I'm not sure if this is related to the OPNsense 26.7.x update, but it seems to have started then - After an OPNsense reboot, Unbound starts and runs successfully for about a minute. Then it crashes and I have to manually restart the service to bring it back up. What could be causing this? Is this a 26.7.x bug?
https://forum.opnsense.org/index.php?topic=52429.0
What does Unbound log?
Quote from: dseven on August 28, 2026, 05:07:38 PMhttps://forum.opnsense.org/index.php?topic=52429.0
Thank you for sharing this. I actually forgot that I saw that already and chimed in!
Quote from: Patrick M. Hausen on August 28, 2026, 05:08:06 PMWhat does Unbound log?
This is what ChatGPT helped me come up with. This was in order to submit an issue, which I have not done yet. I figured I would try the forum first.
QuoteRelevant log output
During boot, multiple interface events invoke unbound_configure_do():
2026-08-28T10:48:41-04:00
/usr/local/etc/rc.linkup:
plugins_configure dns (execute task : unbound_configure_do())
2026-08-28T10:49:15-04:00
/usr/local/etc/rc.linkup:
plugins_configure dns (execute task : unbound_configure_do())
2026-08-28T10:49:28-04:00
/usr/local/etc/rc.linkup:
plugins_configure dns (execute task : unbound_configure_do())
2026-08-28T10:49:36-04:00
/usr/local/etc/rc.linkup:
plugins_configure dns (execute task : unbound_configure_do())
At approximately the same time, the Unbound log shows:
2026-08-28T10:49:35-04:00
info: start of service (unbound 1.26.0)
2026-08-28T10:49:36-04:00
info: service stopped (unbound 1.26.0)
2026-08-28T10:49:40-04:00
info: start of service (unbound 1.26.0)
The interface events immediately preceding this include:
2026-08-28T10:48:41-04:00
kernel
vlan0.30: link state changed to UP
2026-08-28T10:48:41-04:00
kernel
vlan0.40: link state changed to UP
2026-08-28T10:48:41-04:00
kernel
vlan0.10: link state changed to UP
2026-08-28T10:48:41-04:00
kernel
vlan0.20: link state changed to UP
There are also repeated IPv6 initialization messages during boot:
dhcp6c: Sending Solicit on igc0
and:
radvd_configure_do(auto) found no suitable IPv6 address on lan(ixl1)
radvd: sendmsg: Can't assign requested address
At present, the WAN and LAN interfaces have only IPv6 link-local addresses and no global/delegated IPv6 prefix.
There are also repeated kernel messages:
pfr_update_stats: assertion failed.
I do not know whether these are related.
Im also running into this issue on one of my routers. Not using any DNSBL.
I happened to look at the statistics and it showed a recursion time of something like 8.8415e-4 which I thought was very strange.
I restarted it and its fine now. It seems to run for at least a couple weeks with no other issues once it is restarted right after the router boots.
So do you see that same message
(unbound), jid xx, uid xx, was killed: failed to reclaim memory
Also, _Mike, are you using blocklists?
Quote from: Tinus on August 31, 2026, 04:34:28 PMSo do you see that same message
(unbound), jid xx, uid xx, was killed: failed to reclaim memory
Also, _Mike, are you using blocklists?
I'm not using blocklists in Unbound. My DNS is AdGuard Home, so my blocklists are there. Unbound is my upstream DNS. I'm not seeing the memory error, so perhaps my issue is different.