One or more atoms are time integrated more than once

Hi guys,
Has anyone seen this warning before?
"One or more atoms are time integrated more than once"
I use a rigid body (by fix rigid command) to add confining presure to my direct shear test. I have explicitly exclude the atoms of that rigid body when doing the time integration, as suggested in the manual.
For example, I want to integrate a group of atoms excluding the rigid body, I assign:
1. region_1 and group_1 for the atoms I want to integrate
2. region_2 and group_2 for the rigid body I don't want to integrate
And then for the integration, I only set: fix 1 group_1 nve/sphere
But it still warns: One or more atoms are time integrated more than once

Does anyone know why? Is it because,during the simulation, the rigid body moves into region_1 which is under intergration? Does it matter a lot to my simulation result?
Thanks very much.

Anson