Method of file system recovery logging

• US 6,584,582 B1
• Filed: 01/14/2000
• Issued: 06/24/2003
• Est. Priority Date: 01/14/2000
• Status: Active Grant

• Current Assignee: Oracle America, Inc. (Oracle Corporation)
• Original Assignee: Sun Microsystems Incorporated (Oracle Corporation)
• Inventors: OʼConnor, Michael A.
• Primary Examiner(s): Iqbal, Nadeem
• Assistant Examiner(s): Bonura, Timothy M.
• Application Number:
  US09/483,102
• Time in Patent Office: 1,257 Days
• Field of Search: 714/15, 714/18, 714/20, 714/21, 639/47, 639/14, 639/53-5-
• US Class Current: 714/21
• CPC Class Codes: G06F 11/1435 using file system or storage system metadata

Abstract

A recovery logging method wherein when a node in a computer network becomes unavailable, file systems which require verification and are locked are logged in a recovery log and checking of other file systems continues. In this manner, the host node effectively utilizes time which would otherwise be spent waiting for a file system to become available. Upon completing available file system verifications, those file systems which were logged are checked for availability via background processing. When a logged file system becomes available, it is then verified. During the time spent waiting for a logged file system to become available, the affected node is available for other processing.

Claims

1. A method of file system recovery logging by a node, said method comprising:

identifying a first file system for checking responsive to initiating a recovery procedure of said node, wherein said first file system corresponds to a file system which was or is attached to said node;

determining if said first file system requires verification, wherein said determining comprises determining if said first file system corresponds to an unclean system interruption;

logging said first file system, in response to detecting said first file system requires verification and is locked;

verifying said first file system, in response to detecting said first file system requires verification and is unlocked;

checking a log file, wherein said checking comprises a background process;

determining if a logged file system is locked, wherein said determining comprises a background process; and

verifying said logged file system, in response to detecting said logged file system is unlocked.

2. The method of claim 1, further comprising:

disabling a primary intent log file of a node;

substituting a backup intent log file of said node, in response to detecting said node is disabled;

synchronizing said primary intent log file with said backup intent log file;

enabling said primary intent log file of said node.

3. The method of claim 2, wherein said disabling is in response to detecting said node is unavailable.

4. The method of claim 2, wherein said substituting comprises said backup intent log file performing the functions of said primary intent log file.

5. The method of claim 2, wherein said enabling is in response to detecting said node is synchronized and is available.

6. The method of claim 1, wherein said determining further comprises checking a state indicator of said first file system.

7. The method of claim 3, wherein said state indicator is stored on said first file system.

8. The method of claim 3, wherein said state indicator is stored on a second file system.

9. The method of claim 1, wherein said logging writing an identifier to a log file, wherein said identifier identifies said first file system.

10. The method of claim 1, wherein said verifying said first file system comprises:

locking said first file system; and

verifying the integrity of said first file system; and

unlocking said first file system.

11. The method of claim 1, wherein said checking comprises:

reading a log file; and

obtaining a file system identifier, in response to detecting an entry in said log file.

12. The method of claim 1, wherein said verifying said logged file system comprises:

locking said first file system; and

verifying the integrity of said first file system; and

unlocking said first file system; and

removing an identifier from said log file.

13. The method of claim 1, wherein said first file system is a clustered file system.

14. The method of claim 1, wherein said first file system is a log structured file system.

15. The method of claim 1, wherein said node and said first file system are part of a storage area network.

16. The method of claim 1, wherein said node and said first file system are part of a local area network.

17. The method of claim 1, wherein said first file system is a universal file system.

18. The method of claim 11, wherein said first file system is a universal file system.

19. The method as recited in claim 1, wherein said verifying said first file system includes verifying links, block, and other structures of said first file system.

20. The method as recited in claim 19, wherein said verifying includes utilizing an fsck utility.

21. A method of file system recovery logging by a node, said method comprising:

identifying a first file system for checking;

determining if said first file system requires verification;

logging said first file system, in response to detecting said first file system requires verification and is locked;

verifying said first file system, in response to detecting said first file system requires verification and is unlocked, wherein said verifying said first file system includes using an fsck utility;

checking a log file, wherein said checking comprises a background process;

determining if a logged file system is locked, wherein said determining comprises a background process; and

verifying said logged file system, in response to detecting said logged file system is unlocked.

22. A method of file system recovery logging by a node, said method comprising:

identifying a first file system for checking;

determining if said first file system requires verification;

logging said first file system, in response to detecting said first file system requires verification and is locked;

verifying said first file system, in response to detecting said first file system requires verification and is unlocked;

checking a log file, wherein said checking comprises a first process which may run concurrently with a second process on said node, wherein said second process includes input and output via a terminal;

determining if a logged file system is locked, wherein said determining comprises a third process which may run concurrently with a fourth process on said node, wherein said fourth process includes input and output via a terminal; and

verifying said logged file system, in response to detecting said logged file system is unlocked.

23. The method of claim 22, further comprising:

disabling a primary intent log file of a node;

substituting a backup intent log file of said node, in response to detecting said node is disabled;

synchronizing said primary intent log file with said backup intent log file; and

enabling said primary intent log file of said node.

24. A method of file system recovery logging by a node, said method comprising:

performing a first phase of a recovery procedure of said node, wherein said first phase comprises;

identifying one or more file systems which are or were attached to said node prior to initiation of said recovery procedure;

logging a logged file system indication corresponding to a first file system of said file systems, in response to detecting said first file system is dirty and is locked;

sequentially locking, verifying, and unlocking said first file system, in response to detecting said first file system is dirty and is unlocked;

performing a second phase of said recovery procedure in response to determining no dirty and unlocked file system remains among said file systems, and in further response to determining a logged file system list is not empty, wherein said second phase comprises;

periodically checking whether a logged file system is unlocked, wherein said checking is performed as a background process;

sequentially locking, verifying, unlocking, and removing a second file system from said list of logged file systems, in response to detecting said second file system is a logged file system and is unlocked.

25. The method as recited in claim 24, wherein during said checking said node is available for other processing tasks, including accessing file systems not included in said list.

26. A method of file system recovery logging by a node, said method comprising:

identifying a first file system for checking;

determining if said first file system requires verification;

logging said first file system, in response to detecting said first file system requires verification and is locked;

verifying said first file system, in response to detecting said first file system requires verification and is unlocked, wherein said verifying said first file system includes verifying links, block, and other structures of said first file system;

checking a log file, wherein said checking comprises a background process;

determining if a logged file system is locked, wherein said determining comprises a background process; and

verifying said logged file system, in response to detecting said logged file system is unlocked.