Bypass mount/system request at boot time? - VMS

This is a discussion on Bypass mount/system request at boot time? - VMS ; I have a remote client that has somehow gotten himself into a mess, and I'm not sure how to get them out of it remotely. They hired a consultant to come in an "work on the network." This - ah ...

+ Reply to Thread
Page 1 of 2 1 2 LastLast
Results 1 to 20 of 24

Thread: Bypass mount/system request at boot time?

  1. Bypass mount/system request at boot time?

    I have a remote client that has somehow gotten himself into a mess,
    and I'm not sure how to get them out of it remotely.

    They hired a consultant to come in an "work on the network." This -
    ah - "brilliant person" apparently convinced them he just had to
    reboot the 2660.

    He booted the 2660 system with a Linux DVD and somehow or another,
    managed to format the second drive (used for data storage) with
    Linux. How is beyond me...

    Anyway, I'm a good 500 miles away from this server physically, and
    there is a mount/system command in the systartup_vms.com file to mount
    this noe Linux formatted disk. Which obviously won't mount.

    The system boot stalls at this mount request. I think I might be able
    to put an /ASSIST qualifier in there and cancel the mount, but I have
    to get past the mount request to do that.

    Is there any easy way around this? Will just pulling the drive enable
    it to bypass?

    I'm overnighting a freshly formatted drive that will mount, but if I
    can get this production box
    up tonight, it would be a nice thing.

    Sneaky tricks invited.

    -Paul


  2. Re: Bypass mount/system request at boot time?

    PR wrote:

    > Anyway, I'm a good 500 miles away from this server physically, and
    > there is a mount/system command in the systartup_vms.com file to mount
    > this noe Linux formatted disk. Which obviously won't mount.


    Using VAX semanics:

    BOOT/1

    SYSBOOT> SET STARTUP_P1 = "MIN"
    SYSBOOT> CONTINUE

    You can then login and edit the systartup_vms.com file and comment out
    the mount command, then go into sysman, set STARTUP_P1 back to nothing
    and reboot.


  3. Re: Bypass mount/system request at boot time?

    On Oct 14, 10:13*pm, PR wrote:
    > I have a remote client that has somehow gotten himself into a mess,
    > and I'm not sure how to get them out of it remotely.
    >
    > They hired a consultant to come in an "work on the network." *This -
    > ah - "brilliant person" apparently convinced them he just had to
    > reboot the 2660.
    >
    > He booted the 2660 system with a Linux DVD and somehow or another,
    > managed to format the second drive (used for data storage) with
    > Linux. *How is beyond me...
    >
    > Anyway, I'm a good 500 miles away from this server physically, and
    > there is a mount/system command in the systartup_vms.com file to mount
    > this noe Linux formatted disk. Which obviously won't mount.
    >
    > The system boot stalls at this mount request. I think I might be able
    > to put an /ASSIST qualifier in there and cancel the mount, but I have
    > to get past the mount request to do that.
    >
    > Is there any easy way around this? Will just pulling the drive enable
    > it to bypass?
    >
    > I'm overnighting a freshly formatted drive that will mount, but if I
    > can get this production box
    > up tonight, it would be a nice thing.
    >
    > Sneaky tricks invited. *
    >
    > -Paul


    Paul,

    When confronted with such a situation (e.g., "system will not boot
    because of an unplanned for scenario that the various startup
    procedures are not designed to deal with"), my standard solutions are
    the two most basic solutions. When one has access to the MP console on
    an Integrity, they are particularly useful remotely.

    In the both cases, a conversational bootstrap can be set from the MP
    console. This will give control to the console in SYSBOOT. At the
    SYSBOOT prompt, one has a wide set of choices to address the problem.
    In this case, I generally find two alternatives most useful:

    - issue a "SET STARTUP_P1 "MIN", then "CONTINUE" to do a minimal
    system startup (see the "OpenVMS System Manager's Manual")
    - issue a "SET /STARTUP=OPA0:", then "CONTINUE" to give control back
    to the operator before the STARTUP process actually starts. At this
    point, one has a variety of options, but care is needed, most of the
    environment is not functional. I particularly recommend an immediate
    SPAWN command (error handling is somewhat lacking and one can
    therefore need to reboot, which is an annoyance),

    In either case, some care would have allowed you to bring the damaged
    volume online (e.g., SYSMAN IO ...), fix the volume label, etc., and
    then continue the normal startup process manually by invoking the
    [SYS0.SYSCOMMON.SYSEXE]STARTUP.COM. Note that this is before all of
    the logical names are defined, so it is necessary to enumerate the
    path directly.

    I hope that this is helpful.

    - Bob Gezelter, http://www.rlgsc.com

  4. Re: Bypass mount/system request at boot time?

    On Oct 14, 10:13*pm, PR wrote:
    > I have a remote client that has somehow gotten himself into a mess,
    > and I'm not sure how to get them out of it remotely.
    >
    > They hired a consultant to come in an "work on the network." *This -
    > ah - "brilliant person" apparently convinced them he just had to
    > reboot the 2660.
    >
    > He booted the 2660 system with a Linux DVD and somehow or another,
    > managed to format the second drive (used for data storage) with
    > Linux. *How is beyond me...
    >
    > Anyway, I'm a good 500 miles away from this server physically, and
    > there is a mount/system command in the systartup_vms.com file to mount
    > this noe Linux formatted disk. Which obviously won't mount.
    >
    > The system boot stalls at this mount request. I think I might be able
    > to put an /ASSIST qualifier in there and cancel the mount, but I have
    > to get past the mount request to do that.
    >
    > Is there any easy way around this? Will just pulling the drive enable
    > it to bypass?
    >
    > I'm overnighting a freshly formatted drive that will mount, but if I
    > can get this production box
    > up tonight, it would be a nice thing.
    >
    > Sneaky tricks invited. *
    >
    > -Paul


    Paul,

    A supplement to my previous post.

    Do put a message in the console log documenting what happened so that
    there is a record of the event.

    - Bob Gezelter, http://www.rlgsc.com

  5. Re: Bypass mount/system request at boot time?

    PR wrote:
    > I have a remote client that has somehow gotten himself into a mess,
    > and I'm not sure how to get them out of it remotely.
    >
    > They hired a consultant to come in an "work on the network." This -
    > ah - "brilliant person" apparently convinced them he just had to
    > reboot the 2660.
    >
    > He booted the 2660 system with a Linux DVD and somehow or another,
    > managed to format the second drive (used for data storage) with
    > Linux. How is beyond me...
    >
    > Anyway, I'm a good 500 miles away from this server physically, and
    > there is a mount/system command in the systartup_vms.com file to mount
    > this noe Linux formatted disk. Which obviously won't mount.
    >
    > The system boot stalls at this mount request. I think I might be able
    > to put an /ASSIST qualifier in there and cancel the mount, but I have
    > to get past the mount request to do that.
    >
    > Is there any easy way around this? Will just pulling the drive enable
    > it to bypass?
    >
    > I'm overnighting a freshly formatted drive that will mount, but if I
    > can get this production box
    > up tonight, it would be a nice thing.
    >
    > Sneaky tricks invited.
    >
    > -Paul
    >


    Short of concealing a poisoned needle in the power button, there isn't a
    heck of a lot you can do.

    I would write a letter to each of your people who were involved. The
    letter would threaten instant termination for tampering in any way with
    your server or permitting others to do so.

    You will still probably have to fire a couple of people before the
    message sinks in!

    --
    draco vulgaris

  6. Re: Bypass mount/system request at boot time?


    "PR" wrote in message
    news:0f71d7a7-45cf-4f1a-8caa-24a40e8a77a2@b31g2000prf.googlegroups.com...
    >I have a remote client that has somehow gotten himself into a mess,
    > and I'm not sure how to get them out of it remotely.
    >
    > They hired a consultant to come in an "work on the network." This -
    > ah - "brilliant person" apparently convinced them he just had to
    > reboot the 2660.
    >
    > He booted the 2660 system with a Linux DVD and somehow or another,
    > managed to format the second drive (used for data storage) with
    > Linux. How is beyond me...
    >
    > Anyway, I'm a good 500 miles away from this server physically, and
    > there is a mount/system command in the systartup_vms.com file to mount
    > this noe Linux formatted disk. Which obviously won't mount.
    >
    > The system boot stalls at this mount request. I think I might be able
    > to put an /ASSIST qualifier in there and cancel the mount, but I have
    > to get past the mount request to do that.
    >
    > Is there any easy way around this? Will just pulling the drive enable
    > it to bypass?
    >
    > I'm overnighting a freshly formatted drive that will mount, but if I
    > can get this production box
    > up tonight, it would be a nice thing.
    >
    > Sneaky tricks invited.
    >
    > -Paul
    >



    Good reason to have all systems locked away, since the rise of PCs/
    departmental/workgroup servers etc people think there is nothing wrong with
    putting their most expensive asset, i.e. . data out where all and sundry can
    get at them. This includes operator consoles etc.

    Although I am these days a Linux man I spent 18 years with VAX systems and
    would never consider letting anyone with no VMS knowledge near one. 2 weeks
    ago I was phoned by a director in panic mode, could I get his VAX up and
    running from 300 miles away. err no, I have never been near this machine or
    a real vax in years(simh for me thee days). Have no accounts etc. I am not
    going to tell him about conversional boots, alternative uaf etc to break in.
    There are times when only those who know the purpose of a system, how and
    why it is configured should be allowed near a system. The system manager
    being off sick is not justification for me to try to hack, or even worse
    talk a management type through the process.

    Strange then that I tend to leave my oracle databases with the default sys
    password on the grounds that it is only something Oracle DBAs would/should
    know and in my absence any DBA should be able to diagnose/ propose a
    solution. Hopefully they are the sort of people though who would back up a
    database before hacking. ie. not a consultant. easily identified as probably
    late 20s, smart suit, knows it all and yet no idea.



  7. Re: Bypass mount/system request at boot time?

    In article <0f71d7a7-45cf-4f1a-8caa-24a40e8a77a2@b31g2000prf.googlegroups.com>, PR writes:

    > Is there any easy way around this? Will just pulling the drive enable
    > it to bypass?


    Pulling the drive will cause the mount to fail. If there is
    sufficient error handling in systartup_vms.com, that will do as
    a quick fix from far away.

    The real fix, of course, would require access to the console.

    I assume your client doesn't know enough to carry that out for
    you over the phone?


  8. Re: Bypass mount/system request at boot time?

    In article <0f71d7a7-45cf-4f1a-8caa-24a40e8a77a2@b31g2000prf.googlegroups.com>, PR writes:
    >
    > Anyway, I'm a good 500 miles away from this server physically, and
    > there is a mount/system command in the systartup_vms.com file to mount
    > this noe Linux formatted disk. Which obviously won't mount.


    In this case the mount should have failed similar to how it would
    if it wasn't there, unless the system was clustered and sent an
    operator request to a disk operator elsewhere?

    Do you know what did happen? Anything in OPERATOR.LOG?


  9. Re: Bypass mount/system request at boot time?

    On Oct 15, 4:25*am, Bob Gezelter wrote:
    > On Oct 14, 10:13*pm, PR wrote:
    >
    >
    >
    > > I have a remote client that has somehow gotten himself into a mess,
    > > and I'm not sure how to get them out of it remotely.

    >
    > > They hired a consultant to come in an "work on the network." *This -
    > > ah - "brilliant person" apparently convinced them he just had to
    > > reboot the 2660.

    >
    > > He booted the 2660 system with a Linux DVD and somehow or another,
    > > managed to format the second drive (used for data storage) with
    > > Linux. *How is beyond me...

    >
    > > Anyway, I'm a good 500 miles away from this server physically, and
    > > there is a mount/system command in the systartup_vms.com file to mount
    > > this noe Linux formatted disk. Which obviously won't mount.

    >
    > > The system boot stalls at this mount request. I think I might be able
    > > to put an /ASSIST qualifier in there and cancel the mount, but I have
    > > to get past the mount request to do that.

    >
    > > Is there any easy way around this? Will just pulling the drive enable
    > > it to bypass?

    >
    > > I'm overnighting a freshly formatted drive that will mount, but if I
    > > can get this production box
    > > up tonight, it would be a nice thing.

    >
    > > Sneaky tricks invited. *

    >
    > > -Paul

    >
    > Paul,
    >
    > When confronted with such a situation (e.g., "system will not boot
    > because of an unplanned for scenario that the various startup
    > procedures are not designed to deal with"), my standard solutions are
    > the two most basic solutions. When one has access to the MP console on
    > an Integrity, they are particularly useful remotely.
    >
    > In the both cases, a conversational bootstrap can be set from the MP
    > console. This will give control to the console in SYSBOOT. At the
    > SYSBOOT prompt, one has a wide set of choices to address the problem.
    > In this case, I generally find two alternatives most useful:
    >
    > - issue a "SET STARTUP_P1 "MIN", then "CONTINUE" to do a minimal
    > system startup (see the "OpenVMS System Manager's Manual")
    > - issue a "SET /STARTUP=OPA0:", then "CONTINUE" to give control back
    > to the operator before the STARTUP process actually starts. At this
    > point, one has a variety of options, but care is needed, most of the
    > environment is not functional. I particularly recommend an immediate
    > SPAWN command (error handling is somewhat lacking and one can
    > therefore need to reboot, which is an annoyance),
    >
    > In either case, some care would have allowed you to bring the damaged
    > volume online (e.g., SYSMAN IO ...), fix the volume label, etc., and
    > then continue the normal startup process manually by invoking the
    > [SYS0.SYSCOMMON.SYSEXE]STARTUP.COM. Note that this is before all of
    > the logical names are defined, so it is necessary to enumerate the
    > path directly.
    >
    > I hope that this is helpful.
    >
    > - Bob Gezelter,http://www.rlgsc.com


    It is very helpful, thanks. I printing this and advice from several
    other folks and put it in my OPS notebook.

    Yours,
    -Paul


  10. Re: Bypass mount/system request at boot time?

    On Oct 15, 6:25*am, "Richard B. Gilbert"
    wrote:
    > PR wrote:
    > > I have a remote client that has somehow gotten himself into a mess,
    > > and I'm not sure how to get them out of it remotely.

    >
    > > They hired a consultant to come in an "work on the network." *This -
    > > ah - "brilliant person" apparently convinced them he just had to
    > > reboot the 2660.

    >
    > > He booted the 2660 system with a Linux DVD and somehow or another,
    > > managed to format the second drive (used for data storage) with
    > > Linux. *How is beyond me...

    >
    > > Anyway, I'm a good 500 miles away from this server physically, and
    > > there is a mount/system command in the systartup_vms.com file to mount
    > > this noe Linux formatted disk. Which obviously won't mount.

    >
    > > The system boot stalls at this mount request. I think I might be able
    > > to put an /ASSIST qualifier in there and cancel the mount, but I have
    > > to get past the mount request to do that.

    >
    > > Is there any easy way around this? Will just pulling the drive enable
    > > it to bypass?

    >
    > > I'm overnighting a freshly formatted drive that will mount, but if I
    > > can get this production box
    > > up tonight, it would be a nice thing.

    >
    > > Sneaky tricks invited. *

    >
    > > -Paul

    >
    > Short of concealing a poisoned needle in the power button, there isn't a
    > heck of a lot you can do.
    >
    > I would write a letter to each of your people who were involved. *The
    > letter *would threaten instant termination for tampering in any way with
    > your server or permitting others to do so.
    >
    > You will still probably have to fire a couple of people before the
    > message sinks in!
    >
    > --
    > draco vulgaris


    I wish, but they don't work for me, and they purchased the server and
    software. I am providing software support and upgrades for the custom
    software. This is the first time there has been any problem that
    involved VMS, and of course, this was not VMS' fault!

    I'm seriously debating on whether or not to invoice them, and how much
    to invoice them for. You guys solved the issue and incidentally, made
    me look good to the customer. (Thanks!) I don't really think I should
    invoice them because of that.

    On the other hand, if I don't reinforce the lesson, I run the risk of
    it being repeated. Gads... I am not really sure of the right thing to
    do in this case.

    -Paul



  11. Re: Bypass mount/system request at boot time?

    On Oct 15, 7:29*am, koeh...@eisner.nospam.encompasserve.org (Bob
    Koehler) wrote:
    > In article <0f71d7a7-45cf-4f1a-8caa-24a40e8a7...@b31g2000prf.googlegroups..com>, PR writes:
    >
    > > Is there any easy way around this? Will just pulling the drive enable
    > > it to bypass?

    >
    > * *Pulling the drive will cause the mount to fail. *If there is
    > * *sufficient error handling in systartup_vms.com, that will do as
    > * *a quick fix from far away.
    >
    > * *The real fix, of course, would require access to the console.
    >
    > * *I assume your client doesn't know enough to carry that out for
    > * *you over the phone?


    I was nervous asking them to pull the drive, *after* I powered the
    machine off via the MP console.

    -Paul


  12. Re: Bypass mount/system request at boot time?

    On Oct 15, 3:02*pm, PR wrote:
    > On Oct 15, 6:25*am, "Richard B. Gilbert"
    > wrote:
    >
    >
    >
    > > PR wrote:
    > > > I have a remote client that has somehow gotten himself into a mess,
    > > > and I'm not sure how to get them out of it remotely.

    >
    > > > They hired a consultant to come in an "work on the network." *This -
    > > > ah - "brilliant person" apparently convinced them he just had to
    > > > reboot the 2660.

    >
    > > > He booted the 2660 system with a Linux DVD and somehow or another,
    > > > managed to format the second drive (used for data storage) with
    > > > Linux. *How is beyond me...

    >
    > > > Anyway, I'm a good 500 miles away from this server physically, and
    > > > there is a mount/system command in the systartup_vms.com file to mount
    > > > this noe Linux formatted disk. Which obviously won't mount.

    >
    > > > The system boot stalls at this mount request. I think I might be able
    > > > to put an /ASSIST qualifier in there and cancel the mount, but I have
    > > > to get past the mount request to do that.

    >
    > > > Is there any easy way around this? Will just pulling the drive enable
    > > > it to bypass?

    >
    > > > I'm overnighting a freshly formatted drive that will mount, but if I
    > > > can get this production box
    > > > up tonight, it would be a nice thing.

    >
    > > > Sneaky tricks invited. *

    >
    > > > -Paul

    >
    > > Short of concealing a poisoned needle in the power button, there isn't a
    > > heck of a lot you can do.

    >
    > > I would write a letter to each of your people who were involved. *The
    > > letter *would threaten instant termination for tampering in any way with
    > > your server or permitting others to do so.

    >
    > > You will still probably have to fire a couple of people before the
    > > message sinks in!

    >
    > > --
    > > draco vulgaris

    >
    > I wish, but they don't work for me, and they purchased the server and
    > software. I am providing software support and upgrades for the custom
    > software. This is the first time there has been any problem that
    > involved VMS, and of course, this was not VMS' fault!
    >
    > I'm seriously debating on whether or not to invoice them, and how much
    > to invoice them for. You guys solved the issue and incidentally, made
    > me look good to the customer. (Thanks!) I don't really think I should
    > invoice them because of that.
    >
    > On the other hand, if I don't reinforce the lesson, I run the risk of
    > it being repeated. *Gads... I am not really sure of the right thing to
    > do in this case.
    >
    > -Paul


    Paul,

    You are very welcome!

    The question of billing is a more sensitive one. Since our general
    policy invoices clients by how much time is used on their behalf
    (generally in units of 1/4 hour), the answer here would be to invoice
    for the 15-45 minutes that it took to resolve the problem. However,
    that is a philosophical question.

    - Bob Gezelter, http://www.rlgsc.com

  13. Re: Bypass mount/system request at boot time?

    PR wrote:
    > On Oct 15, 6:25 am, "Richard B. Gilbert"
    > wrote:
    >> PR wrote:
    >>> I have a remote client that has somehow gotten himself into a mess,
    >>> and I'm not sure how to get them out of it remotely.
    >>> They hired a consultant to come in an "work on the network." This -
    >>> ah - "brilliant person" apparently convinced them he just had to
    >>> reboot the 2660.
    >>> He booted the 2660 system with a Linux DVD and somehow or another,
    >>> managed to format the second drive (used for data storage) with
    >>> Linux. How is beyond me...
    >>> Anyway, I'm a good 500 miles away from this server physically, and
    >>> there is a mount/system command in the systartup_vms.com file to mount
    >>> this noe Linux formatted disk. Which obviously won't mount.
    >>> The system boot stalls at this mount request. I think I might be able
    >>> to put an /ASSIST qualifier in there and cancel the mount, but I have
    >>> to get past the mount request to do that.
    >>> Is there any easy way around this? Will just pulling the drive enable
    >>> it to bypass?
    >>> I'm overnighting a freshly formatted drive that will mount, but if I
    >>> can get this production box
    >>> up tonight, it would be a nice thing.
    >>> Sneaky tricks invited.
    >>> -Paul

    >> Short of concealing a poisoned needle in the power button, there isn't a
    >> heck of a lot you can do.
    >>
    >> I would write a letter to each of your people who were involved. The
    >> letter would threaten instant termination for tampering in any way with
    >> your server or permitting others to do so.
    >>
    >> You will still probably have to fire a couple of people before the
    >> message sinks in!
    >>
    >> --
    >> draco vulgaris

    >
    > I wish, but they don't work for me, and they purchased the server and
    > software. I am providing software support and upgrades for the custom
    > software. This is the first time there has been any problem that
    > involved VMS, and of course, this was not VMS' fault!
    >
    > I'm seriously debating on whether or not to invoice them, and how much
    > to invoice them for. You guys solved the issue and incidentally, made
    > me look good to the customer. (Thanks!) I don't really think I should
    > invoice them because of that.
    >
    > On the other hand, if I don't reinforce the lesson, I run the risk of
    > it being repeated. Gads... I am not really sure of the right thing to
    > do in this case.
    >
    > -Paul
    >
    >


    I think you would be more than justified in billing them for any
    services you rendered!

  14. Re: Bypass mount/system request at boot time?

    On Oct 15, 2:02*pm, PR wrote:
    > On Oct 15, 6:25*am, "Richard B. Gilbert"
    > wrote:
    >
    >
    >
    > > PR wrote:
    > > > I have a remote client that has somehow gotten himself into a mess,
    > > > and I'm not sure how to get them out of it remotely.

    >
    > > > They hired a consultant to come in an "work on the network." *This -
    > > > ah - "brilliant person" apparently convinced them he just had to
    > > > reboot the 2660.

    >
    > > > He booted the 2660 system with a Linux DVD and somehow or another,
    > > > managed to format the second drive (used for data storage) with
    > > > Linux. *How is beyond me...

    >
    > > > Anyway, I'm a good 500 miles away from this server physically, and
    > > > there is a mount/system command in the systartup_vms.com file to mount
    > > > this noe Linux formatted disk. Which obviously won't mount.

    >
    > > > The system boot stalls at this mount request. I think I might be able
    > > > to put an /ASSIST qualifier in there and cancel the mount, but I have
    > > > to get past the mount request to do that.

    >
    > > > Is there any easy way around this? Will just pulling the drive enable
    > > > it to bypass?

    >
    > > > I'm overnighting a freshly formatted drive that will mount, but if I
    > > > can get this production box
    > > > up tonight, it would be a nice thing.

    >
    > > > Sneaky tricks invited. *

    >
    > > > -Paul

    >
    > > Short of concealing a poisoned needle in the power button, there isn't a
    > > heck of a lot you can do.

    >
    > > I would write a letter to each of your people who were involved. *The
    > > letter *would threaten instant termination for tampering in any way with
    > > your server or permitting others to do so.

    >
    > > You will still probably have to fire a couple of people before the
    > > message sinks in!

    >
    > > --
    > > draco vulgaris

    >
    > I wish, but they don't work for me, and they purchased the server and
    > software. I am providing software support and upgrades for the custom
    > software. This is the first time there has been any problem that
    > involved VMS, and of course, this was not VMS' fault!
    >
    > I'm seriously debating on whether or not to invoice them, and how much
    > to invoice them for. You guys solved the issue and incidentally, made
    > me look good to the customer. (Thanks!) I don't really think I should
    > invoice them because of that.
    >
    > On the other hand, if I don't reinforce the lesson, I run the risk of
    > it being repeated. *Gads... I am not really sure of the right thing to
    > do in this case.
    >
    > -Paul


    You knew where to look for the answer. You recognized the answer when
    you found it. They could have asked here and bypassed you. Bill them
    at your highest rate. Bill them with a free conscience.

    There's the old joke about the guy called in to fix a company's
    computer. He spends a couple of minutes looking at it, pulls out a
    little hammer from his kit, taps the computer on the side and it
    starts working immediately. He says "That'll be $1,000" and the
    manager says "This is outrageous! I want an itemization of your
    charges!" So the guys fills out an invoice that says: "Computer
    repair, Tapping computer with hammer $10.00, Knowing where to tap
    $990."

  15. Re: Bypass mount/system request at boot time?

    In article
    <4e7d7ecb-53d1-4aba-9a7b-a874889a7d16@u28g2000hsc.googlegroups.com>,
    Doug Phillips writes:

    > There's the old joke about the guy called in to fix a company's
    > computer. He spends a couple of minutes looking at it, pulls out a
    > little hammer from his kit, taps the computer on the side and it
    > starts working immediately. He says "That'll be $1,000" and the
    > manager says "This is outrageous! I want an itemization of your
    > charges!" So the guys fills out an invoice that says: "Computer
    > repair, Tapping computer with hammer $10.00, Knowing where to tap
    > $990."


    I heard this with Edison instead of the company manager and Tesla
    instead of the fixit guy.


  16. Re: Bypass mount/system request at boot time?

    On Tue, 14 Oct 2008 19:13:42 -0700 (PDT), PR
    wrote:

    >I have a remote client that has somehow gotten himself into a mess,
    >and I'm not sure how to get them out of it remotely.
    >
    >They hired a consultant to come in an "work on the network." This -
    >ah - "brilliant person" apparently convinced them he just had to
    >reboot the 2660.
    >
    >He booted the 2660 system with a Linux DVD and somehow or another,
    >managed to format the second drive (used for data storage) with
    >Linux. How is beyond me...
    >
    >Anyway, I'm a good 500 miles away from this server physically, and
    >there is a mount/system command in the systartup_vms.com file to mount
    >this noe Linux formatted disk. Which obviously won't mount.
    >
    >The system boot stalls at this mount request. I think I might be able
    >to put an /ASSIST qualifier in there and cancel the mount, but I have
    >to get past the mount request to do that.
    >
    >Is there any easy way around this? Will just pulling the drive enable
    >it to bypass?
    >


    I'd just like to point out, perhaps unnecessarily, that if the error
    is, in fact, a fatal error on the mount (i.e., "-F-"), then the
    startup process is not actually stalling or hanging on the mount
    request.

    Fatal errors in startup tend to actually *kill* the startup process,
    and there is no continuing from that as the job controller will not
    start; and you will never get to login.

    The suggest to do the conversational boot, set startup_p1 to "min",
    edit the comment-out the startup's mount command for that device, and
    reboot, are the only ways around this type of error.

    It's a little nit, to be sure, but understanding what's going on there
    might help in the future.

    When I worked for a large chemical company back in the 1980s & 90s, we
    developed a startup process that was _way_ better than sysman startup,
    and essentially guaranteed that you'd at the very least be able to
    login to a "$" prompt. Basically, it entails a systartup_vms.com that
    *never* executes any significant startups within it's own process, nor
    within a sub-process of itself.

    Guaranteeing the ability to get to a "$" prompt was essential for
    doing remote support and management, because we didn't always have
    remote console access, and couldn't rely on the ability to lead
    someone onsite through the convo-boot steps, edit, etc.

  17. Re: Bypass mount/system request at boot time?

    none wrote:
    >
    > On Tue, 14 Oct 2008 19:13:42 -0700 (PDT), PR
    > wrote:
    >
    > >I have a remote client that has somehow gotten himself into a mess,
    > >and I'm not sure how to get them out of it remotely.
    > >
    > >They hired a consultant to come in an "work on the network." This -
    > >ah - "brilliant person" apparently convinced them he just had to
    > >reboot the 2660.
    > >
    > >He booted the 2660 system with a Linux DVD and somehow or another,
    > >managed to format the second drive (used for data storage) with
    > >Linux. How is beyond me...
    > >
    > >Anyway, I'm a good 500 miles away from this server physically, and
    > >there is a mount/system command in the systartup_vms.com file to mount
    > >this noe Linux formatted disk. Which obviously won't mount.
    > >
    > >The system boot stalls at this mount request. I think I might be able
    > >to put an /ASSIST qualifier in there and cancel the mount, but I have
    > >to get past the mount request to do that.
    > >
    > >Is there any easy way around this? Will just pulling the drive enable
    > >it to bypass?
    > >

    >
    > I'd just like to point out, perhaps unnecessarily, that if the error
    > is, in fact, a fatal error on the mount (i.e., "-F-"), then the
    > startup process is not actually stalling or hanging on the mount
    > request.
    >
    > Fatal errors in startup tend to actually *kill* the startup process,
    > and there is no continuing from that as the job controller will not
    > start; and you will never get to login. [snip]


    Actually, the JBC -DOES- start. However, little-known fact: until a SET
    LOGINS/INTER=n command is issued at least once, JBC will NOT create
    interactive processes. THAT is why you can't login if STARTUP bombs out
    before SYSTARTUP_* completes.

    This is true for hard-wired terminals, CTERM, LAT and UCX/Telnet. Not
    sure about SSH - never had a chance to test it.

    The exception is Telnet in Multinet. For Telnet sessions, the
    MULTINET_SERVER process is the parent of the interactive Telnet
    sessions. It WILL allow logins as soon as it starts, and is not aware
    that STARTUP failed. (Not sure about TCPware - never had a chance to
    test it.)

    D.J.D.

  18. Re: Bypass mount/system request at boot time?

    On Oct 15, 4:03 pm, PR wrote:
    > On Oct 15, 7:29 am, koeh...@eisner.nospam.encompasserve.org (Bob
    >
    > Koehler) wrote:
    > > In article <0f71d7a7-45cf-4f1a-8caa-24a40e8a7...@b31g2000prf.googlegroups.com>, PR writes:

    >
    > > > Is there any easy way around this? Will just pulling the drive enable
    > > > it to bypass?

    >
    > > Pulling the drive will cause the mount to fail. If there is
    > > sufficient error handling in systartup_vms.com, that will do as
    > > a quick fix from far away.

    >
    > > The real fix, of course, would require access to the console.

    >
    > > I assume your client doesn't know enough to carry that out for
    > > you over the phone?

    >
    > I was nervous asking them to pull the drive, *after* I powered the
    > machine off via the MP console.
    >
    > -Paul


    Is there some reason JF's solution wouldn't work?

    You execute a minimum boot, then edit SYSTARTUP_VMS.COM and comment
    out the command that attempts to MOUNT the messed-up disk, and then
    reboot normally. You will, of course, have to use a line editor, such
    as EDT!

    AEF

  19. Re: Bypass mount/system request at boot time?

    On Oct 18, 6:39*am, AEF wrote:
    > On Oct 15, 4:03 pm, PR wrote:
    >
    >
    >
    > > On Oct 15, 7:29 am, koeh...@eisner.nospam.encompasserve.org (Bob

    >
    > > Koehler) wrote:
    > > > In article <0f71d7a7-45cf-4f1a-8caa-24a40e8a7...@b31g2000prf.googlegroups.com>, PR writes:

    >
    > > > > Is there any easy way around this? Will just pulling the drive enable
    > > > > it to bypass?

    >
    > > > * *Pulling the drive will cause the mount to fail. *If there is
    > > > * *sufficient error handling in systartup_vms.com, that will do as
    > > > * *a quick fix from far away.

    >
    > > > * *The real fix, of course, would require access to the console.

    >
    > > > * *I assume your client doesn't know enough to carry that out for
    > > > * *you over the phone?

    >
    > > I was nervous asking them to pull the drive, *after* I powered the
    > > machine off via the MP console.

    >
    > > -Paul

    >
    > Is there some reason JF's solution wouldn't work?
    >
    > You execute a minimum boot, then edit SYSTARTUP_VMS.COM and comment
    > out the command that attempts to MOUNT the messed-up disk, and then
    > reboot normally. You will, of course, have to use a line editor, such
    > as EDT!
    >
    > AEF


    No, it would have worked just fine. However:

    (1) It was getting pretty late at night, this issue popped up at the
    end of an 17 hour day for me.

    (2) Doing a conversational boot under IA64 is slightly more complex
    than under Alphas, and though I had the instructions,
    it was far easier to have them pull the drive (sqeeze the latch,
    pull out the lever, and pull gently till it moves 3 inches...

    It was also less prone to error on my part.

    (3) I was unsure enough of the right way to do this that I asked the
    folks here, introducing a time delay. I am glad I did by the way.
    But the bad part is the delay caused the client's frustration
    level to rise. If I had done the conversational boot 5 minutes after
    they
    called, all would have been well.

    In my defense there, I did not find out the facts for near on 30
    minutes; I had never seen a drive fail in such a way as to cause this
    particular kind of behavior. Either a drive fails and the system
    bypasses it, or it works. This looked like the drive was working, but
    the system could not mount it. It did not make sense until I
    found about the goofball who linux formatted it. By then, *I* was
    tired and
    a little confused.

    (4) Pulling the drive involved the client (just a little and
    allowed a near panicked client to feel that they had personally
    been involved in fixing it.

    The last point has nothing to do with the technical solution of
    course, but it never really hurts to help someone feel a bit better.
    Especially true if that someone is a client.

    By the way, I took the consensus view here and sent the client a bill
    yesterday for an hour of time.

    The offer stands for anyone here who helped; if you would enjoy a six
    pack of your favorite brew on me, just drop me a private e-mail with
    your address - or a paypal address to send the cost of a six pack to.

    Paul




  20. Re: Bypass mount/system request at boot time?

    In article <634e7f9c-502b-4099-a7bf-62c72a74dab4@k16g2000hsf.googlegroups.com>, PR writes:
    >{...snip...}
    >(2) Doing a conversational boot under IA64 is slightly more complex
    >than under Alphas, and though I had the instructions,
    > it was far easier to have them pull the drive (sqeeze the latch,
    >pull out the lever, and pull gently till it moves 3 inches...


    Huh?

    I have aliases established on my Itanium for boot:

    Shell> alias
    b : fs0:\efi\vms\vms_loader.efi
    bo : fs0:\efi\vms\vms_loader.efi
    boo : fs0:\efi\vms\vms_loader.efi
    boot : fs0:\efi\vms\vms_loader.efi

    So booting would be no different than on Alpha:

    IA64: Shell> boot -flags 0,1
    ALpha: >>> boot -flags 0,1

    You could probably just define an alias such as 'conboot' to simplify the
    conversation booting of the Itanium.

    I have also added items to the boot menu. Each of my 3 drives contain a
    version of OpenVMS Itanium (V8.2, V8.3 and V8.3-1H1). I have menu items
    for each version as well as conversational variants. I almost never use
    them as I have the EFI Shell [Built In] defined first, so the system will
    always drop to the Shell> prompt.

    Save for the obnoxious 'chalkboard' color scheme of the Itanium console,
    I don't find it anymore difficult than that of the Alpha SRM.

    --
    VAXman- A Bored Certified VMS Kernel Mode Hacker VAXman(at)TMESIS(dot)COM

    .... pejorative statements of opinion are entitled to constitutional protection
    no matter how extreme, vituperous, or vigorously expressed they may be. (NJSC)

    Copr. 2008 Brian Schenkenberger. Publication of _this_ usenet article outside
    of usenet _must_ include its contents in its entirety including this copyright
    notice, disclaimer and quotations.

+ Reply to Thread
Page 1 of 2 1 2 LastLast