Kernel-2.6.11-1.1267_FC4 Boot Problem

[RUSTY]

When I boot with Kernel-2.6.11-1.1267_FC4 I get the this error
Uncompressing Kernel.....
------------------------------------
Ran out of input date

System Halt

Is that a kernel problem or should I change something to make it work ?

[Finalzone]

Possibly a problem with the kernel on your system. Did that problem happen when using with the old kernel?

[RUSTY]

No it has never happened

[AndyGreen]

This is almost certainly a corrupted vmlinuz image in /boot.... could you have run out of disk space during the kernel install? Or have you been having filesystem problems? Power loss / reboot / panic shortly after the kernel rpm install?

[RUSTY]

This is almost certainly a corrupted vmlinuz image in /boot.... could you have run out of disk space during the kernel install? Or have you been having filesystem problems? Power loss / reboot / panic shortly after the kernel rpm install?

No I did not have any of these problems kernel installs fine and everything seems to be smooth until the reboot
Actually I just tried Kernel-2.6.11-1.1268 and I have the same problem. My boot space is only 22% used

[Joffer]

I just installed FC4T2 on a new AMD64 computer and I get the same error with the kernel supplied: 2.6.11-1.1226_FC4 :(

Any workaround?

[RUSTY]

I just installed FC4T2 on a new AMD64 computer and I get the same error with the kernel supplied: 2.6.11-1.1226_FC4 :(

Any workaround?


hmm well maybe it is a kernel problem then

[AndyGreen]

Is it a SATA drive? Perhaps the problem is an inability of grub to retreive the whole vmlinuz image, not that it isn't there.

[Joffer]

Is it a SATA drive? Perhaps the problem is an inability of grub to retreive the whole vmlinuz image, not that it isn't there.
I use a SATA drive.. Using a VIA K8T800 chipset (or perhaps it was K8T800 PRO).

[RUSTY]

nope mine is 400gb ATA seagate baracuda drive

[Joffer]

If I edit grub at boot time and remove the acpi=off in the kernel line, the kernel boots further, until it actually asks me to enter a runlevel, complaining it can't find a inittab file... problem reading the root??

[RUSTY]

If I edit grub at boot time and remove the acpi=off in the kernel line, the kernel boots further, until it actually asks me to enter a initlevel, complaining it can't find a inittab file... problem reading the root??
I'm gonna check this out ,, what if I put acpi=on

-----------------------------------------------------------------------
Ok I just checked and I don't see any acpi=off in grub

[Joffer]

acpi=on gives me the same error (asking me for runlevel) as just removing the whole acpi option

[Joffer]

Ok I just checked and I don't see any acpi=off in grub
I probably got it since I installed using acpi=off as an option during bootup of installer

[RUSTY]

acpi=on gives me the same error (asking me for runlevel) as just removing the whole acpi option

I guess we have to wait for a new kernel :eek:

[Joffer]

I guess we have to wait for a new kernel :eek:
:mad: :eek: I don't have time for that :( Crap.. when is Test3 going to be released?

As a side note: I installed FC3 yesterday and installed the updates. I got the same error (it hung at bootup) with the new kernel that got installed as I did with FC4T2. I didn't test removing acpi from grub though...

[RUSTY]

:mad: :eek: I don't have time for that :( Crap.. when is Test3 going to be released?

As a side note: I installed FC3 yesterday and installed the updates. I got the same error (it hung at bootup) with the new kernel that got installed as I did with FC4T2. I didn't test removing acpi from grub though...


Well according to fedora schedule the fc4 test3 is coming out on May 9 if I'm not mistaken :)
hopefully there will be working kernel :rolleyes:

[Joffer]

Well according to fedora schedule the fc4 test3 is coming out on May 9 if I'm not mistaken :)
hopefully there will be working kernel :rolleyes:
:( :eek: I really need this to work by monday :( I've mailed the fedora test mailinglist at redhat too, to see if they know a clever workaround/fix.

[RUSTY]

:( :eek: I really need this to work by monday :( I've mailed the fedora test mailinglist at redhat too, to see if they know a clever workaround/fix.
try to use the previous working kernel until the fixed one comes out :< :(

[Joffer]

try to use the previous working kernel until the fixed one comes out :< :(
Looks like I have to revert back to FC3 and exclude kernel upgrades. :eek:

[RUSTY]

Looks like I have to revert back to FC3 and exclude kernel upgrades. :eek:

Don't you have kernel-2.6.11-1.1261_FC4 installed ? I usually leave a previous version in case if something doesn't work so I can go back :) so if you do have it then just unistall the new one and go back to that version there is no need to go to FC3 unless you need stability. :D

[Joffer]

Don't you have kernel-2.6.11-1.1261_FC4 installed ? I usually leave a previous version in case if something doesn't work so I can go back :) so if you do have it then just unistall the new one and go back to that version there is no need to go to FC3 unless you need stability. :D
No? I just installed FC4T2 and it won't boot.

BUT i think I have found the solution.. well.. I was told it on the fedora test mailinglist.

Apparently the x86_64 FC4 kernel has been broken with frequent reboots/deadlocks/panics since April 22nd due to Bug #155827 which also affects upstream 2.6.12-rc3. If you are running x86_64 FC4 please help us test http://people.redhat.com/davej/kernels/Fedora/FC4/ 1275+ which will probably be the base kernel of FC4test3.
So now I'm trying to get this kernel onto my system.. not that easy.. The rescueCD crashes when I try to enable network :( I'm booting a gentoo live CD now and going to start ssh and scp the new kernel rpm file over to the system.

[RUSTY]

No? I just installed FC4T2 and it won't boot.

BUT i think I have found the solution.. well.. I was told it on the fedora test mailinglist.


So now I'm trying to get this kernel onto my system.. not that easy.. The rescueCD crashes when I try to enable network :( I'm booting a gentoo live CD now and going to start ssh and scp the new kernel rpm file over to the system.

get the kernel-devel too I use it cause my nvidia card needs when installing nvidia driver maybe there are some other reasons why you may need it, correct me if Im wrong

[RUSTY]

ok I got the new kernel-2.6.11-1.1275_FC4 and solved the problem everything works fine.

[Joffer]

It worked I see. Great! I have a nvidia card too, so I'm gonna get it (-devel) tomorrow. I didn't have time to test it before I left work. It will probably fixing my problems too :)

[RUSTY]

It worked I see. Great! I have a nvidia card too, so I'm gonna get it (-devel) tomorrow. I didn't have time to test it before I left work. It will probably fixing my problems too :)

Damn after the second reboot it said kernel panic not syncing...........
I already uninstalled the previous working kernel cause I thought this one was working fine....

sucks now gotta find a way to fix..........

[Joffer]

:( I finally got the files over to my FC4T2 box, but I can't install the kernel rpms using the rescue CD. It complains about not having initscripts installed :( (I've chrooted to /mnt/sysimage).

[Joffer]

Damn after the second reboot it said kernel panic not syncing...........
I already uninstalled the previous working kernel cause I thought this one was working fine....

sucks now gotta find a way to fix..........It's not supposed to be easy I guess. Having problems myself getting it installed

[RUSTY]

The problem is in the kernel itself .... Well I don't understand why we need FC4 test3 if the kernel is not even working in FC4 test2 that is stupid very stupid I'm sorry to say that but it's that way I guess.....

3 different kernels I changed and the problem has not been solved! Great job. Instead of making GUI nice make the system work! I'm very frustrated I don't wanna look for the old kernel and installed again going through agony... I hope before FC4 test3 is released we'll see a working kernel.

[Joffer]

Today I reinstalled and now it actually works! Something very odd must have happened when I installed last time.

[GerryJMcF]

Has anyone ever seen this error message "error 2 reading header:cpio Bad magic" My system crashed and now I can't even re-load Fedor 3 I hang up with that message. (Damn power outages)

[AndyGreen]

Out of filesystem space?

df -h

[jdjennings]

I'm running on a Celeron with kernel 2.6.11-1.1276_FC4, which works ok. I get a couple messages at boot time:
mknod: failed to create /dev/console: 17
mknod: failed to create /dev/null: 17
mknod: failed to create /dev/zero: 17
but then it boots successfully.

I've tried kernel 2.6.11-1.1284_FC4 and 2.6.11-1.1286_FC4, and they result in the same messages as above, and then:
Starting udev: MAKEDEV: mkdir: File exists
[OK] :(
Initializing hardware...

and then it just hangs. I've tried with devfs mounted and unmounted, it doesn't seem to make much difference. I've done a "yum upgrade" using the devel repositories, so I'm up to date with FC4T2, I think.

- jeff

[AndyGreen]

If you track development branch you can expect miseries like this from time to time. IIRC it was on the FC3 development there was a filesystem-corrupting kernel went out and bit many people (including me).

My humble advice is use stable FC if your box is important to you or has important data. I guess it's fine to use development branch if you have to have the latest prettiest stuff, but don't expect anything except it to break regularly and have to be reinstalled.

[jdjennings]

sure, i agree that tracking devel will result in stuff like this. but should i just shrug my shoulders and not tell anyone about it, or see if anyone is seeing the same problems?

[AndyGreen]

For this kind of stuff it actually makes sense to bugzilla it. Since it's an artefact of the very latest kernel the devs might be highly interested to hear about it. But I always found that someone else had entered stuff like this already when I was tracking development. (I also found it highly useful to monitor fedora-devel-list ). Somehow other people were always on a faster mirror, which was good when the shrieks of pain emitting from a day or two ahead allowed me to skip updates for a week or two.

[jdjennings]

just today a series of messages on the devel list says it's a bug in udev, there's a fix in udev-057-4. didn't say when it would make it out to the download sites.

- jeff

[jdjennings]

i just installed the new udev-057-4, and everything works fine with the latest kernel.

[RUSTY]

I updated udev and now I'm getting /sbin/udevstart exited abnormally PID(xxx)

[AlaxMAX]

I have just finished to install Fedora Core 4 Test 3. Everything installed without error, but on next boot I have the same message:

Uncompressing Kernel.....
------------------------------------
Ran out of input data

System Halt

The kernel coming with the distribution is 2.6.11-1.1286_FC4. I have rebooted the computer from the DVD in recue mode, I have activated the network and I have mounted the disk partitions. Then, I have tried to download the last kernel package in the development branch, 2.6.11-1.1305_FC4 on a separate computer and transfer it through nfs on my computer. The package hasn't wanted to install, rpm saying it hasn't been able to execute the scriptlets. Eventually, I have managed to force the installation without rpm excuting the scriptlets, but doing it manually:

Preinstall step: /sbin/modprobe loop
Postinstall step: /usr/sbin/module_upgrade followed by /sbin/new-kernel-pkg --package kernel --mkinitrd --depmod --install 2.6.11-1.1305_FC4.

But this didn't help either. On the next boot, it said something about a crc error right after the message Uncompressing Kernel.....

[Sepht]

I get the same problem. but on a 1.8ghz P4 on Fedora Core 4STABLE.
Media checked all 4 disks.

I think it could be the fact I ran reiserfs and removed all the LVM...

[jallain]

I googled my problem and found this thread about running out of imput data. I just installed FC4 stable and had the same problem; running out of input data. I reinstalled three times with the same result. I have already been running Xandros on the machine and I had added it to the grub menu. I rebooted into Xandros so I could research the problem some more. I was very surprised to discover that after I rebooted (after having been in Xandros) Xandros had overwritten grub with it's own lilo. It had never done that before to any distro. On the menu was Fedor Core 4, so I selected it and lo and behold, it booted all the way up into fc4. For some reason, fc4 would not boot using its own grub, but booted just fine when Xandros had added it to its own lilo. So, maybe it's not a kernel problem, but a grub problem. Or I could be all wet.

[Sepht]

Maybe I'll try writing to the 1st sector instead of the MBR...