There are three NVMe controller log pages as follows:
SMART/health information is gathered over the life of the controller and is retained across power cycles. It includes critical warnings about the controller and device status, such as temperature threshold, available spare, device life status, and various I/O statistics used for calculating I/O performance.
Error information is extended error information for commands. A number, which designates the error ID, must be specified with this command.
Vendor Log information is the vendor-specific NVMe log implementation. Use this log when working with Oracle Service to troubleshoot errors.
nvmeadm getlog -h [controller_name]
The following is an example for the nvmeadm getlog -h command:
nvmeadm getlog -h SUNW-NVME-1
SUNW-NVME-1
SMART/Health Information:
Critical Warning: 0
Temperature: 300 Kelvin
Available Spare: 100 percent
Available Spare Threshold: 10 percent
Percentage Used: 0 percent
Data Unit Read: 0x746da4 of 512k bytes.
Data Unit Written: 0x2d0 of 512k bytes.
Number of Host Read Commands: 0xeacba
Number of Host Write Commands: 0x27
Controller Busy Time in Minutes: 0x0
Number of Power Cycle: 0x10d
Number of Power On Hours: 0x3c8
Number of Unsafe Shutdown: 0xfa
Number of Media Errors: 0x0
Number of Error Info Log Entries: 0x0
nvmeadm getlog -e error_id [controller_name]
nvmeadm getlog -s directory [controller_name]
For example, for a controller named SUNW-NVME-1:
nvmeadm getlog -s /logs SUNW-NVME-1
For a Samsung controller, the files crashdump_SUNW-NVME-1 and memorydump_SUNW-NVME-1 are generated and placed under /logs.
For an Intel controller, the files eventlog_SUNW-NVME-1 and nlog_SUNW-NVME-1 are generated and placed under /logs.