pfSens 2.7.0 Crashing in Proxmox 8 VM

Got 8 VMs using ProxMox that have a pfSense that on the new version 2.7.0 is creshing the VM.
The VM have on the console errors about sata0 slot not responding but cant see more as the VM are non responsiv.
Eache pfSense that is crashing it taks 2-3 weaks and it will crash.
Has any one enconter this problem ?

Initialy i thot it has to do with Qemu 8.0 but i downgraded to 7.2 and 7.1 and it is still happening.

All the ons runing pfSense 2.7.0 have crashed atlist ons since they were upgraded.
Some setings that i got on the VM’s are:


i Can trye to get more info as the Machines are crashing and i get more info.

Hardware used:

in the hardware section what is your ‘processor’ set to?
are you using passthrough NICs or bridges?
NUMA setting?

NIC virtio in most cases because using e1000 NIC is reducing networks and internet speed when internet speed is more than 500Mbps (found it out with trial and error)

1 Like

click the ‘Enable NUMA’ box and install the qemu-guest-agent package (in the VM, i think it is compatible with PFsense now). My guess is you are having an ACPI issue and this should ‘fix’ it. (in theory)

1 Like

OK NUMA enable on the ons that crash the most now to w8 and see :slight_smile:

Side note: I am running OPNsense with installed qemu-guest-agent package.
Works flawless but I am using virtio NICs to get 10 gig network.

1 Like

as apposed to Virtio, i use PCI Pass-through and actually give physical NICS to the VM. this is not always an option for everyone, but it has fixed the occasional ‘that is weird’ issues that crop up from virtualizing a router.

1 Like

yeah, that would also work - but you would loose the ability to migrate the VM to another Proxmox node (just in case of).

In general I also played around with PCIe-passthrough but didn’t see any performance difference between virtio and pcie-passthrough.

correct, there are some very specific issues with virtualized DCs, or other multihomed servers and even some multi-vlan configs where a PCIe-passthrough NIC is just better, though.

Yeah I agree. I had some issues with my PVE servers in the past when I added in a new PCIe card the NIC received a new PCIe-id and PVE was no longer booting the VM due to that. Not a big deal but annoying.

1 Like

Same using VirtIO for the 10G but the 1g e1000 is not running 1Gbps correctly is doing some slowdowns
Can’t use PCI Pasture as i am using same port in multiple VM’s

You could have a look if your NIC supports SR-IOV.
If yes, you could create some VFs and pass those through.

I just installed a pfsense 2.7 on a proxmox 8 mini PC server.
So new that I don’t even think is has more than 2 weeks uptime.

At least so far I haven’t seen any issues.

ANd as a difrent box that it is not runing in VM s mall Engeniring Board that i am runing at home and that one head a werd crash to there box was responding and all but internet was not geting routed true after a internet outige for shoort time; trye to replicat problem but could not and since that it was all ok.
Main main problem is the on runing in VM as thouse are spred out and cant get to eazy to them to reboot in most cases.

Upgraded soem to 2.7.1 No Crash yet :slight_smile:

Upgraded some to 2.7.2; No Crash yet.
Head DHCP Server problem using the new Kea DHCP went back to ISC DHCP.