No relevant resource is found in the selected language.

This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Read our privacy policy>

Reminder

To have a better experience, please upgrade your IE browser.

upgrade

Handling LINK Alarms for the Optical Channel Diagnosis Panel in an RH5485

Publication Date:  2015-06-18 Views:  88 Downloads:  0
Issue Description
Hardware configuration:
RH5485 server, four CPUs, and two QuickPath Interconnect (QPI) wrap cards.

Symptom:

The LINK alarm indicator is lit on the optical channel diagnosis panel in the RH5485.

Figure 1 and Figure 2 shows an optical channel diagnosis panel. The LINK alarm indicator is in the red box.

Figure 1 Optical channel diagnosis panel



Figure 2 Diagram of the optical channel diagnosis panel

Handling Process
The LINK indicator is lit, indicating that server QPI links are abnormal. Figure 3 shows the RH5485 architecture. If over two microprocessors are installed in the server, two QPI wrap cards are required.

Figure 3 RH5485 architecture



1.  If QPI links are abnormal, check the QPI wrap cards. Exchange the left and right QPI cards. If the fault symptom changes, QPI cards are abnormal. If the fault symptom persists, check other components. Check the microprocessors, and focus on whether the CPU socket on the main board has twisted pins. If the CPU socket on the main board has twisted pins, the main board is abnormal. Replace the main board.

2.  Exchange CPU 1 or 2 and CPU 3 or 4. If the fault occurs on the same QPI port, the main board is faulty. If the fault symptom changes on the QPI port that reports errors, the CPU is abnormal.

NOTE:
From the main board structure, CPUs 1 and 4 are in the same channel, and CPUs 2 and 3 are in the same channel. In addition, the QPI cards on the rear of the shelf are respectively responsible for data exchange between CPU 1 or 2 and CPU 3 or 4. The mappings between the QPI link and the CPU are reverse, that is, QPI links 1, 2, 3, and 4 respectively correspond to CPUs 4, 3, 2, and 1.
Root Cause
LINK alarms occur on the optical channel diagnosis panel, indicating that server QPI links are abnormal.

Solution
Check the hardware related to the QPI links, and the hardware includes the CPU, main board, I/O board, and QPI wrap card. After identifying faulty hardware, replace the faulty hardware.

END