Linux 2.4.27-1-386
kde 3.2.1
xserver 4.3.0
Once i start kde it keep hanging on the splash screen under initializing
services....after few mins....it jump out to the grey background...and sit
there.........till i kill the x server.
Tried also to start with gnome....and same result.

Any idea what can it be?

Bye
pepper

On Thu, 13 Jan 2005 20:14:32 +0100, pepper staggered into the Black Sun
and said:
> Linux 2.4.27-1-386 kde 3.2.1 xserver 4.3.0

2.4.27-1 is a release candidate, not the final 2.4.27. Fine if you like
to live on the bleeding edge, but it might cause instabilities. Which
version of which distro are you using, anyway? I ask because KDE 3.2.1
(Redhat) may have significant differences from KDE 3.2.1 (Mandrake) or
KDE 3.2.1 (Gentoo).

> Once I start KDE it [keeps] hanging on the splash screen under
> "initializing services". After [a] few mins, it [jumps] out to the
> grey background and [sits] there till I kill the X server. Tried also
> to start with Gnome, and [I got the] same result.

There isn't enough information in this to make a good diagnosis. Look
in /var/log/kdm.log for any error messages. If you find any, post them
in a followup. Look at the output from dmesg for anything suspicious.
If you find anything, post it in a followup.

FWIW, I had a problem with KDE and kernel 2.6.10 on my desktop a few
weeks ago. Something during KDE startup probes all removable-media
devices. This probe tried to load the kernel module for floppy disks,
which caused a kernel OOPS for some reason. (The machine didn't have a
floppy drive at all.) KDE loaded, but the OOPS made things unstable and
I got a system hang at least once. dmesg showed me where the problem
was, so I worked around it by adding the line "alias block-major-2 off"
to /etc/modules.conf . I don't know if your problem is similar, but if
you post the error messages from dmesg and kdm.log, someone may be able
to help.

--
Matt G|There is no Darkness in Eternity/But only Light too dim for us to see
Brainbench MVP for Linux Admin / mail: TRAP + SPAN don't belong
http://www.brainbench.com / Hire me!
-----------------------------/ http://crow202.dyndns.org/~mhgraham/resume

> 2.4.27-1 is a release candidate, not the final 2.4.27. Fine if you like
> to live on the bleeding edge, but it might cause instabilities. Which
> version of which distro are you using, anyway? I ask because KDE 3.2.1
> (Redhat) may have significant differences from KDE 3.2.1 (Mandrake) or
> KDE 3.2.1 (Gentoo).

Debian sarge on board

> There isn't enough information in this to make a good diagnosis. Look
> in /var/log/kdm.log for any error messages. If you find any, post them
> in a followup. Look at the output from dmesg for anything suspicious.
> If you find anything, post it in a followup.
>

mm strange, theresn't any kdm.log.

the only interesting output from dmesg is that :

pci_hotplug: PCI Hot Plug PCI Core version: 0.5
ACPI-1133: *** Error: Method execution failed [\_SB_.LNKA._STA] (Node
f7c38e00), AE_NOT_EXIST
ACPI-0154: *** Error: Method execution failed [\_SB_.LNKA._STA] (Node
f7c38e00), AE_NOT_EXIST
ACPI-1133: *** Error: Method execution failed [\_SB_.LNKB._STA] (Node
f7c38f00), AE_NOT_EXIST
ACPI-0154: *** Error: Method execution failed [\_SB_.LNKB._STA] (Node
f7c38f00), AE_NOT_EXIST
ACPI-1133: *** Error: Method execution failed [\_SB_.LNKC._STA] (Node
f7c311e0), AE_NOT_EXIST
ACPI-0154: *** Error: Method execution failed [\_SB_.LNKC._STA] (Node
f7c311e0), AE_NOT_EXIST
ACPI-1133: *** Error: Method execution failed [\_SB_.LNKD._STA] (Node
f7c312e0), AE_NOT_EXIST
ACPI-0154: *** Error: Method execution failed [\_SB_.LNKD._STA] (Node
f7c312e0), AE_NOT_EXIST
ACPI-1133: *** Error: Method execution failed [\_SB_.LNKE._STA] (Node
f7c313e0), AE_NOT_EXIST
ACPI-0154: *** Error: Method execution failed [\_SB_.LNKE._STA] (Node
f7c313e0), AE_NOT_EXIST
shpchp: acpi_shpchprm:\_SB_.PCI0 evaluate _BBN fail=0x5
shpchp: acpi_shpchprm:get_device PCI ROOT HID fail=0x5
pci_hotplug: PCI Hot Plug PCI Core version: 0.5
ACPI-1133: *** Error: Method execution failed [\_SB_.LNKA._STA] (Node
f7c38e00), AE_NOT_EXIST
ACPI-0154: *** Error: Method execution failed [\_SB_.LNKA._STA] (Node
f7c38e00), AE_NOT_EXIST
ACPI-1133: *** Error: Method execution failed [\_SB_.LNKB._STA] (Node
f7c38f00), AE_NOT_EXIST
ACPI-0154: *** Error: Method execution failed [\_SB_.LNKB._STA] (Node
f7c38f00), AE_NOT_EXIST
ACPI-1133: *** Error: Method execution failed [\_SB_.LNKC._STA] (Node
f7c311e0), AE_NOT_EXIST
ACPI-0154: *** Error: Method execution failed [\_SB_.LNKC._STA] (Node
f7c311e0), AE_NOT_EXIST
ACPI-1133: *** Error: Method execution failed [\_SB_.LNKD._STA] (Node
f7c312e0), AE_NOT_EXIST
ACPI-0154: *** Error: Method execution failed [\_SB_.LNKD._STA] (Node
f7c312e0), AE_NOT_EXIST
ACPI-1133: *** Error: Method execution failed [\_SB_.LNKE._STA] (Node
f7c313e0), AE_NOT_EXIST
ACPI-0154: *** Error: Method execution failed [\_SB_.LNKE._STA] (Node
f7c313e0), AE_NOT_EXIST
pciehp: acpi_pciehprm:\_SB_.PCI0 evaluate _BBN fail=0x5
pciehp: acpi_pciehprm:get_device PCI ROOT HID fail=0x5
eth0: link up, 100Mbps, full-duplex, lpa 0x45E1
blk: queue f882bc9c, I/O limit 4095Mb (mask 0xffffffff)
blk: queue f882c408, I/O limit 4095Mb (mask 0xffffffff)
blk: queue f882cdec, I/O limit 4095Mb (mask 0xffffffff)
blk: queue f882ccb0, I/O limit 4095Mb (mask 0xffffffff)
blk: queue f882c544, I/O limit 4095Mb (mask 0xffffffff)

spurious 8259A interrupt: IRQ7.


Thanks for your help.
Bye
pepper

In article
(Thu,
13 Jan 2005 20:00:53 +0000), Dances With Crows wrote:

> On Thu, 13 Jan 2005 20:14:32 +0100, pepper staggered into the Black Sun
> and said:
>
> 2.4.27-1 is a release candidate, not the final 2.4.27. Fine if you like
> to live on the bleeding edge, but it might cause instabilities.

I'm not the OP. I have the same problems with kernels v2.6.9 and v2.6.10.
The KDE initialization screen hands at "Initializing Services"; trying to
start GNOME yields a blank screen.

> Which version of which distro are you using, anyway?

Debian Sid, with XFree 4.3.0, GDM 2.6.0.4, KDE 3.3.1, and GNOME 2.8.

>> Tried also to start with Gnome, and [I got the] same result.

Same here. Personally, I suspect GDM because neither GNOME nor KDE work
immediately. I found an annoying workaround; see below.

> There isn't enough information in this to make a good diagnosis.

I looked through every log file I could find, without any success.

> Look in /var/log/kdm.log for any error messages.

I'm using GDM -- and the OP didn't say. /var/log/gdm/:0.log has these
messages (trimmed for space)

Skipping ".../libGLcore.a:m_debug_clip.o": No symbols found
Skipping ".../libGLcore.a:m_debug_norm.o": No symbols found
Skipping ".../libGLcore.a:m_debug_xform.o": No symbols found
Skipping ".../libGLcore.a:m_debug_vertex.o": No symbols found
Skipping ".../libspeedo.a:spencode.o": No symbols found

and

MGA: Failed to load module "mga_hal" (module does not exist, 0)

To the OP and anyone else who's interested:

I can work around the problem by starting PPP. As soon as I get a network
connection, I can start KDE and/or GNOME. The error is 100% reproducible
and this workaround is 100% effective.

--
"We are very much on track for a 2007 delivery [of 'Longhorn']."
-- Bob Muglia. Director, Windows Servers. Microsoft Corporation.
3 January 2004

Hamilcar Barca wrote:

> I can work around the problem by starting PPP. As soon as I get a network
> connection, I can start KDE and/or GNOME. The error is 100% reproducible
> and this workaround is 100% effective.


This sounds like you have a badly configured hostname/hosts, I have seen
people who has removed the two localhost/localhost.localdomain from their
/etc/hosts

127.0.0.1 localhost.localdomain localhost

This can cause problems, you may have to add your hostname to that first line,
specially if you have a dynamically generated ip-number.

127.0.0.1 myhost localhost.localdomain localhost


Not 100% sure if this would help, but worth to try.
Note, if you change hostname, your X System Winodws will fail to work untill
it's restarted.


//Aho

In article <34pb95F4dcvkkU1@individual.net> (Fri, 14 Jan 2005 08:42:40
+0100), J.O. Aho wrote:

> Hamilcar Barca wrote:
>
>> I can work around the problem by starting PPP. As soon as I get a network
>> connection, I can start KDE and/or GNOME. The error is 100% reproducible
>> and this workaround is 100% effective.
>
> This can cause problems, you may have to add your hostname to that first line,
> specially if you have a dynamically generated ip-number.
>
> 127.0.0.1 myhost localhost.localdomain localhost

Is it okay to use two lines?

127.0.0.1 localhost localhost.localdomain
127.0.0.1 lain

I have no idea why I did it this way.

> Not 100% sure if this would help, but worth to try.

Thanks for the follow-up.

--
"Open sourcing [sic] Solaris is a big deal."
-- Scott McNealy. CEO, Sun Microsystem.

Hamilcar Barca wrote:

> Is it okay to use two lines?
>
> 127.0.0.1 localhost localhost.localdomain
> 127.0.0.1 lain

No, you should only have one line.


//Aho

In article <34rst0F4fv1i1U1@individual.net> (Sat, 15 Jan 2005 07:55:54
+0100), J.O. Aho wrote:

> Hamilcar Barca wrote:
>
>> Is it okay to use two lines?
>>
>> 127.0.0.1 localhost localhost.localdomain
>> 127.0.0.1 lain
>
> No, you should only have one line.

I corrected this oversight, but was still unable to start GNOME or KDE
before establishing a PPP connection. I even tried purging GDM and
installing KDM in its place. (For me, this is good. My GDM configuration
was one of the last three pieces of Libranet left on my system. It's not
that I didn't like Libranet, but I'm aiming for pure Debian.)

But, without the hint and nudge you gave me, I wouldn't have looked
further to fix this annoying (non-critical) problem.

The reason I was unable to start GNOME and KDE was my boot-time iptables
configuration. I added

/sbin/iptables -P OUTPUT ACCEPT

to my 'local' script run from init, and I can log in immediately again.

Something in my "firewall" configuration (GuardDog) is obviously enabling
the local network transmissions GNOME and KDE were sending when they
started. It seems to me that this is safe at boot because I can't have
any outside network connection until I start PPP. I don't know exactly
how I broke it in the first place, but it's fixed now.

Thanks for your input.

--
"Ironically, Microsoft's efforts to deny interoperability of Windows with
legitimate non-Microsoft applications have created an environment in which
Microsoft's programs interoperate efficiently only with Internet viruses."
-- Daniel Geer.