Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.


We look after a Fujitsu RX300 S4 server that has 6 x 500GB SATA drives in a RAID-6 array, running from an LSI MegaRAID card (built into the motherboard).

A couple of weeks ago, one hard drive flagged itself as being faulty (orange light on the drive bay, MegaRAIDcli software shows a firmware status of "Failed"). We ordered and replaced the drive, but after the rebuild started, a different drive flagged itself as faulty.

This has happened 3 times now - twice it flagged up different drives that had a fault, and once it has flagged up a drive that we have already replaced.

At the moment, two drives are showing faults - we don't know if the drives are actually failing, or whether the backplane or RAID card is at fault.

Has anyone experienced this before? Any tips on what to do next? We have a call into Fujitsu, but wondered if anyone out there had any pointers....

share|improve this question
    
Did you buy all 6 hard disks at the same time? If you did, it's feasible to believe that they're all from the same batch, and all liable to fail at the same time. –  Tom O'Connor Sep 8 '10 at 10:59
    
Yeah, all 6 drives came from the same source at the same time. However, we've replaced 3 drives with new units that have come from different sources, 6 months after the initial build, and they are showing faults. –  fistameeny Sep 8 '10 at 11:29

2 Answers 2

I feel for you. This kind of hardware problems are extremely stressful and annoying to debug.

Back in 2002 I had a "joy" of debugging a similar problem. After wayyyyy too much "Let's replace a HD" and similar server massaging the backplane was the actual fault. But that was an IBM server and a completely different story, anyway.

If possible, test the "faulty" drives with another server and see if they are functioning normally there. My guts tell me in your case it's not about the drives, something else is broken. Drives tend not to break like that.

share|improve this answer
1  
Just come off the phone from Fujitsu. They are aiming to replace the backplane and upgrade the firmware on the controller and possibly the drives. Will post back with an update –  fistameeny Sep 8 '10 at 11:28
    
I was just going to suggest that you call them and check on the firmware! Ensure your backups are good, but I wouldn't "test the drives elsewhere" until you know that it's good. Putting the drives in another array will break their association with your current array. If you can live without that server for a while, I'd consider shutting it down. I'd also ask Fujitsu if your data is at risk and if shutting it down would help; escalate to a tech or engineer a few steps up - this isn't a question for the helpdesk! –  gWaldo Sep 8 '10 at 12:14

This could be a faulty controller. It could be unreliable power. It could be bad SATA cables. It could just be extremely bad luck.

share|improve this answer
    
Turns out that for us, it was the XFS filesystem that was the issue. Once we ditched this, the problems went away and hasn't returned. –  fistameeny Aug 29 '11 at 8:13
    
Wow, I tried to list everything and I missed the culprit! –  David Schwartz Aug 29 '11 at 12:02
    
Certain filesystems causing drive problems is still indicative of broken RAID firmware or hidden hardware defects. One thing that remains to try is downgrading the PHY speed on the SATA channels - hardly any non-SSD or 15k (which would usually be sold a SAS only) hard drive will saturate SATA-150 ANYWAY unless the drive is write caching, which it should not be doing or allowed to do ever when attached to a RAID controller. Higher data rates mean less tolerance for cable/connector imperfections. –  rackandboneman May 17 '12 at 15:07

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.