Thanks for the insightful take. I appreciate it.
Kindly provide your mesh file (in this case .xml file) with a minimum working example so that I can reproduce the error and could possibly give any suggestion.
According to @dokken in this post PETSc Krylov Solver RuntimeError with fine mesh , this information means that the residual is zero at the first iteration, which usually indicates that the initial guess solves your problem. This was also reported for another case at the old Q&A forum.
Kindly go through PETSc Krylov Solver RuntimeError with fine mesh and PETSc Krylov solver failed to converge with big neumann boundary conditions posts where similar error is addressed, from my understanding, you should try few things like keeping the mesh coarser everywhere else except in the vicinity of crack tip and its expected crack path, try using umfpack or lu instead or mumps and also consider using iterative solvers with preconditioners, one of the setup which works for me is Bitbucket.
I hope this helps.