Bad sectors - Solaris

This is a discussion on Bad sectors - Solaris ; What is the impact of bad sector existence on : -Solaris Installation -Pakages Installation Did Solaris remark bad sector for not writing in future ?...

+ Reply to Thread
Results 1 to 15 of 15

Thread: Bad sectors

  1. Bad sectors

    What is the impact of bad sector existence on :
    -Solaris Installation
    -Pakages Installation

    Did Solaris remark bad sector for not writing in future ?


  2. Re: Bad sectors

    -----BEGIN PGP SIGNED MESSAGE-----
    Hash: SHA1

    Andrew Gabriel wrote:
    > In article <1142971568.840177.263500@i39g2000cwa.googlegroups. com>,
    > ehabaziz2001@gmail.com writes:
    >> What is the impact of bad sector existence on :
    >> -Solaris Installation
    >> -Pakages Installation
    >>
    >> Did Solaris remark bad sector for not writing in future ?

    >
    > For the last 15-20 years, this has been handled automatically
    > by intelligent controllers embedded in the disk drive.
    >


    I must contradict you. You are right that most failures are normally
    handled by the disk controller firmware. But I just have a disk under
    SVM that went to state Maintainance, because of bad sectors. I called
    Sun and the support guy told me that if the number of sectors with
    errors is low (in my case three), then one can run format->analyze->read
    or refresh to get the bad sectors assigned to the secondary defect list.
    The primary defect list (created before the product is shipped to the
    customer) usually holds hundreds to thousands of sectors that are bad.
    You can take a look at both lists with the format command, too.

    Now I am wondering, how long my disk will run without generating new
    defects...

    Tom
    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.2.4 (SunOS)
    Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

    iD8DBQFEIHjD6U+hp8PKQZIRArInAJ0dFdQjZ+T5j+ojBjMuop uy4JSQ/gCgt+kX
    HFYukxSu4Zl4Qgaj7g+GraQ=
    =yKqS
    -----END PGP SIGNATURE-----

  3. Re: Bad sectors

    In alt.solaris.x86 Thomas Maier-Komor wrote:
    > -----BEGIN PGP SIGNED MESSAGE-----
    > Hash: SHA1
    >
    > Andrew Gabriel wrote:
    >> In article <1142971568.840177.263500@i39g2000cwa.googlegroups. com>,
    >> ehabaziz2001@gmail.com writes:
    >>> What is the impact of bad sector existence on :
    >>> -Solaris Installation
    >>> -Pakages Installation
    >>>
    >>> Did Solaris remark bad sector for not writing in future ?

    >>
    >> For the last 15-20 years, this has been handled automatically
    >> by intelligent controllers embedded in the disk drive.
    >>

    >
    > I must contradict you. You are right that most failures are normally
    > handled by the disk controller firmware. But I just have a disk under
    > SVM that went to state Maintainance, because of bad sectors. I called
    > Sun and the support guy told me that if the number of sectors with
    > errors is low (in my case three), then one can run format->analyze->read
    > or refresh to get the bad sectors assigned to the secondary defect list.
    > The primary defect list (created before the product is shipped to the
    > customer) usually holds hundreds to thousands of sectors that are bad.
    > You can take a look at both lists with the format command, too.
    >
    > Now I am wondering, how long my disk will run without generating new
    > defects...


    If you're getting errors it's time for a new drive either way.

  4. Re: Bad sectors

    In article ,
    Thomas Maier-Komor writes:
    > -----BEGIN PGP SIGNED MESSAGE-----
    > Hash: SHA1
    >
    > Andrew Gabriel wrote:
    >> In article <1142971568.840177.263500@i39g2000cwa.googlegroups. com>,
    >> ehabaziz2001@gmail.com writes:
    >>> What is the impact of bad sector existence on :
    >>> -Solaris Installation
    >>> -Pakages Installation
    >>>
    >>> Did Solaris remark bad sector for not writing in future ?

    >>
    >> For the last 15-20 years, this has been handled automatically
    >> by intelligent controllers embedded in the disk drive.
    >>

    >
    > I must contradict you. You are right that most failures are normally
    > handled by the disk controller firmware. But I just have a disk under
    > SVM that went to state Maintainance, because of bad sectors. I called


    The drive controller will have remapped the sectors when
    the fault was discovered. The problem that arises is that
    in most cases (where the error was serious enough to return),
    the contents of the sectors will have been lost. In this case,
    the drive controller must keep returning an error on reading
    the sector, even though it has been remapped, because it
    doesn't have any valid data to return. The problem will go
    away next time the sector is written to.

    --
    Andrew Gabriel

  5. Re: Bad sectors

    In alt.solaris.x86 Andrew Gabriel wrote:
    > In article ,
    > Thomas Maier-Komor writes:
    >> -----BEGIN PGP SIGNED MESSAGE-----
    >> Hash: SHA1
    >>
    >> Andrew Gabriel wrote:
    >>> In article <1142971568.840177.263500@i39g2000cwa.googlegroups. com>,
    >>> ehabaziz2001@gmail.com writes:
    >>>> What is the impact of bad sector existence on :
    >>>> -Solaris Installation
    >>>> -Pakages Installation
    >>>>
    >>>> Did Solaris remark bad sector for not writing in future ?
    >>>
    >>> For the last 15-20 years, this has been handled automatically
    >>> by intelligent controllers embedded in the disk drive.
    >>>

    >>
    >> I must contradict you. You are right that most failures are normally
    >> handled by the disk controller firmware. But I just have a disk under
    >> SVM that went to state Maintainance, because of bad sectors. I called

    >
    > The drive controller will have remapped the sectors when
    > the fault was discovered. The problem that arises is that
    > in most cases (where the error was serious enough to return),
    > the contents of the sectors will have been lost. In this case,
    > the drive controller must keep returning an error on reading
    > the sector, even though it has been remapped, because it
    > doesn't have any valid data to return. The problem will go
    > away next time the sector is written to.


    The problem is you've already lost data at that point. Remapping doesn't
    bring back lost data. When you get such errors, it's time for a new drive.

  6. Re: Bad sectors

    Thomas Maier-Komor wrote:

    > -----BEGIN PGP SIGNED MESSAGE-----
    > Hash: SHA1
    >
    > Andrew Gabriel wrote:
    >
    >>In article <1142971568.840177.263500@i39g2000cwa.googlegroups. com>,
    >> ehabaziz2001@gmail.com writes:
    >>
    >>>What is the impact of bad sector existence on :
    >>>-Solaris Installation
    >>>-Pakages Installation
    >>>
    >>>Did Solaris remark bad sector for not writing in future ?

    >>
    >>For the last 15-20 years, this has been handled automatically
    >>by intelligent controllers embedded in the disk drive.
    >>

    >
    >
    > I must contradict you. You are right that most failures are normally
    > handled by the disk controller firmware. But I just have a disk under
    > SVM that went to state Maintainance, because of bad sectors. I called
    > Sun and the support guy told me that if the number of sectors with
    > errors is low (in my case three), then one can run format->analyze->read
    > or refresh to get the bad sectors assigned to the secondary defect list.
    > The primary defect list (created before the product is shipped to the
    > customer) usually holds hundreds to thousands of sectors that are bad.
    > You can take a look at both lists with the format command, too.
    >
    > Now I am wondering, how long my disk will run without generating new
    > defects...


    I'd be really careful about making backups if I were you. The disk may
    be all right but it could also be very near the end of its useful life.
    If it develops more errors in the next few days give serious
    consideration to replacing it.


  7. Re: Bad sectors

    -----BEGIN PGP SIGNED MESSAGE-----
    Hash: SHA1

    Richard B. Gilbert wrote:
    > Thomas Maier-Komor wrote:
    >
    >> -----BEGIN PGP SIGNED MESSAGE-----
    >> Hash: SHA1
    >>
    >> Andrew Gabriel wrote:
    >>
    >>> In article <1142971568.840177.263500@i39g2000cwa.googlegroups. com>,
    >>> ehabaziz2001@gmail.com writes:
    >>>
    >>>> What is the impact of bad sector existence on :
    >>>> -Solaris Installation
    >>>> -Pakages Installation
    >>>>
    >>>> Did Solaris remark bad sector for not writing in future ?
    >>>
    >>> For the last 15-20 years, this has been handled automatically
    >>> by intelligent controllers embedded in the disk drive.
    >>>

    >>
    >>
    >> I must contradict you. You are right that most failures are normally
    >> handled by the disk controller firmware. But I just have a disk under
    >> SVM that went to state Maintainance, because of bad sectors. I called
    >> Sun and the support guy told me that if the number of sectors with
    >> errors is low (in my case three), then one can run format->analyze->read
    >> or refresh to get the bad sectors assigned to the secondary defect list.
    >> The primary defect list (created before the product is shipped to the
    >> customer) usually holds hundreds to thousands of sectors that are bad.
    >> You can take a look at both lists with the format command, too.
    >>
    >> Now I am wondering, how long my disk will run without generating new
    >> defects...

    >
    > I'd be really careful about making backups if I were you. The disk may
    > be all right but it could also be very near the end of its useful life.
    > If it develops more errors in the next few days give serious
    > consideration to replacing it.
    >


    Yes, I'll do. I am also having doubts concerning this approach, but as
    the disk is only the submirror of a raid-1, I don't see too much of a
    problem. format->analyze->refresh fixed one bad sector, now it is
    resyncing. I'll just wait and see what happens. If bad sectors happen to
    come up again, I won't care what the support guy says...
    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.2.4 (SunOS)
    Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

    iD8DBQFEIQto6U+hp8PKQZIRAhQDAKD6oH3X/XgoO4tad/P4Dh7APGPi3wCfTSrX
    gei/sVQRZOm/aDmmm3l7CeA=
    =rFrP
    -----END PGP SIGNATURE-----

  8. Re: Bad sectors

    Ok in installation does solaris recognize it is bad sectore then id
    doest not writer over it ?


  9. Re: Bad sectors

    -----BEGIN PGP SIGNED MESSAGE-----
    Hash: SHA1

    ehabaziz2001@gmail.com wrote:
    > Ok in installation does solaris recognize it is bad sectore then id
    > doest not writer over it ?
    >


    As I said before: if an error occurs during writing (e.g. installation
    of Solaris), the disk-controller firmware handles bad sector mapping. If
    a sector gets broken afterwards, Solaris will get a _read_ error that
    _cannot_ be handled with a single disk.

    If you are concerned about data integrity (you probably should be),
    setup raid-1s, make backups, and expect disks to fail. Handle harddisks
    like the V-belt of you car. It will work for a certain time, but might
    die unexpectedly if you don't exchange it on time. You don't have to
    exchange disks providently, because failures in raid-1s can be handled
    by the system. But expect that these failures _will_ occur.
    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.2.4 (SunOS)
    Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

    iD8DBQFEIU/16U+hp8PKQZIRAlyPAKCTlKXcb5xyARwNBgY/GHT+WgiJPACfXhAa
    zlWERXMBiZIs0sIrqpa0XQg=
    =1b+l
    -----END PGP SIGNATURE-----

  10. Re: Bad sectors

    In article ,
    Thomas Maier-Komor wrote:
    >-----BEGIN PGP SIGNED MESSAGE-----
    >Hash: SHA1
    >
    >Richard B. Gilbert wrote:
    >> Thomas Maier-Komor wrote:
    >>
    >>> -----BEGIN PGP SIGNED MESSAGE-----
    >>> Hash: SHA1
    >>>
    >>> Andrew Gabriel wrote:
    >>>
    >>>> In article <1142971568.840177.263500@i39g2000cwa.googlegroups. com>,
    >>>> ehabaziz2001@gmail.com writes:
    >>>>
    >>>>> What is the impact of bad sector existence on :
    >>>>> -Solaris Installation
    >>>>> -Pakages Installation
    >>>>>
    >>>>> Did Solaris remark bad sector for not writing in future ?
    >>>>
    >>>> For the last 15-20 years, this has been handled automatically
    >>>> by intelligent controllers embedded in the disk drive.
    >>>>
    >>>
    >>>
    >>> I must contradict you. You are right that most failures are normally
    >>> handled by the disk controller firmware. But I just have a disk under
    >>> SVM that went to state Maintainance, because of bad sectors. I called
    >>> Sun and the support guy told me that if the number of sectors with
    >>> errors is low (in my case three), then one can run format->analyze->read
    >>> or refresh to get the bad sectors assigned to the secondary defect list.
    >>> The primary defect list (created before the product is shipped to the
    >>> customer) usually holds hundreds to thousands of sectors that are bad.
    >>> You can take a look at both lists with the format command, too.
    >>>
    >>> Now I am wondering, how long my disk will run without generating new
    >>> defects...

    >>
    >> I'd be really careful about making backups if I were you. The disk may
    >> be all right but it could also be very near the end of its useful life.
    >> If it develops more errors in the next few days give serious
    >> consideration to replacing it.
    >>

    >
    >Yes, I'll do. I am also having doubts concerning this approach, but as
    >the disk is only the submirror of a raid-1, I don't see too much of a
    >problem. format->analyze->refresh fixed one bad sector, now it is
    >resyncing. I'll just wait and see what happens. If bad sectors happen to
    >come up again, I won't care what the support guy says...


    My experience has been that if a fairly new disk (less than 6 months old)
    gets a few bad sectors and then does not soon get any more, the disk is
    OK. If an older disk (over 3 years old) that has been stable gets a few
    bad sectors, it will soon get more and more and die quickly (in a few
    weeks at most, possibly in a few days). If the disk is old, at least
    get a new one in house so that you can act quickly when it fails.

    >-----BEGIN PGP SIGNATURE-----
    >Version: GnuPG v1.2.4 (SunOS)
    >Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
    >
    >iD8DBQFEIQto6U+hp8PKQZIRAhQDAKD6oH3X/XgoO4tad/P4Dh7APGPi3wCfTSrX
    >gei/sVQRZOm/aDmmm3l7CeA=
    >=rFrP
    >-----END PGP SIGNATURE-----



    --
    Tom Schulz
    schulz@adi.com

  11. Re: Bad sectors

    In article ,
    schulz@adi.com (Thomas Schulz) writes:
    >
    >
    > My experience has been that if a fairly new disk (less than 6 months old)
    > gets a few bad sectors and then does not soon get any more, the disk is
    > OK. If an older disk (over 3 years old) that has been stable gets a few
    > bad sectors, it will soon get more and more and die quickly (in a few
    > weeks at most, possibly in a few days). If the disk is old, at least
    > get a new one in house so that you can act quickly when it fails.
    >


    My experience has been that disk failures from bad sectors are really
    rare. Most of my disk failures in production environments have been
    catostrophic failures. (ie. complete electronc failure, head crash, etc.)

    bill

    --
    Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves
    bill@cs.scranton.edu | and a sheep voting on what's for dinner.
    University of Scranton |
    Scranton, Pennsylvania | #include

  12. Re: Bad sectors

    Bill Gunshannon wrote:
    > In article ,
    > schulz@adi.com (Thomas Schulz) writes:
    >
    >>
    >>My experience has been that if a fairly new disk (less than 6 months old)
    >>gets a few bad sectors and then does not soon get any more, the disk is
    >>OK. If an older disk (over 3 years old) that has been stable gets a few
    >>bad sectors, it will soon get more and more and die quickly (in a few
    >>weeks at most, possibly in a few days). If the disk is old, at least
    >>get a new one in house so that you can act quickly when it fails.
    >>

    >
    >
    > My experience has been that disk failures from bad sectors are really
    > rare. Most of my disk failures in production environments have been
    > catostrophic failures. (ie. complete electronc failure, head crash, etc.)
    >

    Bad sector errors form an otherwise well behaved drive can be a sign of
    things to come, I've lost several that way. I used to patch them up
    with format, now I run my truck over them and get a new one.

    --
    Ian Collins.

  13. Re: Bad sectors

    In comp.unix.solaris Andrew Gabriel wrote:
    > In article ,
    > Cydrome Leader writes:
    >> In alt.solaris.x86 Andrew Gabriel wrote:
    >>> In article ,
    >>> Thomas Maier-Komor writes:
    >>>> -----BEGIN PGP SIGNED MESSAGE-----
    >>>> Hash: SHA1
    >>>>
    >>>> Andrew Gabriel wrote:
    >>>>> In article <1142971568.840177.263500@i39g2000cwa.googlegroups. com>,
    >>>>> ehabaziz2001@gmail.com writes:
    >>>>>> What is the impact of bad sector existence on :
    >>>>>> -Solaris Installation
    >>>>>> -Pakages Installation
    >>>>>>
    >>>>>> Did Solaris remark bad sector for not writing in future ?
    >>>>>
    >>>>> For the last 15-20 years, this has been handled automatically
    >>>>> by intelligent controllers embedded in the disk drive.
    >>>>>
    >>>>
    >>>> I must contradict you. You are right that most failures are normally
    >>>> handled by the disk controller firmware. But I just have a disk under
    >>>> SVM that went to state Maintainance, because of bad sectors. I called
    >>>
    >>> The drive controller will have remapped the sectors when
    >>> the fault was discovered. The problem that arises is that
    >>> in most cases (where the error was serious enough to return),
    >>> the contents of the sectors will have been lost. In this case,
    >>> the drive controller must keep returning an error on reading
    >>> the sector, even though it has been remapped, because it
    >>> doesn't have any valid data to return. The problem will go
    >>> away next time the sector is written to.

    >>
    >> The problem is you've already lost data at that point. Remapping doesn't
    >> bring back lost data. When you get such errors, it's time for a new drive.

    >
    > Errors are routine on disks. The embedded controller will usually
    > hide them from you as it will retry and eventually read the sector
    > before remapping it, and you won't know this happened (well, it is
    > signalled back to the OS, but usually goes unnoticed). You only get


    Correct, drives will correct the majority of errors themselves, with
    notice to the OS. There are sometimes recoverable errors that an OS is
    made aware of. These maybe weeks, months of years apart. I keep an eye on
    these, but won't immediately replace such a disk unless the sectors that
    are erroring are always the same or the happen more than several times in
    a row.

    Unrecoverable errors are a clear sign a disk is failing. Once it can no
    longer correct and remap sectors something is very wrong. A damaged media
    surface can be the cause, and may introduce garbage in the drive that may
    or may not get caught in the air filter. If it doesn't, it will just
    destroy more and more of the disk.

    Disks are really cheap these days, and there is very little reason to keep
    running off a known bad drive. They never fix themselves or get better,
    only worse.

    > to see the occasional one where the retries also fail. They only
    > become a cause for concern if the rate of them suddenly increases.


    Which it always does.


  14. Re: Bad sectors

    In comp.unix.solaris Bill Gunshannon wrote:
    > In article ,
    > schulz@adi.com (Thomas Schulz) writes:
    >>
    >>
    >> My experience has been that if a fairly new disk (less than 6 months old)
    >> gets a few bad sectors and then does not soon get any more, the disk is
    >> OK. If an older disk (over 3 years old) that has been stable gets a few
    >> bad sectors, it will soon get more and more and die quickly (in a few
    >> weeks at most, possibly in a few days). If the disk is old, at least
    >> get a new one in house so that you can act quickly when it fails.
    >>

    >
    > My experience has been that disk failures from bad sectors are really
    > rare. Most of my disk failures in production environments have been
    > catostrophic failures. (ie. complete electronc failure, head crash, etc.)


    My experience is the opposite. IDE disks aside, I don't recall the last
    time I flagged a SCSI disk bad that wasn't able to spin up at all, or
    simply wasn't detected on the SCSI chain. Read/write errors seem to be
    more common, especially for aged drives. I have these swapped before they
    completely fail. If left in use, it's possible they would completely
    destroy themselves and never turn on.

    IDE drives seem to fail light light bulbs, they' just dead all of a
    sudden. SCSI drives seem a bit more graceful about failing.

    >
    > bill
    >


  15. Re: Bad sectors

    -----BEGIN PGP SIGNED MESSAGE-----
    Hash: SHA1

    ehabaziz2001@gmail.com wrote:
    > Ok in installation does solaris recognize it is bad sectore then id
    > doest not writer over it ?
    >


    As I said before: if an error occurs during writing (e.g. installation
    of Solaris), the disk-controller firmware handles bad sector mapping. If
    a sector gets broken afterwards, Solaris will get a _read_ error that
    _cannot_ be handled with a single disk.

    If you are concerned about data integrity (you probably should be),
    setup raid-1s, make backups, and expect disks to fail. Handle harddisks
    like the V-belt of you car. It will work for a certain time, but might
    die unexpectedly if you don't exchange it on time. You don't have to
    exchange disks providently, because failures in raid-1s can be handled
    by the system. But expect that these failures _will_ occur.
    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.2.4 (SunOS)
    Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

    iD8DBQFEIRZX6U+hp8PKQZIRAmGXAKCbWhS2ra36qv2QcIuIie ByqLxAOACgrqyz
    LnAUCb2hvts/VeyGOSlAYpg=
    =I1S5
    -----END PGP SIGNATURE-----

+ Reply to Thread