I have a simple ESXi setup. A single host with three VMs, a vCenter Appliance, and a Data Protection appliance. The VDP appliance is being ran from a 5-drive RAID5 NAS using NFSv3. I have one standard vSwitch with two VMkernels. One for VM traffic and one for NFS traffic. The VM traffic is on 192.168.2.x subnet and NFS is on 192.168.3.x subnet. The host NIC and NFS NIC are the only two IP addresses on the 3.x subnet.
My problem is that I'm having SERIOUS issues with latency; to the point that my VDP is not stable. Backups are not completing and integrity checks are not running. I routinely get alerts that VDP services are not running. Looking at the NFS datastore latency, average read latency is usually between 100-200ms and write latency is between 300-400ms with peaks well over 1000ms and sometimes over 2000ms!!! I am not sure how to proceed. Running a traceroute from ESXi shell shows only one hop to the NAS.
The host and NAS are using the same switch that the entire office is using, which is approximately 30 people. Also, it is only a 100Mbps network. I have a feeling these two things combined are contributing to my issues. A dedicated gigabit NIC on the host with a dedicated gigabit switch would help, but not sure if that is the end-all solution. Something seems off. I really appreciate any help you can give.
Thanks!