×

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

Bi-monthly release with minor bug fixes and improvements

Dark Library acquisition crashes kstars

  • Posts: 1009
  • Thank you received: 133
After it's cloudy outside (but the telescope is up due to some tests) I wanted to have a look at the new Dark library.
However, that one is constantly crashing kstars hard.  Usually once it has collected all frames for one datapoint.
Starting kstars from the terminal, I see
FITSIO status = 207: illegal character in keyword
 
FITSIO status = 202: keyword not found in header
��������������������������������������������������������������������������
Name of keyword no. 37 contains illegal character(s): ��������������������������
Failed to read the required primary array header keywords.
ffopen could not interpret primary array header of file: 
/home/pit/.local/share/kstars/darks/darkframe_2021-04-18T22-28-39.fits
Character 9 in this keyword is illegal. Hex Value = FFFFFF80
It also keeps trying to load old files from that directory that I had deleted (they kept crashing kstars as soon as I activated that camera in the dark library window).

And side note: I wanted to do darks for -15 degrees, this is the only T I use.  The application doesn't allow that and insists on doing at least two T points.  No such restriction for exposure time :((

This is kstars from git today.
 
2 years 11 months ago #70243

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

So crash might be related to old files? Need to check. I just fixed the temperature issue. Git pull and retest.
2 years 11 months ago #70244

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

  • Posts: 1009
  • Thank you received: 133
Yes, if there's an old file I get error messages. But also if the old ones are deleted, and I create a new set, once that is written EKOS wants to read it again, and crashes during that, see attached logfile.  Last line is just when it crashes.
Will continue tomorrow...

 
2 years 11 months ago #70246
Attachments:

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

Alright, I pushed a fix that should clear all previous entries. But it's too late for you now unless you want to edit ~/.local/share/kstars/userdb.sqlite and delete all database entries in the darkframe table.
2 years 11 months ago #70262

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

  • Posts: 1009
  • Thank you received: 133
Just trying it out, though with a different camera here on my desk.
a) it doesn't crash anymore. But setup is different, so I'll repeat it (later) with exactly the same one.

b) Single-T seems to work, but I have the impression it doesn't check T now: I did a sequence with T=0 and 3 exposure values. Camera was already at 0. That went fine. Then I wanted to re-do it at T=-1. But pressing 'Start' it immediately started capturing with T at +0.5. While writing this, I notice that now T=-1. So it's set, but not waited for it to settle (Enforce T is checked in the capture tab). Even before that I wondered if a 'settle' period would be a good idea in that module, so wait until T is reported as wanted, then wait an additional period...

c) This camera is an OSC one. I noticed that - although the camera is set to 16bit RAW - the result is stored as [X,Y,3] array. Is that intentional? Is it just duplicating the single dark 3 times, or doing some de-bayering?

More once I've tested with the remote setup at the telescope....
2 years 11 months ago #70264

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

  • Posts: 1009
  • Thank you received: 133
Ah, an addition: Looks the missing T stabilization is linked to using the capture tab. The dark library adds entries to the sequence, but doesn't remove those if you, e.g., press stop. The next start will then add the new sequences, and execute - but that will start wherever it was aborted before....
So I think the sequences should be cleaned out after the run, also (especially) when it was interrupted?
2 years 11 months ago #70265

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

Good point, will add that now.
2 years 11 months ago #70266

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

  • Posts: 1009
  • Thank you received: 133
So I just connected via my mount computer, and to the ASI1600, just as yesterday.  It still had the (no longer existing) dark masters from yesterday in the database, and kstars reported warnings about files not found.  No crash though.  And I could easily remove those entries from the list in the darklib module master browser.
I quit and restarted again, no more warnings.  Taking new darks also succeeded, just like with the ASI183MC locally before.

So to be sure I'm not nuts I downgraded to the old version I used yesterday.  It directly crashes as soon as I try to connect to the ASI1600, with the same error message as yesterday (Character 9 in this keyword is illegal. Hex Value = FFFFFF80).
So either it was some error that got removed/fixed by your changes, or there was something fishy with my build.  But as it seems to work fine now, I assume it's not worth to further investigate this?
 
2 years 11 months ago #70267

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

  • Posts: 1119
  • Thank you received: 182
I don't know whether that is related, but I had repeated crashes while trying to build (not install) a brand new Kstars from the most recent master yesterday on a mini-PC with a stable Kstars installation from March 1. The build repeatedly crashed at the same point, saying there was a problem with a darklibrary file.

I don't recall exactly the name of that file, and I did not explore further since I had the stable Kstars version on the machine and I wanted to image, but reading this thread I am now wondering whether the issues are related. I can try to reproduce later tonight, if that is useful, and post the exact error message.

Jasem, please advise if you want me to do that.

Jo
2 years 11 months ago #70273

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

  • Posts: 1009
  • Thank you received: 133
Jasem,
I'm taking 'real' data today, after having played with the dark library yesterday.  Starting the sequence, I noticed that the received images are no longer displayed in the FitsViewer.  Indeed, the option in the configuration was disabled.  Is that set by the dark library sequences?  If so, please re-enable it afterwards.  Less because I want to see the images, but the statistics tab doesn't display HFR/Ecc if they are not loaded in the viewer....

I've also enabled 'Darks' for the guide camera today, and it flawlessly seems to use the ones I acquired yesterday - nice :D
 
2 years 11 months ago #70306

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

Thanks for the feedback. I'll look into the FITSViewer issue, I think I've seen that.
2 years 11 months ago #70322

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

  • Posts: 985
  • Thank you received: 160
My attempts to build a master darks library were unsuccessful (yesterday's GIT). Crashes occured repeatedly when the master dark was computed regardless of whether or not older files existed. I guess I had ~10 attempts. One succeeded, all others failed and crashed KStars/Ekos altogether.
Last edit: 2 years 9 months ago by Alfred.
2 years 9 months ago #72249

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

Time to create page: 1.140 seconds