Awesome
Proxmox VE TIPS
Here are some tips for creating VMs on a Proxmox VE. For more information on performance tweaks, please refer here.
Sample Configurations and Miscellaneous for Mobile Network
<a id="toc"></a>
Table of Contents
<a id="trouble"></a>
Troubleshooting
<a id="mongodb_crash"></a>
When starting MongoDB, an illegal instruction occurs and MongoDB crashes
If an illegal instruction occurs and MongoDB crashes when starting MongoDB, refer to this and select host as the Type in the Edit: Processors dialog to enable AVX instruction for CPU of VM.
<img src="./images/pve_edit_processor.png" title="./images/pve_edit_processor.png" width=600px></img>
<a id="srsran_build_crash"></a>
While building srsRAN_Project, an error occurs due to missing F16C, FMA and BMI instructions
If an error occur while building srsRAN Project due to missing F16C, FMA and BMI instructions, select host as the Type in the Edit: Processors dialog to enable these instructions for CPU of VM.
<img src="./images/pve_edit_processor.png" title="./images/pve_edit_processor.png" width=600px></img>
<a id="error_related_to_secure_boot"></a>
An error occurs on VM in related to Secure Boot
If an error occurs on VM in related to Secure Boot, try selecting SeaBIOS in the Edit: BIOS dialog.
<img src="./images/pve_edit_bios.png" title="./images/pve_edit_bios.png" width=250px></img>
<a id="perf_imp"></a>
Performance Improvement
<a id="mount_opt"></a>
Filesystem mount option
I have added noatime to the mount option of SSD storages in a Proxmox VE machine.
And not using swap. Also, I have configured as the same for VMs.
<a id="select_disk_format_type"></a>
Selecting the format type of disk when creating a VM
When creating a VM, the format type of disk Raw disk image (raw) is a little faster than QEMU image format (qcow2).
<img src="./images/pve_disk.png" title="./images/pve_disk.png" width=600px></img>
<a id="select_disk_cache_type"></a>
Selecting the cache type of disk
According to this (in Japanese), with regard to the cache type of disk, Write back for HDD is a little faster than Write back (unsafe).
Write back (unsafe) for SSD is a little faster than Write back.
It is also recommended to turn on the IO thread.
For HDD
<img src="./images/pve_hdd_cache.png" title="./images/pve_hdd_cache.png" width=500px></img>
For SSD
Also turn on SSD emulation.
<img src="./images/pve_ssd_cache.png" title="./images/pve_ssd_cache.png" width=500px></img>
<a id="virtio_nic"></a>
VirtIO paravirtualized NIC
For maximum performance, a VirtIO paravirtualized NIC is recommended. You may use the network with DPDK and XDP on a Linux OS that supports VirtIO. For example, you may use N3 and N6 interfaces with UPG-VPP (with DPDK) or eUPF (eBPF/XDP native driver mode) on VirtIO.
A quick measurement using iperf3 on a VirtIO network between two Ubuntu 24.04 VMs on a Proxmox VE machine with an Intel Core-i5 10400 CPU showed the following results:
| A--B | 1) TCP<br>throughput | 2) UDP<br>throughput | 2) UDP<br>packet loss | 3) RTT<br>(msec) |
|---|---|---|---|---|
| 192.168.16.151 -- 192.168.16.152 | S:25.6 Gbps<br>R:25.5 Gbps | S:2.99 Gbps<br>R:2.83 Gbps | 6.9 % | 0.260 |
# iperf3 -c 192.168.16.152
Connecting to host 192.168.16.152, port 5201
[ 5] local 192.168.16.151 port 49214 connected to 192.168.16.152 port 5201
[ ID] Interval Transfer Bitrate Retr Cwnd
[ 5] 0.00-1.00 sec 2.97 GBytes 25.5 Gbits/sec 0 2.63 MBytes
[ 5] 1.00-2.00 sec 2.96 GBytes 25.4 Gbits/sec 0 2.91 MBytes
[ 5] 2.00-3.00 sec 2.94 GBytes 25.2 Gbits/sec 0 3.52 MBytes
[ 5] 3.00-4.00 sec 2.98 GBytes 25.6 Gbits/sec 0 3.52 MBytes
[ 5] 4.00-5.00 sec 2.98 GBytes 25.6 Gbits/sec 0 3.52 MBytes
[ 5] 5.00-6.00 sec 2.97 GBytes 25.5 Gbits/sec 0 3.52 MBytes
[ 5] 6.00-7.00 sec 2.97 GBytes 25.5 Gbits/sec 0 3.52 MBytes
[ 5] 7.00-8.00 sec 3.00 GBytes 25.8 Gbits/sec 0 3.72 MBytes
[ 5] 8.00-9.00 sec 2.99 GBytes 25.7 Gbits/sec 0 3.72 MBytes
[ 5] 9.00-10.00 sec 2.99 GBytes 25.7 Gbits/sec 0 3.72 MBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bitrate Retr
[ 5] 0.00-10.00 sec 29.8 GBytes 25.6 Gbits/sec 0 sender
[ 5] 0.00-10.01 sec 29.8 GBytes 25.5 Gbits/sec receiver
iperf Done.
# iperf3 -c 192.168.16.152 -u -b 5G
Connecting to host 192.168.16.152, port 5201
[ 5] local 192.168.16.151 port 58877 connected to 192.168.16.152 port 5201
[ ID] Interval Transfer Bitrate Total Datagrams
[ 5] 0.00-1.00 sec 316 MBytes 2.65 Gbits/sec 228619
[ 5] 1.00-2.00 sec 317 MBytes 2.66 Gbits/sec 229287
[ 5] 2.00-3.00 sec 315 MBytes 2.65 Gbits/sec 228422
[ 5] 3.00-4.00 sec 364 MBytes 3.05 Gbits/sec 263482
[ 5] 4.00-5.00 sec 380 MBytes 3.19 Gbits/sec 275106
[ 5] 5.00-6.00 sec 374 MBytes 3.14 Gbits/sec 271097
[ 5] 6.00-7.00 sec 375 MBytes 3.14 Gbits/sec 271209
[ 5] 7.00-8.00 sec 378 MBytes 3.17 Gbits/sec 274073
[ 5] 8.00-9.00 sec 372 MBytes 3.12 Gbits/sec 269176
[ 5] 9.00-10.00 sec 372 MBytes 3.12 Gbits/sec 269731
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bitrate Jitter Lost/Total Datagrams
[ 5] 0.00-10.00 sec 3.48 GBytes 2.99 Gbits/sec 0.000 ms 0/2580202 (0%) sender
[ 5] 0.00-9.84 sec 3.24 GBytes 2.83 Gbits/sec 0.002 ms 176852/2580202 (6.9%) receiver
iperf Done.
# ping 192.168.16.152 -c 10
PING 192.168.16.152 (192.168.16.152) 56(84) bytes of data.
64 bytes from 192.168.16.152: icmp_seq=1 ttl=64 time=0.263 ms
64 bytes from 192.168.16.152: icmp_seq=2 ttl=64 time=0.266 ms
64 bytes from 192.168.16.152: icmp_seq=3 ttl=64 time=0.222 ms
64 bytes from 192.168.16.152: icmp_seq=4 ttl=64 time=0.235 ms
64 bytes from 192.168.16.152: icmp_seq=5 ttl=64 time=0.295 ms
64 bytes from 192.168.16.152: icmp_seq=6 ttl=64 time=0.293 ms
64 bytes from 192.168.16.152: icmp_seq=7 ttl=64 time=0.238 ms
64 bytes from 192.168.16.152: icmp_seq=8 ttl=64 time=0.275 ms
64 bytes from 192.168.16.152: icmp_seq=9 ttl=64 time=0.246 ms
64 bytes from 192.168.16.152: icmp_seq=10 ttl=64 time=0.274 ms
--- 192.168.16.152 ping statistics ---
10 packets transmitted, 10 received, 0% packet loss, time 9193ms
rtt min/avg/max/mdev = 0.222/0.260/0.295/0.023 ms