[slurm-users] partition with several nodes not following name pattern

[Vincent Berenz]

Hi,

For example, this configuration in slurm.conf works fine:

   NodeName=kilimanjaro CPUs=16 RealMemory=80419 Sockets=1 
CoresPerSocket=8 ThreadsPerCore=2 State=UNKNOWN
   PartitionName=slurmtest Nodes=kilimanjaro Default=YES 
MaxTime=INFINITE State=UP

This configuration works also:

   NodeName=falken CPUs=8 SocketsPerBoard=1 CoresPerSocket=4 
ThreadsPerCore=2 RealMemory=64358 State=UNKNOWN
   PartitionName=slurmtest Nodes=falken Default=YES MaxTime=INFINITE 
State=UP

I would like now to use kilimanjaro and falken in the same partition. I 
can not change their hostname. I tried:

   NodeName=n1 NodeHostName=kilimanjaro CPUs=16 RealMemory=80419 
Sockets=1 CoresPerSocket=8 ThreadsPerCore=2 State=UNKNOWN
   NodeName=n2 NodeHostName=falken CPUs=8 SocketsPerBoard=1 
CoresPerSocket=4 ThreadsPerCore=2 RealMemory=64358 State=UNKNOWN
   PartitionName=slurmtest Nodes=n[1-2] Default=YES MaxTime=INFINITE 
State=UP

But then job fails with error:

   srun: error: Task launch for 58.0 failed on node n1: Invalid job 
credential
   srun: error: Application launch failed: Invalid job credential
   srun: Job step aborted: Waiting up to 2 seconds for job step to finish.
   srun: error: Timed out waiting for job step to complete

Anything I am doing wrong ?

Many thanks

Vincent