Adjusting timestamps to preserve update timing information for cached data objects

• US 6,842,825 B2
• Filed: 08/07/2002
• Issued: 01/11/2005
• Est. Priority Date: 08/07/2002
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Nick, Jeffrey Mark, Geiner, Robert Vaughn, Kern, Ronald Maynard, Micka, William Frank, Shackelford, David Michael, McBride, Gregory Edward, Yudenfriend, Harry Morris, Petersen, David B., Zobre, Douglas M., Perry, Larry Richard, Slone, Harold Glen
• Primary Examiner(s): Padmanabhan, Mano
• Assistant Examiner(s): Song, Jasmine
• Application Number:
  US10/214,842
• Publication Number:
  US 20040030837A1
• Time in Patent Office: 888 Days
• Field of Search: 711/100, 711/118, 711/112, 711/133, 711/141, 711/161, 711/162, 707/201, 707/204, 714/6, 713/400, 713/501, 713/502
• US Class Current: 711/133
• CPC Class Codes: G06F 11/2064 while ensuring consistency
• CPC Class Codes: G06F 11/2074 Asynchronous techniques
• CPC Class Codes: G06F 11/2076 Synchronous techniques
• CPC Class Codes: G06F 12/0866 for peripheral storage systems, e.g. disk cache
• CPC Class Codes: Y10S 707/99952 Coherency, e.g. same view to multiple users
• CPC Class Codes: Y10S 707/99953 Recoverability

Abstract

In a system including a host, a primary storage subsystem coupled to the host, a cache coupled to the host and separate from the primary storage system, a secondary storage subsystem, and a data mover coupling the primary and secondary storage systems, data is temporarily cached for future storage in the primary storage subsystem so as to preserve timestamp information and maintain data consistency for asynchronously mirroring the data at a secondary subsystem.

Claims

1. A method for using timestamps to manage data, the method comprising the following operations:

storing some data objects in a cache and other non-cached data objects in a primary storage;

for each data object stored in the primary storage, writing a current timestamp to the primary storage representing a time that the data object was stored in the primary storage;

for each data object stored in the cache, placing in the cache an original timestamp corresponding to the data object, the original timestamp representing a time that the data object was stored in the cache;

writing data objects from the cache and corresponding advanced timestamps to the primary storage, where each advanced timestamp is equal to an original timestamp plus a prescribed time delta;

responsive to requests to apply a predetermined consistency group formation routine to identified data objects in the primary storage, for each data object, reading the advanced timestamp corresponding to the identified data object and subtracting the time delta therefrom to compute a restored timestamp, and then applying the routine to the identified data object in conjunction with the restored timestamp instead of the advanced timestamp.

2. The method of claim 1, further comprising repeating the writing operation at time intervals that do not exceed a prescribed maximum time interval.

3. The method of claim 1, the operations further comprising:

at least as often as a prescribed maximum time interval, determining if the cache contains any additional data objects that have not been written to the primary storage;

if so, writing the additional data objects from the cache and corresponding advanced timestamps to the primary storage, where each advanced timestamp is equal to an original timestamp plus a prescribed time delta;

if not, writing an empty record and a corresponding advanced timestamp to the primary storage and a sidefile, where the advanced timestamp is at least as large as the time delta.

4. The method of claim 1, where the operation of writing the data object and the corresponding advanced timestamp to the primary storage comprises:

invoking an input/output routine configured to write data objects to the primary storage along with a normal timestamp indicating the time the input/output routine is executed to write the data object to the primary storage, and passing the input/output routine a command instructing the input/output routine to write the corresponding advanced timestamp to the primary storage instead of the normal timestamp.

5. The method of claim 1, where the operations further comprise:

writing data objects and their corresponding advanced timestamps from the cache to a sidefile of the primary storage;

writing the non-cached data objects and their corresponding current timestamps to a sidefile of the primary storage;

marking a given set of one or more sidefiles as idle if (1) no data objects have been stored to the given set during a prescribed group time interval, and (2) no timestamp corresponding to a data object in given set is an advanced timestamp.

6. A method of temporarily caching data prior to storage in a primary storage subsystem for asynchronous mirroring at a secondary storage subsystem without sacrificing timestamp information, the method performed in a system including a host, the primary storage subsystem coupled to the host, a cache coupled to the host and separate from the primary storage subsystem, a secondary storage subsystem, and a data mover coupling the primary and secondary storage subsystems, the method comprising operations of:

the host identifying various data objects to store in the primary storage subsystem;

the host storing some of the identified data objects in the cache in association with original timestamps correlated to times of storage in the cache;

the host sending other of the identified data objects to the primary storage subsystem for storage therein, and the primary storage subsystem responding by storing the other data objects in association with current timestamps correlated to times of storage in the primary storage subsystem;

according to a predetermined schedule, the host sending cached data objects to the primary storage subsystem for storage therein and also including commands for the primary storage subsystem to store each data object in association with an advanced timestamp comprising the original timestamp associated with the data object increased by a predetermined margin, the predetermined margin being of sufficient amount so as to be recognizable as an advanced timestamp and an amount permitting derivation of the original timestamp from the advanced timestamp.

7. The method of claim 6, the operations further comprising:

responsive to the data mover commencing a prescribed process of preparing data objects in the primary storage subsystem for storage in the secondary storage subsystem, where the process is applied to a designated set of one or more data objects in the primary storage subsystem, performing operations comprising;

reviewing each data object in the set to determine whether its associated timestamp is an advanced timestamp or current timestamp;

if advanced, subtracting the predetermined margin from the advanced timestamp to restore the original timestamp, and conducting future processing of the data object utilizing the restored original timestamp;

if current, conducting future processing of the data object utilizing the current timestamp.

8. The method of claim 6, the operations further comprising:

the primary storage subsystem storing data objects in various sidefiles;

processing the sidefiles in order to form consistency groups, where the processing includes marking each sidefile as IDLE under conditions including the sidefile having at least one data object associated with an advanced timestamp.

9. At least one signal-bearing medium tangibly embodying a program of machine-readable instructions executable by at least one digital processing apparatus to perform a method for using timestamps to manage data, the method comprising the following operations:

storing some data objects in a cache and other, non-cached data objects in a primary storage;

for each data object stored in the primary storage, writing a current timestamp to the primary storage representing a time that the data object was stored in the primary storage;

for each data object stored in the cache, placing in the cache an original timestamp corresponding to the data object, the original timestamp representing a time that the data object was stored in the cache;

writing data objects from the cache and corresponding advanced timestamps to the primary storage, where each advanced timestamp is equal to an original timestamp plus a prescribed time delta;

responsive to requests to apply a predetermined consistency group formation routine to identified data objects in the primary storage, for each data object, reading the advanced timestamp corresponding to the identified data object and subtracting the time delta therefrom to compute a restored timestamp, and then applying the routine to the identified data object in conjunction with the restored timestamp instead of the advanced timestamp.

10. The medium of claim 9, further comprising repeating the writing operation at time intervals that do not exceed a prescribed maximum time interval.

11. The medium of claim 9, the operations further comprising:

at least as often as a prescribed maximum time interval, determining if the cache contains any additional data objects that have not been written to the primary storage;

if so, writing the additional data objects from the cache and corresponding advanced timestamps to the primary storage, where each advanced timestamp is equal to an original timestamp plus a prescribed time delta;

if not, writing an empty record and a corresponding advanced timestamp to the primary storage and a sidefile, where the advanced timestamp is at least as large as the time delta.

12. The medium of claim 9, where the operation of writing the data object and the corresponding advanced timestamp to the primary storage comprises:

invoking an input/output routine configured to write data objects to the primary storage along with a normal timestamp indicating the time the input/output routine is executed to write the data object to the primary storage, and passing the input/output routine a command instructing the input/output routine to write the corresponding advanced timestamp to the primary storage instead of the normal timestamp.

13. The medium of claim 9, where the operations further comprise:

writing data objects and their corresponding advanced timestamps from the cache to a sidefile of the primary storage;

writing the non-cached data objects and their corresponding current timestamps to a sidefile of the primary storage;

marking a given set of one or more sidefiles as idle if (1) no data objects have been stored to the given set during a prescribed group time interval, and (2) no timestamp corresponding to a data object in given set is an advanced timestamp.

14. At least one signal-bearing medium tangibly embodying a program of machine-readable instructions executable by at least one digital processing apparatus to perform operations of caching data prior to storage in a primary storage subsystem for asynchronous mirroring at a secondary storage subsystem without sacrificing timestamp information, the operations performed in a system including a host, the primary storage subsystem coupled to the host, a cache coupled to the host and separate from the primary storage subsystem, a secondary storage subsystem, and a data mover coupling the primary and secondary storage subsystems, the operations comprising:

the host identifying various data objects to store in the primary storage subsystem;

the host storing some of the identified data objects in the cache in association with original timestamps correlated to times of storage in the cache;

the host sending other of the identified data objects to the primary storage subsystem for storage therein, and the primary storage subsystem responding by storing the other data objects in association with current timestamps correlated to times of storage in the primary storage subsystem;

according to a predetermined schedule, the host sending cached data objects to the primary storage subsystem for storage therein and also including commands for the primary storage subsystem to store each data object in association with an advanced timestamp comprising the original timestamp associated with the data object increased by a predetermined margin, the predetermined margin being of sufficient amount so as to be recognizable as an advanced timestamp and an amount permitting derivation of the original timestamp from the advanced timestamp.

15. The medium of claim 14, the operations further comprising:

responsive to the data mover commencing a prescribed process of preparing data objects in the primary storage subsystem for storage in the secondary storage subsystem, where the process is applied to a designated set of one or more data objects in the primary storage subsystem, performing operations comprising;

reviewing each data object in the set to determine whether its associated timestamp is an advanced timestamp or current timestamp;

if advanced, subtracting the predetermined margin from the advanced timestamp to restore the original timestamp, and conducting future processing of the data object utilizing the restored original timestamp;

if current, conducting future processing of the data object utilizing the current timestamp.

16. The medium of claim 14, the operations further comprising:

the primary storage subsystem storing data objects in various sidefiles;

processing the sidefiles in order to form consistency groups, where the processing includes marking each sidefile as IDLE under conditions including the sidefile having at least one data object associated with an advanced timestamp.

17. A data storage system, comprising:

a cache;

a primary storage;

a host programmed to perform operations comprising;

storing some data objects in a cache and some data objects in the primary storage;

for each data object stored in the cache, placing in the cache an original timestamp corresponding to the data object, the original timestamp representing a time that the data object was stored in the cache;

writing data objects from the cache and corresponding advanced timestamps to the primary storage, where each advanced timestamp is equal to an original timestamp plus a prescribed time delta;

the primary storage programmed to perform operations comprising;

for each data object stored in the primary storage, writing a current timestamp to the primary storage representing a time that the data object was stored in the primary storage;

a data mover, programmed to perform operations comprising, responsive to requests to apply a predetermined consistency group formation routine to identified data objects in the primary storage, for each data object, reading the advanced timestamp corresponding to the identified data object from the sidefile and subtracting the time delta therefrom to compute a restored timestamp, and then applying the routine to the identified data object in conjunction with the restored timestamp instead of the advanced timestamp.

18. A data storage system configured to temporarily cache data pending storage in a primary storage subsystem asynchronously mirrored at a secondary storage subsystem without sacrificing timestamp information, comprising:

a host;

a primary storage subsystem coupled to the host;

a cache coupled to the host and separate from the primary storage subsystem;

a secondary storage subsystem;

a data mover coupling the primary and secondary storage subsystems;

where the host, primary storage subsystem, and data mover are programmed to perform operations comprising;

the host identifying various data objects to store in the primary storage subsystem;

the host storing some of the identified data objects in the cache in association with original timestamps correlated to times of storage in the cache;

the host sending other of the identified data objects to the primary storage subsystem for storage therein, and the primary storage subsystem responding by storing the other data objects in association with current timestamps correlated to times of storage in the primary storage subsystem;

according to a predetermined schedule, the host sending cached data objects to the primary storage subsystem for storage therein and also including commands for the primary storage subsystem to store each data object in association with an advanced timestamp comprising the original timestamp associated with the data object increased by a predetermined margin, the predetermined margin being of sufficient amount so as to be recognizable as an advanced timestamp and an amount permitting derivation of the original timestamp from the advanced timestamp.

19. The system of claim 18, the operations further comprising:

responsive to the data mover commencing a prescribed process of preparing data objects in the primary storage subsystem for storage in the secondary storage subsystem, where the process is applied to a designated set of one or more data objects in the primary storage subsystem, performing operations comprising;

reviewing each data object in the set to determine whether its associated timestamp is an advanced timestamp or current timestamp;

if advanced, subtracting the predetermined margin from the advanced timestamp to restore the original timestamp, and conducting future processing of the data object utilizing the restored original timestamp;

if current, conducting future processing of the data object utilizing the current timestamp.

20. The system of claim 18, the operations further comprising:

the primary storage subsystem storing data objects in various sidefiles;

processing the sidefiles in order to form consistency groups, where the processing includes marking each sidefile as IDLE under conditions at least a condition where the sidefile includes at least one data object having an advanced timestamp.

21. A data storage system, comprising:

a cache;

a primary storage;

host means for performing operations comprising;

storing some data objects in a cache and some data objects in the primary storage;

for each data object stored in the cache, placing in the cache an original timestamp corresponding to the data object, the original timestamp representing a time that the data object was stored in the cache;

writing data objects from the cache and corresponding advanced timestamps to the primary storage, where each advanced timestamp is equal to an original timestamp plus a prescribed time delta;

a primary controller in the primary storage for performing operations comprising, for each data object stored in the primary storage, writing a current timestamp to the primary storage representing a time that the data object was stored in the primary storage;

data mover means for performing operations comprising, responsive to requests to apply a predetermined consistency group formation routine to identified data objects in the primary storage, for each data object, reading the advanced timestamp corresponding to the identified data object from the sidefile and subtracting the time delta therefrom to compute a restored timestamp, and then applying the routine to the identified data object in conjunction with the restored timestamp instead of the advanced timestamp.