opnsense freezes and needs reboot

[scream]

Any news to this issue?
I think I‘m running into a simillar case as well.

opnSense VM on ESX 7.0.1 with vmxnet3 cards.
VMWare tools are installed. All pakets are up to date.

On my case it doesn‘t occour daily but sometimes when I take a snapshot or when using vMotion to move the VM to another Host.

Issue shows as WAN gateway has latency over 1000ms and a lot of packet loss. Didn‘t found another solution except reboot until now. :-/

[mgrue]

What fixed my problems was:
- Enabling all hardware accelerations/offloads under Interfaces / Settings
- Moving to faster hardware with 10 GbE NICs
- Updating to VMware ESXi 7.0 U1

I can't not 100% tell what really fixed the problems, but they are gone

[scream]

Tried now with enabled hardware offloading... didn‘t help for me.
ESX is up to date with the latest patches released a few days ago.
I already use 2*10GbE (Intel X722) as uplinks with LACP configured on dvSwitch.

Which NICs do you use? (E1000E or VMXNET3?)

[mgrue]

VMXNET3, Broadcom 57810 NICs, no LACP, Standard vSwitch
My VM has 1.5 GB of memory now. Before I had 1.0 GB and ran out of memory occasianally which also created very sluggish routing behaviour.

[scream]

Okay. May I will test without LACP first. My VM is running with 8GB RAM as I use Sensei with Elasticsearch.
But with 20.1.9 it was running without any issue for many month.

[nykaer]

Seems to be solved on my environment now.
On VMware ESXi 6.7U3.

Interfaces -> Settings: Enabled all hardware offloading + VLAN hardware filtering.

The last part seemed to do the trick, and have been running for more than 48 hours now.  Before I needed to reboot every 6-10 hours or so..

[Rajstopy]

Hi there,

Looks like I've a very similar issue here... OPNSense was running well for months but suddenly interfaces begun to be stuck. Rebooting OPNSense usually solves temporarily the problem, but if I reboot the hypervisor itself then I'm quiet for several days. This issue just makes me nuts because I've a lot of services relying on my network connection.

I suspected Wireguard, but seems to occurs even if the service if off...

An answer I received this morning told me about another VM that could cause the system NIC to freeze. I remember my problem appeared suddenly one day, without having changed anything on the system... but perhaps a new VM

Do you remember if you noticed this issue after having added a new VM?

R.

[tmueko]

I now have the problem with 21.1.6.
FreeBSD-12.x and pfSense are working fine.
I have a OpnSense-Cluster on two Dell R630 on 10GB-Links. Sometimes, both VMs freezed within one hour :-(

Tried all combinations of +/- lro,  +/- tso, +/- (rxcsum, txcsum) and vlanhwtag: Nothing worked.

Any new idea on this?