<br><br><div class="gmail_quote">On Sat, Jul 7, 2012 at 11:28 AM, Jay Ashworth <span dir="ltr"><<a href="mailto:jra@baylink.com" target="_blank">jra@baylink.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
----- Original Message -----<br>
> From: "Alexander Fisher" <<a href="mailto:alex@alexfisher.me.uk">alex@alexfisher.me.uk</a>><br>
<br>
> I've also had some odd behavior from one of my non-enterprise Seagate<br>
> drives. It's firmware locked up the other day, the drive not coming<br>
> back without a power cycle.<br>
<br>
I believe you had the same problem that I've had on 5 different Seagates<br>
over the last 4 years:<br>
<br>
There's a black hole somewhere on the drive. If you never touch it, the drive<br>
works just fine. But the second you even read one of those sectors, the<br>
drive adapter falls off line, and Linux can't touch it anymore, and it doesn't<br>
recover until you power the drive off. Once you do, everything's fine until<br>
you hit that spot again.<br></blockquote><div><br><br>My test for new drives is now to put them in a ZFS mirror, and fill them to 100% capacity, then to a zpool scrub on them to make sure they are holding the data properly. I've found a few bad drives this way that wouldn't have shown up with normal testing unless I happened to hit that bad block. I usually repeat the test a few times, by destroying the pool, then starting over. It takes a long time, but it beats finding out the hard way months later. <br>
</div></div>