-c Storage 41, 42 Host 53, 56 Virtual storage

Claims (3)

A virtualization system connected to a plurality of storages and virtualizing the physical capacity of the plurality of storages to a logical capacity,
A first virtualization switch that executes the virtualization according to first configuration information that defines a configuration of the logical capacity;
A second virtual system configured redundantly with the first virtualization switch and executing the virtualization according to second configuration information defining a configuration of the logical capacity while the first virtualization switch is in operation; Switch
When third configuration information that defines the configuration of the logical capacity is stored, and any one of the first, second, and third configuration information does not match the other two configuration information A management device that changes any one of the configuration information so as to match the other two configuration information;
A virtualization system comprising:   When the first or second configuration information does not exist, the management device generates the same configuration information as the other two existing configuration information, and the first or second virtualization without the configuration information The virtualization system according to claim 1, wherein the generated configuration information is transmitted to a switch. A first virtualization switch connected to a plurality of storages and executing the virtualization according to first configuration information defining configuration information for virtualizing a physical capacity of the plurality of storages to a logical capacity; and A second virtualization switch configured redundantly with the first virtualization switch and executing the virtualization according to second configuration information defining a configuration of the logical capacity while the first virtualization switch is in operation; A virtualization restoration method in a virtualization system having a management device that stores third configuration information that defines the configuration of the logical capacity and manages the first and second configuration information,
The management device requesting acquisition of the first and second configuration information;
When any one of the first, second, and third configuration information does not match the other two configuration information, the management device matches the other two configuration information. Changing any one of the configuration information to:
Storing the changed configuration information in the virtualization switch or management device having configuration information inconsistent with the other two configuration information;
A recovery method characterized by comprising:

Images