Methods to automatically correct and improve system recovery and replication processes
AssigneeCristie Software Ltd
InventorsIan Cameron, James Reynolds, Scott Sterry
Life of this patent
Estimating the term…
Abstract
A backup recovery and testing method is disclosed. The recovery method restores backed-up systems according to a stored configuration, and collects logs as the restored systems start up. The collected logs are compared with stored baseline logs, by a process which includes parsing the logs into templates and parameters, and identifying anomalies automatically for example by principal component analysis. Anomalies identified may be used to modify the stored configuration.
Claims18
A method of recovering a backed-up computer system, the method comprising:
- creating a virtual environment in which to restore the system, the virtual environment being created according to a stored configuration;
- restoring the system from backup or replicated data to the virtual environment;
- starting up the restored system in the virtual environment, and collecting logs from the restored system;
- automatically comparing the collected logs to stored baseline logs to identify anomalous events corresponding to error conditions;
- making changes to the stored configuration according to the identified anomalous events;
- repeating the above steps according to the new stored configuration, until no anomalous events corresponding to error conditions are identified.
The method of recovering multiple backed-up computer systems, the method comprising:
- creating a virtual environment in which to restore the systems, the virtual environment being created according to a stored configuration;
- for each system, restoring the system from backup files to the virtual environment;
- starting up the restored systems in the virtual environment according to a dependency graph forming part of the stored configuration, in which each system is started if and only if any required systems are already started and healthy, and for each system collecting logs from the restored system;
- for each system, automatically comparing the collected logs to stored baseline logs to identify anomalous events corresponding to error conditions;
- for each system, making changes to the stored configuration according to the identified anomalous events;
- repeating the above steps according to the new stored configuration, until no anomalous events corresponding to error conditions are identified, and when no anomalous events are identified, issuing a notification that the system is healthy to trigger startup of any dependent systems.
Dependencies are inferred from the claim text.