Closed rhakbari closed 3 years ago
Can you please check the log file /var/log/pihole-FTL.log
for a crash report?
Look for lines with !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
I'm getting crashes after 5-10 minutes.
Latest log lines of FTL, tons of sql errors here.
Oh, you're still using an old version of FTL, this makes sense because the underling database structure has evolved since back in the days.
[2020-12-01 17:20:56.630 3439M] FTL branch: release/v5.1.1
Please run
pihole checkout master
to get back on track with the most recent versions of Pi-hole
@DL6ER Sorry, you're right, that's after I tried to downgrade pihole - the newest version still crashes for me after 5 minutes.
With the same SQL errors?
On latest master, tailing the log file:
Had to also restart FTL couple of times:
This does not yet contain anything pointing us to the source of the problem. Is there anything interesting in /var/log/pihole.log
?
This does not yet contain anything pointing us to the source of the problem. Is there anything interesting in
/var/log/pihole.log
?
I was looking, but no obvious crashing logs. DNS queries just completely fail until I successfully run pihole restartdns
.
I'm trying to look at any logs in /var/log/pihole.log, but nothing interesting really.
What did you do in between these messages:
[2020-12-01 17:31:36.998 5151M] Reloading DNS cache
[2020-12-01 17:31:36.998 5151M] Blocking status is enabled
[2020-12-01 17:31:37.869 5151/T5155] Compiled 2 whitelist and 2 blacklist regex filters for 11 clients in 23.8 msec
<----->
[2020-12-01 17:31:55.624 5151M] Reloading DNS cache
[2020-12-01 17:31:55.625 5151M] Blocking status is enabled
I suppose you ran pihole -g
.
Is there anything you can do to reliably trigger this or is Pi-hole just getting unresponsive after having worked correctly for some time?
Is there anything you can do to reliably trigger this or is Pi-hole just getting unresponsive after having worked correctly for some time?
It turned unresponsive and websites wouldn't load because dns wasn't resolved.
Currently it's running fine for the last 15 minutes (of course it is after I comment on a github issue :P), will post here again if it crashes.
Okay, let's see. Point is that I haven't seen anything really pointing towards a crash. Maybe it is rather like a deadlock scenario, however, I haven't seen one in a very long time (this does not mean there are none left, maybe you just hit the super-edgecase).
DNS queries failed just now, there's one FATAL log line but upon googling it seems to be fixed in https://github.com/pi-hole/FTL/pull/671
Here's the tail of /var/log/pihole-FTL.log
and here's the pihole debug log (I did it while DNS queries weren't working):
Interrupted system call (-1)
is something new. #671 did not fix this. Reading on this error in the context of shared memory suggests that this has to do with running out of inter-process memory. However, this doesn't seem to be the case looking at
/dev/shm: 4.5MB used, 483.2MB total
However, looking at the rest of your log, we see this a couple of times in different places:
[2020-12-03 14:10:01.537 5694/T5695] IPv4 telnet error: Interrupted system call (4)
or
[2020-12-03 11:36:04.338 5694/T5695] IPv4 telnet error: Interrupted system call (4)
and
[2020-12-03 06:22:01.998 5694/T5695] IPv4 telnet error: Interrupted system call (4)
Interrupted system call
means something went wrong in the lower-level subroutines of your operating system. Did you do anything fancy? How are your disks mounted? Do you have some remote filesystems attached to your system?
Such system errors are typically hard to debug, however, we will surely try our best!
I realise I've rather hijacked OP's issue :P
In terms of the interrupted system call messages, I did do a Raspian upgrade a month ago, but pihole was running alright before updating to the latest version a couple of days ago. No fancy disk mounting.
I also saw some logs regarding cloudflared.service stopped working in journalctl -xe
.
As my pi only runs a couple of things, I basically reimaged my SD card to the latest Raspian version, and reinstalled pihole on it. Hasn't crashed since, but will keep an eye out for it.
Versions
Platform
Actual behavior / bug
The pihole FTL service crashes and then the webpage says unable to connect to API and then i have to restart the DNS resolver in order to fix it temporarily. This has start doing this after this latest update before that it was working alright. A clear and concise description of what the bug is.
Debug Token