Unable to solve the issue b3244

[f4phlyer]

Haven't been able to get any version to work stable since b3235. I've been trying every rendition (including beta's) along the way since b3235. I've always been able to start the version and get it to upload via FTP. I've tried the PHP upload but it appears my host isn't to keen to allow the method. So I've stuck to FTP uploads.

I've considered that it might be a network or hardware issue but out of the 3 Pi's only one has worked reliably. The problem systems are running 64 bit Bullseye. The one working is a Pi 3 with O/S Buster and uses a WeatherFlo Tempest weather station. It's been rock solid.

The combinations that I've tried are two different rPi Compute Module 4's on the rPi CM4 I/O boards and a rPi 4. I'm pretty sure that both of the CM4's at one time or another had thermal issues, I solved the thermal issue with a heatsink/fan that keeps the temp at a solid ~50c. They were set up with other minor apps running on them so rebuilt a bare O/S install. I've tried upgrading from existing versions (b3235), I've tried white sheet installation, I've tried complete fresh O/S installation and scratch start on CumulusMX. Everything past b3235 runs for 3-4 hours but thrn stops in various ways.

I'm using my ISS with Davis Vantage II with iplogger and WLL. All seems pretty much the same. However, once I revert to b3235 it runs fine. I've tried all the advertised methods of upgrade/revert but prefer the "bash" script from Hans (cumulusmx.sh -u).

I've attached the most recent 3244 debug log. Wonder if you might see something or maybe I haven't looked in the right place.

[HansR]

I've tried all the advertised methods of upgrade/revert but prefer the "bash" script from Hans (cumulusmx.sh -u).

1)
I did make an installation program but would that Hans be me?

I think you mean @jank and the CumulusMX Start|Stop - Management Script. As far as I know that script is no longer maintained since CMX can run as a daemon. I would advise to discontinue that script and install CMX as service and starting and stopping CMX like that.

Update of CMX would than take place by simply extracting straight into CMX install directory overwriting all existing files (doing it manually or using my tool above). Make a backup before doing so. Check all files can be overwritten and are overwritten making sure the update/installation is complete.

It may not solve your issue but it takes out a no longer maintained tool.


2)
Is that the logfile after it stops? Because there is no crash, no error, no evidence of stopping whatsoever.
Is there anything on the console when it stops (see: ServiceConsoleLog.txt)?

[f4phlyer]

Hans, thanks for the reply. My bad, yes the update script I’ve been using is @jank’s, and yes I realize that it’s not been updated. I do use the start/stop CMX as a service and manually stop/start before doing the update. I’ve been putting off using the new tool but will get up to speed.

“ Check all files can be overwritten and are overwritten making sure the update/installation is complete.”

I pretty much eliminated the copy issue by doing a “new” install and startup.

“ 2)
Is that the logfile after it stops? Because there is no crash, no error, no evidence of stopping whatsoever.”

Yes it is there has been something, it didn’t have much meaning to me, it was very brief and I didn’t save the report. I’ll recreate the problem and post the report.

[f4phlyer]

Ok so here’s the “status” message on fail.

cumulusmx.service - CumulusMX service
Loaded: loaded (/etc/systemd/system/cumulusmx.service; enabled; vendor preset: enabled)
Active: failed (Result: signal) since Sat 2023-05-27 15:17:17 CDT; 1h 33min ago
Docs: https://cumuluswiki.org/a/Main_Page
Process: 491 ExecStart=/usr/bin/mono-service -d:/home/pi/CumulusMX CumulusMX.exe -service (code=exited, status=0/SUCCESS)
Process: 5180 ExecStopPost=/bin/rm -f /tmp/CumulusMX.exe.lock (code=exited, status=0/SUCCESS)
Main PID: 493 (code=killed, signal=KILL)
CPU: 45min 20.980s

May 27 11:17:05 rPi-Time-Machine systemd[1]: Starting CumulusMX service...
May 27 11:17:05 rPi-Time-Machine systemd[1]: Started CumulusMX service.
May 27 15:17:17 rPi-Time-Machine systemd[1]: cumulusmx.service: Main process exited, code=killed, status=9/KILL
May 27 15:17:17 rPi-Time-Machine systemd[1]: cumulusmx.service: Failed with result 'signal'.
May 27 15:17:17 rPi-Time-Machine systemd[1]: cumulusmx.service: Consumed 45min 20.980s CPU time.

[HansR]

OK. So the only thing you have is the status when you call systemctl, nothing in the MXdiags and nothing in the ServiceConsoleLog.txt. That means there is not a regular stop message in the logs (it just quit in the middle of something) and indicates most likely there is some (system) library missing or outdated.

On your site I see your system-info has nothing so could you tell which mono version you are running (6.12.0.182 required) and if you are using the setting to email on alarms?

Even though you tried it already several times, you need to re-install again afaics, make sure mono is the correct version and please do: apt update and apt full-upgrade.

If you are not successful I think somebody else has to look.

[f4phlyer]

Yes, things have been a work in progress, working through issues has forced some of the issues to the surface.

Trying to install mono I couldn’t find how to install 6.12.0.182 on “Bullseye” so I’ve adapted the “Debian 10” and am waiting.

One of the reasons I’ve waited so long to post is that I couldn’t see anything obvious, and I was hoping that someone else might see what I had missed.

We’ll see. Thanks again.

[HansR]

OK. Raspbian 10 in this recipe should be OK.

[f4phlyer]

I was suspect of that as the hardware designator is ‘armhf’, the Debian includes ‘arm64’. But I guess the kicker is that both are mono-devel, I had mono-complete prior.

We’ll see.

Cheers

[HansR]

I was suspect of that as the hardware designator is ‘armhf’, the Debian includes ‘arm64’. But I guess the kicker is that both are mono-devel, I had mono-complete prior.

We’ll see.

Cheers

mono-complete is what you need, that should do it. Just change that in the command line!
Remove mono first if necessary.