Remote data duplexing

• US 5,504,861 A
• Filed: 02/22/1994
• Issued: 04/02/1996
• Est. Priority Date: 02/22/1994
• Status: Expired due to Term

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Jaworski, Debra L., Kern, Ronald M., Crockett, Robert N.
• Primary Examiner(s): Beausoliel, Jr., Robert W.
• Assistant Examiner(s): CHUNG, PHUNG M
• Application Number:
  US08/199,449
• Time in Patent Office: 770 Days
• Field of Search: 395/575, 395/425, 395/725, 371/68.2, 371/10.1, 371/9.1, 371/11.1, 371/40.1
• US Class Current: 714/13
• CPC Class Codes: G06F 11/1435 using file system or storage system metadata
• CPC Class Codes: G06F 11/2064 while ensuring consistency
• CPC Class Codes: G06F 11/2074 Asynchronous techniques

Abstract

A remote data shadowing system provides storage based, real time disaster recovery capability. Record updates at a primary site cause write I/O operations in a storage subsystem therein. The write I/O operations are time stamped and the time, sequence, and physical locations of the record updates are collected in a primary data mover. The primary data mover groups sets of the record updates and associated control information based upon a predetermined time interval, the primary data mover appending a prefix header to the record updates thereby forming self describing record sets. The self describing record sets are transmitted to a remote secondary site wherein consistency groups are formed such that the record updates are ordered so that the record updates can be shadowed in an order consistent with the order the record updates cause write I/O operations at the primary site. The secondary site optimizes the record updates of each consistency group such that record updates to a same target track are included in a single I/O CCW chain. Additionally, the record updates are written according to full consistency group recovery rules such that should the primary site be unavailable, the secondary site can recover a consistency group.

Claims

1. A method for shadowing record updates, the record updates transmitted asynchronously from a primary site to a secondary site, the secondary site including secondary storage devices and providing disaster recovery capabilities for the primary site, the record updates assembled in consistency groups for shadowing the record updates in real time and in a sequence consistent order, the method comprising steps of:

(a) determining a set of full consistency group recovery rules specific to the secondary storage devices;

(b) optimizing a shadowing of a current consistency group of record updates to the secondary storage devices by sorting the current consistency group of record updates according to secondary storage device destinations and chaining pluralities of input/output control command words (I/O CCWs) according to the sorted record updates, the sorted record updates being written to the secondary storage devices according to the full consistency group recovery rules; and

(c) repeating step (b) for each consistency group.

2. The method according to claim 1 wherein the optimizing step (b) is dependent upon a pattern of various primary site write I/O operations of the record updates at the primary site.

3. The method according to claim 2 wherein the secondary storage devices are direct access storage devices (DASDs) and the optimizing step (b) sorts the current consistency group of record updates according to target volumes.

4. The method according to claim 3 wherein the optimizing step (b) chains the I/O CCWs according to target tracks.

5. The method according to claim 4 wherein a plurality of I/O CCWs are executed in parallel for each direct access storage device of the secondary storage devices.

6. The method according to claim 5 wherein secondary volumes establish remote duplex pairs with corresponding primary volumes.

7. The method according to claim 4 further comprising a step (d) re-executing an I/O CCW chain if an I/O error occurs before the I/O CCW chain completes.

8. The method according to claim 4 further comprising a step (e) re-applying the current consistency group to the secondary storage devices if the current consistency group did not complete successfully.

9. The method according to claim 8 wherein the secondary site performs step (e) without re-receiving record updates from the primary site.

10. A secondary system for providing a disaster recovery capability for a primary system, the primary system running applications generating record updates that are stored to a primary storage subsystem, a primary data mover collecting sets of record updates and corresponding record set informations into self describing record sets, the self describing record sets further being assembled into consistency groups wherein each record update is sequence ordered according to a primary system synchronized time-stamp, the primary data mover asynchronously transmitting the self describing record sets to the secondary system in real time, the secondary system comprising:

a secondary processor for receiving the transmitted self describing record sets;

a secondary storage subsystem for storing the record updates; and

secondary data mover means coupled to the secondary processor, the secondary data mover means optimizing secondary I/O operations for each consistency group of record updates for storing each record update therein in sequence consistent order, the I/O operations optimized by organizing record updates according to secondary storage subsystems destinations, wherein input/output channel command words (I/O CCWs) are chained according to the record updates secondary storage subsystems destinations, the record updates written to the secondary storage subsystem such that full consistency group recovery is maintained.

11. The secondary system according to claim 10 wherein the secondary storage subsystem comprises:

a plurality of secondary storage controllers; and

a plurality of secondary storage devices coupled to the plurality of secondary storage controllers.

12. The secondary system according to claim 11 wherein the secondary storage devices are direct access storage devices.

13. The system according to claim 12 wherein the secondary data mover means accesses full consistency group recovery rules stored in the secondary system, the full consistency group recovery rules governing the storage of the record updates such that an entire consistency group is re-written without errors.

14. The system according to claim 13 wherein the record updates are organized according to target volumes within the plurality of secondary storage devices, and further organized within target volumes according to target tracks.

15. The secondary system according to claim 14 wherein the secondary data mover means causes a single START I/O command to be issued for those record updates being written to a single volume.

16. The secondary system according to claim 15 wherein the secondary data mover means causes an I/O CCW chain to re-execute if an I/O error occurs in the secondary system such that the I/O CCW Chain can complete normally independently of the corresponding applications running in the primary system.

17. The secondary system according to claim 15 wherein a plurality of I/O CCW chains are executing concurrently to different secondary storage devices of the plurality of secondary storage devices.

18. The secondary system according to claim 10 wherein the consistency groups are formed in the primary system.

19. The secondary system according to claim 10 wherein the consistency groups are formed in the secondary system.

20. An asynchronous remote data duplexing system providing storage based, real time data shadowing, including a primary site running applications generating record updates and having a secondary site, remote from the primary site, the secondary site shadowing the record updates and providing disaster recovery for the primary site, wherein the secondary site maintains a set of full consistency group recovery rules for describing record update shadowing parameters for ensuring each consistency group is recoverable in an event that the primary site is unavailable, the asynchronous remote data duplexing system comprising:

a sysplex timer for synchronizing time dependent processes in the primary site;

a primary processor at the primary site for running the applications, the primary processor having a primary data mover therein;

a plurality of primary storage controllers for issuing write I/O operations for each record update, each primary storage controller write I/O operation synchronized to the sysplex timer by the primary processor;

a plurality of primary storage devices coupled to said plurality of primary storage controllers for receiving the write I/O operations and storing the record updates therein accordingly,

wherein the primary data mover collects record set information from the plurality of primary storage controllers for each record update and appends a prefix header to a predetermined group of record set informations, the prefix header and predetermined group of record set informations forming self describing record sets, each record set information including a primary device address,

a record update sequence number, a write I/O type, a search argument, and a record update time, and wherein the prefix header includes a total data length, an operational time stamp, a time interval group number, and a records read time;

a secondary processor at the secondary site having a secondary data mover, the secondary data mover receiving the self describing record sets from the primary site;

a plurality of secondary storage controllers coupled to the secondary processor; and

a plurality of secondary storage devices coupled to said plurality of secondary storage controllers for storing the record updates,

wherein the secondary data mover determines whether the transmitted self describing record sets are complete and forms consistency groups from the self describing record sets and provides the record updates from each consistency group to the plurality of secondary storage controllers for writing to the plurality of secondary storage devices in an order consistent with a sequence that the record updates were written to the plurality of primary storage devices, the secondary data mover means optimizing secondary I/O operations for each consistency group of record updates for storing each record update therein in sequence consistent order, the I/O operations optimized by organizing record updates according to secondary storage subsystems destinations, wherein input/output channel command words (I/O CCWs) are chained together according to the organized record updates, the record updates written to the secondary storage subsystem according to the full consistency group recovery rules.