Re: nfs server lockup (again)

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

On Tue, Jun 28, 2011 at 08:48:17AM -0700, Hisashi T Fujinaka wrote:
> On Fri, 24 Jun 2011, Manuel Bouyer wrote:
> 
> >On Fri, Jun 24, 2011 at 09:45:19AM -0700, Hisashi T Fujinaka wrote:
> >>>To see if it's the problem you'll need to enter ddb and see what the
> >>>per-cpu soft interrupt threads are doing.
> >>
> >>OK, do I need to add kernel debugging, etc? I've done this in the past
> >>but it's been a while.
> >
> >ddb should be enough. after entering ddb, ps to see what thread
> >are waiting on what; you can get stack trace for each blocked threads
> >with tr/a or tr/t (warning: ps show PID and LID in decimal but tr/t
> >takes input in hex by default. tr/t 0t1234 will take input as decimal)
> 
> Arrgh. My stupid motherboard seems to have intermittent issues with the
> USB keyboard and I can't get it into DDB. And I can't complete a build
> without nfs locking up.
> 
> Is there a workaround? I'll try disabling nfsd...

If it's the same issue as mine, setting nfsd_use_loan to 0 should work
around it.

-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--