I run a pretty extensive lab for work related endeavors, but generally as of late the VMs have come to a screech halt... below are my specs, let me know what you what consider looking into...
Storage
FreeNas 11.3-U4.1 Running on a SuperMicro x9DR3-F with 2 intel E52620 Xeons and 128G of DDR3 Ram, 10G network for storage, 1G network for management. 20, 4TB SAS 6G HGST HUS726040ALS214 drives
Hyperv Hosts
Windows Server 2016, running on SuperMicro X9DRi-LN4+ with 2 Xeon E5-2630 CPUs, and 256G DDR3 ram, boots from Samsung 850 SSD, 10G NIC for storage, 1G for management
Unifi 10G (fiber) switch for storage, cisco 2960 for management
I run anywhere between 10 and 50 hyperV vms, generally I assign 4G of ram, and 2 vCPU... VMs run anything from windows 7 to Ubuntu 20!
over the last six months when I log in, the VMs have been getting more and more unresponsive. I have done several VM rebuilds nothing changed, I have done full patching, and short of a 100% rebuild I'm just lost. I Thought this might be related to 10G bottleneck, but I never hit more than 30 or 40MB on the ports\over the switch. I even moved VMs to local storage and no real change!
so, what are you thoughts? I don't want to change hypervisors due to customer needs(I replicate my customers infrastructure for testing and dev)
Storage
FreeNas 11.3-U4.1 Running on a SuperMicro x9DR3-F with 2 intel E52620 Xeons and 128G of DDR3 Ram, 10G network for storage, 1G network for management. 20, 4TB SAS 6G HGST HUS726040ALS214 drives
Hyperv Hosts
Windows Server 2016, running on SuperMicro X9DRi-LN4+ with 2 Xeon E5-2630 CPUs, and 256G DDR3 ram, boots from Samsung 850 SSD, 10G NIC for storage, 1G for management
Unifi 10G (fiber) switch for storage, cisco 2960 for management
I run anywhere between 10 and 50 hyperV vms, generally I assign 4G of ram, and 2 vCPU... VMs run anything from windows 7 to Ubuntu 20!
over the last six months when I log in, the VMs have been getting more and more unresponsive. I have done several VM rebuilds nothing changed, I have done full patching, and short of a 100% rebuild I'm just lost. I Thought this might be related to 10G bottleneck, but I never hit more than 30 or 40MB on the ports\over the switch. I even moved VMs to local storage and no real change!
so, what are you thoughts? I don't want to change hypervisors due to customer needs(I replicate my customers infrastructure for testing and dev)