CHKDSK SYS3175 error? - OS2

This is a discussion on CHKDSK SYS3175 error? - OS2 ; Anyone ever seen CHKDSK for OS/2 come back with a SYS3175 error when attempting a run on a 2GB HPFS hard disk partition. With a bad partition and system shutdown error showing from improper shutdown? From a successful OS/2 Utility ...

+ Reply to Thread
Results 1 to 10 of 10

Thread: CHKDSK SYS3175 error?

  1. CHKDSK SYS3175 error?

    Anyone ever seen CHKDSK for OS/2 come back with a SYS3175 error when attempting
    a run on a 2GB HPFS hard disk partition. With a bad partition and system
    shutdown error showing from improper shutdown? From a successful OS/2 Utility
    Floppy Diskette boot run that can see the C: boot drive but can't access the D:
    drive sited?

    CHKDSK can't access the partition without SYS3175 error, though it works fine
    on drive C:, the boot partition, for example.

    Note .. This is Warp 4 and not MCP2. It may be possible a few hours from now
    to do a floppy diskette boot run for MCP2 on this same Adaptec SCSI drive
    system and try that CHKDSK.

    The system not here but a 'victim' of Hurricane Eduardo power failure and huge
    power issues even with APC UPS support.

    Thoughts?


    --


    --> Sleep well; OS2's still awake!

    Mike Luther

  2. Re: CHKDSK SYS3175 error?

    [A complimentary Cc of this posting was sent to
    Mike Luther
    ], who wrote in article <4898c8a5$0$4024$bbae4d71@news.suddenlink.net>:

    > Anyone ever seen CHKDSK for OS/2 come back with a SYS3175 error when attempting
    > a run on a 2GB HPFS hard disk partition.


    I think I have seen it. Do not remember what I needed to do (I had no
    full backup)...

    I would (in this order)

    a) check the partition table for sanity; (DFSee?)

    b) Is not there a chkdsk in DfSee?

    c) Set the "OK" flag manually, and copy files which are salvageable
    (unless you have a backup);

    d) reformat.

    Yours,
    Ilya

  3. Re: CHKDSK SYS3175 error?

    On 08/06/08 12:26 pm, Ilya Zakharevich wrote:
    > [A complimentary Cc of this posting was sent to
    > Mike Luther
    > ], who wrote in article<4898c8a5$0$4024$bbae4d71@news.suddenlink.net>:
    >
    >> Anyone ever seen CHKDSK for OS/2 come back with a SYS3175 error when attempting
    >> a run on a 2GB HPFS hard disk partition.

    >
    > I think I have seen it. Do not remember what I needed to do (I had no
    > full backup)...
    >
    > I would (in this order)
    >
    > a) check the partition table for sanity; (DFSee?)
    >
    > b) Is not there a chkdsk in DfSee?
    >
    > c) Set the "OK" flag manually, and copy files which are salvageable
    > (unless you have a backup);
    >
    > d) reformat.
    >


    Instead of d) and if desperate you could do a quick format (format x:
    /fs:hpfs) then run chkdsk /f:3 against the drive. This will recover most
    all files into \FOUND0 though you will lose all EAs and file names will
    be truncated to 15 chars.
    Dave

  4. Re: CHKDSK SYS3175 error?

    Thanks!

    Ilya Zakharevich wrote:
    > [A complimentary Cc of this posting was sent to
    > Mike Luther
    > ], who wrote in article <4898c8a5$0$4024$bbae4d71@news.suddenlink.net>:
    >
    >> Anyone ever seen CHKDSK for OS/2 come back with a SYS3175 error when attempting
    >> a run on a 2GB HPFS hard disk partition.

    >
    > I think I have seen it. Do not remember what I needed to do (I had no
    > full backup)...


    A bit more research. From the copy of CHKDSK and CHKDSK32 on this system they
    are earlier than the current releases. I note there were two IBM APAR's that
    address failure of CHKDSK with Warp 4 like this, including one with floppy disk
    booted systems.

    > I would (in this order)
    >
    > a) check the partition table for sanity; (DFSee?)


    That was my thought too. On your confirmation thought, I booted to OS/2 from a
    set of floppy diskettes and ran DFSEE. I chose to check the HPFS file system
    with it. Following that I check the partitioning. Following that I chose to
    deliberately clear the dirty byte flag.
    >
    > b) Is not there a chkdsk in DfSee?
    >


    I didn't get that far. In that this was not the boot partition, I then
    immediately got a start up. There were several still apparently 'in use'
    applications that had never been shut down. A Win-OS2 application. A Smart
    Suite for OS/2 application. Two DOS applications. I shut them down.

    From that point with the hard disk booted system, I was able to drop to an
    OS/2 command window. I was then able to run CHKDSK D: /F, which at that point
    produced no errors at all on this partition.

    Problem solved.

    > c) Set the "OK" flag manually, and copy files which are salvageable
    > (unless you have a backup);
    >
    > d) reformat.


    Fortunately, the files were still all clean, as far as we can tell. As well
    there was no apparent need to reformat.

    It looks like this whole show erupted over the un-attended half dozen sharp
    power surge and total off slices in the facility from Tropical Storm Eduaord
    that even the APC UPS couldn't cope with.

    Once the dirty bit got set, and a couple passes of lockups were attempted on
    the reboot, this SYS3175 error with CHKDSK could only be by-passed with the use
    of DFSEE to clean up the mess by somewhat more crude brute force.

    Thank you for your support!


    --


    --> Sleep well; OS2's still awake!

    Mike Luther

  5. Re: CHKDSK SYS3175 error?

    On Tue, 5 Aug 2008 21:39:51 UTC in comp.os.os2.setup.storage, Mike Luther
    wrote:

    > Anyone ever seen CHKDSK for OS/2 come back with a SYS3175 error when attempting
    > a run on a 2GB HPFS hard disk partition. With a bad partition and system
    > shutdown error showing from improper shutdown? From a successful OS/2 Utility
    > Floppy Diskette boot run that can see the C: boot drive but can't access the D:
    > drive sited?


    There are various versions of chkdsk that you can try - if the partition is not
    large then you can try the old old Warp 3 chkdsk from before it was rewritten to
    give decent chkdsk times. Old code but well tested so may take a long time but
    succeed. If you have an hpfs386 license then perhaps you can try with a boot
    diskette set using that IFS too though you still end up using uhpfs.dll to
    actually do the chkdsk itself. Otherwise various fixpack versions of
    hpfs.ifs/uhpfs.dll might help.

    --
    Trevor Hemsley, Brighton, UK
    Trevor dot Hemsley at ntlworld dot com

  6. Re: CHKDSK SYS3175 error?

    Hi Trevor .. and others who helped. More findings.

    Trevor Hemsley wrote:


    > There are various versions of chkdsk that you can try - if the partition is not
    > large then you can try the old old Warp 3 chkdsk from before it was rewritten to
    > give decent chkdsk times. Old code but well tested so may take a long time but
    > succeed. If you have an hpfs386 license then perhaps you can try with a boot
    > diskette set using that IFS too though you still end up using uhpfs.dll to
    > actually do the chkdsk itself. Otherwise various fixpack versions of
    > hpfs.ifs/uhpfs.dll might help.


    I got this same box back two days ago to recover again! But this time, the
    facility administrator didn't like the way a workstation was printing
    something. So she just slam shutdown the server plus everything else by just
    turning off the power switches!! Server first!

    In this case, the OS/2 boot partition was clean (enough?) and would get to the
    boot point where CHKDSK was attempting to check the Drive D: that had given
    trouble before! What was happening was that CHKDSK would simply fail and the
    screen would go silent, never finishing the boot run. I didn't try something I
    should have but only learned much later. A at that point.

    I did boot the box from the utility floppy disks for it. Then tried to run
    CHKDSK from that on the D: partition. Same story! But I did 'accidentally'
    learn that if I did a there, CHKDSK would simply 'fail' and leave me
    with a functional system. back to the A:\ prompt. But still a dirty drive,
    somehow. Though no SYS3175 errors like I saw before.

    With what I'd learned before, I again used DFSEE to check the HPFS file system,
    and so on, plus reset the dirty byte to clean. Presto! Box booted again
    normally. And more curiously, from a Desktop Connections folder request to
    CHKDSK this drive, it all came out perfect! As well, so with a command line
    try at this on the booted box.

    But, surprise! Every time I'd try to CHKDSK this drive with a floppy disk
    boot, as well as the dirty byte forced CHKDSK for it, I'd hit this same curious
    lockup and CHKDSK would not finish. In keeping with Ilya's and your thoughts,
    I checked the release date, time and size of the CHKDSK files. Apparently OK
    here. But then, for test purposes, I chose to try an MCP2 floppy disk #3
    CHKDSK from the Warp 4 diskette boot run.

    It completed! Duhh, with an EARLIER date release for CHKSDK.COM, but the same
    byte size for the file and the same CHKDSK32.DLL file size and date on the Warp
    4 box so floppy booted. Totally puzzled, I decided to try something else. I
    manually copied CHKDSK32.DLL from the system on to the Disk #3 of the Warp 4
    Utility Boot diskettes. I then jammed it in to a floppy boot run and ran
    CHKDSK D: from it!

    POOF! Totally successful CHKDSK from this clean drive! But without the
    CHKDSK32.DLL on the floppy diskette, no success. Interesting.

    Trying to learn further about this, in line with what I'd never known about the
    T####### 'files' which are crazy Desktop images I'd seen on this and MANY other
    older systems in use for years, I took a look at this same Drive D: in the
    connections folder. Surprise! Absolutely no way to see this in any command
    line or File Freedom listing for what was on the drive, but in the Connections
    Folder object .. there were:

    TWO duplicate Drive C: Drives images!

    I cloned the whole drive to a backup drive for safety. Then I deleted those
    two unwanted curious Drive C: Drives images and cleaned this with Unimaint.
    Supposedly, that 'fixed' this. But then, after learning about things like this
    I also ran a complete Checkini /c run. Surprise! Even though Unimaint thought
    things were fine, here were MORE of those corrupt objects that Checkini cleaned
    out! OK, I then tested this whole mess again.

    Which .. from a floppy diskette booted CHKDSK without the CHKDSK32.DLL on the
    actual Disk #3, correctly finished the operation. And .. in theory, but I have
    no way of testing this, I'm guessing would NOT have produced the whole re-start
    lockup that caused them to bring the box back up here for round two of this.

    Thoughts? I've seen the APAR's on CHKDSK ref to the SYS3175 error. And not,
    at this time, done any GFC compare against the two CHKDSK executables to see if
    despite the same byte size and so on, there may be a difference on them so that
    the EARLIER date one doesn't do what the LATER date one does! In short, what
    might we do to prevent a future power slam off deal from failing this way?

    Inquiring mind wants to know!

    --


    --> Sleep well; OS2's still awake!

    Mike Luther

  7. Re: CHKDSK SYS3175 error?

    [A complimentary Cc of this posting was sent to
    Mike Luther
    ], who wrote in article <48cc14af$0$4033$bbae4d71@news.suddenlink.net>:
    > But, surprise! Every time I'd try to CHKDSK this drive with a floppy disk
    > boot, as well as the dirty byte forced CHKDSK for it, I'd hit this same curious
    > lockup and CHKDSK would not finish.


    Some versions of CHKDSK algorithms (with/without _32???) need an
    enormous amount of memory. When swapping is available, they have a
    good chance to succeed. When you run from floppies, or (?) when
    chkdsk is auto-run from IFS, swapping is not available.

    Hope this helps,
    Ilya

  8. Re: CHKDSK SYS3175 error?

    Ilya Zakharevich schrieb:
    >
    > [A complimentary Cc of this posting was sent to
    > Mike Luther
    > ], who wrote in article <48cc14af$0$4033$bbae4d71@news.suddenlink.net>:
    > > But, surprise! Every time I'd try to CHKDSK this drive with a floppy disk
    > > boot, as well as the dirty byte forced CHKDSK for it, I'd hit this same curious
    > > lockup and CHKDSK would not finish.

    >
    > Some versions of CHKDSK algorithms (with/without _32???) need an
    > enormous amount of memory. When swapping is available, they have a
    > good chance to succeed. When you run from floppies, or (?) when
    > chkdsk is auto-run from IFS, swapping is not available.
    >


    Maybe he needs the help of chkdsk.sys ?

    Hendrik

  9. Re: CHKDSK SYS3175 error?

    Thanks Hendrik .. more thoughts?

    Hendrik Schmieder wrote:

    > Ilya Zakharevich schrieb:


    >> Some versions of CHKDSK algorithms (with/without _32???) need an
    >> enormous amount of memory. When swapping is available, they have a
    >> good chance to succeed. When you run from floppies, or (?) when
    >> chkdsk is auto-run from IFS, swapping is not available.
    >>

    >
    > Maybe he needs the help of chkdsk.sys ?
    >
    > Hendrik


    Is this the the CONFIG.SYS line;

    BASEDEV=CHKDSK.SYS

    that you are suggesting? I thought this was only for very large huge HPFS
    drive partition sizes, and this box only has three 2GB partitions total on it.
    Although the number of files on the D: drive is quite large for application
    reasons I do not control

    Due to Hurricane Ike now, I've had another pass at this. Massive power
    flickering and on and off did it again to this server. Another pass at a
    locked box that got shut down hard. This is, again, a Warp 4 box, fixpack 17
    level. This time it came back on in dirty boot mode. It successfully
    completed the CHKDSK forced cleanup on drive C:. It jammed, again, on Drive D:
    and wouldn't complete.

    The tech on site then used the new set of floppy boot diskettes I'd built for
    the box that had the exact same issue before, but this time, with the exact
    same CHKDSK32.DLL that is on this box in the OS2\DLL directory copied to the A:
    floppy diskette #3, CHKDSK will complete and cleans up the drive just fine!
    It did for this technician.

    Interesting. Now what I wonder, is if I were to manually copy the CHKDSK32.DLL
    into the C:\OS2\BOOT directory, if this happens again, would the dirty byte
    CHKDSK run complete on the hard disk reboot, the same way as it does if it is
    in the A: floppy from which the same CHKSDSK is run that way?

    More information that might be applicable here. In this case, SWAPPER.DAT has
    been moved off of the boot drive C: to that D: HPFS drive. I'm wondering. If
    it were put back on C: as installed, would that result in a 'locked' cleanup of
    Drive C: when this happens next? And that maybe, the reason it cleaned up C:
    from the dirty byte set mode, is related to the absence of the SWAPPER.DAT
    being in the installed location on C: drive?

    I also could move it to a much less used and populated Drive E: in this case as
    well. In that case, I'm wondering that if this happens again we'd get an auto
    cleanup of both C: and D:, and because E: is much less congested with files and
    has more free space, would we maybe avoid this with E: that way? Or would we
    see the same thing with E: that we are now seeing with D:?

    Yes, I know this is still Warp 4 and not MCP2. But at the moment, I'm not in a
    position to 'upgrade' this box to MCP1 and then MCP2, which is the only way
    I've got to handle the huge number of files and so on. Which I have done many
    times very successfully with Warp 4 to MCP1 and thence MCP2, but never been
    able to ever do with any of these systems trying to move to any form of eCs
    operations.

    Thoughts? Thanks!


    --


    --> Sleep well; OS2's still awake!

    Mike Luther

  10. Re: CHKDSK SYS3175 error?

    Mike Luther schrieb:
    >
    >
    > Hendrik Schmieder wrote:
    >
    > > Ilya Zakharevich schrieb:

    >
    > >> Some versions of CHKDSK algorithms (with/without _32???) need an
    > >> enormous amount of memory. When swapping is available, they have a
    > >> good chance to succeed. When you run from floppies, or (?) when
    > >> chkdsk is auto-run from IFS, swapping is not available.
    > >>

    > >
    > > Maybe he needs the help of chkdsk.sys ?
    > >
    > > Hendrik

    >
    > Thanks Hendrik .. more thoughts?


    Sorry , no.

+ Reply to Thread