kexec/ kdump setup problem

kexec/ kdump setup problem

[Mrunal Gawade]

[-- Attachment #1.1: Type: text/plain, Size: 901 bytes --]

Hi,

I have setup the system with two kernels (2.6.25) for kexec/kdump utility. I
am using this command to load kernel. Please comment if its correct. vmlinux
is the dump kernel.

kexec -p /boot/vmlinux --initrd=2.6.25-default-kdump --args-linux
--append="root=/dev/sda1 init 1 irqpoll"

After I execute this command. Should I expect any prompt that kernel loaded
successfully? Because if I execute the sequence alt-SysReq-c after this to
force crash my system hangs. Nothing seems to proceed further. How do I
check if the kernel is loaded successfully? Because if I execute the "-u"
unload option I do not get any error either that kernel is not loaded or
something. Nothing seems to happen over there.

Your timely help would be much appreciated as I need to get this fixed as
soon as possible as my project progress is stopped as I am not able to
analyse the reason for crash.


Thank you,
Mrunal

[-- Attachment #1.2: Type: text/html, Size: 965 bytes --]

[-- Attachment #2: Type: text/plain, Size: 143 bytes --]

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: kexec/ kdump setup problem

[Sachin P. Sant]

Mrunal Gawade wrote:
> After I execute this command. Should I expect any prompt that kernel 
> loaded successfully?
Check the /sys/kernel/kexec_crash_loaded file. Value "1" means success.

Thanks
-Sachin


_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: kexec/ kdump setup problem

[Mrunal Gawade]

[-- Attachment #1.1: Type: text/plain, Size: 575 bytes --]

Thanks Sachin.

The value over there is a "0" which means I have a failure. How do I
diagnose what is the problem? I was able to load a normal kernel and then
use "e" option to reboot into it. But crash dump kernel seemed to be giving
problem always.


Thank you,
Mrunal

On Tue, Apr 29, 2008 at 4:13 AM, Sachin P. Sant <sachinp@in.ibm.com> wrote:

> Mrunal Gawade wrote:
>
> > After I execute this command. Should I expect any prompt that kernel
> > loaded successfully?
> >
> Check the /sys/kernel/kexec_crash_loaded file. Value "1" means success.
>
> Thanks
> -Sachin
>
>

[-- Attachment #1.2: Type: text/html, Size: 1032 bytes --]

[-- Attachment #2: Type: text/plain, Size: 143 bytes --]

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: kexec/ kdump setup problem

[Neil Horman]

On Tue, Apr 29, 2008 at 03:40:13PM -0700, Mrunal Gawade wrote:
> Thanks Sachin.
> 
> The value over there is a "0" which means I have a failure. How do I
> diagnose what is the problem? I was able to load a normal kernel and then
> use "e" option to reboot into it. But crash dump kernel seemed to be giving
> problem always.
> 
> 
> Thank you,
> Mrunal
> 

check /var/log/messages, to see if there is any message there.  It wouldn't hurt
to enable debugging in the kexec binary either, although if the load is actually
failing, I would think you would get something on stderr.  First guess would be
that you don't have a crashkernel area specified on your command line (or you
do, and its reservation failed, which also will show up in the logs most likely)

Neil

> On Tue, Apr 29, 2008 at 4:13 AM, Sachin P. Sant <sachinp@in.ibm.com> wrote:
> 
> > Mrunal Gawade wrote:
> >
> > > After I execute this command. Should I expect any prompt that kernel
> > > loaded successfully?
> > >
> > Check the /sys/kernel/kexec_crash_loaded file. Value "1" means success.
> >
> > Thanks
> > -Sachin
> >
> >

> _______________________________________________
> kexec mailing list
> kexec@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/kexec


-- 
/***************************************************
 *Neil Horman
 *Software Engineer
 *Red Hat, Inc.
 *nhorman@redhat.com
 *gpg keyid: 1024D / 0x92A74FA1
 *http://pgp.mit.edu
 ***************************************************/

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: kexec/ kdump setup problem

[Mrunal Gawade]

[-- Attachment #1.1: Type: text/plain, Size: 2566 bytes --]

Hi,

I just checked /sys/kernel/kexec_crash_loaded file after giving the kexec -p
command for loading panic kernel. And the value was 1. So it was loaded
successfully. I then crashed kernel and it hanged. I rebooted manually and
checked the value in the file again and it was "0". So across reboot it got
reset. So if I assume that the kernel loading was successful, the question
remains if it got loaded successfully at the designated space 65M@16M. How
do I know that? I am running SUSE 2.6.25 on a VMWare workstation. Does
running on VM could change any thing such as memory layout or anything
related to panic handling?

You specified taking a look at the /sys/kernel/kexec_crash_loaded file. Is
there a central document which documents all these error resolving
guidelines? Also I checked in the /var/log/messages and there is no message
related to this printed over there.


Thank you,
Mrunal

On Tue, Apr 29, 2008 at 5:17 PM, Neil Horman <nhorman@redhat.com> wrote:

> On Tue, Apr 29, 2008 at 03:40:13PM -0700, Mrunal Gawade wrote:
> > Thanks Sachin.
> >
> > The value over there is a "0" which means I have a failure. How do I
> > diagnose what is the problem? I was able to load a normal kernel and
> then
> > use "e" option to reboot into it. But crash dump kernel seemed to be
> giving
> > problem always.
> >
> >
> > Thank you,
> > Mrunal
> >
>
> check /var/log/messages, to see if there is any message there.  It
> wouldn't hurt
> to enable debugging in the kexec binary either, although if the load is
> actually
> failing, I would think you would get something on stderr.  First guess
> would be
> that you don't have a crashkernel area specified on your command line (or
> you
> do, and its reservation failed, which also will show up in the logs most
> likely)
>
> Neil
>
> > On Tue, Apr 29, 2008 at 4:13 AM, Sachin P. Sant <sachinp@in.ibm.com>
> wrote:
> >
> > > Mrunal Gawade wrote:
> > >
> > > > After I execute this command. Should I expect any prompt that kernel
> > > > loaded successfully?
> > > >
> > > Check the /sys/kernel/kexec_crash_loaded file. Value "1" means
> success.
> > >
> > > Thanks
> > > -Sachin
> > >
> > >
>
> > _______________________________________________
> > kexec mailing list
> > kexec@lists.infradead.org
> > http://lists.infradead.org/mailman/listinfo/kexec
>
>
> --
> /***************************************************
>  *Neil Horman
>  *Software Engineer
>  *Red Hat, Inc.
>  *nhorman@redhat.com
>  *gpg keyid: 1024D / 0x92A74FA1
>  *http://pgp.mit.edu
>  ***************************************************/
>

[-- Attachment #1.2: Type: text/html, Size: 3474 bytes --]

[-- Attachment #2: Type: text/plain, Size: 143 bytes --]

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: kexec/ kdump setup problem

[Sachin P. Sant]

Mrunal Gawade wrote:
> Hi,
>
> I just checked /sys/kernel/kexec_crash_loaded file after giving the 
> kexec -p command for loading panic kernel. And the value was 1. So it 
> was loaded successfully. I then crashed kernel and it hanged. I 
> rebooted manually and checked the value in the file again and it was 
> "0". So across reboot it got reset. So if I assume that the kernel 
> loading was successful, the question remains if it got loaded 
> successfully at the designated space 65M@16M. How do I know that? 
The file /sys/kernel/kexec_crash_loaded will have value 1 if the kdump kernel has been
loaded successfully. In your case when you executed kexec -p the kernel was loaded at
the designated space and value 1 in kexec_crash_loaded reflects that. This value won't
be retained across reboots. One will have to again execute the kexec -p command to
load the kdump kernel after reboot. [ If kdump service is enabled (chkconfig kdump on)
kexec -p will get executed after every reboot. ]

I think in your case the problem is that after a successful load, the kdump kernel
does not boot on a system crash.[alt-sysrq-c]. you could try adding console
related options to the kexec command and see if that will give some information
on where the kernel hangs.

Thanks
-Sachin


_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: kexec/ kdump setup problem

[Neil Horman]

On Tue, Apr 29, 2008 at 07:14:36PM -0700, Mrunal Gawade wrote:
> Hi,
> 
> I just checked /sys/kernel/kexec_crash_loaded file after giving the kexec -p
> command for loading panic kernel. And the value was 1. So it was loaded
> successfully. I then crashed kernel and it hanged. I rebooted manually and
> checked the value in the file again and it was "0". So across reboot it got
> reset. So if I assume that the kernel loading was successful, the question
> remains if it got loaded successfully at the designated space 65M@16M. How
> do I know that? I am running SUSE 2.6.25 on a VMWare workstation. Does
> running on VM could change any thing such as memory layout or anything
> related to panic handling?
> 

As Sachin mentioned, kexec_crash_loaded gets reset accross reboots.

As for use on a VMWare guest, its behavior is anyones guess.  Best thing I could
think to do would be to add earlyprintk to the kdump kernel command line, and
watch the serial console for mesages about where you're hanging (see
Documentations/kernel-parameters.txt for details about earlyprintk).

Neil

> You specified taking a look at the /sys/kernel/kexec_crash_loaded file. Is
> there a central document which documents all these error resolving
> guidelines? Also I checked in the /var/log/messages and there is no message
> related to this printed over there.
> 
> 
> Thank you,
> Mrunal
> 
> On Tue, Apr 29, 2008 at 5:17 PM, Neil Horman <nhorman@redhat.com> wrote:
> 
> > On Tue, Apr 29, 2008 at 03:40:13PM -0700, Mrunal Gawade wrote:
> > > Thanks Sachin.
> > >
> > > The value over there is a "0" which means I have a failure. How do I
> > > diagnose what is the problem? I was able to load a normal kernel and
> > then
> > > use "e" option to reboot into it. But crash dump kernel seemed to be
> > giving
> > > problem always.
> > >
> > >
> > > Thank you,
> > > Mrunal
> > >
> >
> > check /var/log/messages, to see if there is any message there.  It
> > wouldn't hurt
> > to enable debugging in the kexec binary either, although if the load is
> > actually
> > failing, I would think you would get something on stderr.  First guess
> > would be
> > that you don't have a crashkernel area specified on your command line (or
> > you
> > do, and its reservation failed, which also will show up in the logs most
> > likely)
> >
> > Neil
> >
> > > On Tue, Apr 29, 2008 at 4:13 AM, Sachin P. Sant <sachinp@in.ibm.com>
> > wrote:
> > >
> > > > Mrunal Gawade wrote:
> > > >
> > > > > After I execute this command. Should I expect any prompt that kernel
> > > > > loaded successfully?
> > > > >
> > > > Check the /sys/kernel/kexec_crash_loaded file. Value "1" means
> > success.
> > > >
> > > > Thanks
> > > > -Sachin
> > > >
> > > >
> >
> > > _______________________________________________
> > > kexec mailing list
> > > kexec@lists.infradead.org
> > > http://lists.infradead.org/mailman/listinfo/kexec
> >
> >
> > --
> > /***************************************************
> >  *Neil Horman
> >  *Software Engineer
> >  *Red Hat, Inc.
> >  *nhorman@redhat.com
> >  *gpg keyid: 1024D / 0x92A74FA1
> >  *http://pgp.mit.edu
> >  ***************************************************/
> >

-- 
/***************************************************
 *Neil Horman
 *Software Engineer
 *Red Hat, Inc.
 *nhorman@redhat.com
 *gpg keyid: 1024D / 0x92A74FA1
 *http://pgp.mit.edu
 ***************************************************/

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec