-
-
Notifications
You must be signed in to change notification settings - Fork 200
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
pihole-FTL stuck at 100% CPU #2154
Comments
Happened again overnight. Same Current backtrace is very similar, just different hostname being resolved:
New tricorder link: https://tricorder.pi-hole.net/urnNSqKO/ |
I prepared a patch hopefully fixing this and already sent it upstream to the
edit The CI is taking longer than usual to build the pre-built FTL binaries, I will update here once it finished edit 2 done |
Thanks! I've deployed this branch. Let's see how it goes. |
I'm afraid it is not fixed. I double checked that I am running the branch with the fix, which I was able to confirm. Had both my instances crash again today in relatively short succession. I had to reboot one of them to restore proper functioning of my network (I probably need to fix my SSH server configs to not break when DNS is down). The only thing happening around this time (within a few minutes of the crash) on the system, that I could see from the logs, is the receipt of a new DHCPv6 lease. I am currently running my DHCPv6 server with a very short lease time of 10 minutes for debugging purposes unrelated to pihole. So this is a frequent occurrence on both systems and is quite likely a red herring. Just thought I'd mention it. Another observation: If I Still getting:
EDIT: If it makes your life easier, I'm happy to also post a screenshot of the colorized gdb output in addition to the raw text version. Just let me know. Unfortunately GitHub markdown doesn't support shell escape colors. |
The condition for this to be triggered is a query whose response during DNSSEC validation is too large to fit into the UDP reply (typically a DNSKEY). I pushed another iteration to the special branch which should fix the root cause rather than trying to band-aid fix the symptoms. Please try again, sorry for the troubles. edit Hash |
Thanks! I have deployed commit |
I think the last fix worked. No runaway threads in 3 days. |
The fix has just been merged into
to continue receiving updates |
Versions
Platform
Expected behavior
pihole-FTL does not hang and responds to DNS queries.
Actual behavior / bug
pihole-FTL hangs after an unspecified amount of time, eventually hangs the entire system.
Steps to reproduce
I don't know what triggers this state, so I cannot reproduce it reliably.
Debug Token
Additional context
htop
sees it like this:gdb backtrace from the stuck thread (pid 19509):
Attaching
strace
to the thread (pid 19509) rapidly spams these two lines over and over:A
kill -9 19509
broughtpihole-FTL
back to life and it operates normally after that.The text was updated successfully, but these errors were encountered: