[OpenAFS] Re: Fileserver and bad I/O to underlying disk

Andrew Deason adeason@sinenomine.net
Mon, 27 Jun 2011 16:46:58 -0500

On Mon, 27 Jun 2011 16:29:14 -0500
"Ryan C. Underwood" <nemesis-lists@icequake.net> wrote:

> The fileserver process is in D state presumably wedged in I/O.
> Sending it kill -9 has no effect.  Is there something in the design of
> the fileserver that would prevent it from failing and dying cleanly if
> something evil happens to the underlying data store?

This is ... Linux?

We shouldn't be able to do anything that can affect a kill -9. However,
if you want to see what it is stuck on, doing a
'echo t > /proc/sysrq-trigger' and finding the fileserver process in
syslog would probably help say where exactly it's stuck (or crashtool,
depending on what distro you're on). You could also try looking at the
userspace stack with pstack or gdb, but I'd expect you may have trouble
getting anything useful out of the problematic thread if it won't come
out of D.

Andrew Deason