We are using nobrl in a production environment with 3k workstations.
This fixed also problems with CIFS clients on low bandwidth connections. There a kernel bug led to corrupted files by race conditions between application writes and the kernel flushing buffers in kernel mode - without looking at the locks by the application write (different writing processes).
We are using nobrl in a production environment with 3k workstations.
This fixed also problems with CIFS clients on low bandwidth connections. There a kernel bug led to corrupted files by race conditions between application writes and the kernel flushing buffers in kernel mode - without looking at the locks by the application write (different writing processes).
Maybe that is fixed in newer kernels ...