anal/disk/shadow weirdnes - VMS

This is a discussion on anal/disk/shadow weirdnes - VMS ; I get a parity error with ANAL/DISK/SHADOW. This puts errors in the error log for both disks in the shadow set at the same LBN ANAL/DISK/SHADOW is suppose to only check blocks in files according to the HELP. But when ...

+ Reply to Thread
Results 1 to 9 of 9

Thread: anal/disk/shadow weirdnes

  1. anal/disk/shadow weirdnes

    I get a parity error with ANAL/DISK/SHADOW. This puts errors in the
    error log for both disks in the shadow set at the same LBN

    ANAL/DISK/SHADOW is suppose to only check blocks in files according to
    the HELP.

    But when I run DFU on the LBN of the error in the log it does not
    identify a file.

    And when I run ANAL/DISK/READ, it has no parity errors.

    The shadow disk itself DSA0: logs no errors.

    So is the block with the parity error in a file or not?


  2. Re: anal/disk/shadow weirdnes

    ANAL/DISK/SHADOW

    Examines the entire contents of a shadow set or a specified range
    of blocks in a shadow set for discrepancies.

    ANAL/DISK/READ only reads blocks allocated to files.

    So the bad blocks are NOT in a file.

    Volker.

  3. Re: anal/disk/shadow weirdnes

    In article <87e22c1f-ca1a-4fb3-95be-1395b828a415
    @e23g2000prf.googlegroups.com>, volker_halle@hotmail.com says...
    > ANAL/DISK/SHADOW
    >
    > Examines the entire contents of a shadow set or a specified range
    > of blocks in a shadow set for discrepancies.
    >
    > ANAL/DISK/READ only reads blocks allocated to files.
    >
    > So the bad blocks are NOT in a file.
    >
    > Volker.
    >


    $ anal/disk/shad/blocks=file_system

    apparently does what the OP claims ana/disk/shadow is supposed
    to do.

    (I've never had occasion to use this.)

    --
    John

  4. Re: anal/disk/shadow weirdnes

    On Feb 25, 2:15*pm, Volker Halle wrote:
    > ANAL/DISK/SHADOW
    >
    > Examines the entire contents of a shadow set or a specified range
    > * * * * *of blocks in a shadow set for discrepancies.
    >
    > ANAL/DISK/READ only reads blocks allocated to files.
    >
    > So the bad blocks are NOT in a file.
    >
    > Volker.


    But HELP says /BLOCKS=FILE_SYSTEM is the default.

    Check HELP. I am running 7.3.2

  5. Re: anal/disk/shadow weirdnes

    On Feb 25, 2:25*pm, John Santos wrote:
    > In article <87e22c1f-ca1a-4fb3-95be-1395b828a415
    > @e23g2000prf.googlegroups.com>, volker_ha...@hotmail.com says...
    >
    > > ANAL/DISK/SHADOW

    >
    > > Examines the entire contents of a shadow set or a specified range
    > > * * * * *of blocks in a shadow set for discrepancies.

    >
    > > ANAL/DISK/READ only reads blocks allocated to files.

    >
    > > So the bad blocks are NOT in a file.

    >
    > > Volker.

    >
    > $ anal/disk/shad/blocks=file_system
    >
    > apparently does what the OP claims ana/disk/shadow is supposed
    > to do.
    >
    > (I've never had occasion to use this.)
    >
    > --
    > John


    What do you mean by OP.

    Blocks=file is the default, according to help.

    But I will try it.

  6. Re: anal/disk/shadow weirdnes

    In article <2bf2c280-5ed6-47e4-8855-ed0065a6bbd0
    @q78g2000hsh.googlegroups.com>, tadamsmar@yahoo.com says...
    > On Feb 25, 2:25*pm, John Santos wrote:
    > > In article <87e22c1f-ca1a-4fb3-95be-1395b828a415
    > > @e23g2000prf.googlegroups.com>, volker_ha...@hotmail.com says...
    > >
    > > > ANAL/DISK/SHADOW

    > >
    > > > Examines the entire contents of a shadow set or a specified range
    > > > * * * * *of blocks in a shadow set for discrepancies.

    > >
    > > > ANAL/DISK/READ only reads blocks allocated to files.

    > >
    > > > So the bad blocks are NOT in a file.

    > >
    > > > Volker.

    > >
    > > $ anal/disk/shad/blocks=file_system
    > >
    > > apparently does what the OP claims ana/disk/shadow is supposed
    > > to do.
    > >
    > > (I've never had occasion to use this.)
    > >
    > > --
    > > John

    >
    > What do you mean by OP.
    >


    OP == Original Poster, I.e. You!

    > Blocks=file is the default, according to help.
    >


    I don't see anywhere where it says it's the default on either Alpha
    V7.3-2 or on Alpha V8.3. Maybe it used to say that, and was wrong,
    and an ECO fixed the help text? ... Oh, wait, under FILE_SYSTEM
    under /BLOCKS, it does say: This is the default. But I think that
    only applies if you say anal/disk/shadow/blocks without specifying
    either =all or =file_system... But /BLOCKS *is not* the default for
    anal/disk/shadow, so it doesn't matter what the default style of
    /blocks is if you don't specify it at all.

    In my experience, it (without the /blocks) has always compared
    everything.


    > But I will try it.
    >


    --
    John

  7. Re: anal/disk/shadow weirdnes

    On Feb 25, 5:29*pm, John Santos wrote:
    > In article <2bf2c280-5ed6-47e4-8855-ed0065a6bbd0
    > @q78g2000hsh.googlegroups.com>, tadams...@yahoo.com says...
    >
    >
    >
    >
    >
    > > On Feb 25, 2:25*pm, John Santos wrote:
    > > > In article <87e22c1f-ca1a-4fb3-95be-1395b828a415
    > > > @e23g2000prf.googlegroups.com>, volker_ha...@hotmail.com says...

    >
    > > > > ANAL/DISK/SHADOW

    >
    > > > > Examines the entire contents of a shadow set or a specified range
    > > > > * * * * *of blocks in a shadow set for discrepancies.

    >
    > > > > ANAL/DISK/READ only reads blocks allocated to files.

    >
    > > > > So the bad blocks are NOT in a file.

    >
    > > > > Volker.

    >
    > > > $ anal/disk/shad/blocks=file_system

    >
    > > > apparently does what the OP claims ana/disk/shadow is supposed
    > > > to do.

    >
    > > > (I've never had occasion to use this.)

    >
    > > > --
    > > > John

    >
    > > What do you mean by OP.

    >
    > OP == Original Poster, I.e. You!
    >
    > > Blocks=file is the default, according to help.

    >
    > I don't see anywhere where it says it's the default on either Alpha
    > V7.3-2 or on Alpha V8.3. *Maybe it used to say that, and was wrong,
    > and an ECO fixed the help text? *... *Oh, wait, under FILE_SYSTEM
    > under /BLOCKS, it does say: This is the default. *But I think that
    > only applies if you say anal/disk/shadow/blocks without specifying
    > either =all or =file_system... *But /BLOCKS *is not* the default for
    > anal/disk/shadow, so it doesn't matter what the default style of
    > /blocks is if you don't specify it at all.
    >
    > In my experience, it (without the /blocks) has always compared
    > everything.
    >
    > > But I will try it.

    >
    > --
    > John- Hide quoted text -
    >
    > - Show quoted text -


    I tried /blocks=file

    got the same result. It failed with a parity error. It produced an
    error on each disk in the error log at the same LBN. But the LBN for
    the error is not in a file according to DFU.

    So the explanation of you and Volker don't hold water. Still a
    mystery why anal/disk/shad/block=file get a parity error on a block
    that is not in a file according to DFU.

  8. Re: anal/disk/shadow weirdnes

    tadamsmar wrote:
    > On Feb 25, 5:29 pm, John Santos wrote:
    >
    >>In article <2bf2c280-5ed6-47e4-8855-ed0065a6bbd0
    >>@q78g2000hsh.googlegroups.com>, tadams...@yahoo.com says...
    >>
    >>
    >>
    >>
    >>
    >>
    >>>On Feb 25, 2:25 pm, John Santos wrote:
    >>>
    >>>>In article <87e22c1f-ca1a-4fb3-95be-1395b828a415
    >>>>@e23g2000prf.googlegroups.com>, volker_ha...@hotmail.com says...

    >>
    >>>>>ANAL/DISK/SHADOW

    >>
    >>>>>Examines the entire contents of a shadow set or a specified range
    >>>>> of blocks in a shadow set for discrepancies.

    >>
    >>>>>ANAL/DISK/READ only reads blocks allocated to files.

    >>
    >>>>>So the bad blocks are NOT in a file.

    >>
    >>>>>Volker.

    >>
    >>>>$ anal/disk/shad/blocks=file_system

    >>
    >>>>apparently does what the OP claims ana/disk/shadow is supposed
    >>>>to do.

    >>
    >>>>(I've never had occasion to use this.)

    >>
    >>>>--
    >>>>John

    >>
    >>>What do you mean by OP.

    >>
    >>OP == Original Poster, I.e. You!
    >>
    >>
    >>>Blocks=file is the default, according to help.

    >>
    >>I don't see anywhere where it says it's the default on either Alpha
    >>V7.3-2 or on Alpha V8.3. Maybe it used to say that, and was wrong,
    >>and an ECO fixed the help text? ... Oh, wait, under FILE_SYSTEM
    >>under /BLOCKS, it does say: This is the default. But I think that
    >>only applies if you say anal/disk/shadow/blocks without specifying
    >>either =all or =file_system... But /BLOCKS *is not* the default for
    >>anal/disk/shadow, so it doesn't matter what the default style of
    >>/blocks is if you don't specify it at all.
    >>
    >>In my experience, it (without the /blocks) has always compared
    >>everything.
    >>
    >>
    >>>But I will try it.

    >>
    >>--
    >>John- Hide quoted text -
    >>
    >>- Show quoted text -

    >
    >
    > I tried /blocks=file
    >
    > got the same result. It failed with a parity error. It produced an
    > error on each disk in the error log at the same LBN. But the LBN for
    > the error is not in a file according to DFU.
    >
    > So the explanation of you and Volker don't hold water. Still a
    > mystery why anal/disk/shad/block=file get a parity error on a block
    > that is not in a file according to DFU.


    Another possibility: Judging by the number of I/O's required to do a
    full shadow comapare, ana/disk/shadow is using very big buffers and
    reading a lot with each I/O. Maybe it still does the same with
    /blocks=file, and what you are seeing is bad blocks in the same "chunk"
    as some file, though not necessarily in a file.

    A third possibility: A bug.


    --
    John Santos
    Evans Griffiths & Hart, Inc.
    781-861-0670 ext 539

  9. Re: anal/disk/shadow weirdnes

    Does ANAL/DISK/READ also report those parity errors ?

    I also interpret ANAL/DISK/SHADOW help in such a way, that you can
    limit the data it's comparing with the /BLOCKS qualifier. If you
    specifiy just /BLOCKS, FILE_SYSTEM is the default option for /BLOCKS.
    If you don't specify /BLOCKS (or specify /BLOCKS=ALL), all disk blocks
    will be compared.

    Volker.

+ Reply to Thread