HDD Bad Sectors

[drnetsys]

Just want to know what you all think. I have 2 segate sata drives 1 is a 250gig (model ST3250318AS) and a 2tb (model ST3160815AS) both show many bad sectors in the disk utility smart data. If I boot into windows. It shows both drives as healthy. When I scan with seatools the drives pass. Is this a bug or should I really think about replacing the drives?

[Adunaic]

Personally I would ignore anything windows says on disk health. In my experience windows has said dying drives are healthy and Linux has spotted the errors early on.

I would replace them at the first sign of trouble.

[drnetsys]

That was my thought as well. I just was wondering because it was only the seagate drives and the other 6 WD drives are fine

[stoat]

If money is of no concern what so ever, then I guess you could ignore the Windows and Seagate utilities and go buy new drives. I wish all of my drives had raw values of zero for reallocated sectors in their S.M.A.R.T. data, but the fact is that some of them don't. And I don't want to get rid of them for that reason (not yet, anyway). For me (and this is for only me), I have chosen to monitor the S.M.A.R.T. data for new reallocated sectors myself and decide what to do about it myself. Some drives I check often (the ones already showing reallocated sectors), but I check all of them about quarterly. I keep the smartctl reports on each drive and compare new reports to the previous ones looking at the raw values, the normalized values, and the thresholds for each monitored attribute (the normalized value is sort of a score that lowers toward the threshold as the attribute's raw value gets worse). I also have smartd running and configured to notify me of new events. And so on.

So far, none of my drives, including ones with bad sectors, have changed since I have been doing this. Yet several of them will send Palimpsest into a messaging convulsion if I allow Palimpsest to run. For all I know, those drives left the factory that way. And so far, I have never had a drive go bad from too many reallocated sectors. For me, it has always been loud horrible frightening noise, but I always got the data off the drive the moment that started. So far, I have never lost any data of any kind ever.

Anyway, I guess everybody should do what they think is best, but I for one don't get too excited about what Palimpsest (or whatever the current incarnation of that is) says. I intend to continue using my drives as-is, continue monitoring the S.M.A.R.T. data with smartmontools utilities, and continue making weekly partition image backups (which in fact I did today). IMO, the backups to external drives spread the risk over multiple drives. I burn backups to DVDs about quarterly. I have a couple of thick notebooks full of those.

P.S.: My box of bad drives has fairly equal representation by Seagate, Maxtor, WD, Hitachi, and Samsung.

[jpollard]

It really depends on what the errors are, and how many there are.

The what is important if they are fatal read/write errors. These are usually handled by the drive itself doing block replacement.

The "how many" get important when the manufacturers list of replacement blocks get used up. After that... you get read/write errors propagated back to the host, and start getting disk corruption.

Another significant number is how often you get the errors. This can point to a disk that has started failing, and will definitely give you some pain trying to recover a non readable disk.

If the errors occur once every other month or so, then it isn't too bad - you usually have time to get a replacement before final failure.

After that, you get the fuzzy logic on "disk is healthy" as in what defines "healthy". Some think windows overestimates what is "healthy". Sometimes, you get a lemon for a disk...