Crash during vmcore_init

[Américo Wang]

On Wed, Nov 16, 2011 at 6:57 AM, Tim Hartrick <tim at edgecast.com> wrote:
>
> Americo,
>
> I downloaded the source for kexec-tools 1:2.0.2-1ubuntu3 and rebuilt it
> with -DDEBUG and installed.  Below are the results using the
> 2.6.38-8-server Ubuntu kernel as base and crashkernel.  Note that the
> crashkernel size has been increased to 479M.


Thanks for testing!

>
> While I have the list's attention, I would like to mention another
> problem in this area.  The 2.6.38-8-server kernel cannot create a
> crashkernel area greater than 479M in size (e.g. crashkernel=480M).  If
> the crashkernel is relocated (e.g. crashkernel=480M at 8G), then the kernel
> will create the area but kexec (1:2.0.1-1ubuntu3) fails while attempting
> to load crashkernel into the reserved area.


I am afraid kexec can't load the kernel to memory above 4G.

>
> This is a fatal problem since the 2.6.38-8-server kernel running as
> crashkernel requires more than 479M of crashkernel area to successfully
> take a dump on our systems.  I will be happy to provide additional
> information about this one as well.

Well, this is not kernel problem, it is a problem of your kdump tool
on Ubuntu, the initrd of the second kernel should do as minimum things
as possible.


> Elf header: p_type = 4, p_offset = 0xdb74000000000000 p_paddr = 0xdb74000000000000 p_vaddr = 0x0 p_filesz = 0x400 p_memsz = 0x400
> Elf header: p_type = 4, p_offset = 0xdb74000000000000 p_paddr = 0xdb74000000000000 p_vaddr = 0x0 p_filesz = 0x400 p_memsz = 0x400
> Elf header: p_type = 4, p_offset = 0xdb74000000000000 p_paddr = 0xdb74000000000000 p_vaddr = 0x0 p_filesz = 0x400 p_memsz = 0x400
> Elf header: p_type = 4, p_offset = 0xdb74000000000000 p_paddr = 0xdb74000000000000 p_vaddr = 0x0 p_filesz = 0x400 p_memsz = 0x400
> Elf header: p_type = 4, p_offset = 0xdb74000000000000 p_paddr = 0xdb74000000000000 p_vaddr = 0x0 p_filesz = 0x400 p_memsz = 0x400
> Elf header: p_type = 4, p_offset = 0xdb74000000000000 p_paddr = 0xdb74000000000000 p_vaddr = 0x0 p_filesz = 0x400 p_memsz = 0x400
> Elf header: p_type = 4, p_offset = 0xdb74000000000000 p_paddr = 0xdb74000000000000 p_vaddr = 0x0 p_filesz = 0x400 p_memsz = 0x400
> Elf header: p_type = 4, p_offset = 0xdb74000000000000 p_paddr = 0xdb74000000000000 p_vaddr = 0x0 p_filesz = 0x400 p_memsz = 0x400

You have 8 cpus, but it is strange that these PT_NOTE program headers
are all same...

So, can you show us the output of the following commands on your machine?

#hexdump -C /sys/kernel/crash_notes
#for i in /sys/devices/system/cpu/cpu*/crash_notes; do hexdump -C $i; done

Thanks.