Welcome to the MacNN Forums.

IYork · Feb 1, 2005, 11:05 AM

I have two drives attached to a PCI controller (SIIG ATA 100/133) One of the drives sounds like it's planning to soon kick the bucket. It would be nice if I could run SMART diagnostics for these drives, but there doesn't seem to be any program that can read SMART data from non-motherboard controlled drives. (not Diskwarrior, not Apple's Disk Utils)
If anyone knows differently, I'd much appreciate hearing about it.
The drives are both larger than 127GB, so I'm hesitant to attach them to the main controller to check them out. Would anyone advise differently? I presume that if I did attach them, even if I wrote no data to them, the OS would do so and the disks would be, to use a technical term, 'hosed'?
Thoughts appreciated.

reader50 · Feb 1, 2005, 11:26 AM

OSX will not write data to a non-boot drive unless a finder window is opened from that drive, in which case it may update the .DS_Store file for that window. If you have a few free blocks within the first 128GB of the drive, that isn't likely to damage anything.

OS9 might try to update the desktop database, which involves a scan of the entire drive, that could produce unhelpful results. So don't boot into OS9 while trying this, and turn Classic off just in case.

A drive normally attached to a PCI card is likely to have SCSI drivers, so it may not mount when attached to the onboard ATA anyway.

SMART is a function of the hardware, so it will not matter if the drive mounts or not. Close all Finder windows from that drive, shut down, move to ATA bus, doublecheck jumpers, reboot. Let Disk Utility check the drive's SMART status on the drive, mounted or not.

tooki · Feb 1, 2005, 12:12 PM

I probably wouldn't risk it.

Do you have any reason to suspect the drives are failing?

tooki

IYork · Feb 1, 2005, 01:40 PM

Reader50 -- Thanks for the reply. Did as advised with no apparent trouble.

Tooki -- One of my four drives has started to make an occasional "Whoop-click" sound. Whatever it is, it can't be good, although the SMART diags reported no trouble.

With all this controller swapping, I've been working on the assumption that a drive's SMART diagnostic system writes any error into non-volatile memory, so that cutting power does not reset any error flag that's been generated.

Can anyone confirm this?

Thanks.

Dr. Wahnsinn · Feb 1, 2005, 11:30 PM

Interesting. I have two drives in my system running OS X; one connected to the onboard ATA controller, and one connected to an ATA PCI card. The SMART status of both drives can be verified with either Disk Utility or Tech Tool Pro 4. Disk Utility even identifies the one on the ATA card as connected to an ATA bus.

System Profiler used to see this ATA card a SCSI, but now identifies it correctly. The drive connected to it, however, doesn't show up under ATA devices.

Big Mac · Feb 2, 2005, 04:20 AM

Usually the advice is, if you suspect your drive is failing backup and replace post haste. SMART apparently isn't a great indicator of drive health. People have often reported that their drives have failed despite clean bills of health from SMART. Drives are cheap enough and your data is not worth risking.

IYork · Feb 2, 2005, 12:54 PM

Mac,
I agree. The problem is I have four drives and no idea of which one is the guilty party. The sound is intermittent, very much so at the moment, so unplugging drives doesn't seem very practical. Sticking my head into the case has been fun, but inconclusive, although now my two year old wants to have a turn...

reader50 · Feb 2, 2005, 08:51 PM

Stop by an auto parts store and pick up a stethoscope. It's what the auto shop guys use to track down stray noises.

OreoCookie · Feb 3, 2005, 05:56 AM

Just install smartmontools for the command line (e. g. via fink). They will read the smart status of your drive.

If you want to be sure, use the hd tools of your harddrive manufacturer. Note that some hds have a three-year warranty (e. g. Samsung), so you can have them fix it for you for (basically) free.