fatal error: bracketing interval incorrect

[junjun.yu]

I'm doing optimization calculations using PAW/PBE. In the output file, it shows this error message and the calculation crushed:

ZBRENT: fatal error: bracketing interval incorrect
please rerun with smaller EDIFF, or copy CONTCAR
to POSCAR and continue

I tried to restart calculation by copying CONTCAR to POSCAR. But the problem remains.

How to solve this problem? Thank you!

Here is the INCAR files:
System = Mo-V-Te-NB-O 3L-1V singlet optimization OH-frozen

! 0=new, 1=wave (d), 2=same cutoff&latt (1&2 need WAVECAR file)
ISTART=1
! 0=wave (d), 1=CHGCAR
ICHARG=0

NPAR = 8
! how is parallelization done
LPLANE = .TRUE.

LREAL = auto
! use real space projection of PPs
!ROPT = 2e-4 2e-4
! prec of real space projection; repeat for each species
PREC =Normal
! sets cutoff and FFT grid; Normal | Accurate | High; if High, sets ROPT=2e
-4 automatically
ALGO = Fast
NSIM =5


GGA=PE !PBE

ENCUT= 283.00 eV ! energy cut-off for the calculation (optional); these are num
bers for C
ENAUG= 700.00 eV ! energy cut-off for the augmentation charges; default is 645
eV





ISPIN=2 !singlet
NUPDOWN=0
!MAGMOM=2 2 2 2 2 2 2 2



NSW=120 !!!! Do not relax for now !!!
IBRION = 2 ! 1: quasi-NR, 2: CG algorithm for ions; use 2 for first run
ISIF = 2 ! 0: relax ions, 1,2:relax ions,calc stresses, 3:relax ion+cell
POTIM = 0.4 ! time step
EDIFFG=-0.05
!NFREE = 10 ! number of DIIS vectors to save



! write wavefunction
LWAVE=.TRUE.
! write chg and chgcar
LCHARG=.TRUE.

[admin]

please have a look at the OUTCAR file, at the forces of the last ionic step before Brent's algorithm fails
-- if the calculation is very well-converged already, please proceed using IBRION=1; ADDGRID=.TRUE. and a larger ENCUT
-- if this error appears at an early stage of the relaxation, pelase check if your input geometry was reasonable, eg by having a look at the interatomic distances (written in OUTCAR) or the forces of the first geometries. It may then help to increase the accuracy (PREC= Accurate and a higher ENCUT) and to decrease POTIM to avoid too large stepwidhts in the first relaxation step.