[Pi-hole V6] Ftl crash because of database?

[Gontier-Julien]

[Pi-hole V6] Ftl crash because of database?

[Gontier-Julien]

Also the web interface is still up when it happen, so it need to be manually restarted when it happens

[DL6ER]

Can this be triggered reliably somehow? As in: Does the crash repeatedly happen after the container restart?

If so, it'd be tremendously helpful is you could follow https://docs.pi-hole.net/ftldns/debugging/ and attach the gdb debugger to pihole-FTL so a backtrace can reveal where the crash is happening. The fact that you are running Pi-hole inside a container may make this more tricky but should still be straightforward enough to not cause too much additional trouble.

[Gontier-Julien]

Can this be triggered reliably somehow? As in: Does the crash repeatedly happen after the container restart?

If so, it'd be tremendously helpful is you could follow https://docs.pi-hole.net/ftldns/debugging/ and attach the gdb debugger to pihole-FTL so a backtrace can reveal where the crash is happening. The fact that you are running Pi-hole inside a container may make this more tricky but should still be straightforward enough to not cause too much additional trouble.

And it doesn't happens only when the contain restart, it happens even if it was the first time spinning it up. I need to restart the container manually since it doesn't stop by itself to restart.

I've already tried to debug it a little, and try to crash it reliably but with no success unfortunately :(
It really happens at random, i tried to see if a domain or multiple domain crash it nothing, see if it a specific client maybe nothing.

The only thing i could thing of maybe it could be related to the absurdly high client number?

But i really doubt it? From what i've understand in #1884 ?
It show the same as in the PR from which client it is.

If you have any idea that i could test or anything tell me i'll try it

And i'll see for the debugging to trace it down tomorrow as i'm going to sleep now ^^

[Gontier-Julien]

Should I try to delete all database to see if it fix it?

[rdwebdesign]

Can you please generate a debug log and post here only the Token?

Using the web interface:

• click on Tools > Generate Debug Log menu.
• then mark the second checkbox to upload the log and click on the big button.
• when finished, copy only the token and paste here.

[Gontier-Julien]

I don't see the option on pi-hole V6?
But i've generated a pi-hole debug log that i can upload

I'll think to do on again if it crash again.

[rdwebdesign]

Sorry... It was late and I used a saved reply with v5 text.

You can generate the log running pihole -d inside the container:

docker exec -it <container_name> pihole -d

When asked, answer "Y" to upload the log and then copy and paste only the Token.

[Gontier-Julien]

Sorry... It was late and I used a saved reply with v5 text.

You can generate the log running pihole -d inside the container:

docker exec -it <container_name> pihole -d

When asked, answer "Y" to upload the log and then copy and paste only the Token.

No worry's ^^

Here the token then ^^

B6ouWIHF

[Gontier-Julien]

If i see no crash until the end of the month i'll close that issue then, but so far i didn't encounter it again

[Gontier-Julien]

Closing this as it never happen since then 🤷