Managing cache backup and restore for continuous data replication and protection

• US 9,317,375 B1
• Filed: 03/30/2012
• Issued: 04/19/2016
• Est. Priority Date: 03/30/2012
• Status: Active Grant

• Current Assignee: LenovoEMC Limited (Lenovo Group Ltd.)
• Original Assignee: LenovoEMC Limited (Lenovo Group Ltd.)
• Inventors: Gruttadauria, Brian R., Kalidindi, Suresh Kumar, Sadhu, Vamsikrishna
• Primary Examiner(s): Queler, Adam M
• Assistant Examiner(s): Ta, Trang
• Application Number:
  US13/435,295
• Time in Patent Office: 1,481 Days
• Field of Search: 707/649, 711/162, 711/103
• US Class Current: 1/1
• CPC Class Codes: G06F 11/1451 by selection of backup contents

Abstract

A method is used for managing cache backup and restore for continuous data replication and protection. I/O operations are quiesced at a cache module. A first snapshot of a storage object and a second snapshot of an SSD cache object are taken. The I/O operations at the cache module are unquiesced. A single backup image comprising the first snapshot and the second snapshot is created. The single backup image is sent to a first data protection appliance (DPA) and recorded in a journal.

Claims

1. A method, comprising:

quiescing I/O operations at a cache;

taking a first snapshot of a storage object and a second snapshot of a cache object;

unquiescing I/O operations at the cache;

creating, at least one backup image comprising the first snapshot and the second snapshot; and

storing the backup image;

wherein creating the backup image comprises;

calculating a backup image size;

writing, to the backup image, a backup storage object data blocks from the storage object;

updating the backup header with storage object metadata and cache object metadata;

writing the backup header to the backup image;

reading, from the cache object, cache superblock and dirty region logging (DRL) metadata and reading, from the cache object, cache data blocks determined to have a DRL state that is dirty; and

writing, to the backup image, the cache superblock and dirty region logging (DRL) metadata and the cache data blocks determined to have a DRL state that is dirty.

2. The method of claim 1, further comprising:

reading data from the backup image; and

restoring the cache object from the backup image at least in part by loading the cache object with data present in the cache object when the backup image was created.

3. The method of claim 1, wherein creating the backup image further comprises:

performing a dump of metadata of clean blocks of the cache.

4. The method of claim 1, wherein the method comprises:

creating the backup image within a threshold time of unquiescing I/O operations at the cache.

5. A system, comprising:

a processor; and

storage accessible to the processor and bearing instructions executable by the processor to;

quiesce I/O operations at a cache;

take a first snapshot of at least a portion of data in storage and a second snapshot of at least a portion of data in the cache;

unquiesce I/O operations at the cache;

create, within a threshold time of unquiescing I/O operations at the cache, at least one backup image comprising the first snapshot and the second snapshot; and

store the backup image;

wherein creation of the backup image comprises;

writing, to the backup image, a backup of at least one storage data block from the storage;

including, in a backup header, storage metadata;

writing, to the backup image, the backup header;

reading, from the cache, at least one cache data block determined to have a dirty region logging (DRL) start that is dirty; and

writing, to the backup image, the at least one cache data block determined to have a DRL state that is dirty.

6. The system of claim 5, wherein the instructions are executable to:

read data from the backup image; and

restore the portion of data from the cache from the backup image at least in part by loading the cache with data resent in the cache when the backup image was created.

7. The system of claim 5, wherein creation of the backup image further comprises:

creating the backup header.

8. The system of claim 5, wherein creation of the backup image further comprises:

including, in the backup header, cache metadata.

9. The system of claim 5, wherein creation of the backup image further comprises:

calculating a backup image size for the backup image.

10. The system of claim 5, wherein creation of the backup image further comprises:

reading cache superblock and DRL metadata from the cache; and

writing, to the backup image, the cache superblock and DRL metadata.

11. The system of claim 5, wherein creation of the backup image further comprises:

performing a dump of metadata of clean blocks of the cache.

12. The system of claim 5, wherein creation of the backup image further comprises:

writing DRL metadata for at least one block of the cache to the end of the cache data blocks in the backup image.

13. The system of claim 5, wherein creation of the backup image further comprises:

writing a copy of the header to the storage.

14. The system of claim 5, wherein creation of the backup image further comprises:

determining a backup type for creating the backup image;

in response to determining that an incremental backup is the backup type, writing to the backup image a backup of at least one dirty storage data block from the storage using a storage DRL bitmap; and

in response to determining that a full backup is the backup type, writing to the backup image a full backup of at least one storage data block from the storage.

15. The system of claim 5, wherein creation of the backup image further comprises:

writing, to the backup image, the at least one cache data block determined to have a DRL state that is dirty by overwriting at least one respectively matching storage block in the backup image with the at least one cache data block determined to have a DRL state that is dirty.

16. A computer readable storage medium that is not a transitory signal, the computer readable storage medium comprising instructions executable by a processor to:

quiesce I/O operations at a cache;

take a first snapshot of at least a portion of a storage area and a second snapshot of at least a portion of the cache;

unquiesce I/O operations at the cache;

create, within a threshold time of unquiescing I/O operations at the cache, at least one backup image comprising the first snapshot and the second snapshot; and

store the backup image;

wherein creation of the backup image comprises;

writing, to the backup image, a backup of at least one storage data block from the storage area;

including, in a backup header, storage area metadata;

writing, in the backup image, the backup header;

reading, from the cache, at least one cache data block determined to have a dirty region logging (DRL) state that is dirty; and

writing, to the backup image, the at least one cache data block determined to have a DRL state that is dirty.

17. The computer readable storage medium of claim 16, wherein creation of the backup image further comprises:

writing a copy of the header to the storage area.

18. The computer readable storage medium of claim 16, wherein the instructions are executable to:

read data from the backup image; and

restore the portion of the cache from the backup image at least in part by loading the cache with data present in the cache when the backup image was created.

19. The computer readable storage medium of claim 16, wherein creation of the backup image further comprises:

determining a backup type for creating the backup image;

in response to determining that an incremental backup is the backup type, writing to the backup image a backup of at least one dirty storage data block from the storage area using a storage area DRL bitmap; and

in response to determining that a full backup is the backup type, writing to the backup image a full backup of at least one storage data block from the storage area.

20. The computer readable storage medium of claim 16, wherein creation of the backup image further comprises:

writing, to the backup image, the at least one cache data block determined to have a DRL state that is dirty by overwriting at least one respectively matching storage block in the backup image with the at least one cache data block determined to have a DRL state that is dirty.