×

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

Bi-monthly release with minor bug fixes and improvements

Ekos Scheduler Skipping exposures

  • Posts: 67
  • Thank you received: 3
I've been loving Ekos more and more of late but I am having some problems with the scheduler .

After completing a job, when it gets another job with similar requirements, these are already complete - I can't understand why this is happening at the moment.
Astroberry Kstars Version 3.4.2 Build: 2020-04-13T21:01:23Z

eg: Schedule 1 job:
Target M51, loading a capture file called M51_1 which has 6x300s R,9x300s G,7x300s B

Schedule Job 2: 
Target M51, loading a capture file called M51_2 which has 6x300s R,6x300s G,6x300s B,6x300s L,6x300s Ha .

Result is job 1 runs for full sequence but job 2 thinks the R,G and B are already done and starts on L  - no idea why this is occuring? Something I have set wrong or workflow? File contents below in spoilers

Scheduler ESL File
Warning: Spoiler!


Capture Sequence M51_1.esq

Warning: Spoiler!


Capture Sequence M51_2.esq
Warning: Spoiler!
3 years 10 months ago #54637

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

  • Posts: 194
  • Thank you received: 20
In my limited experience, I'd say the output directories are the same for both sequences, so it sees the frames that are already there, and thinks it is done.

Thanks,
Dave
3 years 10 months ago #54639

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

  • Posts: 32
  • Thank you received: 6
If you have “Remember Job Progress” enabled, try turning it off. I’ve had similar issues when it thought the job had already been ran.
3 years 10 months ago #54646

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

  • Posts: 67
  • Thank you received: 3
So this problem hasn't gone away.

I decided to revisit a target tonight and grab some additional SII data. Should have been as simple as load the scheduler saved file and click go right?

I have "Remember Job Progress" turned off
I loaded the file, tried to set it going and I got an abort.

2020-06-29T00:32:18 Job 'IC_5070' has not been processed upon scheduler stop, marking aborted.
2020-06-29T00:32:18 Shutdown complete.
2020-06-29T00:32:17 No jobs left in the scheduler queue.
2020-06-29T00:32:16 Scheduler started.

I tried creating a new CCD exposure file (different number of subs in it) and tried to use that as the requence, still an abort.
I renamed the location it saves to (SII_2 instead of SII) and still an abort
I changed the FITS file it uses for plate solving, again an abort.

So I have wasted over 40 mins trying to get this working tonight, in the end I am imaging manually after align, calibrating and focusing in separate steps.

It's quite frustrating as it's time lost I could be imaging and I sort of expected it to work how I have tried using it. Is there something I am doing wrong?

My expectation of the behaviour is that the scheduler should have started, solved, aligned and calibrated PHD2 then started the requested sequence of subs with the file counter picking up from the last one in the folder it is instructed to save the images to. Is my expectation wrong, does it work some other way?

Cheers in advance.
3 years 8 months ago #56192
Attachments:

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

No logs?
3 years 8 months ago #56219

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

  • Posts: 67
  • Thank you received: 3
Yes but they only show the same as what I've already posted, is there a way to get more verbose logging of the org.kde.kstars.ekos.* events?
Other problem is total log is 337Mb! (I think I need to turn off the QHY debug!)

If you can let me know a more verbose log option I am fairly sure I can replicate the issue. I might see if I can do so with the qhy debug turned off and see what size that results in.
3 years 8 months ago #56236

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

  • Posts: 67
  • Thank you received: 3
I found how to turn on the scheduler logging and turned off the QHY debug to get a smaller file :)
Replicated the issue again here:

And a second run same result:
3 years 8 months ago #56237

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

  • Posts: 1185
  • Thank you received: 370
Without logs its difficult to analyze, but it looks like you still have "Remember job progress" selected and the directory you are using contains already sufficient images of that type.

But maybe a hint how you could use the scheduler slightly differently. I personally make extensive usage of the "Repeat for" option in combination of a standard set of frames. Typically, I am trying to shoot a target with 60 x 480s Lum and 15 x 240 RGB bin 2x2. My setup is the following:
  • one job 15x the sequence LLLRGB
  • one job 30x LL
I have "Remember job progress" turned on and use this setup over several nights.

With setup I go even further for those nights where frame collecting of a target reaches the limit, I simply put a second target with the same logic as jobs #3 and #4. That way I ensure that I get enough frames of the first one and as soon as that is finished, the scheduler continues with the second target.

HTH
Wolfgang
The following user(s) said Thank You: jiberjaber
3 years 8 months ago #56258

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

  • Posts: 67
  • Thank you received: 3

Thanks Wolfgang,
The logs are in the post above your reply, as I was able to recreate the issue on the simulator but nothing stands out however!.
I certainly have "Remember job progress" turned off which is why I came back to this post to raise the issue again.
I usually stick with one filter and shoot through the night as I find the auto-focus eats into imaging time and it's not be 100% all the time (when it works, it is great, but sometimes I have to keep any eye on the first subs coming in and double check, so not I focus one filter, confirm with bratinhov mask and use that all night for everything (align and imaging).
I also make use of the option to repeat a scheduled sequence until a certain time (usually 02:45 here as that is when it starts to get too bright here at the moment) so I have a standard sequence of 6x600s for each NB filter at the right temperature and binning. This however still seems to be linked to the target because of the file naming .

When I imaged the Ha filter, this was over more than one night and I was able to use the same sequence file but had to create a new schedule entry for that night. This latest one with SII filter just wouldn't work at all.

I see the guys have released a beta debugging assistant, I might see if I can give that a go in the hope that it will produce more info to resolve this, as it seems crazy that I can't repeat something without having to recreate it every time.
3 years 8 months ago #56273

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

  • Posts: 67
  • Thank you received: 3
I couldn't get the debugger to compile so stuck with the logs noted above... Am I doing something wrong with the scheduler?
3 years 8 months ago #56410

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

  • Posts: 1185
  • Thank you received: 370
Frankly spoken, I'm afraid: yes.

The log shows that the scheduler finds enough existing images and that the scheduler therefore decides that there is nothing to do. Shift your images from /home/astroberry/AAAastro_images/IC_5070/ to another place (or delete them) and check again. Then the scheduler should start.

I still am sure that you have selected the option "Remember job progress".

Wolfgang
3 years 8 months ago #56412

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

  • Posts: 67
  • Thank you received: 3
No, remember jobs was not selected when these logs were produced.

However tonight (this morning) I noticed an issue with the scheduler.

I was running a job with a repeat sequence until selected time (03:20) rule. At around 03:05 I decided to use Kstars to check where Comet Neowise would be relative to my horizon, I decided to accelerate the kstars time to see when it would clear a nearby hedgerow and selected 1 hour increments. When I advanced Kstars to 05:00 it stopped the scheduler as it went past 03:20 and started to park the scope and warm the CCD. I immediately went to the scheduler and tried to re-run the sequence where it exhibited the skipping and going straight to job complete.
Local time was still around 03:05, whilst I have all time driven from Kstars I am not expecting the scheduler to take it's time from the accelerated clock of the planetarium !?!?!

I suspect the issue with scheduler skipping exposures is potentially linked to taking the time from Kstars rather than system clock?

I need to head to bed now but will aim try and pull the logs tomorrow when I get up.
3 years 8 months ago #56738

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

Time to create page: 1.468 seconds