Method for highly available transaction recovery for transaction processing systems

• US 7,152,181 B2
• Filed: 01/13/2003
• Issued: 12/19/2006
• Est. Priority Date: 02/22/2002
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: BEA Systems Incorporated (Oracle Corporation)
• Inventors: Somogyi, Alexander J., Fung, Priscilla C.
• Primary Examiner(s): Bonzo, Bryce P.
• Assistant Examiner(s): Maskulinski, Michael C.
• Application Number:
  US10/341,207
• Publication Number:
  US 20030177150A1
• Time in Patent Office: 1,436 Days
• Field of Search: 714/4, 714/10
• US Class Current: 714/4.11
• CPC Class Codes: G06F 11/1662 the resynchronized component or unit being a persistent storage device re-synchronization of failed mirror storage G06F11/2082; rebuild or reconstruction of parity RAID storage G06F11/1008
• CPC Class Codes: G06F 11/2028 eliminating a faulty processor or activating a spare
• CPC Class Codes: G06F 11/203 using migration
• CPC Class Codes: G06F 11/2038 with a single idle spare processing component
• CPC Class Codes: G06F 11/2046 where the redundant components share persistent storage G06F11/2043 takes precedence
• CPC Class Codes: H04L 1/22 using redundant apparatus to increase reliability

Abstract

A highly available transaction recovery service migration system in accordance with one embodiment of the present invention implements a server'"'"'s Transaction Recovery Service (TRS) as a migratable service. In one embodiment of the present invention, the TRS is a server instance or software module implemented in JAVA. The TRS migrates to an available server that resides in the same cluster as the failed server. The migrated TRS obtains the TLOG of the failed server, reads the transaction log, and performs transaction recovery on behalf of the failed server. The migration may occur manually or automatically on a migratable services framework. The TRS of the failed server migrates back in a fail back operation once the failed primary server is restarted. Failback operation may occur whether recovery is completed or not. This expedites recovery and improves availability of the failed server thereby preserving the efficiency of the network and other servers.

Claims

1. A method for recovering transactions comprising:

first performing failover migration of a transaction recovery service from a failed primary server to a back-up server, the primary server having a transaction log, the transaction recovery service having ownership of the transaction log;

second performing recovery for the failed primary server, the recovery performed on the backup server;

wherein said second performing recovery for the failed primary server includes;

reading the transaction log of the primary server by the transaction recovery service while the transaction recovery service resides in the back-up server; and

enabling the back-up server to read and coordinate transactions from the primary server'"'"'s transaction log; and

third performing failback migration of the transaction recovery service from the back-up server to the primary server.

2. A method for recovering transactions comprising:

first performing failover migration of a transaction recovery service from a failed primary server to a back-up server, the primary server having a transaction log, the transaction recovery service having ownership of the transaction log;

second performing recovery for the failed primary server, the recovery performed on the backup server; and

third performing failback migration of the transaction recovery service from the back-up server to the primary server;

wherein said third performing failback migration includes;

initiating failback migration of the transaction recovery service to the primary server; and

completing failback migration of the transaction recovery service to the primary server;

wherein completing failback migration includes moving the transaction recovery service from the back-up server to the primary server before primary server transaction recovery is complete.

3. A method for performing transaction recovery comprising:

(a) detecting a first server is down, the first server having a first transaction log on a memory and a first transaction recovery service;

(b) moving the first transaction recovery service to a second server;

(c) activating the first transaction recovery service on the second server;

(d) performing transaction recovery on behalf of the first server by the first transaction recovery service while the first transaction recovery service resides on the second server;

(d.1) initiating restart of the first server before transaction recovery on behalf of the first server is complete, wherein deactivating the first transaction recovery service is performed before completing transaction recovery;

(e) deactivating the first transaction recovery service on the second server;

(f) moving the first transaction recovery service to the first server; and

(g) activating the first transaction recovery service on the first server.

4. The method of claim 3 wherein said first transaction recovery service and second transaction recovery service are each a server instance implemented in java.

5. A method for recovering transactions comprising;

first performing failover migration of a transaction recovery service from a failed primary server to a back-up server, the primary server having a transaction log, the transaction recovery service having ownership of the transaction log;

wherein said first performing failover migration includes initiating failover migration by an administrator;

second performing recovery for the failed primary server, the recovery performed on the backup server;

wherein said initiating failover migration by an administrator includes an administrator issuing a migration command; and

third performing failback migration of the transaction recovery service from the back-up server to the primary server;

wherein the issuing a migration command includes a user manually issuing a migration command using a JMX MBean.

6. A method for recovering transactions comprising:

first performing failover migration of a transaction recovery service from a failed primary server to a back-up server, the primary server having a transaction log, the transaction recovery service having ownership of the transaction log;

second performing recovery for the failed primary server, the recovery performed on the backup server; and

third performing failback migration of the transaction recovery service from the back-up server to the primary server;

wherein the back-up server has multiple transaction recovery services residing on the back-up server, each transaction recovery service corresponding to a different primary server and a corresponding transaction log;

wherein the backup server includes its own transaction recovery service, and wherein the backup server'"'"'s transaction recovery service functions as usual during migration failover and failback.

7. A method for performing transaction recovery comprising:

(a) detecting a first server is down, the first server having a first transaction log on a memory and a first transaction recovery service;

(b) moving the first transaction recovery service to a second server

(c) activating the first transaction recovery service on the second server

(d) performing transaction recovery on behalf of the first server by the first transaction recovery service while the first transaction recovery service resides on the second server;

(e) deactivating the first transaction recovery service on the second server

(f) moving the first transaction recovery service to the first server; and

(g) activating the first transaction recovery service on the first server;

wherein the second server includes a second transaction log, access to the memory, and a second transaction recovery service.

8. A method for manually performing failover migration from a failed primary server to a backup server comprising:

manually migrating a TRS from a failed primary server to a backup server;

calling a JAVA method of the TRS while the TRS resides on the backup server, the JAVA method configured to activate the TRS on the backup server;

reading at least one of a plurality of TLOG files of the primary server by the TRS;

instantiating and placing transactions in the backup server'"'"'s transaction map by the TRS; and

performing recovery for the primary server on the backup server by the TRS, recovery including driving prepared transactions to completion and performing resource recovery.

9. A method for manually performing failback migration from a backup server to a primary server before recovery on a primary server is complete comprising:

performing recovery for a failed primary server by a TRS associated with the primary server and residing on the backup server;

notifying the backup server that the primary server has restarted, the notification including an administrative MBean;

requesting a migratable framework to migrate the TRS from the backup server to the primary server, the TRS making the request to the migratable framework;

deactivating the TRS residing on the backup server by the migratable framework, wherein deactivating the TRS includes the migratable framework calling a JAVA method residing in the TRS to deactivate the TRS on the backup server; and

migrating the TRS from the backup server to the primary server by the migratable framework, wherein upon a primary server restart, the primary server regains ownership of a TLOG corresponding to the primary server and completes recovery of the primary server.

10. The method of claim 9 wherein deactivating the TRS includes the TRS suspending recovery processing for the primary server, checkpointing the TLOG for the primary server, purging transactions in the transaction map of the backup server that originated from the TLOG of the primary server, and ending resource recovery performed for the primary server.

11. A method for automatically performing failover migration from a failed primary server to a backup server comprising:

detecting the failure of a primary server by a migratable framework;

migrating a TRS associated with the primary server from the failed primary server to a backup server, said migrating a TRS including deactivating the TRS on the primary server by calling an MBean method of the TRS by the migratable framework;

activating the TRS on the backup server, said activating the TRS including reading a TLOG of the primary server, instantiating transactions from the TLOG and moving them to the transaction map of the backup server, and scheduling resource recovery for the failed primary server; and

performing recovery for the primary server by the TRS while the TRS resides on the backup server, said performing recovery including driving prepared transactions to completion and performing resource recovery.

12. A method for automatically performing failback migration from a backup server to a primary server after recovery performed on a primary server is complete, the method comprising:

completing recovery for a primary server by a TRS associated with the primary server, the TRS completing recovery for the primary server while the TRS resides on a backup server, said completing recovery including the TRS checkpointing a TLOG associated with the primary server and relinquishing control of the TLOG of the primary server;

initiating migration of the TRS from the backup server to the primary server, said initiating migration of the TRS including the backup server requesting the migratable framework to migrate the TRS to the primary server;

deactivating the TRS on the backup server, said deactivating the TRS including calling an MBean method of the TRS by the migratable framework; and

migrating the TRS from the backup server to the primary server by the migratable framework, wherein the primary server regains TLOG ownership upon primary server startup.

13. A method for automatically performing failback migration from a backup server to a primary server before recovery performed on a primary server is complete, the method comprising:

restarting a primary server, wherein performing recovery on the primary server is not completed by a TRS residing on the backup server before the primary server is restarted;

initiating migration of a TRS associated with a primary server from a backup server to the primary server, said initiating migration including detecting the startup of the primary server by a migratable framework;

deactivating the TRS residing on the backup server, said deactivating including calling a JAVA method of the TRS to deactivate the TRS, the method configured to;

checkpoint a TLOG associated with the primary server;

purging transactions in the transaction map of the backup server that originated from the

TLOG of the primary server; and

stopping resource recovery being performed for the primary server;

migrating the TRS to the primary server; and

activating the TRS on the primary server, said activating the TRS including calling a JAVA method of the TRS to activate the TRS, wherein the primary server regains ownership of the TLOG and performs recovery as required.

14. The method of claim 13 wherein said transaction recovery service is a JAVA server instance.

15. The method of claim 14 wherein each transaction recovery service is associated with a JAVA element that indicates an alternate server as a migratable target for the transaction recovery service, wherein a user may specify the alternate server within the JAVA method.

16. The method of claim 14 wherein a first JAVA MBean corresponds to each of the backup server and the primary server, each first MBean including runtime information regarding the transaction recovery system instances hosted by the corresponding backup server or primary server.

17. The method of claim 16 wherein the first MBean includes a first JAVA method that returns whether a transaction recovery service instance is currently activated on a nnparticular server.

18. The method of claim 16 wherein performing recovery on a failed server includes reading transactions from a transaction log by the transaction recovery service relating to the failed server, the transaction log corresponding to the failed server, wherein the first MBean includes a second JAVA method configured to return the total number of transactions read from the transaction log by the transaction recovery service.

19. The method of claim 16 wherein the first MBean includes a third JAVA method that returns the percentage of the recovered transactions that are completed by the transaction recovery service.

20. The method of claim 16 wherein a second MBean may access the runtime information of each first MBean.

21. The method of claim 20 wherein the second MBean includes a fourth JAVA method that returns an array of MBeans corresponding to the transaction recovery service instances deployed on a server.

22. The method of claim 16 wherein the second MBean includes a fifth JAVA method that returns the first MBean.

23. A method for performing transaction recovery within a cluster comprising:

performing failover migration of a transaction recovery service from a failed primary server to a backup server, the transaction recovery service implemented as a java instance and owning a transaction log corresponding to the primary server;

activating the transaction recovery service on the backup server by the migratable framework, the migratable framework operable to call an activation method of the transaction recovery service, the activation method implemented in java, activating the transaction recovery service including;

reading the transaction log of the primary server;

placing primary server transaction files in a transaction map of the backup server; and

scheduling resource recovery for the primary server;

performing transaction recovery on behalf of the failed primary server by the transaction recovery service, transaction recovery including driving prepared transactions to completion and performing resource recovery;

initiating migration from the backup server to the primary server upon the occurrence of an event; and

performing failback migration of the transaction recovery service from the backup server to the primary server, failback migration including;

deactivating the transaction recovery service on the backup server by the migratable framework, the migratable framework operable to call a deactivation method of the transaction recovery service, the deactivation method implemented in java;

migrating the transaction recovery service to the primary server; and

activating the transaction recovery service on the primary server by calling the activation method.

24. The method of claim 23 wherein the event is the completion of the recovery performed on behalf of the primary server.

25. The method of claim 23 wherein the event is the restart of the primary server.

26. A method for recovering transactions comprising:

performing failover migration of a transaction recovery service from a failed primary server to a backup server, the primary server having a transaction log, and the transaction recovery service having ownership of the log; and

allowing the backup server to read the transaction log and perform recovery on behalf of the failed server.

27. The method of claim 26 including the step of:

performing recovery on the primary server.

28. The method of claim 27 including the step of:

performing failback migration of the transaction recovery service from the backup server to the primary server.

29. The method of claim 26 wherein said primary server and the backup server are in a cluster.

30. The method of claim 26 wherein transaction recovery is accomplished automatically.

31. The method of claim 26 wherein transaction recovery is accomplished manually.

32. The method of claim 26 wherein transaction recovery is accomplished manually by an administrator.

33. The method of claim 26 including a migratable service framework and wherein said transaction recovery service depends upon the migratable service framework for manual migration support.

34. The method of claim 26 including a migratable service framework and wherein said transaction recovery service depends upon the migratable service framework for automatic migration support.

35. The method of claim 26 wherein said transaction recovery service is a server instance implemented in java.

36. The method of claim 26 wherein said transaction recovery service is manually migrated and monitored by one of an administration console and a JMX API.