[SCore-users-jp] [SCore-users] PM device busy

[Danny Sternkopf]

Hi,

we have a 2-nodes SCore5.4 system based on RH7.3.
We always have to restart pm_ethernet after a job completed.
Otherwise the next job will fail. Here an example:

[esi ＠ score0 esi]$ scrun -nodes=4,file=hf ./a.out
SCOUT: Spawning done.
SCore-D 5.4.0 connected.
<0:0> SCORE: 4 nodes (2x2) ready.
Process 0 of 4 on score0.cluster.org
Process 2 of 4 on score1.cluster.org
Process 3 of 4 on score1.cluster.org
Process 1 of 4 on score0.cluster.org
pi is approximately 3.1416009869231249, Error is 0.0000083333333318
wall clock time = 0.000866
SCOUT: Session done.
[esi ＠ score0 esi]$ scrun -nodes=4,file=hf ./a.out
SCOUT: Spawning done.
<0> SCore-D:PM-PANIC(../network.c:331) pmBindChannel() (16:Device or resource busy)

So it seems when a job has completed the PM device is still busy.
But I don't know why. Do you have any ideas?
Is there any configuration issue?

Best regards,

Danny Sternkopf
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users