Tracker sensing method for regulating synchronization of audit files between primary and secondary hosts

• US 6,446,090 B1
• Filed: 10/08/1999
• Issued: 09/03/2002
• Est. Priority Date: 10/08/1999
• Status: Expired due to Term

• Current Assignee: Unisys Corporation
• Original Assignee: Unisys Corporation
• Inventors: Hart, Donald Ralph
• Primary Examiner(s): Alam, Hosain T.
• Application Number:
  US09/415,333
• Time in Patent Office: 1,061 Days
• Field of Search: 707/10, 707/201-204, 714/2, 714/4, 714/6, 714/8, 714/10, 714/12, 714/15-16, 714/19-20, 709/203, 709/226, 709/237-239, 709/248
• US Class Current: 1/1
• CPC Class Codes: G06F 11/2038 with a single idle spare processing component
• CPC Class Codes: G06F 11/2048 where the redundant components share neither address space nor persistent storage
• CPC Class Codes: G06F 11/2097 maintaining the standby controller/processing unit updated initialisation or re-synchronisation thereof G06F11/1658 and subgroups
• CPC Class Codes: G06F 2201/80 Database-specific techniques
• CPC Class Codes: Y10S 707/99952 Coherency, e.g. same view to multiple users

Abstract

A method for sensing and regulating the level of synchronization between a file of audit blocks generated in a primary host for transfer to a secondary host. The audit files in a primary host are divided into audit blocks which are each identified with an audit block serial number. The individual audit blocks are transmitted through a network from a primary host to a secondary host wherein a tracker mechanism monitors the difference between the number of received audit blocks and the number of generated audit blocks residing at the primary host waiting to be transferred. When the disparity between the audit blocks received and the generated audit blocks waiting to be transmitted reaches a certain critical level, the tracker mechanism will initiate an advance speed up program to expedite the transfer in order to bring the primary and secondary audit files into greater synchronism. Further, the acknowledgment level can be set to indicate the number of audit blocks received in the secondary host for information back to the primary host and the acknowledgment level can be utilized in order to preset a critical value which will indicate that the speed up transfer operation should be initiated.

Claims

1. In a network wherein a primary host is connected to a secondary host in order to set up a secondary host backup database which will maintain data file synchronism with a primary host database, a method for sensing the level of duplication between said primary and secondary databases comprising the steps of:

(a) sensing when audit blocks of sectioned audit files at said secondary host are not on a par with the number of sectioned audit block files residing at said primary host;

(b) initiating a Catchup program to speed up the transfer of said audit blocks of sectioned audit files from primary host to secondary host in an asynchronous transfer fashion when the number of audit blocks at said secondary host falls below the number of audit blocks waiting for transfer from said primary host.

2. The method of claim 1 wherein step (a) includes the steps of:

(a1) recognizing an End-Of-File condition after transfer of audit blocks of sectioned audit files from said primary host to said secondary host;

(a2) scanning the audit block serial numbers of the audit blocks assigned by said primary host and transmitted to said secondary host;

(a3) accessing data from a database support library to indicate any disparity between the audit block serial numbers, assigned in the primary host, to the audit block serial numbers received at said secondary host.

3. The method of claim 2 which includes the step of:

(a4) determining the audit block serial number of the audit block from which re-transmission, from primary to secondary, has to be initiated after an interruption in the transfer of audit blocks from said primary to secondary host.

4. The method of claim 2 wherein step (a2) includes the step of:

(a2a) determining that the serial number of the audit blocks at said primary host and secondary host are on a par so that no speed up of audit block transfers is required.

5. A method of regulating the synchronization level between audit blocks in a primary host, which are awaiting transfer to a secondary host, and audit blocks received in said secondary host, comprising the steps of:

(a) Transferring said primary host audit blocks to said secondary host as permitted by the limitations and delays of an interconnecting network between said primary and secondary hosts;

(b) acknowledging by said secondary host to said primary host each time that a set of n audit blocks are received by said secondary host;

(c) sensing when the number of waiting primary host audit blocks is 2 (n−

1) or greater than the number of received audit blocks in said secondary host.

6. The method of claim 5 which includes the step of:

(d) initiating an audit block transfer speed-up program when said number of received secondary host audit blocks is less than 2 (n−

1).

7. The method of claim 6 which includes the step of:

(e) initiating an audit block transfer speed up program when no acknowledgment is received by said primary host after a pre-designated time-out period.

8. In a system for expediting the transfer of audit files from a primary host to a secondary host, a method for sensing the state of synchronism between sectioned audit files in said primary host awaiting transfer and sectioned audit files positioned as audit blocks in said secondary host which are received from said primary host, said method comprising the steps of:

(a) sectioning said audit files into separate physical audit files designated as audit blocks and assigning a separate identification number to each audit block at said primary host;

(b) placing each audit block in a separate sectioned based buffer at said primary host;

(c) transferring each audit block at said primary host as permitted by the condition of the network connection between said primary and secondary host and establishing a special port connection between said primary and secondary host;

(d) placing each audit block received at said secondary host into a separate sectioned based buffer;

(e) sensing the number of audit blocks at said primary host which have not yet been transferred to said secondary host;

(f) initiating a transfer speed-up program at said secondary host when the audit blocks received at said secondary host is a pre-set number less than the number of sectioned audit blocks ready to be transferred from said primary host.

9. In a system utilizing an audit block write mode method for transfer of audit file data from a primary host to a secondary host, a method for sensing any delay in the transfer of audit block sections from said primary host to said secondary host comprising the steps of:

(a) placing an audit block serial number on each audit block in said primary which is destined for transfer to said secondary host;

(b) transferring said primary host audit blocks over to said secondary host as permitted by the operating condition of interconnecting network communication lines;

(c) calculating the difference value between the highest audit block serial number received at said secondary host and the highest numbered audit block serial number written in said primary host.

10. The method of claim 9 which includes the step of:

(d1) taking no further action if the said difference value is equal to “

1”

or less than 1.

11. The method of claim 9 which includes the step of:

(d2) initiating a speed up program when said difference value is greater than “

1”

.

12. A method for detecting the condition of out-of-synchronization between sectioned audit files of a source host database and a remote host database, comprising the steps of:

(a) assigning an identifying serial number to each audit block in a sectioned audit file at said source host;

(b) checking periodically to compare the highest contiguously written audit block serial number of said source audit blocks with the highest contiguously written audit block serial number of said remote audit blocks;

(c) initiating an expedited transfer of audit blocks from said source database host to said remote database host when the said highest source audit block serial number is greater than said highest remote audit block serial number.

13. A system for re-instituting the transfer of audit blocks after an interrupted or failed network transmission connection between a primary host and a secondary host, and where each audit file block has been assigned a special serial number, said system comprising:

(a) means to determine the highest audit block serial number, H, at said secondary host which has been received in said secondary host;

(b) means to store the audit block serial numbers at said primary host in order to select the serial number, H+1, for initiating a sequence of audit block transmissions from said primary host to said secondary host.