Oracle Server X7-8
 
Oracle Server X7-2
 
Oracle Server X7-2L
 
Oracle Server X7-2c
 

PLA-Support>Sun Systems>Sun_Other>Sun Collections>SN-OTH: Sun PSH
 

In this Document

Purpose

Details

References

Applies to:

Purpose

 Provide additional information for Broadcom 25-GB Network Devices Completion Timeout Errors

Details

Broadcom 25-GB network devices cause completion timeout errors during a warm reset

Type

Defect

fault.io.intel.iio.pcie-fatal

Severity

Minor

Description

The Broadcom 25Gb network devices may cause completion timeout errors to occur, under some circumstances, when the system is performing a warm reset.

The completion timeout errors are logged after the link is brought down and don't have any functional impact on system operation.

A workaround has been added to ILOM to not diagnose a fault when these completion timeout errors occur.

Reference Oracle BugID# 26385235 - ereport.io.intel.iio.pcie-completion-timeout-on-np-transactions

ILOM will not diagnose a fault for completion timeouts observed by downsteam devices,
but this error will be visible in the ereport logs when running "fmdump -eV".

The ereports that match the criteria described below are expected to be logged as a result of this issue.

An example ereport observed on an Oracle Server X7-2's onboard 25Gb NIC is show below:

ereport.io.intel.iio.pcie-completion-timeout-on-np-transactions@/SYS/MB/P0/IIO
1/DEV00/FN0/DEV00/FN0
port = NET 1-2
slot_path = /SYS/MB:/SYS/MB/P0:/SYS/MB/NET1

An example ereport observed on an Oracle Sever X7-2C's obboard 25Gb NIC is show below:

ereport.io.intel.iio.pcie-completion-timeout-on-np-transactions@/SYS/SM0/P0/IIO2/DEV00/FN0/DEV00/FN0
port = NET0/1
slot_path = /SYS/SM0:/SYS/SM0/P0:/SYS/SM0/NET0

This problem may also be observed when using the Broadcom 25Gb Network Adapter Add-in cards.

An example ereport observed with a 25Gb AIC in Slot 1 of an Oracle Server X7-2 is show below:

ereport.io.intel.iio.pcie-completion-timeout-on-np-transactions@/SYS/MB/P1/IIO
2/DEV00/FN0/DEV00/FN0
port = PCIe 1
slot_path = /SYS/MB:/SYS/MB/P1:/SYS/MB/RISER1/PCIE1

The presence of two "/DEVx/FNy" nodes in the device paths of these ereports indicates
the error was detected by the endpoint device and not by the processor's root port.

To confirm that a completion timeout error associated with an Add-in Card slot may be associated with this issue
one then needs to examine the Add-in card inventory and confirm that the slot identified in the ereport is an
"Oracle Dual Port 25 Gb Ethernet Adapter".

This can be done as follows:

-> show /System/PCI_Devices/Add-on/Device_1

/System/PCI_Devices/Add-on/Device_1
Targets:

Properties:
part_number = 7339763
description = Oracle Dual Port 25 Gb Ethernet Adapter
location = RISER1/PCIE1 (PCIE on Riser 1)
pci_vendor_id = 0x14e4
pci_device_id = 0x16d7
pci_subvendor_id = 0x108e
pci_subdevice_id = 0x3044

Automated Response

The completion timeout errors are logged after the link is brought down by the warm reset.

Impact

The completion timeout errors do not have any functional impact on the server as this only occurs during the warm reset process.

Suggested Action for System Administrator

A workaround has been added to ILOM to not diagnose a fault when these completion timeout errors are observed.

Refer to the following document for the latest procedures for displaying event content in preparation
for submitting a service request and applying any post-repair actions that may be required.

PSH Procedural Article for ILOM-Based Diagnosis (Doc ID 1155200.1)

This solution has no attachment