Hi I have a NFS server with 6x100 Gb/s ports. The NFS server connected via two switches to the ESXi hosts. 3 NFS server ports per switch. Many ESXi hosts connected to the same switches using 25 Gb/s ports (4x25 Gb/s ports per single ESXi) The NFS server exports single mount point (raid6, many NVMe SSDs PCIe Gen5) VMs running with Oracle Linux 8.10 (kernel 4.18.0-553) There is a "magic" bandwidth number limited up to 100Gb/s between the NFS server and multiple VMs/ESXis. We tried different numbers of VMs/ESXis, different fio parameters (bs, iodepth, numjobs, reads/writes, etc...), and tried nconnect=16 option with mount command. Total bandwidth between the NFS server and multiple VMs/ESXis limited up to 100 Gb/s. Always. While iperf between the NFS and multiple VMs/ESXis shows hundreds Gb/s . Anton
