DL380G7 Getting "Uncorrectable QuickPath Interconnect Error "

[fmash]

Uncorrectable QuickPath Interconnect Error (Local interface error status 0x00000000, Protocol error status 0x00000001)

I am getting these errors followed by an ASR every few hours on a new DL380 G7 running 2008R2 with PSP 8.60 and MS SQL. I've searched here and Google and I am not getting any bites on this error or where to go from here.
Once I am able to roll the services over to the other node, I'll update any drivers/bios to the absolute latest released after PSP 8.60.

[Zygmunt Krawczyk]

Try to update BIOS to ver 2010.12.01 and restore the factory defaults BIOS settings. If this will not help, contact HP service.

[fmash]

Sorry, it is a 580, not a 380.

I was able to get it updated to the latest firmware (including 2010.12.01 bios) and all of the drivers. Hopefully that works. We have other 580 G7's with an older bios and this is the only one that is doing it.

[kris rombauts]

This error indicates a communication issue between the QPI links on the processors, so either remove/reseat the processors and if that does not address the issue you should place a H/W service call with HP to get a part replaced.


HTH

Kris

[Chris Wotton]

I'm getting a similar issue on the HP DL580 G7, 2008 R2 Enterprise, PSP8.60, FW9.20B - every 3 days or so, I get an error :

Uncorrectable QuickPath Interconnect Error (Local interface error status 0x00000000, Protocol error status 0x00000001)

Followed by an ASR, currently talking to HP hardware support for a fix (so far they have only suggested updating firmware which has already been done).

[Zygmunt Krawczyk]

Chris,
you need to update to BIOS version 2010.12.01. This version is not included in FW9.20B bundle. It is available from

http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareDescription.jsp?lang=en&cc=pl&prodTypeId=15351&prodSeriesId=4142916&prodNameId=4142792&swEnvOID=4049&swLang=8&mode=2&taskId=135&swItem=MTX-7a77f2c4718e45bda62b5dfe7f

[fmash · Accepted solution]

I did the same thing with bios versions, bare configurations, removing memory, reseating cards, even iLo firmware and running memtest etc..

When the ticket finally went to L2 I got this email:

"The Xeon processor let the QuickPath Interconnect (QPI) buffers enter a sleep state to reduce power requirements while the QPI links are not active. HP enables this Intel feature for
G6 and G7 servers through RBSU. Once this feature is enabled, the Intel processor determines when to put the QPI buffers into a sleep state. It appears that QPI power management has no measureable impact on performance.

This error indicates a communication issue between the QPI links on the processors, We need to figure out if the issue is hardware or just RBSU configuration."

The next day, the sys board and two of the four cpus were replaced and its been running for four straight days so far.....