Problem with active space

[Aristotle Papakondylis]

Dear Molpro users
I tried to do a CASSCF calculation for the Oxygen atom
using all valence electrons. This results in a 6e/4orb
CAS. Although the definition of the active space was:

occ,3,1,1,0
closed,1,0,0,0  (C2v)

the program (2000.1 and 2002.1) persistently kept closed
the 2s oxygen orbital.
This problem disappeared when the number of the active
orbitals of the same symmetry was increased, for example:

occ,4,1,1,0
closed,1,0,0,0

in this case the 2s orbital was taken to be active.
The same problem occurs for the S atom and even in the case
of molecular calculations.
Has anybody already encountered this kind of problem?
Is there any solution or this is a bug?
Thank you very much in advance for your answers.
regards
Aristotle

_________________________________
Aristotle Papakondylis
Assistant Professor
Physical Chemistry Laboratory
Department of Chemistry
University of Athens
P.O. Box 64004, 15710 Zografou
Athens, Greece
Phone:(+3010)7274565
Fax:  (+3010)7274527
_________________________________