The hwdiag io nvme_test displays Unable to read FRU after Oracle Flash Accelerator F640 PCIe Card v1 firmware is updated to RD28. Oracle 6.4 TB NVMe SSD displays fault.io.scsi.cmd.disk.dev.rqs.baddrv on RD28 Oracle Flash Accelerator F640 PCIe Card v1s firmware when upgrading firmware to RD28 on all platforms.
Oracle ILOM monitors this condition and reports card fault due to MIC (I2C bus) corruption that occurs during an Oracle firmware upgrade. The MIC (management Interface controller) corruption event is not visible if the system does not use the I2C bus with Oracle ILOM.
An example for Oracle Flash Accelerator F640 PCIe Card v1s follows:
#hwdiag io nvme_test
Checking NVME drive FRU contents...
checking fru on drive PCIe 1.0 NVME OK
checking fru on drive PCIe 1.1 NVME OK
checking fru on drive PCIe 15.0 NVME FAILED, Unable to read FRU for P7_NVME_PCIE1 drive
checking fru on drive PCIe 15.1 NVME FAILED, Unable to read FRU for P7_NVME_PCIE1 drive
NVME drives FRU check: FAILED
Checking NVME drive pcie links...
checking pcie link on drive PCIe 1.0 NVME OK
checking pcie link on drive PCIe 1.1 NVME OK
checking pcie link on drive PCIe 15.0 NVME OK
checking pcie link on drive PCIe 15.1 NVME OK
NVME drives pcie link check: PASSED
Checking NVME drive DSN...
checking DSN on drive PCIe 1.0 NVME OK
checking DSN on drive PCIe 1.1 NVME OK
checking DSN on drive PCIe 15.0 NVME FAILED, Unable to read FRU for P7_NVME_PCIE1 drive
ERROR: Failed reading FRU DSN on drive PCIe 15.0 NVME
checking DSN on drive PCIe 15.1 NVME FAILED, Unable to read FRU for P7_NVME_PCIE1 drive
ERROR: Failed reading FRU DSN on drive PCIe 15.1 NVME
FAILED
NVME test FAILED
Workaround:
During the upgrade, the AIC may fail to respond on the SMbus to ILOM, which will report unable to read FRU. If this happens, the system will need to be AC cycled.
To recover the Oracle Flash Accelerator F640 PCIe Card v1 management controller on all platforms after upgrading drive firmware to RD28, AC power cycle the server, which includes reenergizing the 3Vaux supply. Completely power off the server chassis to power state 0 and disconnect all AC power cords from the server. Refer to the server documentation for instructions.