Unbound repeatedly crashes with SIGSEGV on Hyper-V – clean VM resolved the issue

Started by Canardo96, Today at 08:23:13 AM

Previous topic - Next topic
Hi all,

I am experiencing repeated Unbound crashes on OPNsense 26.7.6_3 running under Microsoft Hyper-V.

The recurring kernel message is:

pid XXXXX (unbound), jid 0, uid 59: exited on signal 11
(no core dump - sugid process denied by kern.sugid_coredump)

Environment:
- OPNsense 26.7.6_3
- FreeBSD 15.1-RELEASE-p4
- Microsoft Hyper-V
- 4 vCPU
- Approx. 12 GB RAM
- Hyper-V network interfaces hn0 / hn1
- Zenarmor 2.7
- Unbound DNS Resolver

I originally had a very unstable OPNsense VM. Besides the Unbound SIGSEGV crashes, I saw syslog-ng aborts, Zenarmor agent crashes and netmap-related issues.

Because it was becoming almost impossible to determine the root cause, I decided not to restore or repair that VM.

Instead, I created a completely new Hyper-V VM and installed OPNsense from scratch.

Important:
- New VM
- Clean OPNsense installation
- Configuration manually recreated
- No old OPNsense configuration backup restored
- Zenarmor freshly installed and configured

The new VM is now operating normally and Zenarmor/netmap also appears to be working correctly.

However, Unbound is crashing again with signal 11.

So the persistent issue that remains after a completely clean rebuild is specifically:

    unbound: exited on signal 11

This makes the Unbound crash particularly interesting, because it has now occurred on both the original installation and a completely new Hyper-V VM without restoring the previous configuration.

Has anyone experienced similar Unbound SIGSEGV crashes with:

- OPNsense 26.7.x
- FreeBSD 15.1
- Microsoft Hyper-V
- hn(4) network interfaces

I would particularly like to determine whether this is a known Unbound/FreeBSD/Hyper-V issue, or whether there is something specific I should collect or enable to obtain a useful Unbound core dump.

At this point I do NOT see the previous Zenarmor/netmap instability on the newly built VM. Zenarmor is currently running normally.

Any suggestions on what Unbound debugging information I should collect next would be very welcome.

Thanks,
Maxim