×

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

Bi-monthly release with minor bug fixes and improvements

Solver crashing Kstars

  • Posts: 388
  • Thank you received: 17
I am running Kstars 3.5.0 Stable (build 2020-12-06T18:17:19Z) on an Rpi 4 with 8 Gb memory. I have logging turned on with a separate log for astrometry, but it has stopped writing any astrometry logs. I have renamed the log file to include the date and set it for _verb, but a few days ago it stopped writing these logs. I get the regular logs and all they show is that everything just stops when the internal StellarSolver starts. It does not do this all the time, some nights it works fine (but still no Astrometry log). Once Kstars crashes, restarting it does not help. I have to reboot the Rpi and even then it may crash unless I set the solver option to only use one thread in solving.

Tonight was an example of this 'unusual' behavior. Everything worked fine thru polar alignment, slewing to a target, imaging and slewing to another target. After about 2 hours of imaging I began a move to another target, the log ended like this:
" 2021-01-18T21:32:07.210 EST INFO ][ org.kde.kstars.ekos.align] - "Image received."
[2021-01-18T21:32:07.214 EST INFO ][ org.kde.kstars.ekos.align] - "Automatically downsampling the image by 1"
[2021-01-18T21:32:07.342 EST INFO ][ org.kde.kstars.ekos.align] - "Evaluating Installed RAM for inParallel Option. Total Size of Index files: 2.74134 GB, Installed RAM: 7.64054 GB, Free RAM: 5.27455 GB"
[2021-01-18T21:32:07.346 EST INFO ][ org.kde.kstars.ekos.align] - "There should be enough RAM to load the indexes in parallel."
[2021-01-18T21:32:07.349 EST INFO ][ org.kde.kstars.ekos.align] - "+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++"
[2021-01-18T21:32:07.352 EST INFO ][ org.kde.kstars.ekos.align] - "Starting Internal StellarSolver Sextractor with the 4-SmallScaleSolving profile . . ."
[2021-01-18T21:32:07.484 EST INFO ][ org.kde.kstars.ekos.align] - "Stars Found before Filtering: 85"
[2021-01-18T21:32:07.487 EST INFO ][ org.kde.kstars.ekos.align] - "Removing the stars with a/b ratios greater than 1.5"
[2021-01-18T21:32:07.490 EST INFO ][ org.kde.kstars.ekos.align] - "Keeping just the 50 brightest stars"
[2021-01-18T21:32:07.493 EST INFO ][ org.kde.kstars.ekos.align] - "Stars Found after Filtering: 50"
[2021-01-18T21:32:07.497 EST INFO ][ org.kde.kstars.ekos.align] - "Starting 4 threads to solve on multiple depths"
[2021-01-18T21:32:07.500 EST INFO ][ org.kde.kstars.ekos.align] - "Child Solver # 1, Depth Low 1, Depth High 13"
[2021-01-18T21:32:07.503 EST INFO ][ org.kde.kstars.ekos.align] - "Child Solver # 2, Depth Low 13, Depth High 25"
[2021-01-18T21:32:07.506 EST INFO ][ org.kde.kstars.ekos.align] - "Child Solver # 3, Depth Low 25, Depth High 37"
[2021-01-18T21:32:07.509 EST INFO ][ org.kde.kstars.ekos.align] - "Child Solver # 4, Depth Low 37, Depth High 49"
[2021-01-18T21:32:07.512 EST INFO ][ org.kde.kstars.ekos.align] - "Child Solver # 5, Depth Low 49, Depth High 61"
[2021-01-18T21:32:07.515 EST DEBG ][ org.kde.kstars.ekos.capture] - setAlignStatus: 4
[2021-01-18T21:32:07.626 EST DEBG ][ default] - WARNING: Phonon::createPath: Cannot connect Phonon::MediaObject ( no objectName ) to Phonon::AudioOutput ( no objectName ).
[2021-01-18T21:32:08.028 EST DEBG ][ org.kde.kstars.ekos.guide] - PHD2: request: "{\"id\":13039,\"jsonrpc\":\"2.0\",\"method\":\"get_connected\"}"
[2021-01-18T21:32:08.029 EST DEBG ][ org.kde.kstars.ekos.guide] - PHD2: defer call "get_app_state"
[2021-01-18T21:32:08.032 EST DEBG ][ org.kde.kstars.ekos.guide] - PHD2: response: "{\"jsonrpc\":\"2.0\",\"result\":true,\"id\":13039}\r\n"
[2021-01-18T21:32:08.032 EST DEBG ][ org.kde.kstars.ekos.guide] - PHD2: request: "{\"id\":13040,\"jsonrpc\":\"2.0\",\"method\":\"get_app_state\"}"
[2021-01-18T21:32:08.034 EST DEBG ][ org.kde.kstars.ekos.guide] - PHD2: response: "{\"jsonrpc\":\"2.0\",\"result\":\"Stopped\",\"id\":13040}\r\n"
[2021-01-18T21:32:08.129 EST INFO ][ org.kde.kstars.ekos.align] - "Scale range: 56.1163 to 112.233 arcmin wide"
[2021-01-18T21:32:08.148 EST INFO ][ org.kde.kstars.ekos.align] - "Image width 1280 pixels; arcsec per pixel range: 2.63045 to 5.2609"
[2021-01-18T21:32:08.152 EST INFO ][ org.kde.kstars.ekos.align] - "+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++"
[2021-01-18T21:32:08.156 EST INFO ][ org.kde.kstars.ekos.align] - "Starting Internal StellarSolver Astrometry.net based Engine with the 4-SmallScaleSolving profile. . ."
[2021-01-18T21:32:08.177 EST INFO ][ org.kde.kstars.ekos.align] - "Scale range: 56.1163 to 112.233 arcmin wide"
[2021-01-18T21:32:08.181 EST INFO ][ org.kde.kstars.ekos.align] - "Image width 1280 pixels; arcsec per pixel range: 2.63045 to 5.2609"
[2021-01-18T21:32:08.184 EST INFO ][ org.kde.kstars.ekos.align] - "+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++"
[2021-01-18T21:32:08.188 EST INFO ][ org.kde.kstars.ekos.align] - "Starting Internal StellarSolver Astrometry.net based Engine with the 4-SmallScaleSolving profile. . ."
[2021-01-18T21:32:08.196 EST INFO ][ org.kde.kstars.ekos.align] - "Scale range: 56.1163 to 112.233 arcmin wide"
[2021-01-18T21:32:08.199 EST INFO ][ org.kde.kstars.ekos.align] - "Image width 1280 pixels; arcsec per pixel range: 2.63045 to 5.2609"
[2021-01-18T21:32:08.203 EST INFO ][ org.kde.kstars.ekos.align] - "+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++"
[2021-01-18T21:32:08.206 EST INFO ][ org.kde.kstars.ekos.align] - "Starting Internal StellarSolver Astrometry.net based Engine with the 4-SmallScaleSolving profile. . ."
[2021-01-18T21:32:08.266 EST INFO ][ org.kde.kstars.ekos.align] - "Scale range: 56.1163 to 112.233 arcmin wide"
[2021-01-18T21:32:08.270 EST INFO ][ org.kde.kstars.ekos.align] - "Image width 1280 pixels; arcsec per pixel range: 2.63045 to 5.2609"
[2021-01-18T21:32:08.274 EST INFO ][ org.kde.kstars.ekos.align] - "+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++"
[2021-01-18T21:32:08.281 EST INFO ][ org.kde.kstars.ekos.align] - "Starting Internal StellarSolver Astrometry.net based Engine with the 4-SmallScaleSolving profile. . ."
[2021-01-18T21:32:08.362 EST INFO ][ org.kde.kstars.ekos.align] - "Scale range: 56.1163 to 112.233 arcmin wide"
[2021-01-18T21:32:08.365 EST INFO ][ org.kde.kstars.ekos.align] - "Image width 1280 pixels; arcsec per pixel range: 2.63045 to 5.2609"
[2021-01-18T21:32:08.369 EST INFO ][ org.kde.kstars.ekos.align] - "+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++"
"
At this point Kstars completely crashed and I had to restart it, but it would then crash on every solve even after a reboot. I finally tried turning off multiple thread solving and then it worked for awhile, but eventually
crashed on another solver later on. Again, there was no Astrometry log written for any of these sessions, even though I explicitly set it to write a separate log before starting my profile in Ekos.

I don't know what I can do to shed any more light on this, if anyone has any ideas please let me know.

thanks,
Ron
3 years 2 months ago #66153

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

  • Posts: 389
  • Thank you received: 15

Replied by John Robison on topic Solver crashing Kstars

Hello,

Thank you for sending your log. One item stands out. This looks ominous as what follows are the app reinitializing.

[2021-01-18T21:32:07.626 EST DEBG ][ default] - WARNING: Phonon::createPath: Cannot connect Phonon::MediaObject ( no objectName ) to Phonon::AudioOutput ( no objectName )

Is the Phonon variable filled in with a recognizable path? This is what I see.
3 years 2 months ago #66182

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

  • Posts: 388
  • Thank you received: 17
I don't know, I saw that also and did not know what Phonon was. It looks like it is a video/music player app that Kstars uses if you play sound. I don't know whether I can just uninstall it or whether I need to reinstall it and make sure it finds what it needs to work. I am not sure why it is coming into play here, but it is there.

I was thinking the crash might have something to do with the solver running multiple threads and not cleaning up it's 'buffers,' because the way I got it to work again was to limit it to a single thread. It would be nice to have a 'restore' function that when Kstars crashes you can clean up any memory it was using, so a restart would leave the Rpi in a start fresh condition. The server continues to run, the equipment all stays connected, PhD2 is still running in the background, just Kstars, Ekos, the Fits viewer and Indi control panel go away. Restarting Kstars/Ekos is not a problem as all the equipment is still connected, so it only takes a minute to get back running. But if the solver continues to crash it, that is where it really becomes annoying.

thanks for that insight,
Ron
3 years 2 months ago #66186

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

  • Posts: 114
  • Thank you received: 5
Further to this issue, I also noticed that if the solver works then it throws out the coordination of the mount. How or why I do not know, but over the last couple of nights after operating the solver the mount will park OK, but is shows the ALT and AZ to be out by the same amount that the solver was used to correct for alignment. Also after solving for the first time it would not solve again (after meridian flip). It would try to solve for a location that was offset by the same amount as the offset to get it align for the first time. On that occasion it would never solve to where I wanted.

The only method I know to reset it was to disconnect the Indi Server then reconnect.
3 years 2 months ago #66191

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

  • Posts: 388
  • Thank you received: 17
Have you tried setting the solver options to solve without using position (so it does a blind solve). Sometimes that has helped my out of this fix, but usually not and I also have to restart everything.
3 years 2 months ago #66217

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

  • Posts: 389
  • Thank you received: 15

Replied by John Robison on topic Solver crashing Kstars

Hello,

Within KSTARS, main window, a sync telescope with KSTARS is there. This requires ensuring AstroBerry and KSTARS are in unison of where they think they are. Look in the lower left hand corner. Which longitude and latitude does it have? Does your Phone GPS agree?

Next, look at the time. AstroBerry and KSTARS must agree to the time and time zone.

If neither are in unison, fixing it is simple and inexpensive. I added an RTC and a GPS to manage location, time zone, and time. The AstroBerry Site suggests both.

As for Phonon, review the settings of your KSTARS and ECHOS. Something must have been selected but not configured.
Last edit: 3 years 2 months ago by John Robison.
3 years 2 months ago #66228

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

  • Posts: 388
  • Thank you received: 17
I checked all over the settings in Kstars and EKos and found nothing to do with sound, only PhD2 has a setting for 'beep on lost star" and it does beep, but on my laptop that I use to connect to the Astroberry Server using a web browser connection. There are no speakers on the Rpi, so perhaps EKOS is not set up like PhD2 to pass the sound output somehow thru the server back to my laptop.

I found some previous info about running these commands:
sudo apt-get install phonon4qt5-backend-vlc
sudo apt-get install phonon-backend-gstreamer-common

and I did so . I still get a warning in the log about PHONON not finding an object name, but the system did not crash last night and ran for several hours, so I don't think that was the issue.

I am thinking it may have something to do with the FITS viewer, as one time it crashed as that was opening. I also read someplace that if you use the zoom features of the FITS viewer that it uses a lot of resources. But I am running on an Rpi4 with 8 Gb of RAM, so I don't see how that could be the issue. I did avoid using zoom in the FITS viewer last night.

This failure mode is very inconsistent; the previous night I had continuous crashes, last night only one at the very beginning and after that I did many solves over several hours as I slewed to various targets! The Astrometry log (it is finally writing one) and the kstars log do not indicate anything interesting; they get to starting the solver and then just stop.

It may be an inconsistence in the way I set things up, although I try not to change much from one night to the next. The Astrometry log overwrites itself unless you change the name after a crash and there is no time-date stamp so it is hard to correlate with the kstars debug log.

I am open to suggestions concerning how to further debug solver crashes, but being inconsistent it is not easy.

thanks for all your work on this project.
3 years 2 months ago #66313

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

  • Posts: 269
  • Thank you received: 53

Replied by Ken Self on topic Solver crashing Kstars

In Kstars > Settings > Configure notifications some events play a sound. Had a similar issue on WSL which has no sound devices. Solution is to go through each notification, and if a play symbol appears agains the notification, untick the Play a Sound checkbox
3 years 2 months ago #66321

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

  • Posts: 114
  • Thank you received: 5
I looked at all of these and all match. This issue has only started in the most recent release of Indi, prior to this I had nothing like this at all. Wouldn't have a clue how to report a bug.
3 years 2 months ago #66329

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

  • Posts: 389
  • Thank you received: 15

Replied by John Robison on topic Solver crashing Kstars

Hello,

If the crashing occurred before applying the Phonon common libraries and QT5 upgrade, normal use is back. To me, that says that while not INDI related, the Phonon isn’t catastrophic in crashing PH2D.

In Linux, a package can be checked for dependencies. It isn’t unrealistic to think LIBINDI 1.8.8 has missing support packages. I have had to augment multiple packages in my past use. Whatever a developer has installed at build time, can creep into a package.

Keeping verbose logging on fills up space quickly when deducing a problem. Be sure verbose and to a file are enabled before each session. INDI, KSTARS, and Drivers used. One of them changed.
3 years 2 months ago #66334

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

  • Posts: 388
  • Thank you received: 17
I did find (thanks again KenS) that the sounds are selected under NOTIFICATIONS in the Kstars settings configurations. I don't think that was the source of the crashes, still having them. I have watched the resources panel of "htop" during solves and the resource use does go up, but not unusually so. The memory does not seem to be a problem as I am using 511M of 7.48M of memory. But I assume that if I keep restarting Kstars only (rather than rebooting) between crashes that more resources are being tied up.

All my logs go to file and I have been going thru them for each crash comparing them against "syslog," etc to see if I can find a pattern or culprit. Probably more to come, we will see.

thanks,
Ron
3 years 2 months ago #66449

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

  • Posts: 1
  • Thank you received: 0

Replied by Vali on topic Solver crashing Kstars

This behaviour is one of the reasons I stopped exploring Indi/Ekos/Kstars. However, due to the sky model build process which is much better than in any other software, I have decided to give it another chance.

Last night I have tested it again. I have removed the sounds but it keeps crashing after solving with internal solver, regardless the method.

The only way it does not crash is if I "Sync" instead of "Sync and Slew". It looks like it is able to solve and when it will try to slew, it crashes. Maybe this information helps checking the code.

I am using AZ GTi in Alt AZ mode, using the Skywatcher Alt Az driver.
By using ASTAP it does not crashes but is much slower and has a higher failure rate.
3 years 2 weeks ago #68373

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

Moderators: Radek Kaczorek
Time to create page: 1.404 seconds