Number of spare sectors on a hard disk

[fjblurt]

Hi folks,

I have a hard drive that's failing. I already got my data off it but
would like to use it for scratch space for a little while. The
symptom is that it has bad sectors which cannot be read. I understand
how sector reallocation works; that when these sectors are rewritten
they will be marked bad, and new sectors will be allocated from a
spare pool that exists for that purpose. I am curious how large that
pool is, and how close I will be to exhausting it.

This is a 250 GiB drive (490234752 sectors), and currently there are
162 bad sectors awaiting reallocation. It's made by Maxtor.

Obviously I don't expect anyone to know the exact number for my disk,
but I presume there is some standard proportion, within an order of
magnitude (one per million sectors? one per thousand?). I didn't find
any info online when I searched, but maybe there is a wizard here that
would know.

Once again, don't worry, I am not going to put anything on this disk
that can't be easily regenerated. Thanks in advance!

 

[meerkat]

Hi folks,

I have a hard drive that's failing. I already got my data off it but
would like to use it for scratch space for a little while. The
symptom is that it has bad sectors which cannot be read. I understand
how sector reallocation works; that when these sectors are rewritten
they will be marked bad, and new sectors will be allocated from a
spare pool that exists for that purpose. I am curious how large that
pool is, and how close I will be to exhausting it.

This is a 250 GiB drive (490234752 sectors), and currently there are
162 bad sectors awaiting reallocation. It's made by Maxtor.

Go read the SMART section here, it gives you what info
is available to us, unless you can get anything directly from
the manufacturer...
http://en.wikipedia.org/wiki/Self-M...orting_Technology#Known_S.M.A.R.T._attributes

 

[kony]

Hi folks,

I have a hard drive that's failing. I already got my data off it but
would like to use it for scratch space for a little while. The
symptom is that it has bad sectors which cannot be read. I understand
how sector reallocation works; that when these sectors are rewritten
they will be marked bad, and new sectors will be allocated from a
spare pool that exists for that purpose. I am curious how large that
pool is, and how close I will be to exhausting it.

This is a 250 GiB drive (490234752 sectors), and currently there are
162 bad sectors awaiting reallocation. It's made by Maxtor.

Obviously I don't expect anyone to know the exact number for my disk,
but I presume there is some standard proportion, within an order of
magnitude (one per million sectors? one per thousand?). I didn't find
any info online when I searched, but maybe there is a wizard here that
would know.

Once again, don't worry, I am not going to put anything on this disk
that can't be easily regenerated. Thanks in advance!

Get the Seagate diagnostics and scan the drive including the
full (read/write?) test. If it doesn't pass, RMA it if
under warranty or otherwise throw it away unless you're a
masochist.

You shouldn't be seeing these bad sectors as the drive
remaps them automatically until it has exhausted it's
spares... so I suspect this number of spares no matter how
many, has already been exhausted.

 

[meow2222]

On 31 Oct 2007 12:34:33 -0700, (e-mail address removed) wrote:


Get the Seagate diagnostics and scan the drive including the
full (read/write?) test. If it doesn't pass, RMA it if
under warranty or otherwise throw it away unless you're a
masochist.

You shouldn't be seeing these bad sectors as the drive
remaps them automatically until it has exhausted it's
spares... so I suspect this number of spares no matter how
many, has already been exhausted.

Quite, all spares exhausted already, which means complete failure
any moment now.


NT

 

[fjblurt]

Get the Seagate diagnostics and scan the drive including the
full (read/write?) test. If it doesn't pass, RMA it if
under warranty or otherwise throw it away unless you're a
masochist.

It is under warranty, and I intend to RMA it. I just have a project
that will require a lot of scratch space, and that I would like to
work on before I get the replacement, which will probably take some
time. And actually, I am kind of a masochist

You shouldn't be seeing these bad sectors as the drive
remaps them automatically until it has exhausted it's
spares... so I suspect this number of spares no matter how
many, has already been exhausted.

I've heard this, but it doesn't make sense to me in this case.
Suppose a single sector on the disk fails in an unrecoverable (but
detectable) manner, so that the data in it cannot be read. How can
the drive "remap" this sector? It can't copy the data somewhere else,
since it can't read it.

If I try to read this sector, the drive must report to me that it
can't do it; it has to throw an error, since silently returning bad
data would be much worse. So this must happen regardless of how many
spare sectors are available.

Now certainly, when I attempt to write that sector, the drive can
remap it and write the data to one of the spare sectors instead. So I
agree that a *write* failure would indicate that the spare sectors are
exhausted. But I don't see how a read failure would indicate that.
Is there something wrong with this logic?

Anyway, thanks for the advice, but I really was just hoping for an
answer to my question.

 

[CBFalconer]

.... snip ...

I've heard this, but it doesn't make sense to me in this case.
Suppose a single sector on the disk fails in an unrecoverable
(but detectable) manner, so that the data in it cannot be read.
How can the drive "remap" this sector? It can't copy the data
somewhere else, since it can't read it.

Yes, it often can. The drive attempts (multiple) rereads, and if
one succeeds it has the correct data. It copies that into a spare
sector, and remaps the bad sector.

 

[kony]

I've heard this, but it doesn't make sense to me in this case.
Suppose a single sector on the disk fails in an unrecoverable (but
detectable) manner, so that the data in it cannot be read. How can
the drive "remap" this sector? It can't copy the data somewhere else,
since it can't read it.

It can retry the read, maybe reading it but it keeps track
if these and based on a manufacturer threshold, at some
point will deem an area too bad to keep using and then
either remap the logical location of the sector and move the
data if it is a recoverable read error.

If I try to read this sector, the drive must report to me that it
can't do it; it has to throw an error, since silently returning bad
data would be much worse. So this must happen regardless of how many
spare sectors are available.

There are more parameters than just can or can't read,
including what tricks the drive tries to read it... things
it shouldn't have to do unless there is a (potential)
problem. There's a bit more here,
http://www.storagereview.com/guide/errorRead.html

Now certainly, when I attempt to write that sector, the drive can
remap it and write the data to one of the spare sectors instead. So I
agree that a *write* failure would indicate that the spare sectors are
exhausted. But I don't see how a read failure would indicate that.
Is there something wrong with this logic?

The failure to read is not indicating it, the bad sector is,
unless it is only logically marked as such which is why you
would want to do the manufacturer's read/write diagnostic
test.

In other words, if you can't read something and the drive
were to remap the sector, it will not now indicate the
sector is bad because it has moved the sector address to a
supposed "good" sector spare location which will not be
considered bad... unless you are out of spares.

The ideal and purpose of the spares is that they are used
proactively before you have massive data loss. If the
reading is more problematic it will be moved without your
ever knowing it happened before there is a degradation bad
enough a lot of sectors are bad. By the time there is a
significant risk the drive is supposed to have reached a
threshold which sets a Smart flag so you are warned.

 

[GT]

Hi folks,

I have a hard drive that's failing. I already got my data off it but
would like to use it for scratch space for a little while. The
symptom is that it has bad sectors which cannot be read. I understand
how sector reallocation works; that when these sectors are rewritten
they will be marked bad, and new sectors will be allocated from a
spare pool that exists for that purpose. I am curious how large that
pool is, and how close I will be to exhausting it.

This is a 250 GiB drive (490234752 sectors), and currently there are
162 bad sectors awaiting reallocation. It's made by Maxtor.

I don't think there are spare sectors to act as backup for failed sectors,
instead I think the bad sectors are marked and never used. The capacity
therefore drops (very) slightly as they fail.

 

[DevilsPGD]

In message <[email protected]> "GT"

I don't think there are spare sectors to act as backup for failed sectors,
instead I think the bad sectors are marked and never used. The capacity
therefore drops (very) slightly as they fail.

You're both right.

The drive itself CAN remap sectors, but only under some circumstances.
The OS won't know about it during a regular read attempt, except for a
substantial slowdown when the remapping happens (due to the drive making
multiple attempts to read the original sector to get the data out).
These are the bad sectors you see in a SMART report.

Second, when a drive does need to return an error, the file system can
flag the sector as bad, which reduces available space somewhat. This is
the bad sectors you see in chkdsk.

 

[ame]

On 31 Oct 2007 12:34:33 -0700, (e-mail address removed) wrote:

>Hi folks,
>
>I have a hard drive that's failing. I already got my data off it but
>would like to use it for scratch space for a little while. The
>symptom is that it has bad sectors which cannot be read. I understand
>how sector reallocation works; that when these sectors are rewritten
>they will be marked bad, and new sectors will be allocated from a
>spare pool that exists for that purpose. I am curious how large that
>pool is, and how close I will be to exhausting it.
>
>This is a 250 GiB drive (490234752 sectors), and currently there are
>162 bad sectors awaiting reallocation. It's made by Maxtor.
>
>Obviously I don't expect anyone to know the exact number for my disk,
>but I presume there is some standard proportion, within an order of
>magnitude (one per million sectors? one per thousand?). I didn't find
>any info online when I searched, but maybe there is a wizard here that
>would know.
>
>Once again, don't worry, I am not going to put anything on this disk
>that can't be easily regenerated. Thanks in advance!


>Get the Seagate diagnostics and scan the drive including the
>full (read/write?) test. If it doesn't pass, RMA it if
>under warranty or otherwise throw it away unless you're a
>masochist.

>You shouldn't be seeing these bad sectors as the drive
>remaps them automatically until it has exhausted it's
>spares... so I suspect this number of spares no matter how
>many, has already been exhausted.

I'm interested of this same question and I can't understand why those smartasses don't read question before answering to it. this is not answering for anything what was asked.