Startx fails on SCO 5.0.5 system - SCO

This is a discussion on Startx fails on SCO 5.0.5 system - SCO ; I'm moving an SCO 5.0.5 system from old hardware to a used Compaq ML 370 server. I did this by restoring the tape backup (Backup Edge 1.1.6) to the Compaq ML 370. The system is booting on the ML 370 ...

+ Reply to Thread
Results 1 to 3 of 3

Thread: Startx fails on SCO 5.0.5 system

  1. Startx fails on SCO 5.0.5 system

    I'm moving an SCO 5.0.5 system from old hardware to a used
    Compaq ML 370 server. I did this by restoring the tape backup
    (Backup Edge 1.1.6) to the Compaq ML 370.

    The system is booting on the ML 370 Ok but I can't start the
    the graphical screen.

    When I execute startx, I get a gray hashed screen with the X
    mouse (moves when the mouse is moved) but after several minutes
    the display returns to just the "#" prompt without an error
    message.

    I searched SCO TA's and found TA 110167 that suggests I look at
    /var/adm/X0msgs
    /var/dt/Xerrors


    X0msg has the following:

    Tue May 8 00:33:49 2007
    /dev/X0S: No such device or address
    X server is already running

    Fatal server error:
    Can't acquire screen now!!

    Can not restore tty mode
    : Bad file number
    PIO_KEYMAP failed.
    : Bad file number

    Wed May 9 18:28:30 2007
    /dev/X0S: No such device or address

    Wed May 9 21:28:09 2007
    /dev/X0S: No such device or address

    Wed May 9 21:36:58 2007
    /dev/X0S: No such device or address

    I've been working on this several days now. I tried
    scologin enable but that exits as well.

    # ls -l /dev/X0*
    crw-rw-rw- 1 root sys 60,254 Feb 16 1999 /dev/X0R
    crw-rw-rw- 1 root sys 60,255 Dec 16 1998 /dev/X0S

    I noted that the system had no patches applied. I
    installed rs505a, and the m64 graphics driver package.
    Now when I run mkdev graphics, and choose add adapter,
    the on board ATI Rage graphics card is detected:

    +------------------------ Video Configuration Manager -------------------------+
    |Current Video Adapter Configuration (1 configured): |
    |+---------------------------------------------------------+ [ Add Adapter... ]|
    ||* - ATI Mach64 PCI (VT/GT) RAGE/II/II+/IIC | [ Modify... ]|
    || Standard VGA | [ Remove... ]|
    || VT/GT 800x600 256 color NI 60Hz |

    But still when I run "startx 2> /tmp/xerror" I see in xerror:

    /dev/X0R: No such device or address
    ...
    /dev/X0R: No such device or address
    ...
    /dev/X0R: No such device or address
    ...
    /dev/X0R: No such device or address
    ...
    /dev/X0R: No such device or address
    ...
    /dev/X0R: No such device or address
    ...
    /dev/X0R: No such device or address
    ...
    giving up.
    xinit: No such device or address (errno 6): unable to connect to X server

    waiting for X server to shut down

    And grep -c X0R /tmp/xerror results in:

    # grep -c X0R /tmp/xerror
    120
    #

    How do I fix this?

  2. Re: Startx fails on SCO 5.0.5 system

    Steve M. Fabac, Jr. wrote:

    > /dev/X0S: No such device or address
    > /dev/X0R: No such device or address
    > # ls -l /dev/X0*
    > crw-rw-rw- 1 root sys 60,254 Feb 16 1999 /dev/X0R
    > crw-rw-rw- 1 root sys 60,255 Dec 16 1998 /dev/X0S


    > xinit: No such device or address (errno 6): unable to connect to X server


    On OSR506, those nodes are created by /etc/conf/node.d/Xsco (well, by
    `idmknod` according to instructions in that file); relating to the
    STREAMS pipe driver "sp".

    Is that driver enabled?

    # cat /etc/conf/sdevice.d/sp
    sp Y 0 0 0 0 0 0 0 0

    If it isn't enabled, I would be worried about the rest of the STREAMS
    subsystem. Run `mkdev streams` and tell it to go about its business.
    Defer any kernel relink.

    Next, is "sp" configured for 256 units? The nodes refer to minor
    numbers 254 & 255, i.e. the 255 & 256th units. This is so regular use
    of "sp" via /dev/spx can automatically allocate units 0 through (256 -
    2 * (# active X servers)) without clashing with the preallocated X
    server nodes.

    # grep NUMSP /etc/conf/cf.d/stune
    /etc/conf/cf.d/stune:NUMSP 256

    If not, run /etc/conf/cf.d/configure, go through the Streams section,
    set NUMSP to 256. No more, no less.

    If you've changed anything, relink (/etc/conf/cf.d/link_unix -y) and
    reboot...

    >Bela<


  3. Re: Startx fails on SCO 5.0.5 system

    Bela Lubkin wrote:
    >
    > Steve M. Fabac, Jr. wrote:
    >
    > > /dev/X0S: No such device or address
    > > /dev/X0R: No such device or address
    > > # ls -l /dev/X0*
    > > crw-rw-rw- 1 root sys 60,254 Feb 16 1999 /dev/X0R
    > > crw-rw-rw- 1 root sys 60,255 Dec 16 1998 /dev/X0S

    >
    > > xinit: No such device or address (errno 6): unable to connect to X server

    >
    > On OSR506, those nodes are created by /etc/conf/node.d/Xsco (well, by
    > `idmknod` according to instructions in that file); relating to the
    > STREAMS pipe driver "sp".
    >
    > Is that driver enabled?
    >
    > # cat /etc/conf/sdevice.d/sp
    > sp Y 0 0 0 0 0 0 0 0


    Yes to the above, the /dev nodes exist as show above.

    >
    > If it isn't enabled, I would be worried about the rest of the STREAMS
    > subsystem. Run `mkdev streams` and tell it to go about its business.
    > Defer any kernel relink.
    >
    > Next, is "sp" configured for 256 units? The nodes refer to minor
    > numbers 254 & 255, i.e. the 255 & 256th units. This is so regular use
    > of "sp" via /dev/spx can automatically allocate units 0 through (256 -
    > 2 * (# active X servers)) without clashing with the preallocated X
    > server nodes.
    >
    > # grep NUMSP /etc/conf/cf.d/stune
    > /etc/conf/cf.d/stune:NUMSP 256
    >
    > If not, run /etc/conf/cf.d/configure, go through the Streams section,
    > set NUMSP to 256. No more, no less.


    I found NUMSP set to 128. I increased it to 256, relinked and rebooted but
    the problem still existed.

    >
    > If you've changed anything, relink (/etc/conf/cf.d/link_unix -y) and
    > reboot...
    >
    > >Bela<



    Bela,

    Thanks for the above information. I was pressed for time before leaving for
    vacation last week and needed to ship the system back to the client.

    The resolution to the startx problem: I took a closer look as custom and noted
    that the system is 5.0.5 with NONE of the required patches. When I transferred the
    root-only backup to the new hardware, I just set about installing EFS-5.50 and
    had a lot of trouble getting the NIC to work as well as moving /dev/u from
    hd00 to hd02 divvy table: Each time I would relink the kernel, rebuild the
    environment, and reboot, the /dev/u major minor would change from 1,104 to
    1,43 (the old u file system now renamed to u1).

    Well the solution to the startx problem was to remove EFS5.5, install rs505a,
    0ss497c, oss600a, oss652a, and m64 graphics drivers. I then reinstalled EFS 5.5
    and startx now works.

    My problem was that I was focused on moving the root backup to the new hardware
    and I did not pay attention to the patches (not) applied to the old system.


    --

    Steve Fabac
    S.M. Fabac & Associates
    816/765-1670

+ Reply to Thread