CCL:G: Problem during QCISD calculations in Gaussian 09

["Ambrish Kumar Srivastava" <ambrishphysics::gmail.com>]

 Sent to CCL by: "Ambrish Kumar Srivastava" [ambrishphysics=gmail.com]
 Dear all,
 I have performed calculations on small alkali halide clusters with QCISD method
 in Gaussian 09. The route section of my job reads:
 %NProcShared=20
 %mem=42gb
 # opt freq aug-cc-pvtz polar qcisd
 The job fails with an error which can be seen from the bottom of output below:
  Iteration Nr.  12
  **********************
  DD1Dir will call FoFMem   1 times, MxPair=        56
  NAB=    20 NAA=    10 NBB=     6.
  DE(Z)  -0.28201232D+00          E(Z)=    -0.11475119242D+03
  NORM(A)=   0.10305469D+01
  Discarding MO integrals.
  S**2, projected HF & approx projected MPn energies after annihilation of
  unwanted spin states (see manual for definitions):
  spins       (S**2,0) (S**2,1)    PUHF        PMP2        PMP3        PMP4
  annihilated
  s+1         0.75000  0.75000 -114.469210 -114.756186 -114.750801
  s+1,s+2     0.75000  0.75000 -114.469210 -114.756186 -114.750801
  s+1 to s+3  0.75000  0.75000 -114.469210 -114.756186 -114.750801
  s+1 to s+4  0.75000  0.75000 -114.469210 -114.756186 -114.750801
 Erroneous write. Write -1 instead of 1500424200.
 fd = 4
 orig len = 1500424200 left = 1500424200
 g_write
 Any help to resolve this problem will be appreciated. Also, can i obtain
 polarizability and hyperpolarizabity at QCISD method by keywords given as above?
 Thanks in advance!