Unsolved
This post is more than 5 years old
4 Posts
0
539
December 24th, 2016 16:00
Need help identifying problem with POST error message on PE2900-III please...
Last night our PE2900-III server had some strange I/O errors - was unable to unmount an external USB drive, etc. This system runs CentOS 7 with KVM and was host to several VMs.
I decided to investigate this morning, was able to successfully shutdown all the VMs. Shutdown services, and was going to reboot the server but it got stuck during the shutdown unmounting some of the file systems. Using the remote DRAC console, I reset power on it and it started POSTing, but then halted with the following error:
degraded link width error: embedded i/o bridge device 1
expected link width is x4
actual link width is x1
At first, I thought it might be the USB 3.0 PCIe card that was connected to the external USB drive - since that seemed to relate to the 1st event noticed last night. So, I removed that card entirely, but still got the same error. That USB3.0 card was in slot 5, (PCIE_X4_5 according to Dell PE2900 manual). There's another card in slot 6 PCIE_X4_6 that is a SATA controller connected to the boot drive SSDs (mirrored pair). I tried re-seating that card, but same error.
Honestly, I'm not sure which device is "embedded I/O bridge device 1",... I read somewhere that misbehaving PS2/USB adapters can cause similar issues and have unplugged all USB devices; still same error.
My main request here is for help identifying what "embedded I/O bridge device 1" is... can't seem to locate it in the Dell PE2900 manual. If it's some embedded part of the motherboard, at least I can identify what might need replacing here. Or, is it really the SATA controller or the integrated RAID card slot? In my search, the integrated RAID card slot seems to identify itself as such, not as "embedded i/o bridge device"... so I didn't think that would be it.
Would appreciate it if anyone could help me identify the issue.
Daniel My
10 Elder
•
6.2K Posts
0
December 25th, 2016 08:00
Hello
The embedded I/O bridge device 1 does not seem to correlate to a specific slot. According to the system board block diagram it appears to be the south bridge. There is an x4 link from the north bridge to the south bridge. The issue could be with the link between the north and south bridge or one of the links on the south bridge. There are two x4 links connected to the south bridge. Both of them go to the LAN on motherboard.
I would suggest taking the system to minimum-to-POST by removing all unnecessary hardware. If the issue persists the system board will be the most likely component.
Thanks
kmasuda
4 Posts
0
January 26th, 2017 22:00
Daniel: one more thing, although I'm not sure it is worth replacing the motherboard on this old system, but if I decided to go that route, can you help me identify the appropriate Dell part number for this system's motherboard? Specifically, this is a Gen-III of the PE2900, and I know the previous Gen-I and II had different motherboards so I don't want to get confused searching for PE2900 motherboards. Would most appreciate it if you could tell which part numbers to search for.
kmasuda
4 Posts
0
January 26th, 2017 22:00
Thanks Daniel for the reply. I think you were right in that after removing all expansion devices, including the RAID card in the integrated PCI slot, I still got the same error. So, it looks like this motherboard has gone bad all of a sudden.
Daniel My
10 Elder
•
6.2K Posts
0
January 27th, 2017 09:00
The system board part number is NX642. The service kit that contains the system board and the necessary parts for an onsite tech to replace the board is WP039.
Thanks