Windows 2016 VMs with poor performance

[Shadow Sysop]

I have 2 Proxmox nodes in my home lab. The first is a Dell Poweredge r710 w h700 RAID, 72 gb ram, SSD disks. This is the machine I am having issues with. Linux based VMs run perfectly, with no issue. Window 2016 VMs misbehave in odd ways, most notably terrible network performance as regards to Upload speeds (barely reaching 2MBPS, download speeds are around 200). The internet connection is a cable line at 960Mbps down to 100Mbps up. Windows 10 and Linux VMs achieve proper network speeds. I've even noticed the disk speed in Windows 2016 VMs can be a bit choppy too.

My 2nd node is an HP Proliant DL360, a similar server but not as powerful as the Poweredge. Windows 2016 VMs on this machine behave properly, achieving ideal network and disk use. So these feels odd to me.

I have installed the Virtio drivers. My windows VM configuration is as follows.

agent: 1
boot: order=virtio0;ide2;net0
cores: 4
ide2: iso1:iso/WIN2016.iso,media=cdrom
machine: pc-i440fx-5.2
memory: 8192
name: Win2016
net0: virtio=82:59:9F:C2:7E:96,bridge=vmbr0,firewall=1
numa: 0
ostype: win10
scsihw: virtio-scsi-pci
smbios1: uuid=b5783987-2733-4f97-b352-d892bcb62e34
sockets: 1
virtio0: local-lvm2:vm-211-disk-0,size=70G
vmgenid: bacd0cb3-3cf5-40e0-bbdd-990a61a58034

If I can get this ironed out, I plan on getting a subscription for support and moving into small scale production. Does anyone have any recommendations?

 

[squirell]

try set cpu = host, disk cache type writeback

 

[Shadow Sysop]

Whenever I do that in a Windows VM before the OS installation, it results in VM boot errors. “Host” works great for Linux VMs but not for Windows. I can usually change to “host” post OS installation but never before the install. Is this normal?

 

[entilza]

How is iperf from the host? from one pve to other for example.

 

[Shadow Sysop]

I'm not familiar with iperf. How can I get this information?

 

[entilza]

install iperf on both pve's "apt install iperf3"

on one run "iperf3 -s" to be server

on second run "iperf3 -c IP_OF_SERVER"

then reverse test. see results. You can then go from the guest to hosts from windows for example... iperf3 is for windows too.

 

[Shadow Sysop]

These are the results. I did as instructed.


Server listening on 5201
-----------------------------------------------------------
Accepted connection from 47.22.170.222, port 44268
[ 5] local 47.22.170.221 port 5201 connected to 47.22.170.222 port 44270
[ ID] Interval Transfer Bitrate
[ 5] 0.00-1.00 sec 112 MBytes 940 Mbits/sec
[ 5] 1.00-2.00 sec 112 MBytes 941 Mbits/sec
[ 5] 2.00-3.00 sec 112 MBytes 941 Mbits/sec
[ 5] 3.00-4.00 sec 112 MBytes 941 Mbits/sec
[ 5] 4.00-5.00 sec 112 MBytes 941 Mbits/sec
[ 5] 5.00-6.00 sec 112 MBytes 941 Mbits/sec
[ 5] 6.00-7.00 sec 112 MBytes 941 Mbits/sec
[ 5] 7.00-8.00 sec 112 MBytes 941 Mbits/sec
[ 5] 8.00-9.00 sec 112 MBytes 941 Mbits/sec
[ 5] 9.00-10.00 sec 112 MBytes 941 Mbits/sec
[ 5] 10.00-10.00 sec 287 KBytes 987 Mbits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bitrate
[ 5] 0.00-10.00 sec 1.10 GBytes 941 Mbits/sec

root@server1:~# iperf3 -c 47.22.170.221
Connecting to host 47.22.170.221, port 5201
[ 5] local 47.22.170.222 port 44270 connected to 47.22.170.221 port 5201
[ ID] Interval Transfer Bitrate Retr Cwnd
[ 5] 0.00-1.00 sec 115 MBytes 961 Mbits/sec 0 486 KBytes
[ 5] 1.00-2.00 sec 112 MBytes 940 Mbits/sec 0 486 KBytes
[ 5] 2.00-3.00 sec 113 MBytes 946 Mbits/sec 0 515 KBytes
[ 5] 3.00-4.00 sec 112 MBytes 940 Mbits/sec 0 515 KBytes
[ 5] 4.00-5.00 sec 112 MBytes 943 Mbits/sec 0 515 KBytes
[ 5] 5.00-6.00 sec 111 MBytes 935 Mbits/sec 0 515 KBytes
[ 5] 6.00-7.00 sec 113 MBytes 948 Mbits/sec 0 542 KBytes
[ 5] 7.00-8.00 sec 112 MBytes 940 Mbits/sec 0 542 KBytes
[ 5] 8.00-9.00 sec 112 MBytes 940 Mbits/sec 0 542 KBytes
[ 5] 9.00-10.00 sec 112 MBytes 939 Mbits/sec 0 542 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bitrate Retr
[ 5] 0.00-10.00 sec 1.10 GBytes 943 Mbits/sec 0 sender
[ 5] 0.00-10.00 sec 1.10 GBytes 941 Mbits/sec receiver

 

[entilza]

So we're just verifying the network speeds from the hosts. You can test the other way as well.

Next is to install iperf3 on the windows 2016 server try to iperf3 from it to the second pve.

At least you can test network performance.

 

[Shadow Sysop]

This is a great tool. I will use it to check on the VM next and deal with it accordingly. Thank you so much for the information and your prompt responses

 

[entilza]

No prob. Remember this just tests the network, once hard disks and VM's are involved it's separate but can help pinpoint the issue. Good luck.