Symptoms
The show disk general command is executed on the CLI of the disk array to display the Health Mark of a disk, as shown in figure 1. The Toolkit is used to collect information about the disk by one click and a disk health analysis tool is used to calculate the health mark, as shown in figure 2. However, the displayed marks are different using the two methods.
Figure 1 Disk health mark (CLI)
Figure 2 Disk health mark (Toolkit)
Cause
The disk health analysis tool analyzes the disk offline using logs (using the offline evaluation standard; the result indicates the disk status when information was collected). This calculation method is used to identify risky disks so that the disks can be replaced in time. The Health Mark displayed on the CLI by running the show disk general command is calculated online by the DHA module of the disk array based on disk parameters (using the online evaluation standard; calculation is performed at 02:00 every day when the disk array is running). This calculation method is used to identify risky disks so that the disks can be pre-copied online. As the two calculation methods are different in terms of the evaluation mode, scoring standard, and parameter values, and some parameter values cannot be obtained offline (such as I/O error code that is reported in real time and disk utilization and pressure), so the marks are different, but this has minor impact on the system.
Solution
Use both calculation methods to identify risky disks. A disk is identified as a risky one when either of the methods finds its mark exceeding the threshold. Figure 3 shows the thresholds of the CLI calculation method (thresholds are set based on the service years; replace the disk if its mark is lower than the respective threshold). Figure 4 shows the thresholds of the calculation method using the disk health analysis tool.
Figure 3 Thresholds of the health mark (CLI)
Figure 4 Thresholds of the health mark (Toolkit)
In OceanStor V300R003C10 and later, the offline and online evaluation functions are combined into one. The prediction accuracy is the same for SAS disks, though the marks are different (for example, the marks for disk service pressure and running time displayed on the CLI are lower than those displayed on the disk health analysis tool).
If a large number of disks have low marks or the mark difference is big (more than 30) using the two calculation methods, send the logs collected by one click to the disk department for further analysis.
Application Scope
OceanStor V300R001, OceanStor V300R002, OceanStor V300R003 and OceanStor V300R005
