:> twice, so if you do this a few times you can panic a machine. I've
:> submitted a gnats report (kern/9970).
:
:I can confirm that this also halts my crashes with nmh's spost command.
:I'll leave it up to an NFS guru to tell whether it's the *right* fix,
:but it's certainly something that needs l
David Malone writes:
> We think we've sorted out this problem. Trying to make a cross
> device link to an NFS filesystem decreases the reference count
> twice, so if you do this a few times you can panic a machine. I've
> submitted a gnats report (kern/9970).
I can confirm that this also halts my
> I've seen identical panics when using nmh's spost command to send
> mail. Instant panic saying "negative ref cnt". This is using an NFS
> mounted home directory (containing the draft mail to be sent).
> Unfortunately, I didn't have time to investigate further, so I just
> switched over to using
> I've seen identical panics when using nmh's spost command to send
> mail. Instant panic saying "negative ref cnt". This is using an NFS
> mounted home directory (containing the draft mail to be sent).
> Unfortunately, I didn't have time to investigate further, so I just
> switched over to using
David Malone writes:
> We're getting lots of negative reference counts for 3.0-STABLE. We've been
> getting them since long before Christmas. We have 3 SMP machines, all heavy
> NFS clients, which are dieing about 1 per day with this panic.
>
> Several of these hangs have been provoked by me logg
We're getting lots of negative reference counts for 3.0-STABLE. We've been
getting them since long before Christmas. We have 3 SMP machines, all heavy
NFS clients, which are dieing about 1 per day with this panic.
Several of these hangs have been provoked by me logging out - you can see
the ^D in