On Thu, Dec 14, 2000 at 02:59:47AM -0200, Marcelo Tosatti wrote:
>
>
> On Wed, 13 Dec 2000, David Gould wrote:
>
> > On Wed, Dec 13, 2000 at 07:51:03PM -0200, Marcelo Tosatti wrote:
> > >
> > >
> > > On Wed, 13 Dec 2000, Gabriele Paciucci wrote:
> > >
> > > > Hi all,
> > > >
> > > > I write this email, asking Marcello if he could make a resume of all
> > > > the patches for a stable drdb implementation on a 2.2.18 kernel... I think that
> > > > with th new kernel there are some patches no more needed, and in many person
> > > > there are a lot of confusion.
> > >
> > > 2.2.18 is OK wrt drbd, finally. :)
> >
> > Really? Why do we think this? Has it been tested with an SMP secondary
> > and protocol C?
>
> Oh, I forgot about that bug, sorry.
>
> David, if you have time could you please setup a kernel with the NMI
> oopser (from http://people.redhat.com/mingo/nmi-patches/) on a SMP box and
> see if it gives us oops when the secondary node crashes?
>
> Also is good to remember that drbd resync is not safe with journalling
> filesystems.
The node does not crash, it simply hangs. Sysreq works sort of and seems
to show scheduling back and forth between a pair of drbd threads, so I
assume it is some sort of deadlock or starvation issue, but haven't looked
further. I mean to, but other things keep getting in the way...
-dg
--
David Gould dg@example.com
SuSE, Inc., 580 2cd St. #210, Oakland, CA 94607 510.628.3380
why would you want to own /dev/null? "ooo! ooo! look! i stole nothing!
i'm the thief of nihilism! i'm the new god of zen monks."