Computer system and data migration method

• US 9,213,490 B2
• Filed: 06/18/2012
• Issued: 12/15/2015
• Est. Priority Date: 06/18/2012
• Status: Active Grant

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Jindo, Kazue, Sakaki, Hidetoshi, Muto, Junichi, Hiramatsu, Shinichi, Miyashita, Jun
• Primary Examiner(s): Rones, Charles
• Assistant Examiner(s): DOAN, HAN V
• Application Number:
  US13/519,272
• Publication Number:
  US 20130339647A1
• Time in Patent Office: 1,275 Days
• US Class Current: 1/1
• CPC Class Codes: G06F 11/1443 Transmit or communication errors
• CPC Class Codes: G06F 2201/82 Solving problems relating to consistency ensuring consistency in mirrored systems G06F11/2064
• CPC Class Codes: G06F 3/0617 in relation to availability
• CPC Class Codes: G06F 3/0635 by changing the path, e.g. traffic rerouting, path reconfiguration
• CPC Class Codes: G06F 3/0647 Migration mechanisms
• CPC Class Codes: G06F 3/067 Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]

Abstract

The present invention continues host access and holds data of a migration source volume as the latest data even when a fault occurs in any of a plurality of access paths.

Upon receiving a first write command from a host computer via a first access path, the migration source controller writes first update data to the migration source volume and, upon receiving a second write command from the host computer via a second access path, the migration destination controller transfers second update data to the migration source controller via a storage path and, upon receiving the second update data from the migration destination controller, the migration source controller writes the second update data to the migration source volume.

Claims

1. A computer system, comprising:

a plurality of storage apparatuses which manage a plurality of storage devices; and

a host computer which exchanges information with each of the storage apparatuses,

wherein one storage apparatus among the plurality of storage apparatuses is managed as a migration source storage apparatus, another storage apparatus is managed as a migration destination storage apparatus, a logical volume which is constructed in a storage area of a storage device of the migration source storage apparatus is managed as a migration source volume, a logical volume which is constructed in a storage area of a storage device of the migration destination storage apparatus is managed as a migration destination volume, and the migration source volume and the migration destination volume are interconnected via a storage path,

wherein the migration source storage apparatus comprises a migration source controller which manages the migration source volume as an access target of the host computer,

wherein the migration destination storage apparatus comprises a migration destination controller which virtualizes the migration destination volume as the migration source volume and manages the virtualized migration destination volume as an access target of the host computer,

wherein, in a case where a first access path linking the host computer to the migration source volume and a second access path linking the host computer to the migration destination volume are each online and where the migration destination volume is virtualized as the migration source volume, the host computer manages each of the first access path and the second access path as alternate paths for accessing the migration source volume,

wherein, in a case where a first write command is received by the migration source controller from the host computer via the first access path during a period up until the first access path is offline after formation copy processing to migrate data of the migration source volume to the migration destination volume via the storage path has ended, the migration source controller writes first update data appended to the first write command to the migration source volume and executes an update copy to copy the first update data stored in the migration source volume to the migration destination volume, and

wherein, in a case where a second write command is received by the migration destination controller from the host computer via the second access path during the period up until the first access path is offline after formation copy processing has ended, the migration destination controller determines whether the second write command indicates a processing mode to be one of;

a host frame transfer mode;

a duplex transfer mode; and

a simplex transfer mode,

upon determining the processing mode to be the host frame transfer mode, the migration destination controller transfers the second write command to the source destination controller without processing a frame of the second write command and the source destination controller stores second update data, appended to the second write command, in the migration source volume and sends an update completion notification to the migration destination controller that the second update data has been stored in the migration source volume,

upon determining the processing mode to be the duplex transfer mode, the migration destination controller processes the frame of the second write command and transfers the second update data appended to the second write command to the migration source controller via the storage path, as a first update copy command, and responsive to receiving the first update copy command, the migration source controller stores the second update data stored in the migration source volume and transfers a copy command to the migration destination controller to copy, in the migration destination volume, the second update data stored in the migration source volume, the migration destination controller writes the second update data to the migration destination volume, and sends a copy completion notification to the migration source controller and upon receiving the copy completion notification, the migration source controller sends a command completion notification prompting the migration destination controller to notify the host that the second write command is completed, and upon determining the processing mode to be the simplex transfer mode, the migration destination controller processes the frame of the second write command and transfers the second write command to the migration source controller as a second update copy command, the migration destination controller executes update copy processing to copy the second update data to the migration source volume and holds the update data in the migration destination volume, and upon receiving the second update copy command, the migration source controller stores the second update data in the migration source volume and sends a command completion notification to the migration destination controller which stores the second update data held in the migration destination volume and issues completion notification to the host that the second write command is completed.

2. The computer system according to claim 1, wherein, upon receiving the first write command, the migration source controller transfers first update data which is appended to the first write command to the migration destination controller via the storage path, wherein, upon receiving the first update data from the migration source controller, the migration destination controller writes the received first update data to the migration destination volume.

3. The computer system according to claim 1, wherein, upon receiving the first write command, the migration source controller transfers the first update data appended to the first write command to the migration destination controller via the storage path, wherein, upon receiving the first update data from the migration source controller, the migration destination controller writes the received first update data to the migration destination volume.

4. The computer system according to claim 1, wherein, if the first access path, the second access path, and the storage path are each formed in a plurality, the host computer accesses the migration destination volume via any second access path of the plurality of second access paths and, under the condition that an abnormal response to the effect that a path fault has occurred in the plurality of storage paths is received from the migration destination controller, the host computer accesses the migration source volume via any first access path among the plurality of first access paths.

5. The computer system according to claim 1, wherein, if the first access path, the second access path, and the storage path are each formed in a plurality, the host computer accesses the migration destination volume via any second access path of the plurality of second access paths under the condition that a fault in the plurality of first access paths is detected.

6. The computer system according to claim 1, wherein, if the first access path, the second access path, and the storage path are each formed in a plurality, the host computer accesses the migration source volume via any first access path of the plurality of first access paths under the condition that a fault in the plurality of second access paths is detected.

7. A data migration method for a computer system which comprises:

a plurality of storage apparatuses which manage a plurality of storage devices; and

a host computer which exchanges information with each of the storage apparatuses, wherein one storage apparatus among the plurality of storage apparatuses is managed as a migration source storage apparatus, another storage apparatus is managed as a migration destination storage apparatus, a logical volume which is constructed in a storage area of a storage device of the migration source storage apparatus is managed as a migration source volume, a logical volume which is constructed in a storage area of a storage device of the migration destination storage apparatus is managed as a migration destination volume, and the migration source volume and the migration destination volume are interconnected via a storage path,

wherein the migration source storage apparatus comprises a migration source controller which manages the migration source volume as an access target of the host computer,

wherein the migration destination storage apparatus comprises a migration destination controller which virtualizes the migration destination volume as the migration source volume and manages the virtualized migration destination volume as an access target of the host computer,

wherein, in a case where a first access path linking the host computer to the migration source volume and a second access path linking the host computer to the migration destination volume are each online and where the migration destination volume is virtualized as the migration source volume, the host computer manages each of the first access path and the second access path as alternate paths for accessing the migration source volume,

the data migration method comprising;

a step in which, in a case where a first write command is received by the migration source controller from the host computer via the first access path during a period up until the first access path is offline after formation copy processing to migrate data of the migration source volume to the migration destination volume via the storage path has ended, the migration source controller writes first update data appended to the first write command to the migration source volume and executes an update copy to copy the first update data stored in the migration source volume to the migration destination volume;

a step in which, in a case where a second write command is received by the migration destination controller from the host computer via the second access path during the period up until the first access path is offline after formation copy processing has ended, the migration destination controller determines whether the second write command indicates a processing mode to be one of;

a host frame transfer mode;

a duplex transfer mode; and

a simplex transfer mode,

upon determining the processing mode to be the host frame transfer mode, the migration destination controller transfers the second write command to the source destination controller without processing a frame of the second write command and the source destination controller stores second update data, appended to the second write command, in the migration source volume and sends an update completion notification to the migration destination controller that the second update data has been stored in the migration source volume,

upon determining the processing mode to be the duplex transfer mode, the migration destination controller processes the frame of the second write command and transfers second update data appended to the second write command to the migration source controller via the storage path

as a first update copy command, and responsive to receiving the first update copy command, the migration source controller stores the second update data stored in the migration source volume, and transfers a copy command to the migration destination controller to copy, in the migration destination volume, the second update data stored in the migration source volume, the migration destination controller writes the second update data to the migration destination volume and sends a copy completion notification to the migration source controller and upon receiving the copy completion notification, the migration source controller sends a command completion notification prompting the migration destination controller to notify the host that the second write command is completed; and

a step in which upon determining the processing mode to be the simplex transfer mode, the migration destination controller processes the frame of the second write command and transfers the second write command to the migration source controller as a second update copy command, the migration destination controller executes update copy processing to copy the second update data to the migration source volume and holds the update data in the migration destination volume, and upon receiving the second update copy command, the migration source controller stores the second update data in the migration source volume and sends a command completion notification to the migration destination controller which stores the second update data held in the migration destination volume and issues completion notification to the host that the second write command is completed.

8. The data migration method according to claim 7, comprising:

a step in which, upon receiving the first write command, the migration source controller transfers first update data which is appended to the first write command to the migration destination controller via the storage path;

a step in which, upon receiving the first update data from the migration source controller, the migration destination controller writes the received first update data to the migration destination volume.

9. The data migration method according to claim 7, comprising:

a step in which, upon receiving the first write command, the migration source controller transfers the first update data appended to the first write command to the migration destination controller via the storage path;

a step in which, upon receiving the first update data from the migration source controller, the migration destination controller writes the received first update data to the migration destination volume.