Messages in this thread | | | From | "Mudama, Eric" <> | Subject | RE: Blockbusting news, results are in | Date | Sun, 19 Oct 2003 11:36:31 -0600 |
| |
> Eric, is it true what we tell users, that if a drive can't remap > a bad block it has probably used up all its spares, and that in > turn means that it is wise to buy a new one because the chance of > experiencing additional data corruption on a drive that has used > up all its spares is much higher than the average drive?
Not sure about other vendors, but a fatal write on a maxtor means we couldn't do your write after exhausting all attempts at reallocation, recertification, etc. If you ever get this on a drive, either:
1) the drive is unable to reallocate any more blocks because it has run out of spares
or
2) the drive was attempting those writes under environmental conditions that it was unable to handle. (extreme shock&vibe, <5C, >55C, etc)
> What are the common sources of data corruption, is one of them > that the drive head starts bumping the media more and more often > because a bearing (or something) has started to show signs of wear?
From my understanding, most returns are due to damaged heads (some small percent burn up over time) or operational shock "head/media events" where someone bumped a running drive and the head dug a crater in the media. Any particulate contamination can be struck by the heads causing high-fly write events. (head bounces up off the media in the middle of a write). I haven't heard of bearing wear being a common issue... all drives these days use fluid bearings. Early fluid bearings had outgassing issues at high temperature, but I think those problems were solved by manufacturers long before the first drives using them hit store shelves.
All in all, they're rather delicate. I'm amazed they work at all too.
--eric - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |