Gentoo Archives: gentoo-user

From: Dale <rdalek1967@×××××.com>
To: gentoo-user@l.g.o
Subject: Re: [gentoo-user] Hard drive error from SMART
Date: Tue, 12 Apr 2022 13:01:24
Message-Id: 9428cf38-eb0a-8ae6-56bb-d7844a912931@gmail.com
In Reply to: Re: [gentoo-user] Hard drive error from SMART by Wols Lists
1 Wols Lists wrote:
2 > On 12/04/2022 02:27, Dale wrote:
3 >> The one I aborted was because it was stuck on 10% for well over a day.
4 >> The whole test doesn't take that long, or shouldn't anyway.  I restarted
5 >> it shortly after that.  I might add, the test did take many hours longer
6 >> than it estimated which from my past experience is quite odd.  It's
7 >> usually pretty accurate.  Still, it completed and shows it passed, just
8 >> has a boo boo on it.  I also did a file system check it fixed a couple
9 >> problems and a bunch of little things I see corrected often on bootup.
10 >> Something about length of something.  Seems trivial.
11 >
12 > Given that the firmware SOMETIMES gets its knickers in a twist,
13 > especially consumer drives (not sure what yours are?), and read errors
14 > are a dime a dozen, I wouldn't worry that much about ONE error.
15 >
16 > Do another SMART test after your next reboot. Any NEW errors will be a
17 > red flag, but just this one again? Don't worry.
18
19
20 That seems to be what my google searches revealed.  After all, nothing
21 is perfect.  I'm sometimes surprised that drives aren't shipped with a
22 couple of these.  I'll keep my backups up to date as usual tho.  ;-)
23
24
25 >>
26 >> Given the low number and it showing it corrected that error, and then
27 >> passed a short and long test, is this drive "safe enough" to keep in
28 >> service?  I have backups just in case but just curious what others know
29 >> from experience.  At least this isn't one of those nasty messages that
30 >> the drive will die within 24 hours.  I got one of those ages ago and it
31 >> didn't miss it by much.  A little over 30 hours or so later, it was a
32 >> door stop.  It would spin but it couldn't even be seen by the BIOS.
33 >> Maybe drives are getting better and SMART is getting better as well.
34 >
35 > SMART is a lot better than it was, but remember, it only picks up wear
36 > and tear. Mechanical failure is just as deadly, and usually strikes
37 > out of the blue. I saw some stats somewhere it's something like 1/3,
38 > 2/3 wear and tear picked up by SMART, and mechanical failure
39 > undetectable by smart. Can't remember which stat was which.
40
41 My understanding is that SMART detects media problems and sometimes even
42 when a electronic component is getting out of spec.  However, it is
43 unlikely to detect that the spindle motor or the mechanism that moves
44 the heads is about to go out.  It can detect some things but not
45 everything.  From my understanding, it is mostly about monitoring the
46 magnetic media itself.  It is however, better than nothing at all. 
47
48
49 >>
50 >> Thoughts.  Replace as soon as drive arrives or wait and see?
51 >
52 > If you get a couple of errors, then no more for months, the drive is
53 > probably fine. If you get new errors every time you test, ditch it ASAP.
54 >
55 > Either way, make sure it's backed up!
56 >
57 > Cheers,
58 > Wol
59 >
60 >
61
62
63 Sounds like a plan.  Drive should be here Friday.  I'll keep a eye on
64 it.  It's down to 10% on long selftest and no errors reported yet.  I'll
65 keep the drive unmounted until Friday tho, just in case. 
66
67 Thanks for the opinions. 
68
69 Dale
70
71 :-)  :-)