23.7 upgrade from 23.1 results in DNS issues

[BasilBasil]

I've been running opnsense 23.1 flawlessly I think pretty much from release (I have had the odd issue with opnsense prior to 23.1, but effectively 99.9% uptime for many years since switching from pfsense).

Yesterday I updated to 23.7.3 and immediately starting having issues.

First off was that I was not getting any DNS responses (I could connect via IPs OK) - for some reason "Enable Unbound" was not ticked anymore. Ticked, now getting DNS responses once unbound started.

HOWEVER, how my custom domain does no longer work.

For example, if I had server1.customdomain it would originally have responded with the IP address.

Now I get:

Server:      10.10.1.1
Address:   10.10.1.1#53

** server can't find server1.customdomain: NXDOMAIN


server1.local DOES however work.

All my overrides still work too.



Thankfully as I run this as a VM I rebooted into my 23.1 instance - everything working again.


What has changed in 23.7 that A) meant that unbound DNS did not automatically start because it had been unchecked to do so and B) Why is my custom domain stuff working?

In the mean time I've reverted back to 23.1, as a number of things rely on the device.customdomain tag and thus dont work properly...!

I've had a fiddle, but at the end of the day this is a home setup so I've run out of debug time and skills!

Many Thanks!

[zandrr]

I for one actually had the same experience post-upgrade and also rolled back without too much investigative analysis. VM as well.
Don't really have anything to add sorry, just mirroring your experience. This was back in 23.7.0 though, so first release.

Will keep an eye on this thread for insights. I'm in no rush to upgrade home again, but would like to. Just waiting patiently,

[newsense]

Just waiting patiently

There are currently no open issues regarding DNS post 23.7.3

[BasilBasil]

Just waiting patiently

There are currently no open issues regarding DNS post 23.7.3

This may be the case, but the upgrade for me has caused two different issues, both DNS related.

Some guidance as to why the domain stuff isn’t working (or what to look for) may yield clues, but for now I’m sticking with 23.1.

Edit: Switched between 23.7.3 and 23.1 a few times. Black and white working/not working for me.

23.1.11 device.customdomain works
23.7.3 device.customdomain does NOT work

So there is clearly something between 23.1.11 and 23.7.3 that is causing this, if someone can provide some guidance as to where the problem might lie then I can do some more digging.

[CJ]

I believe a few people had trouble where unbound got disabled during the upgrade but I don't think it was a common thing or that it was ever determined why.

How do you have your custom domain configured and what do the Unbound logs show?

[BasilBasil]

Seems like a pretty significant bug if multiple users have seen it happen 

As for custom domain, I think the only place it is set is here:

SYSTEM: SETTINGS: GENERAL

Domain "customdomain"

I'll have to boot back into 23.7 to get some unbound logs.

[franco]

It's probably fixable under 30 minutes with the right error message attached.

If you have the 23.7 where Unbound is still disabled right after upgrade you can try running:

# /usr/local/opnsense/mvc/script/run_migrations.php
# opnsense-log | grep run_migrations


Cheers,
Franco

[Sekkis]

I had issues starting Unbound after updating to 23.7. The logs gave no clues, but I stumbled on to a post (can't find it ATM) where they discussed a bug when running Unbound on "not all" interfaces, as I was.

The fix was to:
  Go to Services -> Unbound DNS -> General
  Choose to Clear All beneath the Network Interfaces drop down, Apply
  Choose your preffered interface(s) again, and reapply.

Worked for me, hope this helps.

[franco]

Yeah this one was fixed in 23.7.1:

https://github.com/opnsense/changelog/blob/01889aa7eb9c4e75c3aff6dde9abeca18f16ea55/community/23.7/23.7.1#L28


Cheers,
Franco

[BasilBasil]

I’ll spend a bit of time messing about with this this weekend to get answers.

Couldn’t turn the internet off yet, I’d get an unhappy other half. 

[rshakin]

Same issue since updating cant acess any servers on a local hostname... any fixes or should i just to back to old version.

[mic]

@Franco

Hi,

I have the same issue, after upgrade to 23.7.3 Unboud was disabled so I enable it but all my overrides dosn't work.
I run the following command:

Code: [Select]

/usr/local/opnsense/mvc/script/run_migrations.php

** OPNsense\Unbound\Unbound Migration failed, check log for details

and then

Code: [Select]

opnsense-log | grep run_migrations

<147>1 2023-09-13T08:23:04+02:00 localhost config 2076 - [meta sequenceId="29"] #1 /usr/local/opnsense/mvc/script/run_migrations.php(54): OPNsense\Base\BaseModel->runMigrations()
<147>1 2023-09-13T08:27:54+02:00 opnsense-casa.proximanet.net config 80369 - [meta sequenceId="8"] #1 /usr/local/opnsense/mvc/script/run_migrations.php(54): OPNsense\Base\BaseModel->runMigrations()
<147>1 2023-09-13T09:56:22+02:00 opnsense-casa.proximanet.net config 50858 - [meta sequenceId="6"] #1 /usr/local/opnsense/mvc/script/run_migrations.php(54): OPNsense\Base\BaseModel->runMigrations()

Then I tried also to enable the access list and insert my subnets, but it still doesn't work...

Can you help me, please?

Thank you

[franco]

Can you give the full log and the config.xml section of your overrides? That would help reproduce this quickly. You can also PM me or better yet send to franco@opnsense.org


Cheers,
Franco

[mic]

Hi Franco,

I sent you the unbound section of config.xml via email.

Thak you

[franco]

With mic's help we now have a POC:

https://github.com/opnsense/core/pull/6844

I don't want to advertise it too much though. It needs to be discussed internally first and not everyone is at the office at the moment.


Cheers,
Franco