CAPTURING AND RESTORING DATABASE SESSION STATE

• US 20100005097A1
• Filed: 07/01/2008
• Published: 01/07/2010
• Est. Priority Date: 07/01/2008
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: Cao, Feng, Pingte, Rajendra, Liang, Chao
• Granted Patent: US 8,433,680 B2
• Time in Patent Office: Days
• US Class Current: 707/8
• CPC Class Codes: G06F 11/3414 Workload generation, e.g. scripts, playback
• CPC Class Codes: G06F 11/3495 for systems
• CPC Class Codes: G06F 16/27 Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor
• CPC Class Codes: G06F 2201/86 Event-based monitoring
• CPC Class Codes: G06F 2201/87 Monitoring of transactions

Abstract

Techniques are described herein for capturing and restoring database session state. Production database server components save the session state of each of a plurality of database sessions. The components store workload units that are processed in these sessions. The components store updated session states in response to certain events. Thus, the components may capture multiple session states, pertaining to various different points in time, for each session. The captured session states and the captured workload are moved to a test database server. A user selects, from among the time points represented by the session states, a point in time at which the user would like workload replay to begin. Sessions are re-created on the test database server. Session states of these sessions are set to reflect the session states as they existed at the user-selected time point. Workload units are replayed in the sessions relative to the test database server.

Claims

1. A computer-implemented method comprising:

capturing a session state of a database session as the database session exists at a point in time after a specified time point; and

recreating, based on the session state, the database session in a second database server that is separate from the first database server.

2. The method of claim 1, wherein the step of capturing the session state comprises storing the session state on a persistent non-volatile computer-readable storage medium.

3. The method of claim 1, wherein the step of recreating the database session comprises loading the session state from a persistent non-volatile computer-readable storage medium.

4. The method of claim 1, further comprising:

capturing workload that is processed in the database session by a first database server; and

replaying at least a portion of the workload, starting at the point in time after the specified time point, in the recreated database session in the second database server.

5. The method of claim 4, further comprising:

receiving, from a user, an indication of a particular time point from which workload replay should begin; and

replaying at least a portion of the workload, starting at the particular time point, in the recreated database session in the second database server.

6. The method of claim 4, wherein replaying at least a portion of the workload, starting at the point in time after the specified time point, in the recreated database session in the second database server comprises replaying the portion of the workload without replaying any portion of the workload that is associated with a time that is prior to the point in time after the specified time point.

7. The method of claim 4, further comprising:

while capturing the workload, determining that a transaction boundary has been reached in the database session; and

in response to determining that the transaction boundary has been reached in the database session, capturing an updated session state of the database session that is current as of a time that the transaction boundary was reached.

8. The method of claim 4, further comprising:

after replaying the portion of the workload, starting at the point in time after the specified time point, in the recreated database session in the second database server, replaying, in the second database server, another portion of the workload starting at another user-specified point in time that is different than the specified time point.

9. The method of claim 1, further comprising:

receiving, from a user, a selection of two or more database sessions that exist in the first database server;

for each particular database session of the two or more database sessions, suspending the particular database session; and

after each of the two or more database sessions are suspended, capturing and storing a separate database session state for each of the two or more database sessions.

10. The method of claim 1, further comprising:

receiving, from a user, a selection of two or more database sessions that exist in the first database server;

for each particular database session of the two or more database sessions, capturing and storing a session state of the particular database session, and, after capturing and storing the session state of the particular database session, suspending the particular database session; and

after each of the two or more database sessions are suspended, resuming each of the two or more database sessions.

11. The method of claim 1, wherein:

capturing the session state of the database session as the database session exists at the point in time after the specified time point comprises storing, on a non-volatile computer-readable storage medium, one or more variables that are contained in volatile memory of a first computer on which the first database server executes, along with one or more values of those one or more variables; and

recreating, based on the session state, the database session in the second database server comprises loading, from the non-volatile computer-readable storage medium, and into a volatile memory of a second computer on which the second database server executes, the one or more variables along with the one or more values of those one or more variables;

wherein the first computer is separate from the second computer; and

wherein the first database server does not execute concurrently with the second database server.

12. A volatile or non-volatile computer-readable storage medium storing instructions, the instructions including instructions which, when executed by one or more processors, cause the one or more processors to perform the steps of:

capturing a session state of a database session as the database session exists at a point in time after a specified time point; and

recreating, based on the session state, the database session in a second database server that is separate from the first database server.

13. The computer-readable storage medium of claim 12, wherein the step of capturing the session state comprises storing the session state on a persistent non-volatile computer-readable storage medium.

14. The computer-readable storage medium of claim 12, wherein the step of recreating the database session comprises loading the session state from a persistent non-volatile computer-readable storage medium.

15. The computer-readable storage medium of claim 12, wherein the steps further comprise:

capturing workload that is processed in the database session by a first database server; and

replaying at least a portion of the workload, starting at the point in time after the specified time point, in the recreated database session in the second database server.

16. The computer-readable storage medium of claim 15, wherein the steps further comprise:

receiving, from a user, an indication of a particular time point from which workload replay should begin; and

replaying at least a portion of the workload, starting at the particular time point, in the recreated database session in the second database server.

17. The computer-readable storage medium of claim 15, wherein replaying at least a portion of the workload, starting at the point in time after the specified time point, in the recreated database session in the second database server comprises replaying the portion of the workload without replaying any portion of the workload that is associated with a time that is prior to the point in time after the specified time point.

18. The computer-readable storage medium of claim 15, wherein the steps further comprise:

while capturing the workload, determining that a transaction boundary has been reached in the database session; and

in response to determining that the transaction boundary has been reached in the database session, capturing an updated session state of the database session that is current as of a time that the transaction boundary was reached.

19. The computer-readable storage medium of claim 15, wherein the steps further comprise:

after replaying the portion of the workload, starting at the point in time after the specified time point, in the recreated database session in the second database server, replaying, in the second database server, another portion of the workload starting at another user-specified point in time that is different than the specified time point.

20. The computer-readable storage medium of claim 12, wherein the steps further comprise:

receiving, from a user, a selection of two or more database sessions that exist in the first database server;

for each particular database session of the two or more database sessions, suspending the particular database session; and

after each of the two or more database sessions are suspended, capturing and storing a separate database session state for each of the two or more database sessions.

21. The computer-readable storage medium of claim 12, wherein the steps further comprise:

receiving, from a user, a selection of two or more database sessions that exist in the first database server;

for each particular database session of the two or more database sessions, capturing and storing a session state of the particular database session, and, after capturing and storing the session state of the particular database session, suspending the particular database session; and

after each of the two or more database sessions are suspended, resuming each of the two or more database sessions.

22. The computer-readable storage medium of claim 12, wherein the steps further comprise:

capturing the session state of the database session as the database session exists at the point in time after the specified time point comprises storing, on a non-volatile computer-readable storage medium, one or more variables that are contained in volatile memory of a first computer on which the first database server executes, along with one or more values of those one or more variables; and

recreating, based on the session state, the database session in the second database server comprises loading, from the non-volatile computer-readable storage medium, and into a volatile memory of a second computer on which the second database server executes, the one or more variables along with the one or more values of those one or more variables;

wherein the first computer is separate from the second computer; and

wherein the first database server does not execute concurrently with the second database server.

23. A computer-implemented method comprising:

synchronizing multiple sessions at a transaction boundary during a capture process and a restore process while making no change to a database state of a database during the capture and restore process; and

ensuring transactional consistency between concurrently executing transactions during the capture process and the restore process by performing steps comprising;

a coordinator session marking a particular session with a suspend option;

suspending one or more user sessions once the one or more user sessions reach both a capture boundary of a particular transaction and a current call boundary, and when capture criteria are met;

after each selected session of a plurality of selected sessions reaches the capture boundary and each selected session of the plurality of selected sessions is suspended, the coordinator session initiating capture of each selected session of the plurality of suspended sessions, thereby aligning captured session states of each selected session of the plurality of selected sessions to a particular database system change number;

maintaining the particular session in a suspended state after the particular session captures a session state of the particular session, until all the selected sessions of the plurality of selected sessions are done with a capture process and until the coordinator resumes workload;

restoring each selected session of the plurality of sessions during the restore process after each selected session of the plurality of sessions has stabilized;

suspending a restored version of the particular session after the restore process, until all of the selected sessions of the plurality of selected sessions are done restoring;

wherein, throughout the capture process and the restore process, no change is made to the database state of the database;

serializing session states of each selected session of the plurality of selected sessions to external storage;

wherein no change needs to be committed in the database during the capture process and the restore process.