HSG80 errors - DEC

This is a discussion on HSG80 errors - DEC ; Ken Fairfield wrote: > A "soft" error. If this happens only very occassionally, > there may be no problem. Unfortunately, on HSG80's with ACS > firmware prior to V8.8-F2, the bad block replacement won't > occur and you keep getting ...

+ Reply to Thread
Results 1 to 6 of 6

Thread: HSG80 errors

  1. HSG80 errors


    Ken Fairfield wrote:
    > A "soft" error. If this happens only very occassionally,
    > there may be no problem. Unfortunately, on HSG80's with ACS
    > firmware prior to V8.8-F2, the bad block replacement won't
    > occur and you keep getting the same disk blocks flagging
    > errors (sigh). I don't have any first hand experience with
    > HSZ's, just a warnging.


    That's discouraging to hear. I have 4 HSG80's with ACS 8.6F-13 and
    I've noticed a few soft errors. As a hobbyist I was lucky to get 4
    cards with that recent of a version. Finding V8.8 for a reasonable
    price is unlikely. Since I'm using 8GB and 18GB drives I guess it
    would just be cheaper to swap a drive. (double sigh)


  2. Re: HSG80 errors

    johnhreinhardt@yahoo.com wrote:
    > Ken Fairfield wrote:
    >
    >> A "soft" error. If this happens only very occassionally,
    >>there may be no problem. Unfortunately, on HSG80's with ACS
    >>firmware prior to V8.8-F2, the bad block replacement won't
    >>occur and you keep getting the same disk blocks flagging
    >>errors (sigh). I don't have any first hand experience with
    >>HSZ's, just a warnging.

    >
    >
    > That's discouraging to hear. I have 4 HSG80's with ACS 8.6F-13 and
    > I've noticed a few soft errors. As a hobbyist I was lucky to get 4
    > cards with that recent of a version. Finding V8.8 for a reasonable
    > price is unlikely. Since I'm using 8GB and 18GB drives I guess it
    > would just be cheaper to swap a drive. (double sigh)


    It is discouraging, but in practice, it has only been a
    (big) problem with the 146GB Universal drives. The soft
    error rate we've seen on 18GB and 36GB blue-brick drives
    has been small enough to not be a problem, but the 146GB
    drives log _lots_ of soft errors, and those get repeated
    anytime you do a full shadow merge or copy. :-( [Host-
    based minimerge releaves the problem by not rereading or
    rewriting most of the bad blocks, but I got hit pretty hard
    recently doing a shadow copy...]

    BTW, we've taken to setting up a dedicated universal
    shelf with HSG and running DILX against all replacement
    drives received from HP before calling them "good" and
    putting them into production. For the 146GB drives, I let
    it DILX run for 2 weeks...there are a LOT of blocks to
    probe...

    -Ken
    --
    I don't speak for Intel, Intel doesn't speak for me...

    Ken Fairfield
    D1C Automation VMS System Support
    who: kenneth dot h dot fairfield
    where: intel dot com

  3. Re: HSG80 errors

    In article , Ken Fairfield writes:
    > BTW, we've taken to setting up a dedicated universal
    > shelf with HSG and running DILX against all replacement
    > drives received from HP before calling them "good" and
    > putting them into production. For the 146GB drives, I let
    > it DILX run for 2 weeks...there are a LOT of blocks to
    > probe...


    DILXing new drives before putting them in production has been a good idea
    since the first of the HSx controllers shipped over a decade ago. I used to
    run it for 10000 minutes which is about 1 week, but I never had disks bigger
    than 36GB.

    --
    Bob Kaplow NAR # 18L >>> To reply, there's no internet on Mars (yet)! <<<
    Kaplow Klips & Baffle: http://nira-rocketry.org/Document/MayJun00.pdf
    www.encompasserve.org/~kaplow_r/ www.nira-rocketry.org www.nar.org

    Vulcans believe peace should not depend on force. -- Amanda, "Journey to
    Babel," stardate 3842.3

  4. Re: HSG80 errors

    In message
    kaplow_r@encompasserve.org.mars (Bob Kaplow) wrote:

    > In article , Ken Fairfield writes:
    > > BTW, we've taken to setting up a dedicated universal
    > > shelf with HSG and running DILX against all replacement
    > > drives received from HP before calling them "good" and
    > > putting them into production. For the 146GB drives, I let
    > > it DILX run for 2 weeks...there are a LOT of blocks to
    > > probe...

    >
    > DILXing new drives before putting them in production has been a good idea
    > since the first of the HSx controllers shipped over a decade ago. I used to
    > run it for 10000 minutes which is about 1 week, but I never had disks bigger
    > than 36GB.


    From my recollection, the HSC50 shipped more like TWO decades ago, certainly
    before VMS4, as I remember having to wait for VMS4 to build a cluster round
    them.

    Oh, the joys of bleeding edge software...

    Alan

    --
    Alan Adams
    alan.adams@orchard-way.freeserve.co.uk
    http://www.nckc.org.uk/

  5. Re: HSG80 errors

    Alan Adams wrote:
    > In message
    > kaplow_r@encompasserve.org.mars (Bob Kaplow) wrote:
    >
    >
    >>In article , Ken Fairfield writes:
    >>
    >>> BTW, we've taken to setting up a dedicated universal
    >>>shelf with HSG and running DILX against all replacement
    >>>drives received from HP before calling them "good" and
    >>>putting them into production. For the 146GB drives, I let
    >>>it DILX run for 2 weeks...there are a LOT of blocks to
    >>>probe...

    >>
    >>DILXing new drives before putting them in production has been a good idea
    >>since the first of the HSx controllers shipped over a decade ago. I used to
    >>run it for 10000 minutes which is about 1 week, but I never had disks bigger
    >>than 36GB.

    >
    >
    > From my recollection, the HSC50 shipped more like TWO decades ago, certainly
    > before VMS4, as I remember having to wait for VMS4 to build a cluster round
    > them.
    >
    > Oh, the joys of bleeding edge software...
    >
    > Alan
    >


    Yup... the cluster support started dribbling in with 3.5->3.7 or so
    IIRC. Been a long time.

    First they had shared disk on the HSC, then the multirooted system boot
    disk happened... I seem to remember real supported clusters were in the
    last of 3.x or so and 4.x had all the stuff really integrated.

    Bill


    --
    --
    d|i|g|i|t|a|l had it THEN. Don't you wish you could still buy it now!
    pechter-at-ureach.com

  6. Re: HSG80 errors

    In message
    Bill Pechter wrote:

    > Alan Adams wrote:
    > > In message
    > > kaplow_r@encompasserve.org.mars (Bob Kaplow) wrote:
    > >
    > >
    > >>In article , Ken Fairfield writes:
    > >>
    > >>> BTW, we've taken to setting up a dedicated universal
    > >>>shelf with HSG and running DILX against all replacement
    > >>>drives received from HP before calling them "good" and
    > >>>putting them into production. For the 146GB drives, I let
    > >>>it DILX run for 2 weeks...there are a LOT of blocks to
    > >>>probe...
    > >>
    > >>DILXing new drives before putting them in production has been a good idea
    > >>since the first of the HSx controllers shipped over a decade ago. I used to
    > >>run it for 10000 minutes which is about 1 week, but I never had disks bigger
    > >>than 36GB.

    > >
    > >
    > > From my recollection, the HSC50 shipped more like TWO decades ago, certainly
    > > before VMS4, as I remember having to wait for VMS4 to build a cluster round
    > > them.
    > >
    > > Oh, the joys of bleeding edge software...
    > >
    > > Alan
    > >

    >
    > Yup... the cluster support started dribbling in with 3.5->3.7 or so
    > IIRC. Been a long time.
    >
    > First they had shared disk on the HSC, then the multirooted system boot
    > disk happened... I seem to remember real supported clusters were in the
    > last of 3.x or so and 4.x had all the stuff really integrated.


    Yes, sort of. We had, with 3.7, four machines each with its own system disc
    (RM03 I think) and sharing a huge RP05 data disc. When V4 came along, we
    joined them into a cluster, still keeping separate system discs. At this
    point we discovered that:

    DEC seemed to expect everyone to use the shared system disc approach

    Every cluster could be different, and was

    The help desks weren't up to speed.

    It was an interesting month or so before we got the systems reasonably
    stable. V4.1 was a HUGE improvement.

    Oh, yes, and one machine was a 730, and they couldn't join clusters
    until a later version - V4.2? The rest were a 750, 780 and 785. We
    added an 8700 soon afterwards.

    The 730 was running (crawling...) Oracle, as we could only afford a 730
    license.

    Alan

    --
    Alan Adams
    alan.adams@orchard-way.freeserve.co.uk
    http://www.nckc.org.uk/

+ Reply to Thread