X-server can not be started

[Yaroki]

Hello folks,

After rebooted the HP-UX (B.11.11) into a new terminal I encounter the following message:

The X-Server cannot be started on display "hostname"

for details look into Dtlogin.errorLogFile (defult: /var/dt/Xerrors):

error (pid 2709): server for display "hostname":0 cannot be started

ERROR-detected GVID graphics devices but the gvid_info kernel module did not loaded successfully.

Unable to started an X server.

I can only log into not graphical screen.

env | grep TERM

TERM=unknown

I tried to export dtterm but still the message resume. (TERM=unknown).

Please advice,

Yali

[Dennis Handly]

Where are you trying to display?  On your server (what model?) or your PC?

>ERROR-detected GVID graphics devices but the gvid_info kernel module did not loaded successfully.

This is the error you need to fix.

[Yaroki]

Hello,

Im trying to display the HP-UX on the a LCD screen (AOC monitor).

The HP-UX is C-8000, B.11.11.

How can I fix the Error?

[V. Nyga]

>After rebooted ....
Anything change before - like the hostname?

Also check log files in /var/X11/Xserver/logs

V.

[V. Nyga]

>Im trying to display the HP-UX on the a LCD screen

As a normal ws, with keyboard, mouse?
Login at the monitor?
Was there another monitor before?

[Yaroki]

Hello,

Nothing changed before. This HP has been brought from other site to here and when I connected it to the net I have noticed that it has conflict with another PC IP address. For now I disconnected the HP from the net. I will change the IP. Does it concern with the issue?

/var/X11/Xserver/logs/Xstarup.0.log:

"ERROR - Detected GVID graphics devices but the gvid_info kernel module did not load successfully. /usr/bin/X11/X: unable to start an Xserver.

Now, rebooted normally it into an HP LCD monitor and tried to login at monitor but it still have the same problem.

>Was there another monitor before?

No, it was not attached to a monitor before (but via net; remote with Exceed).

I guess it will work with no monitor attached, but this is not a proper solution.

Yali.

[Matti_Kurkela]

An IP address conflict is not at all related to a kernel module problem.

> No, it was not attached to a monitor before

Perhaps someone at the previous site chose to disable all the features that were not required there, including the support for local graphical display. Now you need to find what was done, and re-enable as necessary.

Run "kmadmin -s" and "kmsystem -q gvid". What is the output?

You might also check the kernel configuration using SAM.

Run "sam", go to "Kernel Configuration" -> "Drivers" and find "gvid". If its type is "Static" and current & pending state is "Out", select it, then choose Actions -> Add Driver(s) to Kernel, then Actions -> Process New Kernel.

SAM will then make the necessary changes to your kernel configuration and reboot the system.

[Yaroki]

I run "kmadmin -s" and "mksystem -q gvid". This is the output:

# kmadmin -s
Name            ID      Status          Type
=====================================================
krm             1       UNLOADED        WSIO
gvid_info       2       UNLOADED        WSIO
drmfgl          3       UNLOADED        Misc
drmfglrx        4       UNLOADED        Misc
gvid_him_cons   5       UNLOADED        Misc

# kmsystem -q gvid
Module              Configured          Loadable
=================================================
gvid                Y                   -

> Run "sam", go to "Kernel Configuration" -> "Drivers" and find "gvid". If its type is "Static" and current & pending state is "Out", select it, then choose Actions -> Add Driver(s) to Kernel, then Actions -> Process New Kernel.

It is "Static" but current & pending is "IN" So I did nothing. If it has any meaning, the "gvid_hom_cons" & "gvid_info" fields are in the status of "out for both parameters.

[Yaroki]

Hello,

Does anyone have an idea?

[V. Nyga]

Hi,

I can compare it with mine:
"gvid_hom_cons" is in the status of "out'
"gvid_info" is in the status of "in' - type 'loadable'.

What's about kernel driver 'graph3'? It should be 'in', too.

Can you select a monitor type (graphic mode) at boot-up?
Just after keyboard is activated?

Can you find 'graph3' in your syslog.log?

Is xf86 activated in boot process (Sxxxxf86 in /sbin/rc2.d)?
Does /etc/rc.config.d/xf86 exist?
Check the settings:
gvid_info: parameters are set to '1' all other are set to '0'.

V.

[Yaroki]

> "gvid_hom_cons" is in the status of "out'
   "gvid_info" is in the status of "in' - type 'loadable'.

Mine as well

> What's about kernel driver 'graph3'? It should be 'in', too.

The same as mine.

> Can you select a monitor type (graphic mode) at boot-up?
   Just after keyboard is activated?

I can, What excatly should I do at this stage?

> Can you find 'graph3' in your syslog.log?

Nothing concerning 'graph3' is mentioned on syslog.log

> Is xf86 activated in boot process (Sxxxxf86 in /sbin/rc2.d)?
>  Does /etc/rc.config.d/xf86 exist?

Yes, it exists.

 >  Check the settings:
   gvid_info: parameters are set to '1' all other are set to '0'.

It already exist.

[V. Nyga]

>> "gvid_hom_cons" is in the status of "out'
> "gvid_info" is in the status of "in' - type 'loadable'.

>Mine as well

On Sunday you said:
>If it has any meaning, the "gvid_hom_cons" & "gvid_info" fields are in the status of "out for both parameters.

So "gvid_info" is in the status of "in' now?

BUT from your /var/X11/Xserver/logs/Xstarup.0.log:

>"ERROR - Detected GVID graphics devices but the gvid_info kernel
>module did not load successfully. /usr/bin/X11/X: unable to start an Xserver.

It looks like you have to do something here - maybe you can test or change something in SAM to get it work (Process new kernel?).

Just for interest:
>> Can you select a monitor type (graphic mode) at boot-up?
> Just after keyboard is activated?

>I can, What excatly should I do at this stage?

So during boot-up monitor messages are working?
Until CDE should start?

In syslog.log very soon (line 39) the slot of the graphics card is initialized (10/6/2/0 graph3).

But we should focus to this kernel parameter ...
V.

[Yaroki]

The gvid_him_cons and the gvid_info are in status of "out".

Indeed, I wrote that the gvid_info is in the status of "in" but it is not correct. (I might was wrong or something has been changed. Sorry for that).

I attached the kernel configuration from SAM just to be on a safe side.

Please let me know what parameter should I set on the kernel configuration.

Thank you,

Yali

[V. Nyga]

Configure gvid_info to current status of "in".
Type 'loadable'.

V.

[Yaroki]

I tried to set it via:

Action -> Configure Kernal Logging. Then, I set it to "Enable" but It cannot be config (it is not on the configurable paraneters list). How can I set it?

[V. Nyga]

Hmm, never had to modify this - try the option: 'Add driver to kernel'

V.

[Yaroki]

Ok, that option seems to be the one (add driver to kernel) but after I made the changes as request, I couldn't process new kernel.

I attached the print screen (process new kernel.png) and the /usr/sbin/config (due to the error message).

BR,

Yali

[Yaroki]

Sorry, there is a problem to attach files.

[Yaroki]

In addition, when I try to set the gvid_info via Action -> Add Driver(s) to Kernel, parameters cannot be changed.

The following message received:

"The kadmin command has failed to load driver module gvid_info. The stderr output from the command is showen below:

Kmadmin: gvid_info: Invalid argument"

[Matti_Kurkela]

The screenshot you attached contains an interesting error message:

config: ran out of internal memory(malloc)

 This suggests there is not enough RAM+swap space to configure the kernel.

Please run "dmesg |grep Physical": if the system has not produced too many kernel messages since the last time it was rebooted, the output should be a line that indicates the amount of RAM memory in the system.

Please run "swapinfo -tam" and copy & paste the output to this thread. Knowing your amount of system RAM and swap should allow us to evaluate your current configuration and make better suggestions for next steps.

Adding more swap space should allow you to successfully configure a new kernel. The outputs of "vgdisplay" and "bdf" would help in deciding where to place the new swap space.

[Yaroki]

First thing i did when receiving this HP was run the bdf. As you shell see the results are fine. But I also suspect that you point to the problem.

I attached as per your request the commands' display.

Thank you,

Yali

[Dennis Handly]

>The screenshot you attached contains an interesting error message:
>config: ran out of internal memory(malloc)
> This suggests there is not enough RAM+swap space to configure the kernel.

I noticed that too.  But it appears to be a strings(1) output from the core or executable, I.e. useless and counterproductive.

>Please run "swapinfo -tam". Knowing your amount of system RAM and swap

There seems to be 6 Gb free, so if that malloc message was the problem, it would be maxdsiz being too small.

[Yaroki]

So, what next? Can I supply more information that should help?

[Dennis Handly]

>Can I supply more information that should help?

Where did you get those config error messages?

What is your maxdsiz kernel value?