I have the following code:
 //---------------------------------------------------Solving for etaf
  varf ahKf(etafK, varphi)
    =  int2d(Th[K])(kappa*grad(etafK)'*grad(varphi));
  varf fhKf(etafK, varphi)
    = int2d(Th[K])(f*varphi); 
  varf bhKf(vf, varphi)
    = int2d(Th[K])(vf*varphi);
 
  // local matrices for etaf
  matrix Akf = ahKf(VhK,VhK);
  matrix Bkf = bhKf(V0,VhK);
  real[int] Fkf(ndofVhK);
  Fkf = fhKf(0, VhK);
  // block matrix LHS
  matrix AAkf = [ [ Akf, Bkf ],
                  [ Bkf', 0  ] ];
  real[int] rhsetaf(ndofVhK+nDoFu0), soletafk(ndofVhK+nDoFu0);
  real[int] blmf(nDoFu0), lmf(nDoFu0);
  //second equation to hybrid form equal 0
  blmf = 0;
  // block vector RHS
  rhsetaf = [ Fkf, blmf ];
  set(AAkf, solver=sparsesolver);
  soletafk = AAkf^-1*rhsetaf; //solve linear system
  [etafK[], lmf] = soletafk;   //set the value
  cout << "etaF elem " << K << ": " << etafK[] << "\n";
but it keeps giving me the following error:
 Error umpfack umfpack_di_numeric  status  1
 Error umfpack_di_solve  status  1
etaF elem 0: 6	
	-0.3611303349	-0.3611303349	-0.09386227966	-0.09386227966	0.1877245593
	-0.251181056	
I already tried to change the solver to
set(AAkf,solver="UMFPACK64")
but the error remains. Does anyone have any ideia what it could be?