DL380 G3 Smart Array 5i controller failure?

[Paul Maglinger]

We have a DL380 G3 server with 3 RAID 1 logical drives locally on the server. This morning we're showing a failure on drive 0, 2, and 5 at the same time. This would indicate to me a problem with the controller as the odds of 3 drives failing at the same time are pretty high. I am curious why 2 of the "failed" drives would be on the lower tier and 1 is on the upper tier. Is it feasible the controller could fail in this manner? A report from the ADU is attached.

[wobbe · Accepted solution]

A backuplane failure sounds more likely to me.

[Perry Lund]

I am having the same issue. Intermittent drive failures and no particular rhyme or reason. I will be trying a replacement PCI-X Smart Array 642 controller to verify. If it is a backplane problem, I wonder if a replacement part can be ordered for the DL380 G3?

[Jake Mason]

289552-001 is your part number for the SCSI backplane.

[Ali]

Hi Paul,

I remember there is a known issue with DL380 G3 SCSI backplane where Servers May Erroneously Report that Internal Hot-Plug Hard Drives Have Failed.

This issue was observed only with Foxconn backplane. so identify whether you have this backplane.. if yes there isn't any second thought ... just replace the backplane.

Examine the backplane connector, which is part of the SCSI backplane board, to determine if it is stamped with the word "Foxconn"
Ref. : EL030501_SW01
http://www.erchi.com/hp_sides/oars/DISK1_CURRENT/ADVISORIES/SERVICE/EL030501_SW01.HTML

if your server is not affected with this issue.. then try these steps:

1. Switch off the server, take out all the cables (power, network etc..), unplug all the harddrive ( a little bit.. don't take it out fully)

2. Reseat SCSI backplane and cables.
3. replug all the hard drive securely.
4. Switch On the server and Run the Array configuration utility to verify the status of logical drive and hard drives.
refer to these steps for reseating the components
http://bizsupport1.austin.hp.com/bizsupport/TechSupport/Document.jsp?objectID=c01898772〈=en&cc=us&taskId=&prodSeriesId=316529&prodTypeId=15351

share your observations...

thanks,
Aftab

[Paul Maglinger]

We ended up replacing the server.