Probably doesn't help, but I'm using 2.5.1 on Hyper-V on 2012R2, Gen1, ZFS without issues, either running or upgrading. But I did build myself a fresh 2.5.1 with restored config after a while, just because.

@spacecabbie No you can't use it like that. pfSense does support high-availability configurations but that requires you to run two guest instances of pfSense linked together on different hypervisor hosts. If it's that much of a concern then go pick up an SG-1100 and then you can reboot without losing connectivity.

@ahking19 I set it up early last week to test it to see if that would remain stable. After it’s been up for nearly a week, I didn’t have time this weekend to roll back. I also made a few config changes on in the firewall and a VLAN tag that I would need to replicate from the 2.5.1 router that was crashing and the router on a stick back to the 2.4.5 VM. In this case, it’s not quite as simple as just hitting the “roll back” button.

@matt_alton I followed below process

https://www.youtube.com/watch?v=PwBanc4IGsI

@matt_alton
I saw the same behavior but then on my wan link. I am running physical tho. I ditched 2.5.1 went back to 2.5 and everything was back in order and stable.

I’d love to go back to 2.4.5 p1 but I cannot install latest version of pfblockerng. My advise, stay clear of the 2.5 line of releases as long as you can.

@kiokoman

Thank you so much, I don't know why I had such difficulty figuring this out. Thanks again!

I have checked the suggested optimizations.

default:
vmx0: Using 512 TX descriptors and 512 RX descriptors
vmx0: Using 8 RX queues 8 TX queues

with optimizations:
vmx0: Using 4096 TX descriptors and 2048 RX descriptors
vmx0: Using 8 RX queues 8 TX queues
vmx0: Using MSI-X interrupts with 9 vectors

2.5.1
mean: pf 2,15 Gbit/s / forwarding: 2,74 Gbit/s
mean: pf 4,35 Gbit/s / forwarding: 5,48 Gbit/s
mean: pf 7,42 Gbit/s / forwarding: 8,67 Gbit/s
mean: pf 8,67 Gbit/s / forwarding: 8,61 Gbit/s

The values are even a little bit slower....

Any other ideas ? I will crosscheck with freebsd 12-2. I have checkd
Freebsd 12-2 and Freebsd 13-RC4 bare metal. Factor 2 and Freebsd 13 is much faster.

I solved the issue "the easy way".

I went into Proxmox and changed VMBR0 bridge from old interface NIC to new interface NIC instead of having a new bridge along with setting back the IP and gateway.
And subsequently I changed the same in PfSense back to use the old NIC.
As it works and I'm no network genius, I'll leave it be for now :)

NetDiag-Page-1.jpg

EDIT: just change some DNS Resolver settings, like the outgoing network interfaces (selected only WAN), disable DNSSEC, enabling Forwarding Mode and also disable IPv6 on WAN and LAN on the interfaces setting and now it's working!

@ikifar
If E1000 works what is the issue with that. Check the throughput using iPerf and speed test to ensure it works as expected.
If you don’t see any performance drop happy days.

@KOM I appreciate the insight. I think that will be my backup plan.

I realized my office computer has a wifi adapter. I never used it. I'm thinking I will connect it to my current Netgear router so that network is not taken down. I'll connect my ethernet to Pfsense. I found a program called ForceBindIP, which I'm thinking could force Firefox to use the ethernet adapter, so I can configure PFsense in the web gui. I have to make sure the WAF is stable...😁

@limez17 yes easy tek.png that's the setup the problem is that in aws you can't add specific routes so my next guess is to portforward in the dmz are websites and rdp is a must for admin emp dev network

thx for the help

It seems this could be an ISP issue now, I am still troubleshooting, changed nothing and am now getting around 850 mbps
https://twitter.com/MathesonStep/status/1362429291838001154

@raitd said in vmx NIC ordering for pfSense on vSphere 5.5+:

@jimp I've seen many linux distro's handle having several nics in vmware and handle adding more perfectly fine.

Linux is vastly different in their default naming schemes. The current default names in Linux are based on bus locations which likely wouldn't change in those cases. FreeBSD counts up from 0 for each instance of the driver it finds when probing, so if the probe order changes, so does the NIC assignment order.

Does pfSense / FreeBSD have an option where you can force binding to a particular MAC address? If so I imagine many people would love to have this as a selectable option.

No.

@bullz3y3 Have same issue in 2021, and fixed by changing the Network Type from Virtio to e1000 also. Your finding in 2015 is still helping others. Thanks.

My case is having a PCI Passthrough WAN, and 2 bridge LAN, hosted on KVM on debian 10