Hello.
I have a similar case: communication between VMs using the provider network works fine, but not with the self-service network. Using jumbo frames will improve performance significantly.  

Nguyen Huu Khoi

On Wed, Jul 30, 2025 at 10:36 AM engineer2024 <engineerlinux2024@gmail.com> wrote:

Did this work better with LXB plugin earlier? 

On Wed, 30 Jul 2025, 09:01 Jeff Yang, <yjf1970231893@gmail.com> wrote:

Hi all,

I'm running OpenStack Zed with OVN as the Neutron backend. I've encountered a severe VM network performance degradation issue and would appreciate any insights from the community.

🔹 **Environment**
- OpenStack version: Zed
- Neutron backend: OVN
- Two VMs (each 4 vCPUs, 8 GB RAM) launched on two different compute nodes
- Multi-queue is enabled on the VM vNICs

🔹 **Issue**
The network performance inside the VMs is significantly worse than on the host nodes, with high packet retransmissions observed.

🔹 **Test results**

**Host-to-Host (VTEP IPs)**

```

$ iperf3 -c 192.168.152.152
Connecting to host 192.168.152.152, port 5201
[  5] local 192.168.152.153 port 45352 connected to 192.168.152.152 port 5201
[ ID] Interval           Transfer     Bitrate         Retr  Cwnd
[  5]   0.00-1.00   sec  1.38 GBytes  11.8 Gbits/sec    0   3.10 MBytes
[  5]   1.00-2.00   sec  1.37 GBytes  11.8 Gbits/sec    0   3.10 MBytes
[  5]   2.00-3.00   sec  1.42 GBytes  12.2 Gbits/sec    0   3.10 MBytes
[  5]   3.00-4.00   sec  1.39 GBytes  11.9 Gbits/sec    0   3.10 MBytes
[  5]   4.00-5.00   sec  1.38 GBytes  11.8 Gbits/sec    0   3.10 MBytes
[  5]   5.00-6.00   sec  1.43 GBytes  12.3 Gbits/sec    0   3.10 MBytes
[  5]   6.00-7.00   sec  1.41 GBytes  12.1 Gbits/sec    0   3.10 MBytes
[  5]   7.00-8.00   sec  1.41 GBytes  12.1 Gbits/sec    0   3.10 MBytes
[  5]   8.00-9.00   sec  1.41 GBytes  12.1 Gbits/sec    0   3.10 MBytes
[  5]   9.00-10.00  sec  1.42 GBytes  12.2 Gbits/sec    0   3.10 MBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bitrate         Retr
[  5]   0.00-10.00  sec  14.0 GBytes  12.0 Gbits/sec    0             sender
[  5]   0.00-10.04  sec  14.0 GBytes  12.0 Gbits/sec                  receiver

iperf Done.

```

**VM-to-VM (through overlay network)**

```

$ iperf3 -c 10.0.6.10
Connecting to host 10.0.6.10, port 5201
[  5] local 10.0.6.37 port 56710 connected to 10.0.6.10 port 5201
[ ID] Interval           Transfer     Bitrate         Retr  Cwnd
[  5]   0.00-1.00   sec   499 MBytes  4.19 Gbits/sec  263    463 KBytes
[  5]   1.00-2.00   sec   483 MBytes  4.05 Gbits/sec  467    367 KBytes
[  5]   2.00-3.00   sec   482 MBytes  4.05 Gbits/sec  491    386 KBytes
[  5]   3.00-4.00   sec   483 MBytes  4.05 Gbits/sec  661    381 KBytes
[  5]   4.00-5.00   sec   472 MBytes  3.95 Gbits/sec  430    391 KBytes
[  5]   5.00-6.00   sec   480 MBytes  4.03 Gbits/sec  474    350 KBytes
[  5]   6.00-7.00   sec   510 MBytes  4.28 Gbits/sec  567    474 KBytes
[  5]   7.00-8.00   sec   521 MBytes  4.37 Gbits/sec  565    387 KBytes
[  5]   8.00-9.00   sec   509 MBytes  4.27 Gbits/sec  632    483 KBytes
[  5]   9.00-10.00  sec   514 MBytes  4.30 Gbits/sec  555    495 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bitrate         Retr
[  5]   0.00-10.00  sec  4.84 GBytes  4.15 Gbits/sec  5105             sender
[  5]   0.00-10.05  sec  4.84 GBytes  4.14 Gbits/sec                  receiver

iperf Done.

```

As you can see, there's a sharp drop in performance inside the VMs (from ~12 Gbps down to ~4 Gbps), with a lot of retransmissions in 10 seconds.

Has anyone encountered a similar issue with OVN and VM networking performance? What could be the potential causes?

Any suggestions or debugging tips are welcome.

Thanks in advance!

Best regards,