Simulation of failure recovery within clustered systems
View Patent ↗Failure recovery within clustered systems is simulated. For each of a number of failure conditions for an initial state of a number of computing elements of a computerized system, a failure state of the computing elements is generated that corresponds to the failure condition and that is based on the initial state of the computing elements. A failure condition may include one or more hardware and/or software failures. For each failure state of the computing elements, a recovery state is then generated, or simulated, for the computing elements, based on the failure state and the initial state of the computing elements and on a number of recovery rules for the computing elements. Each recovery state is then output for subsequent analysis, where such analysis may be either with human interaction, or automatically and without human interaction.
1. A method comprising:
for each of a plurality of failure conditions, generating a failure state of a plurality of computing elements of a computerized system corresponding to the failure condition and based on an initial state of the plurality of computing elements of the computerized system, the failure conditions regarding a plurality of failures of the plurality of computing elements,
wherein each failure comprises one of a single hardware failure within the plurality of computing elements and a single software failure within the plurality of computing elements, the failures comprising at least three failures;
wherein the failure conditions comprise all possible combinations of one, two, and three of the failures;
for each failure state of the plurality of computing elements of the computerized system, generating a recovery state of the plurality of computing elements of the computerized system based on the failure state and the initial state of the plurality of computing elements of the computerized system and on a plurality of recovery rules for the plurality of computing elements of the computerized system; and,
outputting each unique recovery state of the plurality of computing elements of the computerized system for subsequent analysis of the recovery state,
wherein one of the recovery rules comprises doing nothing.
2. The method of claim 1 , further comprising initially inputting the initial state of the plurality of computing elements of the computerized system.
3. The method of claim 1 , further comprising initially generating the plurality of failure conditions for the initial state of the plurality of computing elements of the computerized system.
4. The method of claim 1 , wherein at least one of the failure conditions each comprises a hardware failure within at least one of the plurality of computing elements of the computerized system.
5. The method of claim 1 , wherein at least one of the failure conditions each comprises a software failure within at least one of the plurality of computing elements of the computerized system.
6. The method of claim 1 , wherein at least one of the failure conditions each comprises a single failure within one of the plurality of computing elements of the computerized system.
7. The method of claim 1 , wherein at least one of the failure conditions each comprises a multiple failure within one of the plurality of computing elements of the computerized system.
8. The method of claim 1 , wherein each failure state represents a state of the plurality of computing elements of the computerized system based on the corresponding failure condition occurring to the initial state of the plurality of computing elements of the computerized system.
9. The method of claim 1 , wherein each recovery state of the plurality of computing elements of the computerized system represents a state of the plurality of computing elements of the computerized system based on applying the recovery rules to the failure state of the plurality of computing elements of the computerized system.
10. The method of claim 1 , further comprising determining whether each recovery state is appropriate for the failure state to which the recovery state corresponds.
11. A method comprising:
for each of a plurality of failure conditions for an initial state of a plurality of computing elements of a computerized system, generating a failure state of the plurality of computing elements representing a state of the plurality of computing elements corresponding to the failure condition occurring to the plurality of computing elements in the initial state, the failure conditions regarding a plurality of failures of the plurality of computing elements,
wherein each failure comprises one of a single hardware failure within the plurality of computing elements and a single software failure within the plurality of computing elements, the failures comprising at least three failures;
wherein the failure conditions comprise all possible combinations of one, two, and three of the failures;
for each failure state of the plurality of computing elements, generating a recovery state of the plurality of computing elements by applying a plurality of recovery rules to the plurality of computing elements in the failure state, and based on the initial state of the plurality of computing elements, to yield the recovery state; and,
outputting each unique recovery state of the plurality of computing elements for subsequent analysis of the recovery state,
wherein one of the recovery rules comprises doing nothing.
12. The method of claim 11 , further comprising initially:
inputting the initial state of the plurality of computing elements of the computerized system; and,
generating the plurality of failure conditions for the initial state of the plurality of computing elements of the computerized system.
13. The method of claim 11 , wherein each failure condition comprises at least one of:
a hardware failure within at least one of the plurality of computing elements;
a software failure within at least one of the plurality of computing elements;
a single failure; and,
a multiple failure.
14. The method of claim 11 , further comprising determining whether each recovery state is appropriate for the failure state to which the recovery state corresponds.
15. A computerized system comprising:
hardware;
a failure mechanism implemented at least in the hardware and to generate a plurality of failure states of a plurality of computing elements by applying user-specified failure criteria to an initial state of the plurality of computing elements, the failure criteria regarding a plurality of failures of the plurality of computing elements,
wherein each failure comprises one of a single hardware failure within the plurality of computing elements and a single software failure within the plurality of computing elements, the failures comprising at least three failures;
wherein the failure criteria comprise all possible combinations of one, two, and three of the failures; and,
a simulation mechanism implemented at least in the hardware and to generate a plurality of recovery states of the plurality of computing elements corresponding to the failure states by applying a plurality of recovery rules to the failure states and based on the initial state of the plurality of computing elements, and to output the unique recovery states for subsequent analysis,
wherein one of the recovery rules comprises doing nothing.
16. The computerized system of claim 15 , wherein each failure state represents a failure occurring within the plurality of computing elements in the initial state.
17. The computerized system of claim 16 , wherein the failure of each failure state comprises at least one of:
a hardware failure within at least one of the plurality of computing elements;
a software failure within at least one of the plurality of computing elements;
a single failure; and,
a multiple failure.
18. An article of manufacture comprising:
a tangible computer-readable data storage medium; and,
means in the medium for generating a plurality of failure states of a plurality of computing elements of a computerized system from an initial state of the plurality of computing elements by applying user-specified failure criteria to the initial state, for generating a plurality of recovery states of the plurality of computing elements corresponding to the failure states based at least on an initial state of the plurality of computing elements by applying a plurality of recovery rules to the failure states, and for outputting the recovery states for subsequent analysis, the failure criteria regarding a plurality of failures of the plurality of computing elements,
wherein each failure comprises one of a single hardware failure within the plurality of computing elements and a single software failure within the plurality of computing elements, the failures comprising at least three failures;
wherein the failure criteria comprise all possible combinations of one, two, and three of the failures;
and wherein one of the recovery rules comprises doing nothing.