PCA Forums

Anyone having issues presently or in the past with a 75G Western Digital Raptor? I tried to do my monthly Acronis backup, and it stopped at 40% and sat there. After 5 minutes of inactivity, Acronis came back and reported that it could not read the sector.

So I booted with SpinRite, and it also stopped at the same spot. After about 5 minutes of very slow activity, it was able to recover some of the data and marked it as unrecoverable. Is this an indication that the drive will be failing by having more sectors go bad? Reason I'm asking is because this is a Western Digital drive, where in the past I've been using Seagates exclusively.

I had one go bad about a year ago in a raid0 set but it is rare. Back your stuff up and do an over the phone RMA with advanced replacement.

Greg

+1...do a backup right away and then work on the replacement. WD should have a utility you can download to run diagnostics on the drive (probably required for an RMA anyway).

any SMART errors?

If that's the only bad sector, and it's not marked as permanently unusable by Spinrite, then it's possible that you could continue to use the drive without a problem BUT if it was my drive and it was under warranty I would definitely arrange for an RMA just to be sure. With a modern drive, one bad sector is often an early warning that more are coming.

Use WDC's online RMA form. No need to talk to anyone, and usually no questions asked, and you can pay for an advanced RMA which will have them ship a drive to you 2 day air so you don't have any down time.

Spinrite can often recover most of a bad sector. In reality you may often only damage a few bytes, but normally the whole 512 bytes of the sector is thrown out by the drive's error correction routines because that's the way FAT and NTFS was designed. Most disk recovery utilities also throw out the entire sector. Spinrite works very hard to recover whatever part of the sector is still readable. The nice thing is that since sectors are somewhat redundant, you can often lose a couple of bytes and successfully restore the data.

Spinrite tries to read each bad sector 2000 times by overriding the drive's internal error correction and using various techniques to manually read the sector in different ways to try to coax the surviving data out of it. Since a bad sector can take several seconds to read, this is why Spinrite is so slow to recover a badly damaged drive - it can take days sometimes - even weeks).

I could use SpinRite's option #4 (I think), that allows you to perform a format on each sector, then places the data back. Anyone use that option? I never have since I've always used option #2 for emergency recovery.

Level #2 is the one you usually use to repair a drive (if you can't read data or it's unbootable). You should always follow up any Spinrite scan with a Scandisk or CHKDSK (not a full sector scan, just a file system check and repair) because if Spinrite repaired a bad sector, and that sector was a file pointer or part of the directory structure, you could have a corruption of the file system that you now have to fix. Spinrite doesn't touch the file system - it works at a low level on the individual sectors only.

Level #2 is much faster than a Level 4 because it just reads each sector and forces the drive's internal error correction to refresh each sector as it's read. If any sector is unreadable then Spinrite goes into Dynastat mode where it tries to recover as much of the unreadable sector as possible by reading it 2000 times. You can always tell when Spinrite finds a weak or bad sector because it will take a lot more time with it.

Now not all sectors are bad. Sometimes the magnetic imprint is just weak and it needs to be refreshed. This is why often you will Spinrite a drive that won't boot, and Spinrite won't find any bad sectors or recovered sectors, but afterwards the drive boots just fine. There were no bad sectors, just a weak one that Spinrite forced the drive's internal error correction to detect and correct.

Level 4 is much more aggressive (and thus takes much longer). It's not meant to be run on a drive on the verge of failure because it's going to really push that drive very hard which might put it over the edge. Spinrite will actually warn you about this, if it can access the SMART data on the drive, and it reports that the drive is near failure.

Level 4 is meant to be run once every few months or so on a GOOD drive as a PREVENTATIVE. It reads every sector twice, and writes the data back to the sector twice (the first time inverting the data). This "refreshes" the magnetic media under the sector which strengthens the signal of the recording (keep in mind that although the data is digital, the magnetic signal recorded on the media is analog, and it can become so weak as to make it difficult for the drive to tell if a byte is a 0 or a 1).

It takes a LONG time to run a level 4 on a modern large drive. I personally run a level 4 on my main workstation and my bench PC one every 6 months (and I just run it over night), but usually only on the boot and data partitions. I don't bother with any other PC unless it starts having problems that point to weak sectors.

On occasions I've had a client's drive whose data could not entirely be recovered after a Level 2 so I would backup whatever I could (in case the next step kills the drive), then run a Level 4 on it and then recovered more or all of the data.

Another thing to keep in mind is you should ALWAYS run Spinrite on a drive connected directly to the onboard controller (IDE, SATA or ESATA). Spinrite cannot communicate with the low level functions of the drive via USB or Firewire (it can't directly access the drive to do the magic 2000 reads of the sector for instance). It also can't access SMART via USB or Firewire. It will still do the best it can, but it will only use the drive's internal error correction so it won't be as effective, and it will be MUCH slower.

One more thing. Spinrite will monitor the drive temps via SMART if it's available, and stop the process if the drive starts to overheat. The continuous heavy thrashing of Level 4 (and even Level 2) can heat up a drive pretty quickly if it has poor cooling. Make sure you have lots of cooling. I usually open the case and if it's a laptop (I see laptop drives overheat very easily during Spinrite) I raise the laptop a few inches off the desk and have a 12" box fan blowing on it.

On the flip side, a Spinrite Level 4 is a darn good way to test if your rig has adequate drive cooling.

Believe me, it's no fun running Spinrite all night and finding out in the morning that it only got to 25% and stopped due to an overheating drive.