Kernel 3.0 giving me panic :)

pinga123 · 06-01-2011, 04:35 AM

I was very exited to try out new kernel version so i decided to give it a try.
http://www.conceivablytech.com/7683/...-linux-3-0-rc1

So i downloaded the kernel 3.0-rc1 from
http://www.kernel.org/

It was my first kernel compilation everything went smooth until i reboot the system .New kernel is not loading and giving me following error.

The root partition is using LVM and its having no trouble booting from previous 2.6.18-92.el5 kernel .

Code:

Red Hat nash version 5.1.19.6 starting 
reading all physical volumes. This may take while...
No volume groups found
Volume group "VolGroup00" not found
mount: could not find filesystem '/dev/root'
Kernel panic - not syncing :attempted to kill init!

I m not sure why its happening.Is it because the previous kernel was el5 and not rhel5.

Code:

# uname -r
2.6.18-92.el5
# lsb_release -a
LSB Version:    :core-3.1-ia32:core-3.1-noarch:graphics-3.1-ia32:graphics-3.1-noarch
Distributor ID: EnterpriseEnterpriseServer
Description:    Enterprise Linux Enterprise Linux Server release 5.2 (Carthage)
Release:        5.2
Codename:       Carthage

markush · 06-01-2011, 05:34 AM

Quote:

Originally Posted by pinga123

Code:

Red Hat nash version 5.1.19.6 starting 
reading all physical volumes. This may take while...
No volume groups found
Volume group "VolGroup00" not found
mount: could not find filesystem '/dev/root'
Kernel panic - not syncing :attempted to kill init!

It seems your kernel lacks drivers for the filesystem and LVM.

Could you please describe in more detail how you did the configuration of your kernel?

Markus

pinga123 · 06-01-2011, 06:24 AM

Quote:

Originally Posted by markush

It seems your kernel lacks drivers for the filesystem and LVM.

Could you please describe in more detail how you did the configuration of your kernel?

Markus

I did it exactly as described in following blog .I used previous kernel configuration file to avoid any error.My previous kernel has support for lvm so in that regard the new kernel should be having the support for LVM.

http://www.howtoforge.com/kernel_compilation_fedora_p2

markush · 06-01-2011, 06:43 AM

Did you check with

Code:

make xconfig

if the configuration is really correct? You may as well use grep to check your .config file. For example

Code:

grep -i _md .config

Markus

grail · 06-01-2011, 07:07 AM

Also did you read any background on the 3.0 kernel? Could be they have changed the way some options are stored. This may affect the fact that you used the old config and those options were named different or not compatible. Markus' idea about grepping for the desired elements should help.

markush · 06-01-2011, 08:23 AM

Quote:

Originally Posted by grail

Also did you read any background on the 3.0 kernel? Could be they have changed the way some options are stored. This may affect the fact that you used the old config and those options were named different or not compatible...

Well, I've compiled linux-3.0-rc1 and it's running yet

Code:

Linux samsung 3.0.0-rc1 #1 SMP Wed Jun 1 15:06:53 CEST 2011 x86_64 AMD Athlon(tm) II Dual-Core M320 AuthenticAMD GNU/Linux

There are no big differences to the 2.6.38.2 Kernel which I was running before with Slackware64-13.37. I did a

Code:

make oldconfig

(after copying the .config from the 2.6.38.2 kernel) and used all the defaults on new features and checked whith make xconfig.

Markus

grail · 06-01-2011, 08:55 AM

Fair enough ... haven't had a go yet myself so just throwing up some ideas

sundialsvcs · 06-01-2011, 09:27 AM

The root cause of the problem is in the "Volume not mounted" message.

The message, "not syncing - attempted to kill init" is slightly cryptic, but here is what it really means:

"Not syncing" actually means that Linux was not in the middle of a disk-write when the error occurred. It is not a comment on the problem itself. (It does mean, happily, that your filesystem has probably not just been destroyed.)
"Attempted to kill init" would be better stated as, "init just died ... but init is not allowed to die!"

And so it is. init is one of those very-special processes that are "hand-made" by Linux during the startup sequence. It is fundamental to the operation of the system, such that Linux cannot run if this process is not in good health. If this process dies, or fails to start, or gets killed, for any reason at all, the kernel will stop cold ("panic") with this message.

But, as you can now see, this message is a symptom, or perhaps a consequence, of "the real problem" which occurred moments before. Conditions prevailed at startup-time which made it impossible for init either to start or to stay started, and Linux has no choice but to halt if this ever happens.

markush · 06-01-2011, 09:55 AM

Quote:

Originally Posted by sundialsvcs

"Not syncing" actually means that Linux was not in the middle of a disk-write when the error occurred. It is not a comment on the problem itself. (It does mean, happily, that your filesystem has probably not just been destroyed.)
"Attempted to kill init" would be better stated as, "init just died ... but init is not allowed to die!"

...

thanks for the detailed explanation. In my post #2 above I referred to this

Quote:

Originally Posted by pinga123

Code:

mount: could not find filesystem '/dev/root'

I had this message sometimes with Gentoo when the kernel was missing either the driver for the filesystem or the SATA-driver (which I had forgotten to compile into the kernel instead of building it as a module). Maybe (otherwise) the system of the OP needs an initrd which was not yet installed.

Markus

pinga123 · 06-01-2011, 10:06 PM

Quote:

Originally Posted by markush

Did you check with

Code:

make xconfig

if the configuration is really correct? You may as well use grep to check your .config file. For example

Code:

grep -i _md .config

Markus

Code:

# grep -i _md .config
CONFIG_SCTP_HMAC_MD5=y
CONFIG_MD=y
CONFIG_BLK_DEV_MD=y
CONFIG_MD_LINEAR=m
CONFIG_MD_RAID0=m
CONFIG_MD_RAID1=m
CONFIG_MD_RAID10=m
CONFIG_MD_RAID456=m
CONFIG_MD_RAID5_RESHAPE=y
CONFIG_MD_MULTIPATH=m
CONFIG_MD_FAULTY=m
CONFIG_USB_MDC800=m
CONFIG_CRYPTO_MD4=m
CONFIG_CRYPTO_MD5=m

Any clue?

grail · 06-02-2011, 12:07 AM

You could try setting all the 'm' to 'y' as it may be that the driver cannot be a module as Markus mentioned in earlier post.

pinga123 · 06-03-2011, 12:44 AM

Quote:

Originally Posted by grail

You could try setting all the 'm' to 'y' as it may be that the driver cannot be a module as Markus mentioned in earlier post.

I did it and recompiled the kernel still facing the same problem.

However i forget to give you some information which i m revealing now.
I was running short of disk space in / partition so i added a new disk and added to / partition using LVM.

Is this a cause of the trouble.
I have also googled the problem and found something which might be related to the error.
http://www.gentoo.org/doc/en/lvm2.xml
Since this is my first kernel compilation i m little confused on what to do ?

markush · 06-03-2011, 02:54 AM

Hi,

are you sure, that you can boot with your old kernel? After your last post I'm not sure if the problem actually is the new kernel.

BTW, I read the announcement for kernel 3.0-rc1 at linux.kernel and Linus Torvalds states that there is no big changes in the kernel:

Quote:

Originally Posted by Linus Torvalds at linux.kernel on May 30

...
So what are the big changes?

NOTHING. Absolutely nothing. Sure, we have the usual two thirds driver
changes, and a lot of random fixes, but the point is that 3.0 is
*just* about renumbering, we are very much *not* doing a KDE-4 or a
Gnome-3 here. No breakage, no special scary new features, nothing at
all like that. We've been doing time-based releases for many years
now, this is in no way about features. If you want an excuse for the
renumbering, you really should look at the time-based one ("20 years")
instead.
...

Note that if you have any drivers which are compiled with the old kernel, for example firmware-network-drivers, you'll have to rebuild them with the new kernel. But this requires that you can boot your system and has (as well) nothing to do with kernel 3.0

Markus

evo2 · 06-03-2011, 03:12 AM

Just want to point out that while there are only small changes between 2.6.39 and 3.0-rc1, there are some very significant changes between 2.6.18 and 2.6.39.

Evo2.

markush · 06-03-2011, 04:55 AM

Hello evo2,

that's a very good point, I didn't recognize that the OP's old kernel is outdated.

Markus