IP Library Granted Patent US 11,675,545
Granted Patent B2
US 11,675,545 · App. 17/474,562 · Granted Jun 13, 2023

Distributed storage system and storage control method

Inventors: Shugo Ogawa (Tokyo, JP); Akira Yamamoto (Tokyo, JP); Yoshinori Ohira (Tokyo, JP); Ryosuke Tatsumi (Tokyo, JP); Junji Ogawa (Tokyo, JP); Hiroto Ebara (Tokyo, JP)
Assignee: Hitachi, Ltd.
G06F3/067G06F3/0604G06F3/0647G06F3/0655
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 11,675,545
App. No.
17/474,562
Granted
Jun 13, 2023
Kind
B2
Abstract

One or a plurality of physical storage devices that provide a physical storage area are connected to first and second computers. The computer updates metadata indicating the address correspondence relationship between the logical address of the volume and the physical address of the physical storage area in the write processing performed based on a write request designating the volume. The first computer copies the metadata to the second computer while receiving the write request. When the address correspondence relationship indicated by the copied metadata portion is changed during copying of the metadata, the first computer updates the metadata portion and copies the metadata portion to the second computer.

Claims (46)

1. A distributed storage system, comprising:

one or a plurality of physical storage devices; and

a plurality of computers including first and second computers and connected to the one or plurality of physical storage devices, wherein

the plurality of computers provide a volume that is a logical address space, and the volume is migratable between the plurality of computers,

the one or plurality of physical storage devices provide a physical storage area that is a physical address space,

the plurality of computers are configured to perform write processing based on a write request designating the volume, wherein the write processing includes:

allocating a physical address not allocated to any logical address to a logical address specified from the write request, and

writing data accompanying the write request to a physical storage area to which the allocated physical address belongs,

the allocating of a physical address to a logical address includes writing an address correspondence relationship between the logical address and the physical address in metadata included in the first computer,

the plurality of computers are configured to perform a migration while the write processing is being performed, wherein the migration includes:

copying from the first computer the metadata to the second computer

updating an acquisition address pointer corresponding to a progress of the copying step, wherein an acquisition address pointer is associated with a second logical address associated with the volume;

determining whether an address correspondence relationship indicated by a copied metadata portion is changed in the first computer during copying of the metadata, by comparing the acquisition address pointer with the logical address specified from the write request;

updating the metadata portion to a metadata portion indicating a changed address correspondence relationship, and copies the updated metadata portion to the second computer.

2. The distributed storage system according to claim 1 , wherein

the first computer copies a first portion of metadata indicating a correspondence relationship between a logical address and a physical address to the second computer in a pre-determined order regarding the logical address, and

when allocation of a physical address to a logical address of the volume occurs in the first computer during copying of the first portion of metadata, the first computer compares a logical address in the allocation with the acquisition address pointer, and determines whether the logical address in the allocation is a logical address corresponding to the first portion of metadata based on a result of the comparison.

3. The distributed storage system according to claim 1 , wherein

in the write processing,

the writing of data is writing of data in a log structure including compression of data,

the physical address allocated to the logical address specified from the write request is a physical address next to a last physical address among the physical addresses allocated to any of the logical addresses in a pre-determined order, and

the plurality of computers defines a unit of writing as a size of the volume area to which the logical address belongs.

4. The distributed storage system according to claim 3 , wherein

in the write processing, when the logical address specified from the write request is a logical address to which a physical address is allocated, another physical address is allocated to the logical address instead of the allocated physical address, and the allocated physical address becomes an invalid physical address,

the first computer performs garbage collection for releasing an invalid physical address, and

the garbage collection includes allocating a released physical address to the logical address instead of the physical address allocated to the logical address to fill the allocated physical address, and

the first computer does not perform the garbage collection on the volume from when the copy of the metadata is started to when the copy of the metadata is completed.

5. The distributed storage system according to claim 1 , wherein

the first and second computers are connected to a host that issues a write request,

the first computer has a control right for the volume, the control right being an authority to control the volume or access the volume, and

after all of the metadata is copied to the second computer, the control right is transferred from the first computer to the second computer, and an access destination from the host to the volume is changed to the second computer.

6. The distributed storage system according to claim 1 , wherein

the first computer performs background processing in which allocation of a physical address to a logical address is changed by migrating data between storage areas in the physical storage area regardless of whether the write request is received, and

the second computer does not perform the background processing on the volume from start to completion of copying of the metadata.

7. A volume migration method in a storage system that includes a plurality of computers including first and second computers and connected to one or plurality of physical storage devices, wherein

the computer provides a volume that is a logical address space, and the volume is migratable between the computers,

the one or plurality of physical storage devices provide a physical storage area that is a physical address space,

the computer performs write processing based on a write request designating the volume,

the write processing includes allocating a physical address not allocated to any logical address to a logical address specified from the write request, and writing data accompanying the write request to a physical storage area to which the allocated physical address belongs, and

the allocating of a physical address to a logical address includes writing an address correspondence relationship between the logical address and the physical address in metadata included in the computer,

the volume migration method, comprising:

in the migration of the volume from the first computer to the second computer,

copying, by the first computer, the metadata to the second computer while receiving a write request designating the volume;

updating an acquisition address pointer corresponding to a progress of the copying step, wherein an acquisition address pointer is associated with a second logical address associated with the volume;

determining whether an address correspondence relationship indicated by a copied metadata portion is changed in the first computer during copying of the metadata, by comparing the acquisition address pointer with the logical address specified from the write request;

updating, by the first computer, the metadata portion to a metadata portion indicating a changed address correspondence relationship, and copying the updated metadata portion to the second computer.

Assignments (2)
COMPANY SPLIT Recorded Aug 20, 2024
From: HITACHI, LTD.
To: HITACHI VANTARA, LTD.
Reel/Frame 069518/0761 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Sep 14, 2021
From: OGAWA, SHUGO; YAMAMOTO, AKIRA; OHIRA, YOSHINORI; TATSUMI, RYOSUKE; OGAWA, JUNJI; EBARA, HIROTO
To: HITACHI, LTD.
Reel/Frame 057477/0283 →
Priority Claims (1)
JP JP2021-068624 · Apr 14, 2021 · national
Continuity (1)
Related Publication 20220334726A1 · Oct 20, 2022