Hello all,
i have a proxmox install, latest version on my server with a xeon 1220 v6, 32 GB RAM ssd with ZFS for proxmox and a ZFS RAID0 with 2 drives for the vm storages.
I have 7 VM running on my server for about 2 years now.
Average load of the server is 2,5 and 30% of CPU and 5% of IO delay.
For networking i have 2 onboard gigabit ethernet and 2 intel 2port gigabit pci express cards.
Networking at then node and the vms is configured through OVS bonds and bridges and vlans.
One of the vms is running OPNsense as main router/firewall on my network.
Lately i have a problem with latency and packet loss and spikes, i did not have about 6-7 months ago and i would like to know if it might have to do with all updates and cpu microcode updates.
If for example i ping the local address of a vm and at the same time i run speedtests and file downloading, there are spikes from 1ms to 70 ms (local network) and packet loss for internet connection.
At the beginning i though it might be the firewall vm, but even though i gave maximum ram and cpu to the firewall VM and also tried creating a simple vm for an internet pppoe connection taking off the OPNsense VM, but the problem continued.
I would like to know, if changing from ovs to linux bridge might resolve this problem? could it be related to something else?
(i have not tried yet switching to linux bridge because of the complicated ovs setup and the server running at my office)
Thank you in advance
i have a proxmox install, latest version on my server with a xeon 1220 v6, 32 GB RAM ssd with ZFS for proxmox and a ZFS RAID0 with 2 drives for the vm storages.
I have 7 VM running on my server for about 2 years now.
Average load of the server is 2,5 and 30% of CPU and 5% of IO delay.
For networking i have 2 onboard gigabit ethernet and 2 intel 2port gigabit pci express cards.
Networking at then node and the vms is configured through OVS bonds and bridges and vlans.
One of the vms is running OPNsense as main router/firewall on my network.
Lately i have a problem with latency and packet loss and spikes, i did not have about 6-7 months ago and i would like to know if it might have to do with all updates and cpu microcode updates.
If for example i ping the local address of a vm and at the same time i run speedtests and file downloading, there are spikes from 1ms to 70 ms (local network) and packet loss for internet connection.
At the beginning i though it might be the firewall vm, but even though i gave maximum ram and cpu to the firewall VM and also tried creating a simple vm for an internet pppoe connection taking off the OPNsense VM, but the problem continued.
I would like to know, if changing from ovs to linux bridge might resolve this problem? could it be related to something else?
(i have not tried yet switching to linux bridge because of the complicated ovs setup and the server running at my office)
Thank you in advance
Last edited:
