Continuously-snapshotted protection of computer files

• US 5,604,862 A
• Filed: 03/14/1995
• Issued: 02/18/1997
• Est. Priority Date: 03/14/1995
• Status: Expired due to Term

• Current Assignee: Autonomy Incorporated (HP Inc.)
• Original Assignee: Network Integrity, Inc.
• Inventors: Gonsalves, Manuel, Holland, Charles J., Midgely, Christopher W., Webb, John W.
• Primary Examiner(s): Beausoliel, Jr., Robert W.
• Assistant Examiner(s): Decady, Albert
• Application Number:
  US08/403,347
• Time in Patent Office: 707 Days
• Field of Search: 395/181, 395/182.03, 395/182.04, 395/182.05, 395/427, 395/444, 395/445, 395/859, 395/488, 395/489, 371/7
• US Class Current: 714/6.1
• CPC Class Codes: G06F 11/1448 Management of the data involved in backup or backup restore
• CPC Class Codes: G06F 11/3476 Data logging G06F11/14, G06F11/2205 take precedence
• CPC Class Codes: G06F 2201/84 Using snapshots, i.e. a logical point-in-time copy of the data

Abstract

An Integrity Server computer for economically protecting the data of a computer network'"'"'s servers, and providing hot standby access to up-to-date copies of the data of a failed server. As the servers'"'"' files are created or modified, they are copied to the Integrity Server. The invention provides novel methods for managing the data stored on the Integrity Server, so that up-to-date snapshots of files of the protected file servers are stored on low-cost media such as tape, but without requiring that a system manager load large numbers of tapes.

Claims

1. A method for managing copies of a protected set of files on a bounded number of sequential-access volumes, the method being executed by computer and comprising:

(a) from among a plurality of said sequential-access volumes, selecting one as the current volume;

(b) when an external process independent of the sequential-access volumes alters the contents of one of the protected files to produce a new current version of the protected file, snapshotting the new current version of the altered protected file at the end of the current volume;

(c) when the current volume is full to a defined limit, selecting a new volume to be the current volume;

(d) maintaining the population of an active set of said sequential-access volumes at or below said bounded number, said active set being the minimum set of the most-recently-current of said volumes that together contain at least one version of each of said protected files, by;

as said population approaches or equals said bounded number, selecting for compaction one volume of said active set, and

copying from the compaction volume to the current volume those versions of file versions stored on the compaction volume not having a more recent version stored on the active set; and

(e) repeating steps (b)-(d) while the external process continues.

2. The method of claim 1, further comprising:

when the contents of one of the protected files is altered to produce a new current version of the protected file, snapshotting the new current version of the altered protected file to a direct access storage cache; and

queueing in a write queue the cache copy of the file for later writing to said current volume.

3. The method of claim 2, further comprising:

when dequeueing a file from said write queue for writing to said current volume, reviewing said write queue for a later version of the dequeued file, and suppressing writing to said current volume of any version other than the latest queued version of said dequeued file.

4. The method of claim 2, further comprising:

copying a file version from the compaction volume to said storage cache; and

queueing the file from the compaction volume in a write queue as the cache copy of the file for later writing to said current volume.

5. The method of claim 4, further comprising:

when dequeueing a file from said write queue for writing to said current volume, reviewing said write queue for a later version of the dequeued file, and suppressing writing of any version other than the latest queued version of said dequeued file.

6. The method of claim 1, wherein:

recently-compacted volumes are maintained as a legacy set of volumes containing additional copies of current versions and non-current versions of protected files, and

storage records corresponding to the file versions stored on the legacy set are retained allowing prompt retrieval of those copies as requested by the external process.

7. The method of claim 6, further comprising:

concurrently with steps (b)-(c), enqueueing versions of said altered files for writing to an archival set of volumes distinct from said active and legacy volumes;

wherein all volumes of the active volume set are resident in an auto-loader, and all file versions of said active volume set can be retrieved with a relatively small latency; and

wherein filled volumes of said archival set are scheduled for removal from said auto-loader, retaining only a single archival volume in said auto-loader.

8. The method of claim 7, wherein

after a file version is written to the current volume, queuing said file version to be written to the current archival volume in an archival queue.

9. The method of claim 8 further comprising:

retaining a file version in said archival queue for a time; and

when dequeueing a file from said archival queue for writing to said current archival volume, reviewing said archival queue for a later version of the dequeued file, and suppressing writing to said current archival volume of any version other than the latest queued version of said dequeued file.

10. The method of claim 7 further comprising:

maintaining at a small number a population of said archival volumes preceding each said archival volume that taken together form a recovery set, a recovery set being a sequence of consecutive ones of said archival volumes that collectively contain at least one version of every file of the protected set, by copying to said current archival volume those file versions copied from the compaction volume to the current volume.

11. The method of claim 10, further comprising:

for each archival volume, recording a number of archival volumes preceding said each archival volume that together form a recovery set ending with said each archival volume.

12. The method of claim 1, wherein:

all volumes of said active set are loaded in a cartridge tape autoloader.

13. The method of claim 1, wherein:

the altering of the file servers'"'"' protected files includes creation of a file by the external process.

27. The method of claim 2, wherein:

said sequential-access volumes are removable mass storage media in a form suitable for an archive for off-site storage, said snapshotting being carried out at a rate similar to the rate at which said files are altered by an external process, a new snapshot of a given file in said storage cache displacing any older snapshot of said given file in existence in said storage cache;

copying, continuously but at a lower rate than said snapshotting that allows a significant proportion of said snapshotted versions to be displaced from said archive storage cache, non-displaced snapshotted versions from said storage cache to said removable mass storage media, said lower rate of copying being much lower than the rate of snapshotting so that a large fraction of said snapshotted versions are displaced and not written to removable mass storage.

28. The method of claim 27, further comprising:

maintaining at a small number of volumes a population of said volumes of said off-site media preceding each said off-site volume that taken together form a recovery set, a recovery set being a sequence of consecutive ones of said off-site volumes that collectively contain at least one version of every file of the protected set, by periodically refreshing said storage cache with copies of protected files not recently altered, and copying said refreshed files from said storage cache to said removable mass storage media.

29. The method of claim 27, further comprising:

periodically expiring media from said off-site archive, leaving short sequences of consecutively-generated volumes of said off-site archive that taken together store at least one copy of every file of said protected set.

30. The method of claim 2, wherein:

the altering of the proctected set of files by the external process includes creation of new files in the protected set by the external process, and said newly-created ones of said protected files are snapshotted to said storage cache.

31. The method of claim 2, further comprising:

recording that a protection process holds the file open during said snapshotting; and

when said external process requests access to a file, consulting said recording to determine whether the file is currently held open by said protection process, and;

if the file is currently held open by said protection process, blocking said external process until said protection process completes snapshotting of the file, and

if the file is not currently held open by said protection process, or when the file is released by the protection process, proceeding to open the file in accord with the file open protocol of said protected computer.

14. A method for protecting a protected set of files of varying size and stored on direct-access mass storage devices of a plurality of file server nodes of a network of computers, the method comprising:

at a rate similar to the rate at which said files are altered by an external process, snapshotting recently-altered protected ones of said files from said direct-access mass storage devices to an archive storage cache, a new snapshot of a given file in said storage cache displacing any older snapshot of said given file in existence in said storage cache;

copying, continuously but at a lower rate than said snapshotting that allows a significant proportion of said snapshotted versions to be displaced from said archive storage cache, non-displaced snapshotted versions from said storage cache to removable mass storage media.

15. The method of claim 14, further comprising:

periodically verifying the contents of the protected files against the contents of the versions stored on said removable mass storage media.

16. The method of claim 15 wherein

avoiding copying contents of verified files over the network, or reading the removable media, during the verifying, by comparing a summary value of the content of a protected file with a summary value of the content of the stored snapshot.

17. The method of claim 14, wherein:

said lower rate is only a little lower, so that a large fraction of said snapshotted versions survive said displacing and are written to said removable mass storage media.

18. The method of claim 17 wherein an active set, being a minimum set of most-recently-written volumes of said media that together contain at least one version of each of said protected files, is maintained at a bounded number of volumes, by further steps comprising:

(a) from among a plurality of said removable mass storage media, selecting one as the current volume;

(b) when the contents of one of the files is altered to produce a new current version of the file, snapshotting the new current version of the altered file to the current volume;

(c) when the current volume is full to a defined limit, selecting a new volume to be the current volume;

(d) maintaining the population of said active set at no greater than said bounded number, by;

as said population approaches or equals said bounded number, selecting for compaction one volume of said active set, and

copying from the compaction volume to the current volume those versions of file versions stored on the compaction volume not having a more recent version stored on the active set; and

(e) repeating steps (b)-(d) while the external process continues.

19. The method of claim 18, further comprising:

from among the volumes previously compacted, periodically expiring some of said volumes, leaving short sequences of consecutively-generated volumes that taken together store at least one copy of every file of said protected set.

20. The method of claim 14, wherein:

said removable mass storage media form an archive suitable for off-site storage, said lower rate of copying being much lower than the rate of snapshotting so that a large fraction of said snapshotted versions are displaced and not written to removable mass storage.

21. The method of claim 20, further comprising:

maintaining at a small number of volumes a population of said volumes of said off-site media preceding each said off-site volume that taken together form a recovery set, a recovery set being a sequence of consecutive ones of said off-site volumes that collectively contain at least one version of every file of the protected set, by periodically refreshing said storage cache with copies of protected files not recently altered, and copying said refreshed files from said storage cache to said removable mass storage media.

22. The method of claim 20, further comprising:

periodically expiring media from said off-site archive, leaving short sequences of consecutively-generated volumes of said off-site archive that taken together store at least one copy of every file of said protected set.

23. The method of claim 14, wherein:

the altering of the file servers'"'"' files by the external process includes creation of new files in the protected set by the external process, and said newly-created ones of said protected files are snapshotted to said archive storage cache.

24. The method of claim 14, further comprising:

recording that a protection process holds the file open during said snapshotting; and

when said external process requests access to a file, consulting said recording to determine whether the file is currently held open by said protection process, and;

if the file is currently held open by said protection process, blocking said external process until said protection process completes snapshotting of the file, and

if the file is not currently held open by said protection process, or when the file is released by the protection process, proceeding to open the file in accord with the file open protocol of said protected computer.

25. In a computer data protection system, a method comprising the steps of:

traversing a file system of a protected computer by a protection process, snapshotting files of said file system to removable storage media, and

as each file is opened for said snapshotting, recording that the protection process currently holds the file open, and

as the protection process completes said snapshotting, recording that the protection process has released the file; and

when a client process requests access to a file, consulting said recording to determine whether the file is currently open by said protection process, and

if the file is currently held open by said protection process, blocking said client process until said protection process releases the file, and

if the file is not currently held open by said protection process, or when the protection process completes said snapshotting, proceeding to open the file in accord with the file open protocol of said protected computer.

26. The method of claim 25, wherein:

said protected computer is one of a network of computers, and

said client process is hosted on a client computer of said network other than said protected computer.