Kernel crash, possibly raid, and dump but no log available as shutdown.

Kernel crash, possibly raid, and dump but no log available as shutdown.

[Wilson Jonathan]

I have been suffering from a crash and dump for some time now when
shutting down my AMD Threadripper TR4 system.

I think, but am not entirely sure, that the issue is with software
raid.

The system does its usual shut down tasks and then somewhere between
shutting down the virtual machines (none of which were running/had been
started) and the disks doing their shutdowns (not sure what to call it,
but its nearly the last thing that happens before power off but it
mentions each disk) I get a crash dump to the console.

The only camera I have is on a really cheap phone, and its so bad that
I can't pause the scrolling video (or the video is out of focus because
it loses focus when the screen blanks before the output) to see what
the error messages are because its just a blur.

Is there anyway to get/print/transmit the log in real time to some
external "thing" such as another pc on the network as by the time the
crash happens the array (with the log files) must be down because there
is nothing in the log files. No errors, shutdown messages, or shutdown
log (is there even such a thing as a shutdown log?).

Normally when something like this happens (which is increasingly rare)
I just let it ride and usualy some new kernel eventually fixes it. The
problem doesn't seem to corrupt anything so where ever the issue is
(mdadm or otherwise) its not a huge deal... just I'd like it to stop
because I hate seeing the crash dump and thinking "oh, there it goes
again". 

This has been happening for at least 6-12 months (give or take) so its
a long standing issue. (PS its debian testing/bookworm but the same
issue happened with bullseye.)

I guess if there is no such thing as re-directing the output to a
network device, is it possible to move the log files to a non-raid
device or an external usb? Does debian even have a shutdown log file?

Thanks in advance, Jon.

Re: Kernel crash, possibly raid, and dump but no log available as shutdown.

[Paul Menzel]

Dear Jonathan,


Am 11.11.22 um 16:10 schrieb Wilson Jonathan:
> I have been suffering from a crash and dump for some time now when
> shutting down my AMD Threadripper TR4 system.

[…]

> The system does its usual shut down tasks and then somewhere between
> shutting down the virtual machines (none of which were running/had been
> started) and the disks doing their shutdowns (not sure what to call it,
> but its nearly the last thing that happens before power off but it
> mentions each disk) I get a crash dump to the console.

[…]

> Is there anyway to get/print/transmit the log in real time to some
> external "thing" such as another pc on the network as by the time the
> crash happens the array (with the log files) must be down because there
> is nothing in the log files. No errors, shutdown messages, or shutdown
> log (is there even such a thing as a shutdown log?).

Please look into the serial console, a lot of boards still have such a 
header, netconsole, and pstore.


Kind regards,

Paul

Re: Kernel crash, possibly raid, and dump but no log available as shutdown.

[Roman Mamedov]

On Fri, 11 Nov 2022 15:10:47 +0000
Wilson Jonathan <i400sjon@gmail.com> wrote:

> Is there anyway to get/print/transmit the log in real time to some
> external "thing" such as another pc on the network as by the time the
> crash happens the array (with the log files) must be down because there
> is nothing in the log files. No errors, shutdown messages, or shutdown
> log (is there even such a thing as a shutdown log?).

Yes: https://www.kernel.org/doc/Documentation/networking/netconsole.txt

-- 
With respect,
Roman