Proactive detection of data inconsistencies in a storage system point-in-time copy of data

A system for detecting a data inconsistency in a point-in-time copy of data, comprising:

• a computer-readable medium having computer program instructions;
• and
• a processor executing the computer program instructions to:
• receive event information from systems and applications associated with a plurality of point-in-time copies of data, each of said copies being a data image of a storage volume's state at a particular time;
• correlate events from the event information with at least one said copies;
• determine that at least event is associated with a data inconsistency, said determine being based on a probability that there is a data inconsistency with the occurrence of a particular event where the probability is based on an association of the particular event with a previously encountered data inconsistency;
• and
• determine that at least one event is associated with a data inconsistency, including:
• identify, at least one of said copies correlated with an event determined to be associated with a data inconsistency as a candidate point-in-time copy of data, for testing for a data inconsistency;
• select, particular tests for testing said candidate to determine whether said candidate has a data inconsistency, the selecting being based on a correlation between the event associated with the data inconsistency and a probability that each of the tests will detect the data inconsistency;
• schedule testing of said candidate using said tests, the scheduling includes prioritizing each of the selected tests based on an estimated run time;
• test said candidate with at least one of the tests selected to determine if said candidate has a data inconsistency;
• determine that the testing results with at least one data inconsistency, and correct the at least one data inconsistency and provide the corrected point-in-time copy of data;
• monitor performance of the tests and a production application associated with said point-in-time copy of data;
• and
• if performance of the production application is below a pre-determined threshold, the tests are throttled to reduce performance impact on the production application.

A computer program product for detecting a data inconsistency in a point-in-time copy of data, the computer program product comprising a non-transitory computer readable storage medium having computer readable program code embodied therewith, said program code being executable by a processor to:

• receive event information from systems and applications associated with a plurality of point-in-time copies of data, each of said copies being a data image of a storage volume's state at a particular time;
• correlate each event from the event information with at least one said copies;
• determine that at least event is associated with a data inconsistency, said determining by using probability that there is a data inconsistency with the occurrence of a particular event where the probability is based on an association of the particular event with a previously encountered data inconsistency;
• and
• determine that at least one event is associated with a data inconsistency, including:
• identify, at least one of said copies correlated with an event determined to be associated with a data inconsistency as a candidate point-in-time copy of data, for testing for a data inconsistency;
• select, particular tests for testing said candidate to determine whether said candidate has a data inconsistency, the selecting being based on a correlation between the event associated with the data inconsistency and a probability that each of the tests will detect the data inconsistency;
• schedule testing of said candidate using said tests, the scheduling includes prioritizing each of the selected tests based on an estimated run time;
• test said candidate with at least one of the tests selected to determine if said candidate has a data inconsistency;
• determine that the testing results with at least one data inconsistency, and correct the at least one inconsistency and provide the corrected point-in-time copy of data;
• monitor performance of the tests and a production application associated with said point-in-time copy of data;
• and
• if performance of the production application is below a pre-determined threshold, the tests are throttled to reduce performance impact on the production application.