1

I have an older xSeries 306 that has started giving me a SMART Failure Predicted on Hard Disk 2: SATA Port 0 message, a warning about backing up, and a message to Press F1 to continue. It's my understanding that SMART is actually part of the hard drive itself but this happens with hard drives that work fine in other systems. I've put a new hard drive in and am attempting to reinstall server 2003 (still get the SMART warning). The installer sees the hard drive and I get as far as formatting the drive. It just hangs at 0%. I assume this isn't really the the computer sensing a SMART problem but a problem with the server itself. Any ideas other than a bad motherboard of what could cause this?

By the way, this system only has 1 hard drive. I've never noticed it being referenced as "hard disk 2: SATA Port 0". Is that normal?

thanks

Jeff
  • 277
  • 3
  • 13
  • 1
    What causes the problem is a hard disk that is potentially about to go bad. Given the inconvenience of a hard drive failure plus the cost of your hourly rate for the time spent messing about, frankly I think that when you see something like this its worth just replacing the hard disk and not worrying too much more about it. – Rob Moir Dec 22 '11 at 16:02

4 Answers4

1

Those SMART alerts are showing up for a reason.

Yes, the hard drive itself has the SMART circuitry/logic on it. It measures many different performance metrics such as spin-up time, bad sectors, etc. etc. There are myriad free tools online to easily pull this data and see for yourself.

The Windows installer doesn't care about the SMART data.

Ryan Ries
  • 55,011
  • 9
  • 138
  • 197
1

SMART variables are aprt of the drive firmware. When certain parameters go over defined limits, the machine is supposed to start crying. This has NOTHING to ddo on an OS level. It is totalyl in the drive firmware.

This does NOT mean teh disc is crap - it means the disc starts having issues. It can work well for years to come, or fail within days.

So, for pro machines / servers where data integrity counts, it is practice to relpace the disc the moment SMART starts crying IF you validate it is a SMART issue (could be abus / controller issue - seen one disc interfere with other ports ona a SAS bus).

I've never noticed it being referenced as "hard disk 2: SATA Port 0". Is that normal?

Possibly. Check what else is there according to BIOS. For example my main server has 2 keyboards - a real one, one from the lights out management that comes via a "fake" usb port. Maybe you have some "ghost" hd 0? DVD drive?

TomTom
  • 50,857
  • 7
  • 52
  • 134
1

Yes, SMART metrics are gathered by the drive firmware, and as such are independent of which system they are put in.

However, the fact that some systems complain and some do not has nothing to do with any of the above-mentioned reasons.

It is simply the monitoring OS or application that decides what those metrics mean.

EDIT: well, not entirely - SMART also provides a single-value PASS/FAIL status that the OS or application can check, and that status is determined by the drive firmware.

However, any monitoring solution worth its salt will not only use that, but also interpret the actual values from the SMART table.

adaptr
  • 16,479
  • 21
  • 33
0

Others have rightly noted that SMART is an on-drive function. It may be that other chassis don't complain about this HDD because they don't check its SMART status at boot.

As I have noted in an earlier answer, Google famously released a paper some years back about (inter alia) SMART monitoring, and one conclusion was that the presence of SMART errors correlates strongly with incipient drive failure.

Of course, this is only a statistical argument, but if I were you I'd put that HDD into a Linux box, use smartctl to confirm that the drive really is reporting predicted failure, and immediately retire the drive.

Community
  • 1
MadHatter
  • 78,442
  • 20
  • 178
  • 229