×

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

Bi-monthly release with minor bug fixes and improvements

Ekos Scheduler Changes - Need Override to Stop Rescheduling

  • Posts: 112
  • Thank you received: 6
With Kstars 2.9.6 I'm now getting this with my sequences:

2018-06-01T21:28:23 Job evaluation complete.
2018-06-01T21:28:23 Jobs 'T1-1' and 'T4-4' have close start up times, job 'T4-4' is rescheduled to 01/06/18 23:43.
2018-06-01T21:28:23 Jobs 'T1-1' and 'T3-4' have close start up times, job 'T3-4' is rescheduled to 01/06/18 23:38.
2018-06-01T21:28:23 Jobs 'T1-1' and 'T2-4' have close start up times, job 'T2-4' is rescheduled to 01/06/18 23:33.
2018-06-01T21:28:23 Jobs 'T1-1' and 'T1-4' have close start up times, job 'T1-4' is rescheduled to 01/06/18 23:28.
2018-06-01T21:28:23 Jobs 'T1-1' and 'T4-3' have close start up times, job 'T4-3' is rescheduled to 01/06/18 23:23.
2018-06-01T21:28:23 Jobs 'T1-1' and 'T3-3' have close start up times, job 'T3-3' is rescheduled to 01/06/18 23:18.
2018-06-01T21:28:23 Jobs 'T1-1' and 'T2-3' have close start up times, job 'T2-3' is rescheduled to 01/06/18 23:13.
2018-06-01T21:28:23 Jobs 'T1-1' and 'T1-3' have close start up times, job 'T1-3' is rescheduled to 01/06/18 23:08.
2018-06-01T21:28:23 Jobs 'T1-1' and 'T4-2' have close start up times, job 'T4-2' is rescheduled to 01/06/18 23:03.
2018-06-01T21:28:23 Jobs 'T1-1' and 'T3-2' have close start up times, job 'T3-2' is rescheduled to 01/06/18 22:58.
2018-06-01T21:28:23 Jobs 'T1-1' and 'T2-2' have close start up times, job 'T2-2' is rescheduled to 01/06/18 22:53.
2018-06-01T21:28:23 Jobs 'T1-1' and 'T1-2' have close start up times, job 'T1-2' is rescheduled to 01/06/18 22:48.
2018-06-01T21:28:23 Jobs 'T1-1' and 'T4-1' have close start up times, job 'T4-1' is rescheduled to 01/06/18 22:43.
2018-06-01T21:28:23 Jobs 'T1-1' and 'T3-1' have close start up times, job 'T3-1' is rescheduled to 01/06/18 22:38.
2018-06-01T21:28:23 Jobs 'T1-1' and 'T2-1' have close start up times, job 'T2-1' is rescheduled to 01/06/18 22:33.
2018-06-01T21:28:23 Job 'T4-4' is due to run as soon as possible.

This completely defeats the purpose of what I'm trying to do. Was not an issue in 2.9.5. I just need the sequences to run in order with no regard for start and end times, and certainly to not have them automatically rescheduled! It is not moving between the jobs with the ASAP setting, it just hangs until the rescheduled time is met??

Is there a place I can grab 2.9.5 to roll back?

Thanks!

David
Last edit: 5 years 9 months ago by David Rankin.
5 years 9 months ago #26349

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

To get 2.9.5, use the exact same URL you used for 2.9.6, but simply replace .6 with .5

Eric has been making quite a few changes recently so this could be a regression.
Last edit: 5 years 9 months ago by Jasem Mutlaq.
5 years 9 months ago #26350

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

  • Posts: 112
  • Thank you received: 6
No prob, I was able to roll back to 2.9.5 for now.

Cheers!
5 years 9 months ago #26351

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

  • Posts: 1029
  • Thank you received: 301
@rankinstudio could I get your .esq and .esl files to reproduce the issue on my side? This behavior seems related to the checkbox "Sort jobs", which was supposed to sort jobs per altitude and score, but was not giving expected results before 2.9.6. If left unchecked, jobs will not be sorted from their characteristics, and will be rescheduled in order of appearance in the list. If you really want strict control, you need to schedule your jobs manually, but that should not be necessary here. You can also use priority, but that's a hassle in your case too (plus it doesn't appear in the table).

Your other problem, asap jobs not triggering, is indeed a regression I introduced while pushing partial work, and I think is fixed in D12954, or possibly D13009. I'd like to verify that too with your scheduler and sequence files.

-Eric
5 years 9 months ago #26355

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

  • Posts: 112
  • Thank you received: 6
Here is an esq file I'm running right now in 2.9.5. I generate these with another program I wrote to create repeating search fields.

It loads in 2.9.6 but it wont move from the first sequence to the next.

Cheers
The following user(s) said Thank You: Eric
Last edit: 5 years 9 months ago by David Rankin.
5 years 9 months ago #26356
Attachments:

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

  • Posts: 1029
  • Thank you received: 301
I confirm what I said about rescheduling: you certainly have "Sort jobs per altitude and priority" checked in Ekos options, and your job are being reordered. Testing locally without this option schedules jobs in order of appearance.

For the other issue triage, I need to check which change fixes it exactly.

-Eric
5 years 9 months ago #26358

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

  • Posts: 112
  • Thank you received: 6

I do not use that option in the Ekos settings: "Sort jobs per altitude and priority"

In 2.9.5 it runs as expected, down the list one at a time with no delay between jobs.

These are the settings I've used for quite some time now.


Cheers
Last edit: 5 years 9 months ago by David Rankin.
5 years 9 months ago #26359
Attachments:

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

  • Posts: 1029
  • Thank you received: 301
Thanks, I must have made a mistake while testing what I thought was 2.9.6 then :/ let me double-check.

-Eric
5 years 9 months ago #26362

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

  • Posts: 1029
  • Thank you received: 301
@knro could you clarify what is v2.9.6 in the KDE git? I followed the steps for this report, and can't reproduce the issue with f8e1086ac, which I think is close enough.

I could reproduce the issue with ea68c70fc, that's where the regression appears in my changes. But that's old already.

In any case, I confirmed merge 9320fafca is OK, so perhaps that's all that needs to be done right now until 2.9.7. Unless there's something more to this that I did not see.



-Eric
5 years 9 months ago #26394

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

  • Posts: 1029
  • Thank you received: 301
My bad, I could finally reproduce your issue! This one was tricky, related to the stability of the sort algorithm used by the scheduler (what happens when two elements compare equal). Thanks a lot for reporting this one!

-Eric
The following user(s) said Thank You: David Rankin
5 years 9 months ago #26453

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

  • Posts: 112
  • Thank you received: 6
No prob! Thanks to the whole team for making such an awesome free set of programs. Very cool stuff.
5 years 9 months ago #26454

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

  • Posts: 1029
  • Thank you received: 301
5 years 9 months ago #26466

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

Time to create page: 4.226 seconds