Folks,
I have the following hardware:

• RaspberryPi 4 with Stellarmate OS latest stable (as of 2/1).
• KStars/Ekos running locally via VNC
• Waveshare MotorHat Focuser
• QHY183C Main Camera (via USB3 Port)
• QHY5-II Color guide camera (via USB2 Port)
• Losmandy Gemini 2 (Direct Connect Ethernet)
• The Pi and MotorHat are powered from a 50AH LifePO4 battery via the power port on the MotorHat.

From prior experience with my Celestron Mount, I would normally align the mount and then perform the Polar Alignment. When I try to run the "Capture and Solve" after slewing to my target, Kstars would crash. I restart KStars and Ekos and It says that there is an Indi Server already running, do I wish to stop it and start a new one.  If I respond yes, it restarts.  If I respond no, I get a dialog box indicating that the Waveshare MotorHat driver has crashed and do I wish to restart.  Clicking Yes causes the dialog to go away and come right back up.  Clicking no, results in Ekos starting without that driver. 
Questions:

1. Is the fact that I am trying to align before polar aligning causing the issue?
2. Is there a reason why the align module would crash the Waveshare driver, or is that just a red herring?
3. How do I use the Ekos Debugger to provide logs so I can figure out what is actually causing the crash?

Thanks,
Anand

The waveshare crashing is probably a separate issue. Usually I only ever see it crash if I've updated indilib and have not recompiled for it.
You should skip the Celestron mount alignment until after polar aligning because it will ruin the model.
I've personally encountered alignment crashes if my telescope configuration is not saved with a name. So go to the mount tab, enter your focal lengths, give the configuration a name and save it. It will be written to the INDI driver, so save it there too. Maybe that will help you too.

Not having a configuration name should cause a crash. For the Waveshare, you're right, it could be because it was not compiled for the current version of INDI. If we can get the driver in the official repo, I think this would be less of a problem.

You can use Ekos Debugger on the desktop and run KStars and if it ends up crashing for whatever reason, then you can submit the relevant logs.

Jasem,
Yesterday, I recompiled the Waveshare driver and added a name for the telescope in the mount tab. I skipped alignment and went straight into Polar alignment and it did crash once. I've only just seen the link in your signature for submitting logs, so I'll wait until I have a clear night and try again with to capture logs. I'll post here as soon as I have those.

Anand

For Alignment crash, use Ekos Debugger for this purpose. From previous reports, it crashes on the 3rd image when it tries to find WCS Coordinates. Not sure if this is still the case.

This makes sense. I noticed that the messages were saying that WCS was not set. At one point, I ran capture and solve with the "do nothing" option and it set it. I can't be sure, but I don't think it crashed after that. I will make an effort to be more deliberate with the testing to reproduce the crash.

Jasem,
I've attached the logs for the crashes last night.  It's definitely related to the solver and I am certain it goes away when I run a capture and solve with the "sync" option selected. I do have the core dumps if that helps.

Anand

 

Thanks, I'll talk to Robert about this. I think it *might* be related as well to the fact that logging is on. But we'll double check.

I took a quick look at the log, and that is not the verbose astrometry logging and so the astrometry logging should be no problem.  The issue he was referring to should be fixed now, but there used to be an issue where turning on verbose astrometry logging would really bog down the system because it was so much stuff.  I have worked out a way so that it shouldn't bog it down any more, so if you have the latest stellarsolver it shouldn't be an issue anymore, but you didn't have verbose astrometry logging on, so it shouldn't be an issue anyway.

Also note that there was another unrelated bug related to logging I was trying to track down where KStars would crash sometimes (but not always) if CCD logging was turned on.  Did you have CCD logging on?

Another thing I do see here is that you are loading all your indexes in parallel and you are also solving in multiple threads.  I don't suspect the threads will be an issue, but I suspect in parallel option might be in this case.  It looks like your Pi has 8 GB of ram is the correct?  Are you running armhf or arm64?  If it is armhf, you might have an issue with that.  Your computer can give no more than 4 GB of ram to a single process if it is armhf.  If you are loading all the indexes in parallel, and you have a decent amount of indexes, that could quickly run out of memory.  It doesn't look like you have that much in the way of index files, as I see here it says 0.6 GB, so maybe this is not an issue for you at all, but Astrometry.net is extremely sensitive to the in parallel option, so maybe this could be an issue.  It does make the solve much much faster if it is turned on, but if the system can't handle it, it can cause a crash.  I do have a check in there to turn off that option if you don't have enough RAM for it, but it probably thinks you have 8 GB when you really can only use 4 GB.  And there is probably ram taken up by other things too.  You could try turning off the "in parallel" option in the options profile of your choice.  Just a thought of something you can try.



 

Hi Rob,
I have a Pi4 with 8GB of Ram, but it does appear that its armhf archictecture

CCD Logging was off for these tests, so I don't think its related.  I will disable the parallel option and see if that corrects the issue.

Anand
 

Anand, I believe you have StellarMate correct? why don't you use the 64bit 1.7.0 version since you have 8GB of RAM?

Hi Jasem
I didn't know there were different versions. I guess I just followed the download link in the email. How can I download the image for it? When I go to the downloads page on stellarmate.com, it takes me to the orders page.

Anand