I have a computer, purchased in 2007 (or was it 2006). It’s a Dell Dimension C521.
It is slow, compared to newer computers. I am mostly using it as a test machine. If it dies, I can manage without.
Recently, when I powered it on, the BIOS message said “No operating system found”.
Naturally, I tried CTRL-ALT-DEL, and it booted. Running opensuse 13.1, I had one instance where a message was logged saying that the disk response was slow. Apart from that, it has been fine. But, if I power down, the same problem will arise.
From this, I conclude that the disk surface is fine, but something else is failing:
Possibly the SATA controller on the main motherboard is failing;
Perhaps the device electronics is failing (on the disk);
Perhaps the disk is slow to get up to full rotation speed (congealed lubricant?).
For either (2) or (3), a relatively inexpensive disk drive replacement would solve the problem. For (1), it might be more trouble than it is worth.
Perhaps somebody has similar experience, and can tell me which is more likely to be the problem.
And if I do decide to purchase a replacement disk, do I need to worry about SATA II vs. SATA III? Or will a SATA III disk do just fine?
The old disk is 320G. I would probably replace with 500G. Going to 1T might confuse the BIOS.
it could simply be that your BIOS battery is dead so when you first boot the bios is reset to default and may not find the harddrive, but when you CTL-ALT-DEL it is finds the HDD that had time to spin up by now and thus booting as expected.
since the LP is old the harddrive could be slow thus needing the time to spinup
> Naturally, I tried CTRL-ALT-DEL, and it booted. Running opensuse 13.1,
> I had one instance where a message was logged saying that the disk
> response was slow. Apart from that, it has been fine. But, if I power
> down, the same problem will arise.
>
> From this, I conclude that the disk surface is fine, but something else
> is failing:
>
> - Possibly the SATA controller on the main motherboard is failing;
> - Perhaps the device electronics is failing (on the disk);
> - Perhaps the disk is slow to get up to full rotation speed (congealed
> lubricant?).
Run the SMART long test on it, with smartctl.
Or you can download the SeaTools ISO from seagate, burn it, boot it, run
it. It basically runs the same test, but it also does some checking on
the interface. Others brands may have similar tools for checking hard
disks, but that’s the one I’m most familiar with.
–
Cheers / Saludos,
Carlos E. R.
(from 13.1 x86_64 “Bottle” at Telcontar)
On 2014-03-04 00:06, nrickert wrote:
>
> robin_listas;2628349 Wrote:
>> Run the SMART long test on it, with smartctl.
>
> It seems to have PASSED the test. It notes only two errors, which seem
> to be bad sectors noticed at power on. That’s after the long test.
Well, that’s very important.
Notice that the test gives “passed” even if you have thousands of bad
sectors. When you attempt to write into a sector that is bad, the disk
firmware automatically writes and remaps to another sector, reserved for
the purpose during manufacture.
When that pool of spare sectors is spent, then the test gives “FAIL”
(and a warning when it is nearly spent).
When I see even only one bad sector, I rewrite the entire hard disk with
zeros, or anything (doesn’t matter what, as long as you fill absolutely
all sectors).
Then I run the long test again.
If there are more bad sectors, I repeat.
Only if the number of bad sectors is stable, I keep using the disk,
otherwise I discard it.
If I keep the disk, obviously I have to reformat and recover the data
from the backup, because I destroyed it all.
–
Cheers / Saludos,
Carlos E. R.
(from 13.1 x86_64 “Bottle” at Telcontar)
I have a 320-GB WD SATA HD purchased in a laptop in 2007. I removed it more than a year ago because it failed – sort of.
Most of the time, it is not recognized, is totally invisible.
However, booting and rebooting into BIOS, it will finally show up. When it does, I can run every disk diagnostic I can get my hands on at it, including the diagnostics from the manufacturer, WD, and it continually will pass all tests with flying colours! After, on reboot, it most often no longer exists! Very frustrating.
So, the diagnostics are not the final say on the health of the disk.
I have thoroughly tested, and yes, it is the drive itself. The tests that verify this are:
Install in another PC, same symptoms. Try in a couple other PCs, same symptoms.
Try other drives in the laptop, no symptoms.
Problem narrowed down to drive only.
By the way, when it is booted up into a high performance system (I think that is the common factor for this next symptom), it will run for random lengths of time from only a moment, to quite awhile, then disappears.
If I load it (when BIOS sees it, of course) with a diagnostic, or load it with an external system (such as a Puppy boot key), it will remain visible and fully readable at least until the system is shut down or rebooted.
No, I don’t need this solved, it is a failure in electronic design by the manufacturer, and the WD is past warranty.
I have long since replaced the drive and moved on.
I am just mentioning it here so that others with similar HD-like symptoms or problems are aware that the diagnostics can incorrectly pass a failed disk with flying colours.
>
> I have a 320-GB WD SATA HD purchased in a laptop in 2007. I removed it
> more than a year ago because it failed – sort of.
>
> Most of the time, it is not recognized, is totally invisible.
…
> I have thoroughly tested, and yes, it -is- the drive itself. The
> tests that verify this are:
>
> Install in another PC, same symptoms. Try in a couple other PCs, same
> symptoms.
Curious…
> I am just mentioning it here so that others with similar HD-like
> symptoms or problems are aware that the diagnostics can incorrectly pass
> a failed disk with flying colours.
Well… it means that the diagnostic software we have available is
incomplete.
–
Cheers / Saludos,
Carlos E. R.
(from 13.1 x86_64 “Bottle” at Telcontar)
On 2014-03-06 22:46, Fraser Bell wrote:
>
> robin_listas;2628940 Wrote:
>> On 2014-03-06 14:06, nrickert wrote:
>>
>>> The diagnostics are mostly concerned with recording surface problems.
>>> Your unit was suffering from a failure of the drive electronics.
>>
>> Seatools, from Seagate, claims to test also the interfacing.
> Seagate’s diagnostics only work if you have a Seagate (or Seagate
> subsidiary, like Maxtor) HD detected on the system.
Partially. If it does not find a Seagate disk, it does generic tests.
There are two versions. One is a small live CD, using freeedos (I
think), so we can use it on a Linux machine, It is quite limited. It can
print a return form, but it wants to save it as a file somewhere, which
can not do because it is a CD (the older version run from a floppy).
They document an ugly trick to write instead to a usb stick. Very ugly.
Another version is modern, but it runs only under Windows. I had to use
it this week, but the disk I had to test was connected externally, via
usb dock station, to my double boot laptop. The blasted thing insisted
that the disk was not a Seagate, which it was, of course, so it did the
generic test only. It detected the docking box chipset primarily.
> WD also has diagnostics (same prerequisites, need a WD or subsidiary
> disk), and it failed to detect the problem.
Sigh.
>> Well… it means that the diagnostic software we have available is
>> incomplete.
>>
>
> is the closest answer, along with (using the qualifying phrase “Most of
> …”):
>
>>
>> The diagnostics are mostly concerned with recording surface problems.
Some hardware platforms I have worked with (not PCs) have very extensive
self diagnostic facilities. And several digit prices!
> So, just be aware, because it passes the diagnostics we have at our
> disposal, -“it” could still be a problem with the disk-
>
> whenever disk-like symptoms pop up.
Right…
–
Cheers / Saludos,
Carlos E. R.
(from 13.1 x86_64 “Bottle” at Telcontar)
> … which is why I often advise switching disks around (when you can) to
> see if the problem follows the disk or stays with the machine.
>
> I do that with memory, as well.
>
> And, in both cases, it is often far quicker diagnosing than running the
> tests.
Provided you do have the extra hardware.
If I had such a problem with the memory on my desktop, I don’t have any
replacement for swapping. Hard disks, yes, I have a few.
The problem is, that although I have several machines, each of them uses
a different class of chips, and the replacements I have do not match all
machines. Or rather I don’t have ram replacement for all of them. So I
would try to run tests to be sure that ram is at fault first.
It is different when you do hardware maintenance for a living, you
usually have surplus material, or at least several similar computers on
the shop.
–
Cheers / Saludos,
Carlos E. R.
(from 13.1 x86_64 “Bottle” at Telcontar)
Carlos E. R. wrote:
> Provided you do have the extra hardware.
> If I had such a problem with the memory on my desktop, I don’t have any
> replacement for swapping. Hard disks, yes, I have a few.
You don’t need extra hardware, all you need is more than one stick
installed. Then just swap them around and see if the problem is still at
the same address or has moved to the other bank (or gone away entirely
because it was a badly seated stick
On 2014-03-07 11:11, Dave Howorth wrote:
> Carlos E. R. wrote:
>> Provided you do have the extra hardware.
>> If I had such a problem with the memory on my desktop, I don’t have any
>> replacement for swapping. Hard disks, yes, I have a few.
>
> You don’t need extra hardware, all you need is more than one stick
> installed. Then just swap them around and see if the problem is still at
> the same address or has moved to the other bank (or gone away entirely
> because it was a badly seated stick
Yes, sometimes that works.
Buts:
You do not find out which of the two modules work, and which one is bad.
If the problem is not a bad module, but a module type that your
computer does not fully like for some reason, the swap test does
nothing. The worst case is when the system doesn’t detect the problem,
but simply malfunctions in some way. Even worst, intermittently. (As
Fraser Bell described for his hard disk).
–
Cheers / Saludos,
Carlos E. R.
(from 13.1 x86_64 “Bottle” at Telcontar)
Exactly. But, keep in mind, most machines have more than one memory module, so you can still quick-test:
If you have 2 modules, pull one out & test. If problem remains, put that module back in & pull the 2nd module, test again. Of course, this only works if the system is not one of the machines that requires modules in pairs, but I believe few do, any more.
If you have 4 modules, pull the 2nd pair & test. If problem remains, pull first pair, put 2nd pair in the first slots, test again.
In the latter case, you may want to run additional tests to find which module is faulty out of the pair that showed the symptoms.
Drives, of course, are another matter. In your case, you have additional, but most cases I would assume do not have another laptop drive to install for the test. If they don’t have another (desktop or tower) to test the laptop’s drive in, this method is not available to them.
As you confirm, these methods are only good for those who have the resources at hand.
