Nvidia-smi internal error running processes only

I installed CUDA 10.2 (removed CUDA 8.0 first) on a CentOS 7 system with 4 GeForce RTX 2080 Ti cards. nvidia-smi works for the first part but gives internal error trying to find running processes (see below). What is different about the last part? We have 2 identical systems from AdavancedHPC. The other server has CUDA 8.0 and CentOS 7 that they installed. nvidia-smi works fine there and is actually quite responsive. I’ve also noticed that nvidia-smi has a bit of a lag time on the system where I went to CUDA 10.

±----------------------------------------------------------------------------+
| NVIDIA-SMI 375.26 Driver Version: 440.33.01 |
|-------------------------------±---------------------±---------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 GeForce RTX 208… Off | 0000:18:00.0 Off | N/A |
| 25% 39C P0 50W / 250W | 0MiB / 11019MiB | 0% Default |
±------------------------------±---------------------±---------------------+
| 1 GeForce RTX 208… Off | 0000:3B:00.0 Off | N/A |
| 27% 44C P0 65W / 250W | 0MiB / 11019MiB | 1% Default |
±------------------------------±---------------------±---------------------+
| 2 GeForce RTX 208… Off | 0000:86:00.0 Off | N/A |
| 30% 48C P0 62W / 250W | 0MiB / 11019MiB | 1% Default |
±------------------------------±---------------------±---------------------+
| 3 GeForce RTX 208… Off | 0000:AF:00.0 Off | N/A |
| 31% 50C P0 28W / 250W | 0MiB / 11019MiB | 0% Default |
±------------------------------±---------------------±---------------------+

±----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
Internal error