1 |
On Mon, Jul 8, 2013 at 10:58 AM, Alan McKinnon <alan.mckinnon@×××××.com> wrote: |
2 |
> On 08/07/2013 17:39, Paul Hartman wrote: |
3 |
>> On Thu, Jul 4, 2013 at 9:04 PM, Paul Hartman |
4 |
>> <paul.hartman+gentoo@×××××.com> wrote: |
5 |
>>> ST4000DM000 |
6 |
>> |
7 |
>> As a side-note these two Seagate 4TB "Desktop" edition drives I bought |
8 |
>> already, after about than 100 hours of power-on usage, both drives |
9 |
>> have each encountered dozens of unreadable sectors so far. I was able |
10 |
>> to correct them (force reallocation) using hdparm... So it should be |
11 |
>> "fixed", and I'm reading that this is "normal" with newer drives and |
12 |
>> "don't worry about it", but I'm still coming from the time when 1 bad |
13 |
>> sector = red alert, replace the drive ASAP. I guess I will need to |
14 |
>> monitor and see if it gets worse. |
15 |
>> |
16 |
> |
17 |
> |
18 |
> Way back when in the bad old days of drives measured in 100s of megs, |
19 |
> you'd get a few bad sectors now and then, and would have to mark them as |
20 |
> faulty. This didn't bother us then much |
21 |
> |
22 |
> Nowadays we have drives that are 8,000 bigger than that so all other |
23 |
> things being equal we'd expect sectors to fail 8,000 time more (more |
24 |
> being a very fuzzy concept, and I know full well I'm using it loosely :-) ) |
25 |
> |
26 |
> Our drives nowadays also have smart firmware, something we had to |
27 |
> introduce when CHS no longer cut it, this lead to sector failures being |
28 |
> somewhat "invisible" leaving us with the happy delusion that drives were |
29 |
> vastly reliable etc etc etc. But you know all this. |
30 |
> |
31 |
> A mere few dozen failures in the first 100 hours is a failure rate of |
32 |
> (Alan whips out the trust sci calculator) 4.8E-6%. Pretty damn |
33 |
> spectacular if you ask me and WELL within probabilities. |
34 |
> |
35 |
> There is likely nothing wrong with your drives. If they are faulty, it's |
36 |
> highly likely a systemic manufacturing fault of the mechanicals (servo |
37 |
> systems, motor bearing etc) |
38 |
> |
39 |
> You do realize that modern hard drives have for the longest time been up |
40 |
> there in the Top X list of Most Reliable Devices Made By Mankind Ever? |
41 |
|
42 |
An update: the Seagate drives have both continued to spit more |
43 |
unrecoverable errors and find more and more bad sectors. Including |
44 |
some end-to-end errors indicated as critical "FAILING NOW" status in |
45 |
SMART. From what I have read that error means the drive's internal |
46 |
cache did not match the data written to disk, which seems like a |
47 |
serious flaw. The threshold is 1 which means if it happens at all, the |
48 |
drive should be replaced. It has happened half a dozen times on each |
49 |
disk so far (but not at the exact same time, so I don't think it is a |
50 |
host controller problem -- and other disks on the same controller and |
51 |
cable have had no issues). They have also been disconnecting and |
52 |
resetting randomly, sometimes requiring me to pull the drive and |
53 |
reinsert it into the enclosure to make it reappear. It happens even |
54 |
after I disabled APM, so I know it isn't a spin-down/idle timeout |
55 |
thing. Temperatures are actually very good (low 30's) so they are not |
56 |
overheating. |
57 |
|
58 |
I think I will try to trade them in to Seagate for a new pair under |
59 |
warranty replacement. And then probably try to sell the replacements |
60 |
and be rid of them. |
61 |
|
62 |
Meanwhile, during that experiment, I bought 2 brand new Western |
63 |
Digital Red 3TB drives last week. No problems in SMART testing or |
64 |
creating LVM/RAID/Filesystems. I have now been running the destructive |
65 |
write/read badblocks tests for 24+ hours and they have been perfect so |
66 |
far, exactly 0 errors. They are more expensive (3TB for the same price |
67 |
as the 4TB seagate) and slightly slower read/write speed (150MB/sec |
68 |
peak vs 170MB/sec peak), but I value reliability over all other |
69 |
factors. |
70 |
|
71 |
These Seagate drives must have some kind of manufacturing defect, or |
72 |
perhaps were damaged in shipping... UPS have been known to treat |
73 |
packages like a football! |