Armbian Bookworm + Pi-hole - Freezing after a few hours uptime

Please follow the below template, it will help us to help you!

Expected Behaviour:

Tinkerboard S - Fresh install Armbian Bookworm and Pi-hole earlier this week.
Also have Unbound & OpenVPN
DHCP on Router with DNS Options as Pihole

Please let me know if you need any other information

Actual Behaviour:

System is freezing within a couple hours of uptime. Ping responds but can't view HTTP or even SSH. No major events happening when it freezes.

Debug Token:

https://tricorder.pi-hole.net/8fr3gooQ/

Custom DNS 127.0.0.1#5335
Permit All Origins (Because of OpenVPN on tun0)

Using:
Never forward non-FQDN queries
Use DNSSEC
Conditional Forwarding

Not using:
Never forward reverse lookups for private IP ranges

https://tricorder.pi-hole.net/Ex7zjHEO/

It froze again around 6am this morning

https://tricorder.pi-hole.net/PftNIOpB/

Still having daily freezes. Any insight would be greatly appreciated!

Are you trying access the admin page via the ip address? If not does that work? Same for ssh. If those services don't work using the Pis IP Address it would suggest some network / system issue.

This might be a power issue. Have you tried a different power suppy or tried to power it off the gpio pins? I saw a few posts referencing power issues but that may or may not be related to this.

Thanks for the reply.

Yeah, as of lately I am just keeping a constant ping. When it stops responding then HTTP & SSH are unreachable as well. I think HTTP goes down first because I have caught it a few times lagging on refresh but ping and SSH are functional. Within the past week I have added a USB volt meter. The amps don't change when it is fully functional and when it freezes.

It's like the system is freezing, its a blank screen, can't break or login.

You said: "No major events happening when it freezes."

Your log doesn't show any obvious errors or crashes. Pi-hole seems to be working as expected.

If you are unable to access the machine via SSH using an IP, it means the issue is on the network level or OS level, as CallMeCurious suggested.

Tinkerboards forum may have some ideas. I'll just toss out some ideas / thoughts.

It would be hard to catch a change in voltage / amp. All it takes is a momentary drop in voltage to cause a problem and the issue could be in the connector / pd circuit. I would still suggest trying a different power supply / different power supply wired to gpio.

Is there a monitor / keyboard hooked up? if not and you have one to use that might be useful. You could potentially use your tv as a monitor for trouble shooting ( if its near enough ).

You can check dmesg / journalctl for any errors. If you have a monitor setup you can also tail / follow these and see if something pops up. You can also do this through a ssh session. Even if the session fails you may still see something.

Switch interfaces and see if the problem persists. I'm guessing your using the ethernet currently. Try setting up the wifi interface. Make sure to set it up as static and update your router accordingly.

If your running the OS on the eMMC you could try and run it on the SD card slot or visa versa.

I have a plethora of chargers. Currently I am using a 60w PD charger which is overkill but the issue still exists between different power supplies.

I have had it connected to an extra monitor since this trouble started. It is just at a black screen and keyboard non-responsive when I catch it offline.

I have just removed Unbound 127.0.0.1#5335 and just set 1.1.1.1 as the upstream as a test. I am just realizing that this didn't seem to be an issue before I installed Unbound.

It is just strange, I have had the same setup running for years across 2 identical Tinker Boards, I decide to do a clean install and the issue is present on both.

So is the OS running on a SD card or the eMMC? If its the SD card, when you say it happens on both Tinkerboards are you swapping the SD between them? Perhaps the SD is starting to fail?

You mentioned the clean install to Armbian Bookworm. Was this an upgrade from the last OS. I'm guessing so. This could be related to that.

I found a link to something that sounds like a similiar issue ( kernal panics ). It may or maynot be relevant but could give you ideas.

Also, if one of those tinker boards is spare you could try to put the most recent OS on without any other apps and see if continues to lock up or you could also put an older version of Armbian on one and see if it persists.

It is the built in eMMC not an SD card. I was running as pi-hole01 and pi-hole02 with different IP on each. Both are freezing after the Bookworm install at intermittent times.

I was on Buster originally and dragged my feet on the distribution upgrade then they removed the repo for Buster to reach APT. At that point I just decided it would be best to completely wipe both of them and reload fresh with Bookworm. I ran teleport then flashed Bookworm and proceeded with setup.

Thanks for that article and all of your insight. When I get home I will start playing around again.

Thanks again. I was thinking this was a software error because it was a fresh OS install with only Pi-hole.

I downgraded my kernel and it has been running for 2 days now without locking up.

I am running 6.5.13 now instead of 6.6.22. I will have to periodically test 6.6 now on newer releases.

1 Like

You may find the atop package useful for this. It's like a recording version of top or htop. After something has happened you can look back at the loads, resources and processes around that time. I wrote a post with a couple of common commands and a link to a detailed guide.

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.