DL380 G3 "memory parity error" problems

[Miles Ashley Edwards]

My company has a good amount of the DL380 G3 servers deployed throughout the world in support of our work. There is an increaed trend starting to show with a particular error. In the last 60 days, I have had 12 cases opened for an error that cause a blue screen of death stating "Hardware Malfunction error NMI Parity Check Memory. Please contact hardware vendor."

I can reset the server and have it come up and without warning within one hour to a couple days, it will happen again. Each time I call HP for support, they will never tell me whats wrong and their fix is to just send a new motherboard out.

Could someone tell me if they are experiencing the same problems or know what might me causing this. We are running NT4.0 and the only add-in card is a 64-bit PCI-X dual channel SCSI card. Thank you.

[David_971]

With other proliant is was an issue with PCI cards - netflex network cards in particular.

Check you have the latest firmware/drivers installed.

[Miles Ashley Edwards]

Yes, everything is up to date and current. All drivers, firmware, software versions, other than NT that is. Using the built in Gigabit NIC on the server. Only add in card is the SCSI one as stated earlier.

[David_971]

Don't know if this is any use as it's old info regarding a 1600 but I believe it is a PCI problem.
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=423689

Unfortuantly I have never had this problem on a modern server.

[Oleg Koroz]

Miles, try to disable HT (Hyper Treating) from BIOS, also please check what option you have from BIOS in the memory Protection mode, I don't recall exactly if there any option to disable Parity check, I do assume System ROM is Up to date, and latest system management Driver loaded for NT4. Also is there any chance on the one of the NT4 servers to reload with 2000 and monitor Server. DO you have any box loaded with 2000 or 2003 with same HW configuration where no Problem ever occurred?

Here is more about NMI
http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?objectID=PSD_CN0416W

Is all installed memory original, not mixed by ASSY part number?

http://h18023.www1.hp.com/support/files/server/us/romtabl.html

Is any server continue do so after System Board Replaced ?

[Rob Leadbeater]

Hi,

I've not seen this problem on any of our DL380 G3s however I've had one BL20pG2 do it quite regularly.

My fix for this particular server was to swap the memory.

That said, if HP are insisting on swapping out the motherboards rather than the memory, that implies that they're aware of a problem with that particular board. They're not going to go to the expense of replacing a motherboard if the problem could be fixed by swapping memory.

I take it that the machines with the replacement boards work OK...

Cheers,

Rob

[Stephen Dubrosky]

Miles Ashley, Let me know if you have anything more from HP on this error you described. Anything is helpful. If they do not fix this right away, we are going to dump several multi- million dollar contracts with them this year.

[Miles Ashley Edwards]

No, I have heard nothing from them regarding this matter and it keeps occuring more and more. We are still averaging one server every five days having this error. There answer is to always to just swap out the motherboard and give us no indication as to what the error is. I am getting very fustrated with the lack of support on their end as well. I have tried replacing various parts, disabling HT and none of this seems to work.

They have a similar issue with the SCSI backplanes going bad to giving indication that the hard drives have failed. There is a customer advisory related to this:

http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?objectID

I am very disappointed in HP's service and support in regards to these matters and can not recommend them to anyone because of this.

[Stephen Collopy]

We have matched this error with this event.

Abnormal Program Termination (BugCheck, STOP: 0xC0000244 (0xC0000001, 0x00000000, 0x00000000, 0x00000000))

Please check the following microsoft article.

232564 STOP 0xC0000244 when security log full
http://support.microsoft.com/?id=232564


We are following up with Microsoft regarding this error and the crash event. I will post any update when i recieve one.

Regards
Stephen Collopy

[Stephen Dubrosky]

You mentioned you were using NT 4.0, do you have any 2000 servers with the same problem? There are a lot of patches for 2000 I know (related\unrelated) but it appears those have not solved it, according to many other threads with similar or the same exact issue.