×

INDI Library v2.0.6 is Released (02 Feb 2024)

Bi-monthly release with minor bug fixes and improvements

KStars crash during focusing

  • Posts: 50
  • Thank you received: 6
This is the latest version from PPA. Similar crashes have been happening to me over several last months. Though I cannot promise it is the same spot.
On some nights it is going solidly. On others, with the same config, it's tripping multiple times. Last night it crashed after 3 hours of running. I restarted it and it crashed again in 2 hours.
INDI server and drivers keep running after crash, it's only KStars itself gets a segfault.

See the log, the focusing is just started. It is likely starting an exposure or receiving the image.

File Attachment:

File Name: log_00-51-28.txt.gz
File Size:58 KB


Here is the stack from coredump. I can't see more details despite the kstars-bleeding-dbg package is installed.
#0  __memcpy_ssse3 () at ../sysdeps/x86_64/multiarch/memcpy-ssse3.S:131
#1  0x00005644b84ca49f in INDI::BaseDevice::setBLOB(_IBLOBVectorProperty*, xml_ele_*, char*) ()
#2  0x00005644b84caf37 in INDI::BaseDevice::setValue(xml_ele_*, char*) ()
#3  0x00005644b84ceed7 in INDI::BaseClientPrivate::dispatchCommand(xml_ele_*, char*) ()
#4  0x00005644b84cfe4c in INDI::BaseClientPrivate::listenINDI() ()
#5  0x00005644b84d01c9 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<std::_Bind<void (INDI::BaseClientPrivate::*(INDI::BaseClientPrivate*))()> > > >::_M_run() ()
#6  0x00007fd6742f52b3 in std::execute_native_thread_routine (__p=0x5644bf908640) at ../../../../../src/libstdc++-v3/src/c++11/thread.cc:82
#7  0x00007fd673f7cb43 in start_thread (arg=<optimized out>) at ./nptl/pthread_create.c:442
#8  0x00007fd67400ea00 in clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:81

I have a guess from my overall experience. I focus at 2x2 binning (and align at 4x4) to speed up things on my slow computer. Main capture is with no binning. So, Ekos has to switch binning back and forth. Occasionally I see an image with wrong resolution in some module. Like CCD get binned image or Focus gets an unbinned one. Furthermore, sometimes focus images get into main Ekos preview or even trigger a FITS viewer pop up, even though that is disabled. So, there may be some races that are mostly benign, unless it is around changing binning...

I should perhaps enable more logs for capture and focus next night.
1 year 6 months ago #86527
Attachments:

Please Log in or Create an account to join the conversation.

  • Posts: 593
  • Thank you received: 274

Replied by John on topic KStars crash during focusing

There's nothing obvious that I can see in the log. You have focus, scheduler, capture and guiding all active at around the time. These segfaults are hard to diagnose unless you can catch where it is in debug at the time. I run focus at bin 2 but usually capture at bin 1 so if this is related to the problem it must be an edge case. Still, if you think its related you could try focus at bin 1 and see if you still get the issue.

Worth putting verbose mode on in the logging for the active modules to see if you can capture any more clues, and consider running kstars under some form of debugger (if you are able to).
1 year 6 months ago #86535

Please Log in or Create an account to join the conversation.

  • Posts: 50
  • Thank you received: 6
Here is more detailed log from last night. KStars made it almost to the end but still crashed at 5 AM :)

File Attachment:

File Name: log_19-10-35.txt.gz
File Size:971 KB


The stack trace is exactly the same. I can share core dumps if anybody wants to look. Again, when I run GDB on those dumps, no helpful stuff (local variables, links to source files, etc.) appears despite the KStars debug package is installed. Do I miss some important step?
My imaging computer has only 2 GB of RAM. KStars itself takes over 1 GB. I tried running it under Ekos Debugger. GDB takes another 1.5 GB and things get really tough, though it doesn't die right away.

This time it again happened when Focus just started. Another interesting fact is that guiding is still running (Focus has been only doing a test exposure (why do I have it enabled?..)) Apparently, two cameras finished their exposures really close to each other and here the story ends.
My both cameras are ZWO now. Until about a week ago I used SX Lodestar as a guide camera and got similar crashes. I only have terse logs from the past. Here is one.

File Attachment:

File Name: log_19-11-18.txt.gz
File Size:14 KB

It ends during normal capture cycle. Guiding exposure was 1s. I do see a good chance that main and guide cameras finished at the same time.
1 year 6 months ago #86543
Attachments:

Please Log in or Create an account to join the conversation.

Using Ekos Debugger would be a lot more helpful for crashes.
1 year 6 months ago #86580

Please Log in or Create an account to join the conversation.

Time to create page: 0.395 seconds