Absolute nightmare of a DL380 G7

[Michael A. McKenney]

UPS could cause alot of issues. You replaced it. I would go for the server board first. You did the CPUs. Then reduce the RAM to minimal config and do the firmware on the everything.

[BigLupu]

Replaced the motherboard on Saturday and changed BIOS settings per advice from HP support:

- Advance Power option -> change to = HP Static High Performance Mode.

- Minimum Processor Idle Power State -> No C-state

No machine check exceptions so far, but we'll wait several weeks now before jumping into any conclusions.

[Michael A. McKenney]

I want to know did HP adopt Dell standards for quality and reliability? I expect a lot more out of HP. I have been very frustrated with them. I have asked for a manager to call me. No calls.

[Mauricio Salas]

BSOD 0xF4 is when an important process is terminated OR when the HDD are not being accessible, so for example if you lost conection with your Hard Drives because you are booting from iSCSI, MSA, NAS or any other External storage and you loss the conection because of a fail cable/Hardware issue, or if the Smart Array fails in that moment (that will explain why there are no Dump files, because the server is not able to writte on them) so recommend to check cables, update FW of the Smart Array and if you are onsite when that happen go and check if the serve have lights on the HDD, and also check the ADU report! you could see several Hot-Plug logs on it, just like if someone is reseating the drives

[BigLupu]

Our DL380 G7 has been stable for three and half weeks now after replacing the motherboard and switching to these BIOS settings:

- Advance Power option -> change to = HP Static High Performance Mode.

- Minimum Processor Idle Power State -> No C-state

[frankd_1]

We have numerous customers worldwide running these systems and they all see this randomly. I have updated to all the latest FW for BIOS, iLO and P410. I also have the recommended BIOS setting yet they will randomly get a burst of server reboots and then the problem is gone for weeks? I am starting to think that this platform with the P410 is not stable and probably never will be.

[frankd_1]

We have solved this for our servers. There is a bug in the bnx2 public drivers that can be hit if you have unused netwrok ports. So if you system has 4 ports and you are only using two you will see the problem with random server reboots. Just connect the ports to a switch, they need not be configured.

[Wilson010]

Hi Martijn,

I got similiar issue too.

After replace the 1GB BBWS module, is everything okay then?

Thanks.

-Wilson

[battar]

Hi everybody,

Even i am also facing the same problem like others in this blog. I had got total 3nos dl380 g7in which 2nos r running in redhatlinux clustering. Everytime the machine checksum is occurring on either of the server running in clustering  gets rebooted automatically with amberlight. Amberlight gets vanished if rebooted agian manually. But the 3rd machine which is runing out of cluster is fine.

At the same we cannot suspect the os level or cluster config level (as per my assumption and redhat reverted to check with manufacture). Becasue everytime the server reboot it shows machine cheskum error. Attached the iml report.

Workarounds carried out at the site. Firmware 9.30 applied

bios modifciation carried out to Set "Minimum Processor Idle Power State" and enter to "No C states". and HP power profile to "Maximum Performance".still the same problem gets existed.

Presently to be frank psp is not updated, because we think that updating psp may overwrite the kernel of redhat. Is it real, pls clarify.

No part has been replaced still yet.

Os is Redhat Linux 6.1 version.

Advance thanks