<div class="gmail_quote">On Fri, Jul 6, 2012 at 10:34 AM, Adam Tauno Williams <span dir="ltr"><<a href="mailto:awilliam@whitemice.org" target="_blank">awilliam@whitemice.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div class="im">On Fri, 2012-07-06 at 10:13 -0400, Michael Mol wrote:<br>
> On Fri, Jul 6, 2012 at 9:58 AM, Collin Kidder <<a href="mailto:collink@kkmfg.com">collink@kkmfg.com</a>> wrote:<br>
> > On 7/6/2012 9:35 AM, Michael Mol wrote:<br>
</div><div class="im">> > I've since set up rsyslog to do remote logging to a second machine. Of<br>
> > course, now that I've done this the machine has been up for longer than a<br>
> > day so far with no hiccups. But, when it happens again I'll be ready. I have<br>
> > mcelog installed as well so any mce messages will be remote logged.<br>
> My only fear of remote logging is my suspicion it's not quite as<br>
> resilient to things like kernel panics as a serial port. But it's a<br>
> good step.<br>
<br>
</div>It is not very resilient. You *might* get a few of the first messages<br>
[logging via UDP] but quite likely the network stack will be gone and<br>
syslog inoperable before the kernel does the sysrq thing and posts its<br>
dumps.<br>
<div class="im"><br></div></blockquote><div><br></div><div>There's a kernel module called "netconsole" for just this purpose:</div><div><br></div><div><a href="http://www.kernel.org/doc/Documentation/networking/netconsole.txt">http://www.kernel.org/doc/Documentation/networking/netconsole.txt</a></div>
<div><br></div><div>I'm sure it's still not perfect, and under certain conditions it probably doesn't work as well as a serial console. But, I have used it in the past to catch a kernel panic on a headless serial-port-less server, and it seemed to work pretty well. It caught the panic where a remote syslog did not.</div>
<div><br></div><div>-Kevin</div></div>