Allsky not starting automatically or as a service

[brianboru82]

I'm seeing an issue where the allsky service won't run. It says the process is dead, and it doesn't respond when I try systemctl start allsky.

I can start it manually by using ./allsky.sh, but it appears when doing it that way, nothing is recorded in the allsky log.

Trying to generate the support files, I'll upload when I can.

[brianboru82]

Thanks, here's the support files.
support-4591-20250527065049.zip

[EricClaeys]

@brianboru82,
Thanks for the file. FYI, during installation this message appeared:
Latitude (45.304) should contain EITHER a '+' or '-', OR a 'N' or 'S'
It looks like you fixed this?

I can start it manually by using ./allsky.sh, but it appears when doing it that way, nothing is recorded in the allsky log.

Correct. Allsky writes to stdout (standard output) so when you run it from the command line the output goes to your terminal. When run via systemctl the output is redirected to /var/log/allsky.log.
Try this when manually starting Allsky:

nohup ./allsky.sh > /var/log/allsky.log 2>&1 &

That will tell Linux to continue running even after you "hang up" (nohup) as in hanging up a 20-year old modem. Normally when you hang up or exit the window, any process running in the window is killed.
This will also redirect any output and error output to /var/log/allsky.log.
And the process will be put in the background (by adding & to the end of the line) so you can continue using that window.

I didn't find anything explaining why Allsky doesn't start as a service. Someone else had this problem recently so I'll check prior Issues and Discussions. I think we determined it was corruption in some Linux file. In the meantime, please do NOT reinstall Linux or make any major changes that would eliminate our chances of finding (and fixing) the problem.

[brianboru82]

Yes, the latitude thing was corrected - that seems to come back every time I upgrade to a new version, so I'm wondering if the + in front isn't saving in the file for the upgrade script to read, or it's something else. Haven't seen anyone else complain about this behavior, and it's easy enough to re-enter it.

[EricClaeys]

The other user with this problem had a bad SD card so it was corrupting files. He reinstalled Linux on a new card and the problem went away.

Please run this when Allsky is NOT running:

sudo strace --follow-forks --decode-fds=path --output /tmp/strace.txt systemctl start allsky

Give it 20 seconds. If it hasn't finished (and I expect it won't), kill it then attah /tmp/strace.txt.

Thanks - Eric

[brianboru82]

Thanks, I'll try it tonight.

[brianboru82]

Here's the trace file.
strace.txt

[EricClaeys]

@brianboru82,
Thanks for the trace file. It looked very similar to the output I got when I ran strace, but mine successfully stopped some process and yours didn't. I don't know enough about what's happening to understand the difference.

I found some stuff to try:

ls -ld / /usr/lib /usr/lib/systemd /usr/lib/systemd/system

If any of them do NOT have root root run

sudo chown root:root / /usr/lib /usr/lib/systemd /usr/lib/systemd/system
sudo systemctl start allsky   ## kill it if it hangs

If it still hung, run:

sudo truncate -s 0 /var/log/syslog
sudo systemctl log-level debug
sudo systemctl start allsky    # kill it when it hangs
cp /var/log/syslog /tmp/syslog.txt
sudo systemctl log-level info

Then upload /tmp/syslog.txt

[brianboru82]

Here's the status from the first command, it looks like it passed, so I didn't do the chown command.

Ran the debug log, it's attached here.
syslog.txt

[brianboru82]

I ran the systemctl status allsky again after this, and it had some new info, not sure if it's helpful.

[brianboru82]

As an update, I've updated my allsky to the latest version, and this issue is still present. Which I expected, because it seems like this is a system level issue, not an allsky issue. Still, it would be nice to know the solution here other than doing a fresh raspbian install.

[brianboru82]

I think I found some useful info from here:
https://www.suse.com/support/kb/doc/?id=000020261

I ran systmctl list-jobs on my pi, and it shows this:

The only running thing is the wayvnc service. I then ran systemctl status wayvnc, and see this:

I am running RealVNC on my RPi rather than WayVNC that came with the latest RPi OS as WayVNC wasn't playing nicely with my RealVNC client on my laptop. I'm wondering if WayVNC is still trying to bind to the Pi IP address that RealVNC is using, and can't so it's just hung there.

I'll try disabling WayVNC and see what we get.

Update: Yes, that was it. I disabled WayVNC and the allsky service was able to start.

So I would suggest for others in the future if the allsky service is not running, to run systemctl list-jobs and see what is currently running, and troubleshoot that.

Thanks!