[torqueusers] Reservation problem with recent torque

[Ronny T. Lampert]

>>there is yet another thing which bothering me about 2.1.x: If I submit a job 
>>with "-l nodes=3:ppn=1" I used to get 1 processor on 3 nodes each (they are 
>>SMP-HT ones, see my other post). Now I get 3 processors on one node, which is 
>>not quite what I had intended. Have I missed another behind-the-scenes 
>>change?

By any chance, did you miss to set "set node_pack=false" ?
I know that nodes=3:ppn=1 should FORCE the scheduler to do otherwise, but
one never knows.

Seems we have to be a bit careful with the 2.1.x series.
Cheers,
Ronny