Database system providing self-tuned parallel database recovery
A database system providing self-tuned parallel database recovery is described. In one embodiment, for example, in a database system, a method is described for performing recovery operations using an optimal number of recovery threads, the method comprises steps of: (a) spawning an initial recovery thread to perform recovery operations; (b) measuring I/O (input/output) performance with the initial recovery thread; (c) spawning a subsequent recovery thread to perform recovery operations; (d) measuring I/O performance with the subsequent recovery thread; and (e) as long as I/O performance does not degrade beyond a preselected percentage, repeating steps (c) and (d) for spawning a desired number of additional recovery threads. In another embodiment, the database system auto-tunes the cache during performance of database recovery operations to optimize the performance of recovery operations.
1. In a database system, a self-tuning method for performing recovery operations by dynamically adapting how many recovery threads are spawned during recovery, the method comprising:
(a) spawning an initial recovery thread to perform recovery operations for at least one database that may require recovery;
(b) measuring I/O (input/output) performance with the initial recovery thread;
(c) spawning a subsequent recovery thread to perform recovery operations;
(d) measuring I/O performance with the subsequent recovery thread; and
(e) self-tuning how many threads are spawned by continuing, as long as I/O performance does not degrade beyond a preselected percentage, to repeat steps (c) and (d) for spawning a desired number of additional recovery threads and when I/O performance measured for a just-spawned recovery thread degrades beyond the preselected percentage, putting the just-spawned recovery thread to sleep;
wherein a user of the system is able to specify a particular number of concurrent recovery threads, and wherein the system generates an advisory if the particular number of concurrent recovery threads specified can be changed to achieve better I/O performance.
2. The method of claim 1 , wherein I/O performance is measured over a given period of time.
3. The method of claim 2 , wherein the given period of time is about 1 second.
4. The method of claim 1 , wherein steps (c) and (d) are repeated for spawning additional recovery threads, as long as I/O performance degrades by no more than about 15 percent.
5. The method of claim 1 , wherein steps (c) and (d) are repeated such that only a preconfigured maximum number of recovery threads may be generated.
6. The method of claim 5 , wherein the maximum number of recovery threads is limited to not exceed a count of databases that can be opened.
7. The method of claim 5 , wherein the maximum number of recovery threads is limited to not exceed one less than a count of database engines online.
8. The method of claim 1 , further comprising:
after another recovery thread finishes, awaking the thread that has been put to sleep.
9. The method of claim 1 , wherein steps (c) and (d) are repeated up to a configured maximum number of databases that can be recovered concurrently.
10. The method of claim 1 , wherein each recovery thread itself recovers a single database at a time.
11. A computer-readable medium having processor-executable instructions for performing the method of claim 1 .
12. The method of claim 1 , further comprising:
downloading a set of processor-executable instructions for performing the method of claim 1 .
13. A database system performing self-tuning recovery operations by dynamically adapting how many recovery threads are spawned during recovery, the system comprising:
a computer having at least one processor;
a database system having at least one database that may require recovery;
an initial recovery thread that is spawned to perform recovery operations, wherein the system measures I/O (input/output) performance with the initial recovery thread; and
a plurality of additional recovery threads that are spawned to perform recovery operations, wherein the system dynamically adjusts how many recovery threads are spawned based on I/O (input/output) performance with each additional recovery thread that is spawned, and wherein the system ceases spawning additional recovery threads and puts a just-spawned recovery thread to sleep when I/O performance for the just-spawned recovery thread degrades beyond a desired amount;
wherein a user of the system is able to specify a particular number of concurrent recovery threads, and wherein the system generates an advisory if the particular number of concurrent recovery threads specified can be changed to achieve better I/O performance.
14. The system of claim 13 , wherein I/O performance is measured over a given period of time.
15. The system of claim 14 , wherein the given period of time is about 1 second.
16. The system of claim 13 , wherein the system may spawn additional recovery threads as long as I/O performance degrades by no more than about 15 percent.
17. The system of claim 13 , wherein the plurality of additional recovery threads spawned is limited such that only a maximum number of recovery threads may be generated.
18. The system of claim 17 , wherein the maximum number of recovery threads is limited to not exceed a count of databases to be opened.
19. The system of claim 17 , wherein the maximum number of recovery threads is limited to not exceed one less than a count of database engines online.
20. The system of claim 13 , wherein the system awakens the thread that has been put to sleep, upon termination of another thread.
21. The system of claim 13 , wherein a maximum number of recovery threads permitted is limited to a configured maximum number of databases that can be recovered concurrently.
22. The system of claim 13 , wherein each recovery thread recovers a particular database at a given point in time.
23. In a database system, an auto-tuning method for performing database recovery in a manner that dynamically adjusts how many recovery threads are spawned based on current performance, the method comprising:
spawning a thread to perform database recovery and recording statistics about performance associated with that thread; and
during recovery, dynamically adjusting how many threads are spawned by performing substeps of:
attempting to spawn additional threads to perform database recovery and recording statistics about performance associated with each additional thread spawned;
and if the performance measured for a just-spawned thread degrades beyond a desired amount, freezing execution of the just-spawned thread and ceasing any attempt to spawn additional threads for database recovery;
wherein a user of the system is able to specify a particular number of concurrent threads, and wherein the system generates an advisory if the particular number of concurrent threads specified can be chanced to achieve better I/O performance.
24. The method of claim 23 , wherein performance comprises I/O (input/output) performance measured over a given period of time.
25. The method of claim 24 , wherein the given period of time is about 1 second.
26. The method of claim 23 , wherein the desired amount is no more than about 15 percent degradation in performance.
27. The method of claim 23 , wherein only a certain maximum number of threads may be generated for performing database recovery.
28. The method of claim 27 , wherein the maximum number of threads is limited to not exceed a count of databases that may be opened.
29. The method of claim 27 , wherein the maximum number of threads is limited to not exceed one less than a count of database engines online.
30. The method of claim 23 , further comprising:
after another one of the threads finishes, thawing the thread that has been frozen.
31. The method of claim 23 , wherein the system does not attempt to spawn more threads than a maximum number of databases available to be recovered concurrently.
32. The method of claim 23 , wherein each thread recovers a particular database at a time.
33. A computer-readable medium having processor-executable instructions for performing the method of claim 23 .
34. The method of claim 23 , further comprising:
downloading a set of processor-executable instructions for performing the method of claim 23 .