Avoiding frozen-volume write penalties

• US 7,783,598 B1
• Filed: 04/27/2007
• Issued: 08/24/2010
• Est. Priority Date: 04/27/2007
• Status: Active Grant

• Current Assignee: Network Appliance Incorporated (NetApp, Inc.)
• Original Assignee: Network Appliance Incorporated (NetApp, Inc.)
• Inventors: Malaiyandi, Prasanna K., Grunwald, David A., Wu, Stephen
• Primary Examiner(s): Wong; Don
• Assistant Examiner(s): Hwa; Shyue Jiunn
• Application Number:
  US11/741,707
• Time in Patent Office: 1,215 Days
• Field of Search: None
• US Class Current: 707/610
• CPC Class Codes: G06F 11/2064 while ensuring consistency
• CPC Class Codes: G06F 16/184 implemented as replicated file system

Abstract

Methods of operating a computer storage system where an identical copy of a mirror source filesystem is maintained on a mirror destination storage volume managed by a mirror destination server. A modification to the mirror source filesystem is planned and executed, and a corresponding change to the mirror destination filesystem is accomplished by storing data directly to the mirror destination storage volume, bypassing filesystem logic that interprets the data stored on the volume. Systems that implement these methods are also described.

Claims

1. A method comprising:

maintaining a mirror destination filesystem that is an identical copy of a mirror source filesystem, the mirror destination file system stored on a mirror destination storage volume;

freezing the mirror destination filesystem on the mirror destination storage volume to prevent modifications to the mirror destination filesystem;

receiving data describing a change to the mirror source filesystem;

writing, by a computer, the data directly to an unused block on the mirror destination storage volume by bypassing logic that interprets the mirror destination filesystem on the mirror destination storage volume, wherein the writing occurs while the mirror destination filesystem is frozen;

unfreezing the mirror destination file system; and

merging the change to the mirror destination filesystem into an active version of the mirror destination filesystem after unfreezing the mirror destination file system.

2. The method of claim 1 wherein said storing affects a block of the mirror destination storage volume that is marked unused within a logical structure of the mirror destination filesystem.

3. The method of claim 1 wherein the mirror destination storage volume is a container file.

4. The method of claim 1 wherein the mirror destination storage volume is one of a disk or a Redundant Array of Independent Disks (“

RAID array”

).

5. The method of claim 1 wherein the mirror source filesystem and mirror destination filesystem are compatible with one of a New Technology File System (“

NTFS”

) structure, a File Allocation Table (“

FAT”

) structure, a VFAT structure, or a Hierarchical File System (“

HFS”

) structure.

6. A method comprising:

identifying a block of a first mass storage device that is unused by a first filesystem stored on the first mass storage device;

freezing a second filesystem stored on a second mass storage device in a consistent state;

disallowing modifications to the second filesystem while the second filesystem is frozen;

storing, by a computer, new data in a corresponding unused block of the second mass storage device by writing directly to the second mass storage device, in a format compatible with a structure of the second filesystem, wherein the identifying and storing occur while the second filesystem is frozen;

unfreezing the second file system; and

updating a filesystem logic state so that the data merges into the second filesystem, wherein the updating occurs after unfreezing the second file system.

7. The method of claim 6 wherein the second mass storage device is one of a disk drive, a Redundant Array of Independent Disks (“

RAID array”

), or a container file.

8. The method of claim 6, further comprising:

marking the block as “

in-use”

within a data structure stored on the first mass storage device.

9. The method of claim 6 wherein a filesystem stored on the second mass storage device is an identical copy of the filesystem stored on the first mass storage device.

10. A system comprising:

a mass storage device;

filesystem logic to maintain a filesystem structure within data storage space of the mass storage device;

frozen-volume write logic to write data on an unused portion of the mass storage device when the filesystem logic has disabled modifications to the filesystem and to bypass the filesystem logic;

a communication interface to receive a request from a client, wherein the request is to modify a file in the filesystem while the filesystem logic has disabled modifications to the filesystem;

unfreezing logic to enable modifications to the filesystem; and

merging logic to merge the data into an active version of the filesystem after enabling modifications to the filesystem.

11. The system of claim 10, further comprising:

mirror logic to cooperate with the filesystem logic and maintain an identical copy of a mirror source filesystem within the data storage space of the mass storage device.

12. The system of claim 10, further comprising:

container-file filesystem logic to provide data storage in a container file, wherein the filesystem logic is to maintain the filesystem structure in the container file.

13. The system of claim 10, further comprising:

memory to store a cached copy of a root inode of the filesystem structure.

14. The system of claim 10, further comprising:

a Redundant Array of Independent Disks (“

RAID array”

) interface to manage a plurality of mass storage devices as a RAID array.

15. A machine-readable storage medium storing data and instructions to cause a programmable processor to perform operations comprising:

planning a change to a mirror source filesystem, said change to affect unused blocks of the mirror source filesystem;

preparing data to effect the change;

storing the data on a first mass storage subsystem;

freezing a mirror destination filesystem stored on a mirror destination server in a consistent state;

disallowing changes to the mirror destination filesystem while the mirror destination filesystem is frozen;

transmitting the data to the mirror destination server to be written on an unused block in a second mass storage subsystem of said mirror destination server, wherein the storing operation occurs while the mirror destination filesystem is frozen;

unfreezing the mirror destination filesystem;

merging the change into an active version of the mirror destination filesystem after unfreezing the mirror destination filesystem.

16. The machine-readable storage medium of claim 15 wherein the mass storage device is one of a disk drive, a Redundant Array of Independent Disks (“

RAID array”

), or a container file.

17. The machine-readable storage medium of claim 15, containing additional data and instructions to cause the programmable processor to perform operations comprising:

marking the unused blocks as “

in-use”

within a data structure stored on the mass storage device.