Because of previous failed install attempts, I had to start with

sudo apt update --fix-missing
sudo dpkg --configure -a

Then the exact steps you did worked for me.


I have just installed ubuntu 22.04 on a pi4, and confirm that stars 3.5.9 stable installs perfectly from the ppa.

Good news! The performance of stars and ekos under Gnome in jammy seems noticeably more snappy than in 20.04. I haven't tried an actual imaging session yet, but this is really an improvement.


Scott Denning replied to the topic 'Ekos crash in stellarsolver' in the forum. 6 months ago

Thank you for your help!

It's repeatable in that Ekos eventually crashes pretty much every imaging session. Not repeatable in that it runs just fine for hours and then crashes at apparently random times when I'm asleep.

I will continue to try to document the crashes. Perhaps this is the wrong place to post the crash reports?

Should I just provide a link to the last successful saved image as the one that "caused it?"


Scott Denning created a new topic ' Ekos crash in stellarsolver' in the forum. 6 months ago

Here's a crash last night in KStars 3.5.9 beta ubuntu 20.04.3 nightly spa

Thread 14 "Thread (pooled)" received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7fff9ffff700 (LWP 6066)]
tcache_get (tc_idx=<optimized out>) at malloc.c:2937
#0 tcache_get (tc_idx=<optimized out>) at malloc.c:2937
#1 __GI___libc_malloc (bytes=60) at malloc.c:3051
#2 0x00007ffff6a5c873 in SEP::addobjdeep(int, SEP::objliststruct*, SEP::objliststruct*, int) () from /lib/x86_64-linux-gnu/
#3 0x00007ffff6a581da in SEP::Deblend::deblend(SEP::objliststruct*, int, SEP::objliststruct*, int, double, int, SEP::Lutz*) () from /lib/x86_64-linux-gnu/
#4 0x00007ffff6a58a36 in SEP::Extract::sortit(SEP::infostruct*, SEP::objliststruct*, int, SEP::objliststruct*, int, double, double) () from /lib/x86_64-linux-gnu/
#5 0x00007ffff6a5af63 in SEP::Extract::sep_extract(SEP::sep_image*, float, int, int, float*, int, int, int, int, double, int, double, SEP::sep_catalog**) () from /lib/x86_64-linux-gnu/
#6 0x00007ffff6a18508 in InternalExtractorSolver::extractPartition(InternalExtractorSolver::ImageParams const&) () from /lib/x86_64-linux-gnu/
#7 0x00007ffff6a216e9 in non-virtual thunk to QtConcurrent::RunFunctionTask<QList<FITSImage::Star> >::run() () from /lib/x86_64-linux-gnu/
#8 0x00007ffff518cf82 in ?? () from /lib/x86_64-linux-gnu/
#9 0x00007ffff51899d2 in ?? () from /lib/x86_64-linux-gnu/
#10 0x00007ffff6b09609 in start_thread (arg=<optimized out>) at pthread_create.c:477
#11 0x00007ffff4a54163 in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/".


In case anybody else wants to do this, the procedure is to install as Jasem said

sudo apt install -y ekos-debugger

But then run by typing


(without the dash). This starts a GUI from which you can launch KStars and INDI, but which will trap crashes and collect crash data.


Thank you!

And then to run inside the debugger, do I just run ekos-debugger from the shell?


I've been having a lot of crashes lately and would like to run kstars/ekos in the debugger so that I can submit crash reports.

I used to do this but it's been years, and I don't remember how to run it this way. I see the git for the debugger but I need a step-by-step on how to actually run this way.

Running KStars 3.5.8 stable on ubuntu 20.04.3 LTS on amd64 hardware.

Many thanks!


In 3.5.7, the focus filter in a sequence automatically changes to the first filter in the sequence.

I can see why this was done, but it's a problem when imaging a target with dim stars in narrowband. 

I have my L filter set as the "Lock" filter so that even if. my imaging is all Ha or Sii I can get enough stars for autofocusing in a reasonable time. The improved sequence logic OVERRIDES this Lock and swaps my filter against my will when focusing at the start of a sequence in the Scheduler.

Now my autofocus fails because too often I'm trying to autofocus with extremely dim stars in narrowband. It's possible to workaround this feature using very long autofocus exposures (like 20 or 30 seconds) to get enough stars, but I would much rather autofocus with my Lock filter.

Please consider rolling this feature back in 3.5.8 or at worst making it an optional checkbox.