Skip to contentSkip to conversation

kdump kernel tried to load modules not included in the kdump initrd?

DISCUSSION Jay Lan 2 participants 3 messages Jan 19, 2012 – Jan 20, 2012

Subsystems (2)
lore
Share & tools
lore

Jay Lan

author
#0 lore raw

Hi,

I have a 2.6.32-131.6.1.el6 kernel, and kexec-tools-2.0.0-145.el6.x86_64 rpm on a cent6 machine.

When i forced a kdump by 'echo c > /proc/sysrq-trigger', the kdump kernel panicked during the boot. Surprisingly, it tried to load lustre modules that should not be part of the initrd and panicked!

The syslog showed the --initrd specified a kdump.img when kdump was started. I dumped out the kdump.img (with zcat) and verified that no lustre module was included in the initrd image.

So, how did this have happened?

Thanks,
Jay

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

#1 lore raw
Quoted 2 lines · Jay Lan

Hi,

Hi, Jay,

This mailing list is not proper for reporting problems of distro's mkdumprd, please file a bug in bugzilla of centos.

Quoted 8 lines

I have a 2.6.32-131.6.1.el6 kernel, and kexec-tools-2.0.0-145.el6.x86_64 rpm on a cent6 machine.

When i forced a kdump by 'echo c > /proc/sysrq-trigger', the kdump kernel panicked during the boot. Surprisingly, it tried to load lustre modules that should not be part of the initrd and panicked!

Quoted 4 lines

The syslog showed the --initrd specified a kdump.img when kdump was started. I dumped out the kdump.img (with zcat) and verified that no lustre module was included in the initrd image.

lsinitrd can do that too.

Quoted 3 lines

So, how did this have happened?

Odd, el6 mkdumprd picks all modules listed by lsmod in your *first kernel*, so it is not surprised mkdumprd picks it. But if it is not picked into kdump initrd, it is almost impossible that it will be loaded in the second kernel.

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

author
#2 lore raw

Hi Cong,

Thanks for your reply. It is good to know about lsinitrd. :)

I noticed that I ran out of memory during booting of 2nd kernel. After I increased the crashkernel=xxx to 512M, it worked correctly. Maybe OOM killed some processes and
things got confused... Anyway, it works now.

Thanks,
Jay

Quoted 31 lines · Cong Wang
Quoted 2 lines · Jay Lan

Hi,

Hi, Jay,

This mailing list is not proper for reporting problems of distro's mkdumprd, please file a bug in bugzilla of centos.

Quoted 8 lines

I have a 2.6.32-131.6.1.el6 kernel, and kexec-tools-2.0.0-145.el6.x86_64 rpm on a cent6 machine.

When i forced a kdump by 'echo c> /proc/sysrq-trigger', the kdump kernel panicked during the boot. Surprisingly, it tried to load lustre modules that should not be part of the initrd and panicked!

Quoted 4 lines

The syslog showed the --initrd specified a kdump.img when kdump was started. I dumped out the kdump.img (with zcat) and verified that no lustre module was included in the initrd image.

lsinitrd can do that too.

Quoted 2 lines

So, how did this have happened?

Odd, el6 mkdumprd picks all modules listed by lsmod in your *first kernel*, so it is not surprised mkdumprd picks it. But if it is not picked into kdump initrd, it is almost impossible that it will be loaded in the second kernel.

Quoted 4 lines

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec