Predicting and handling of slow disk
Embodiments of the present disclosure provide a computer-implemented method, an electronic device and a computer program product. The method comprises: obtaining historical data of recoverable errors that occurred in a storage disk during a first period in the past. The method also comprises: determining, based on the historical data, a predicted number of recoverable errors to occur in the storage disk during a second period. The first period has a same duration as that of the second period. The method further comprises: in response to the predicted number exceeding a threshold for identifying a slow disk, performing an operation for handling a slow disk on the storage disk.
1. A computer-implemented method, comprising:
obtaining historical data of recoverable errors that occurred in a storage disk during a first period;
determining, based on the historical data, a predicted number of recoverable errors to occur in the storage disk during a second period, wherein the first period having a same duration as that of the second period and wherein the second period is after the first period, and wherein determining the predicted number comprises:
determining, from the historical data, an attribute of the recoverable errors that occurred in the storage disk during the first period; and
determining the predicted number based on an association between the predicted number and the attribute;
in response to the predicted number exceeding a threshold, determining that the storage disk is a slow disk; and
performing, based on the determination, an operation on the storage disk.
2. The method of claim 1 , wherein the association is determined based on historical data of the recoverable errors that occurred in the storage disk prior to the first period.
3. The method of claim 1 , wherein the attribute comprises at least one of:
a number of the recoverable errors that occurred in the storage disk during a period having the duration,
a change rate of the number within a predetermined time window, and
a variance of the number within the predetermined time window.
4. The method of claim 1 , wherein the association is determined by a machine learning model.
5. The method of claim 4 , wherein a length of a time window for determining a change rate of a number of the recoverable errors is used as a hyper-parameter of the machine learning model to optimize the machine learning model.
6. The method of claim 4 , wherein the machine learning model comprises a random forest regression model.
7. The method of claim 1 , wherein obtaining the historical data comprises:
obtaining raw data including configuration information, runtime statistics, and event logs; and
extracting the historical data from the raw data.
8. The method of claim 1 , wherein performing the operation comprises at least one of:
sending alarm information indicating that the storage disk is to become the slow disk;
labeling the storage disk as a failed disk to trigger replacement of the storage disk; and
in response to receiving a request for reading data from the storage disk, providing the data from another storage disk to avoid reading the storage disk.
9. An electronic device, comprising:
at least one processor; and
at least one memory storing computer program instructions, the at least one memory and the computer program instructions being configured, with the at least one processor, to cause the electronic device to:
obtaining historical data of recoverable errors that occurred in a storage disk during a first period;
determining, based on the historical data, a predicted number of recoverable errors to occur in the storage disk during a second period, wherein the first period has a same duration as that of the second period and wherein the second period is after the first period, and wherein determining the predicted number comprises:
determining, from the historical data, an attribute of the recoverable errors that occurred in the storage disk during the first period; and
determining the predicted number based on an association between the predicted number and the attribute;
in response to the predicted number exceeding a threshold, determining that the storage disk is a slow disk; and
performing, based on the determination, an operation on the storage disk.
10. The electronic device of claim 9 , wherein the association is determined based on historical data of the recoverable errors that occurred in the storage disk prior to the first period.
11. The electronic device of claim 9 , wherein the attribute comprises at least one of:
a number of the recoverable errors that occurred in the storage disk during a period having the duration,
a change rate of the number within a predetermined time window, and
a variance of the number within the predetermined time window.
12. The electronic device of claim 9 , wherein the association is represented by a machine learning model.
13. The electronic device of claim 12 , wherein a length of a time window for determining a change rate of a number of the recoverable errors is used as a hyper-parameter of the machine learning model to optimize the machine learning model.
14. The electronic device of claim 12 , wherein the machine learning model comprises a random forest regression model.
15. The electronic device of claim 9 , wherein the at least one memory and the computer program instructions are further configured, with the at least one processor, to cause the electronic device to:
obtain raw data including configuration information, runtime statistics, and event logs; and
extract the historical data from the raw data.
16. The electronic device of claim 9 , wherein performing the operation comprises at least one of:
sending alarm information indicating that the storage disk is to become the slow disk;
labeling the storage disk as a failed disk to trigger replacement of the storage disk; and
in response to receiving a request for reading data from the storage disk, provide the data from another storage disk to avoid reading the storage disk.
17. A computer program product being tangibly stored on a non-volatile computer-readable medium and comprising machine-executable instructions which, when being executed, cause a machine to perform steps of a method, the method comprising:
obtaining historical data of recoverable errors that occurred in a storage disk during a first period;
determining, based on the historical data, a predicted number of recoverable errors to occur in the storage disk during a second period, wherein the first period having a same duration as that of the second period and wherein the second period is after the first period;
in response to the predicted number exceeding a threshold, determining that the storage disk is a slow disk; and
performing, based on the determination, an operation on the storage disk.
18. The computer program product of claim 17 , wherein performing the operation comprises at least one of:
sending alarm information indicating that the storage disk is to become the slow disk;
labeling the storage disk as a failed disk to trigger replacement of the storage disk; and
in response to receiving a request for reading data from the storage disk, providing the data from another storage disk to avoid reading the storage disk.