1 |
Alan McKinnon wrote: |
2 |
> On 26/06/2014 13:20, Dale wrote: |
3 |
>> Alan McKinnon wrote: |
4 |
>>> My experiences aren't worth much in this case, what I had to deal with |
5 |
>>> was data center setups where - the power has never gone off for 6 |
6 |
>>> years - the drives never spin down and just keep on turning year after |
7 |
>>> year - the servers were the nice big ones Dell makes with awesome |
8 |
>>> cooling - the data center feels like a fridge and the ambient temp |
9 |
>>> never varies more than 1 deg - the server power supplies are seriously |
10 |
>>> high grade, the 5V and 12V out of them are solid and do not fluctuate |
11 |
>>> at all Add all this up and it's an almost perfect environment for |
12 |
>>> drives to last a long time. You don't have that, not even close. I |
13 |
>>> have only 1 little bit of anecdotal data: my nas at home has 4 x 3T WD |
14 |
>>> Green drives in it, going on almost 2 years now. My kids hammer the |
15 |
>>> blazes out of that thing, and ZFS scrubs keep it real busy when the |
16 |
>>> kids don't. And those drives just keep on turning and turning and |
17 |
>>> turning, I didn't do anything special. I put it down to statistics - |
18 |
>>> no-one makes bad drives (or cars) these days and I haven't pulled the |
19 |
>>> unlucky card yet. I dunno, go figure |
20 |
>> Well, it does make good points tho. I keep my room here pretty cool. |
21 |
>> It's not as cool as your data center but I have a window A/C and my own |
22 |
>> heater. I don't mind it being a little cool in the winter but don't |
23 |
>> like it warm in the summer either. The cooler the better. |
24 |
>> |
25 |
>> I also have the Cooler Master HAF-932 case with those really nice large |
26 |
>> fans. The hard drives are right in front of the front intake fan. I |
27 |
>> have a power supply that is really to big for what I have running. I |
28 |
>> can't recall the brand and wattage just that it doesn't pull near as |
29 |
>> much power as I thought it would. It pulls less than half what my older |
30 |
>> and much slower puter pulled. Also, I rarely shut this thing down. I |
31 |
>> did the other night to unplug/re-plug all the cables but other than |
32 |
>> that, it is usually because I have lost power from the mains. |
33 |
>> |
34 |
>> So, keep them cool, good clean power and leave them running when ya |
35 |
>> can. Sounds like a plan. ;-) |
36 |
> |
37 |
> You got it :-) |
38 |
> |
39 |
> hard drives are mechanical objects, not electronic ones, and they fail |
40 |
> for mechanical reasons. Motors fail, bearings seize, spindle arms wear |
41 |
> out. Transforming magnetic blobs on the platter into binary bits is very |
42 |
> reliable, as long as the head is in exactly the place it is supposed to |
43 |
> be. So the enemies of disks are environmental; |
44 |
> |
45 |
> - temperature and humidity changes |
46 |
> - frequent spin ups and spin downs |
47 |
> - dust |
48 |
> - power dips/fluctuations and brown-outs |
49 |
> - being dropped, knocked and generally ubused |
50 |
> |
51 |
> etc, etc, etc |
52 |
> |
53 |
> Take care of the environmental factors, and statistics fall in your |
54 |
> favour making the odds good you'll get the life you expect |
55 |
> |
56 |
|
57 |
I think that is one reason I have had some pretty good luck with that. |
58 |
I might also add, I have actually only had one computer that failed. |
59 |
That includes the ones that folks just gave me which is quite a few. |
60 |
Most of them just get to slow to use. The ones I build, I build them |
61 |
like a tank. I put coolers on everything that is even a little warm. |
62 |
My CPU cooler on my current rig is pretty large. Case fans blowing a |
63 |
lot of air, quiet if possible. For this drive that I have going out now |
64 |
to go out, it has to have a issue not related to cooling and such. |
65 |
Unless it was somehow handled badly while being shipped to me, its never |
66 |
been dropped or anything either. This is a desktop, with wheels since |
67 |
it is on carpet, and it rarely goes anywhere. It doesn't get rattled |
68 |
around like a laptop or something. |
69 |
|
70 |
My old rig, AMD 2500+ in a old full tower case still runs good. I |
71 |
booted it a month or so ago. I had a Volcano 11 or 12 on the CPU which |
72 |
is solid copper. I replaced the northbridge cooler with a copper cooler |
73 |
with a fan. The mosfets close to the CPU, I added coolers to them too. |
74 |
It had 5 case fans. It wasn't quiet but it ran cool. The mobo temps |
75 |
were usually just a couple degrees above room temp. CPU never got over |
76 |
100F. Heck, the CPU in my current rig has never seen 110F. The highest |
77 |
I have ever seen was 107F and that was when I was compiling and had |
78 |
power to blink just enough to cut off my A/C for a hour or so. Maybe I |
79 |
need a UPS for my A/C too. :-D |
80 |
|
81 |
It seems the best thing WE can do, good power, good cooling, don't drop |
82 |
it and keep backups. |
83 |
|
84 |
I went back through the error logs and found this: |
85 |
|
86 |
Jun 12 23:30:36 localhost smartd[2688]: Device: /dev/sdc [SAT], 104 |
87 |
Currently unreadable (pending) sectors |
88 |
Jun 12 23:30:36 localhost smartd[2688]: Device: /dev/sdc [SAT], 104 |
89 |
Offline uncorrectable sectors |
90 |
|
91 |
That's the first error I could find. It went from nothing to that in |
92 |
one huge jump. I also found this: |
93 |
|
94 |
Jun 8 03:10:02 localhost sSMTP[7164]: Unable to locate mail |
95 |
Jun 8 03:10:02 localhost sSMTP[7164]: Cannot open mail:25 |
96 |
Jun 8 03:10:03 localhost CROND[7145]: (root) MAIL (mailed 57 bytes of |
97 |
output but got status 0x0001 |
98 |
) |
99 |
|
100 |
It seems it is trying to mail something. I need to work on that when I |
101 |
get the new drive set up. I already have smtp installed. |
102 |
|
103 |
Dale |
104 |
|
105 |
:-) :-) |