Warings of complex samples_logistic regression

[Suh-Ing Amy Hsieh]

Hi listers:

While I create an file of complex samples, it shows warings.

Warings: This procedure does not check the consistency of the working data file with the plan file. We recommend looking at the output talbe or the plan file to check consistency before performng selection or analysis.

While I used logistic regression (binary or multinomial dependent variables) of complex samples for analyzing the clustered data (15.0 SPSS), two types of warings appeared in 3 covariates of SPSS output. Please see below. I try to increase iterations to 2,000 but one one predictor did not appear warnings again. What are causes of warnings? How can I fix it without warnings?

CSLOGISTIC d100_in(LOW) WITH geo_loc_2 geo_loc_3 geo_loc_4
  /PLAN FILE = 'E:\NHIRD_hsct7\Index hsct.csaplan'
  /MODEL geo_loc_2 geo_loc_3 geo_loc_4
  /INTERCEPT INCLUDE=YES SHOW=YES
  /STATISTICS PARAMETER EXP SE CINTERVAL TTEST
  /TEST TYPE=F PADJUST=LSD
  /MISSING CLASSMISSING=EXCLUDE
  /CRITERIA MXITER=100 MXSTEP=5
   PCONVERG=[1e-006 RELATIVE] LCONVERGE=[0] CHKSEP=20
   CILEVEL=95
  /PRINT SUMMARY CLASSTABLE VARIABLEINFO SAMPLEINFO.

Warings: The design-based covariance matrix is singular. The validity of results is uncertain.

CSLOGISTIC d100_in(LOW) WITH dx_5gr_2 dx_5gr_3 dx_5gr_4 dx_5gr_5
  /PLAN FILE = 'E:\NHIRD_hsct7\Index hsct.csaplan'
  /MODEL dx_5gr_2 dx_5gr_3 dx_5gr_4 dx_5gr_5
  /INTERCEPT INCLUDE=YES SHOW=YES
  /STATISTICS PARAMETER EXP SE CINTERVAL TTEST
  /TEST TYPE=F PADJUST=LSD
  /MISSING CLASSMISSING=EXCLUDE
  /CRITERIA MXITER=100 MXSTEP=5
   PCONVERG=[1e-006 RELATIVE] LCONVERGE=[0] CHKSEP=20
   CILEVEL=95
  /PRINT SUMMARY CLASSTABLE VARIABLEINFO SAMPLEINFO.

Warings: The log-likelihood value cannot be increased after the maximum number of steps in the step-halving method. The CSLOGISTIC procedure contues despite the above waring(s). Subsequent results shown are based on the lastiteration. Validity of the model fit is uncertain.

Any suggestions? Thanks.
Amy