Note
For V4.4.1, this parameter is available starting with V4.4.1.
Description
log_storage_warning_trigger_percentage specifies the write performance threshold for triggering a log disk failure.
Privilege requirements
Query the parameter
The
systenant and all user tenants can use theSHOW PARAMETERSstatement or theGV$OB_PARAMETERSview to query this parameter.Modify the parameter
Only the
systenant can modify this parameter. User tenants cannot modify this parameter.
Attributes
| Attribute | Description |
|---|---|
| Type | Int |
| Default value | 0 |
| Value range | [0, 50]. We recommend that you set the value to a value less than 10. |
| Modifiable | Yes. You can use the ALTER SYSTEM SET statement to modify the parameter. |
| Effective upon OBServer node restart | No. |
Considerations
The log disk failure detection algorithm continuously collects baseline data of the log disk write performance, monitors the current log disk write performance, and compares it with the baseline performance. If the current write performance drops significantly compared to the baseline performance and does not recover within a certain period, the log disk is marked as failed.
The relationship between the
log_storage_warning_trigger_percentageparameter and the log_storage_warning_tolerance_time parameter is as follows:- When the value of
log_storage_warning_trigger_percentageis 0, the performance data is not collected. If the log write latency exceeds the value of thelog_storage_warning_tolerance_timeparameter, the log disk is marked as failed. - When the value of
log_storage_warning_trigger_percentageis greater than 0, if the current log disk write performance drops below the baseline performance multiplied by the percentage specified by thelog_storage_warning_trigger_percentageparameter and the performance drop lasts forlog_storage_warning_tolerance_timeseconds, the log disk is marked as failed. log_storage_warning_trigger_percentagespecifies the write performance threshold for triggering a log disk failure. A larger value means a more sensitive failure detection algorithm.log_storage_warning_tolerance_timespecifies the duration of the performance drop. A larger value means a more robust detection algorithm.
- When the value of
Examples
Set the write performance threshold for triggering a log disk failure to 5% and the performance drop duration threshold to 3 seconds. If the log disk write performance drops below 5% of the baseline performance and the performance drop lasts for 3 seconds, the log disk is marked as failed.
obclient> ALTER SYSTEM SET log_storage_warning_trigger_percentage = 5;
obclient> ALTER SYSTEM SET log_storage_warning_tolerance_time = '3s';