kernel dump

[Michael Havens]

My eyes feel like a kaleidoscope after less than 20 minutes of reading!

On Thu, Dec 10, 2015 at 8:38 PM, Michael Havens <bmike1 at gmail.com> wrote:

> thank you so much Jerry. That was quite helpful. made me realize that I
> was getting in way over my head! So now I am reading one of the resources
> <https://www.kernel.org/doc/Documentation/kdump/kdump.txt>.
>
> On Thu, Dec 10, 2015 at 6:30 PM, Jerry Snitselaar <dev at snitselaar.org>
> wrote:
>
>> On Thu Dec 10 15, Michael Havens wrote:
>>
>>>   I want to see why my kernel panics (if it does) so I was lead to:
>>>   [1]https://help.ubuntu.com/lts/serverguide/kernel-crash-dump.html
>>>   sudo cat /etc/default/kdump-tools>> USE_KDUMP=1
>>>   Will the above line append USE_KDUMP=1 to kdump-tools.
>>>   --
>>>   :-)~MIKE~(-:
>>>
>>> References
>>>
>>>   1. https://help.ubuntu.com/lts/serverguide/kernel-crash-dump.html
>>>
>>
>> I've never played with it on Ubuntu.
>>
>> You will need the crashkernel parameter on the boot line so memory
>> gets set aside to the kdump kernel memory image.
>>
>> Also you will need to reboot, so it builds the image if you haven't
>> already done so (according to that page, Red Hat/Fedora does the same
>> thing).
>>
>> You will probably need to install debug symbols for the kernel in
>> question.  Not sure how that is done for Ubuntu.
>>
>> It doesn't look like it explains it on that page, but look to see if
>> there is a kdump.conf file in /etc, that will probably need to be
>> modified to suit your system.
>>
>> Once you have it up and running, when the system panics it will
>> jump to the kdump kernel, bring that up, harvest an image of the
>> memory, then reboot the system and come up again on the regular
>> kernel.
>>
>> Then you would go to /var/crash or wherever it is configured to
>> place the vmcore file, and then:
>>
>> crash /boot/System-map-for-kernel
>> /lib/modules/debug/lib/modules/kernel-rel/vmlinux vmcore
>>
>> The vmlinux part would be whatever location it installed the kernel image
>> which still has
>> debug symbols not stripped.
>>
>> Some commands:
>>
>> bt       - prints backtrace of the current process
>> bt -a    - prints backtrace for processes on all cpus
>> bt -f    - prints the contents of the stack with the backtrace
>>
>> help     - will list available commands and help command, will print
>>           out detailed help for the commands.
>>
>> log      - dumps the in memory system log (what you would normally see
>>           with dmesg command)
>>
>> set #    - set focus to certain pid
>> set -c # - set focus to certain cpu
>> dis function-name - provides disassembled code for the function given
>>
>> mod      - load symbols for a module
>>
>> rd       - read contents of memory
>> struct   - display a struct, (with address provided as well dumps out
>>           formatted struct with values)
>>
>>
>> Unless you are familiar with kernel internals and assembly code, what
>> you'll probably want is to:
>>
>> set scroll off
>> log
>>
>> And look at the end of the log to find the message where it paniced,
>> and post those contents somewhere. If it is a distro kernel your best
>> bet will be contacting those folks.
>>
>> Jerry
>>
>> ---------------------------------------------------
>>> PLUG-discuss mailing list - PLUG-discuss at lists.phxlinux.org
>>> To subscribe, unsubscribe, or to change your mail settings:
>>> http://lists.phxlinux.org/mailman/listinfo/plug-discuss
>>>
>>
>> ---------------------------------------------------
>> PLUG-discuss mailing list - PLUG-discuss at lists.phxlinux.org
>> To subscribe, unsubscribe, or to change your mail settings:
>> http://lists.phxlinux.org/mailman/listinfo/plug-discuss
>>
>
>
>
> --
> :-)~MIKE~(-:
>



-- 
:-)~MIKE~(-:
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.phxlinux.org/pipermail/plug-discuss/attachments/20151210/7c799783/attachment.html>