<br><br><div class="gmail_quote">On Sat, Aug 28, 2010 at 11:00 PM, Yaron <span dir="ltr"><<a href="mailto:tclug@freakzilla.com">tclug@freakzilla.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
On Sat, 28 Aug 2010, Robert Nesius wrote:<br>
<br>
<blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
suppose you could ssh into the box and attach a stack tracer to the daemon handling logins and follow forks... (I haven't tried this, but it could work.)<br>
</blockquote>
<br>
The SMALL problem with that is I'd have to ssh in as root which is just... ewww. But I'd do it on a temporary basis if it'll help solve this.<br>
<br>
The bigger problems are A) What daemon do I attach a stack tracer to, and B) What the heck is a stack tracer? (ok I assume that's strace or truss or ptrace or whatever the Linux equivalent of those is).<br></blockquote>
<div><br>Right - strace or ktrace or truss or... I think it's just strace on linux. <br><br>I should know the answer to question A) but don't off the top of my head. I'll fire up my virtual machine tomorrow and see what I find. BUT - if you log in via ssh as a non-root user you could run strace against the su command. 'strace -f -o /tmp/strace.out su - user'. You should be prompted for a password and you should be able to see what system call does the epic-hang. <br>
<br>-Rob<br></div></div>