IBM Power System 8335-GCA, Руководство пользователя

Страница: 39 / 174

Use the following table to map the slot or GPU number information in the operating system log to the
GPU description and service action. This ends the procedure.
Table 11. Slot numbers, GPU descriptions, and service action for the 8335-GCA or 8335-GTA
Slot number information from the
log
GPU description Service action
Slot5 GPU 2 Replace the GPU indicated in the
GPU description column. Go to
“8335-GCA and 8335-GTA locations”
on page 111 to identify the physical
location and the removal and
replacement procedure.
Slot2 GPU 1
Table 12. GPU numbers, GPU descriptions, and service action for the 8335-GTB
GPU number information from the
log
GPU description Service action
GPU1 GPU 1 Replace the GPU indicated in the
GPU description column. Go to
“8335-GTB locations” on page 121 to
identify the physical location and the
removal and replacement procedure.
GPU2 GPU 2
GPU3 GPU 3
GPU4 GPU 4
Identifying the location of the NVMe Flash adapter
Use this procedure to identify the location of a Non-Volatile Memory Express (NVMe) Flash adapter.
1.
Does the operating system log contain the slot number? For example, the log might contain an error
message similar to the following text:
[131779.752714] EEH: PHB#0 failure detected, location: Slot1
If Then
Yes:
If your system is an 8335-GCA, use Table 13 on page 24 to map the slot number
information in the operating system log to the PCIe adapter description and service action.
If your system is an 8335-GTB, use Table 14 on page 24 to map the slot number information
in the operating system log to the PCIe adapter description and service action. This ends
the procedure.
No: Continue with the next step.
2.
Locate the NVMe Flash adapter by using the PCI address:
a. The operating system log contains information about the NVMe Flash adapter in the form of a PCI
address. Record the PCI address information for the NVMe Flash adapter that has failed. For
example, in the operating system log message
nvme 0006:01:00.0: Failed status: ffffffff,
reset controller , the PCI address of the failing NVMe Flash adapter is 0006:01:00.0.
b. At the command line, type
lscfg -vl pciaddress
, where
pciaddress
is the NVMe Flash adapter
information that you recorded in step 2.a. Then, press Enter.
c. Record the slot number information that is in the location code field.
d. If your system is an 8335-GCA, use Table 13 on page 24 to map the slot number information to the
PCIe adapter description and service action. If your system is an 8335-GTB, use Table 14 on page
24 to map the slot number information to the PCIe adapter description and service action. This
ends the procedure.
Beginning troubleshooting and problem analysis
23