On Thu, 2005-03-17 at 21:28 +0100, Heiko Bauke wrote:
> Dear all,
>
> I'm trying to use LAM/MPI 7.1.1 with Berkeley Lab Checkpoint/Restart
> 0.4.0 and kernel 2.4.26. But I don't get things working correctly. Is
> anybody using BLCR to checkpoint MPI applications?
>
> I can checkpoint and restart sequential programs and programs that use
> POSIX threads without problems. So, BLCR seams to work.
>
> As described in the User's Guide, I start my MPI programs with
>
> $ mpirun -np 3 -ssi rpi crtcp -ssi cr blcr checkpoint_mpi
>
> When I call cr_checkpoint with the pid of mpirun only a single
> checkpoint file with the context of mpirun is saved. But I cannot find
> any context files of my applications. I also tried to linked my
> application directly to libcr.so, but this did not help.
>
> Has anybody an idea, what I could had made wrong?
>
>
> Heiko
I think you are do properly. I have the same problem.
|
