×

INDI Library v2.0.7 is Released (01 Apr 2024)

Bi-monthly release with minor bug fixes and improvements

Offline Astrometry Solver failing reading FITS

  • Posts: 18
  • Thank you received: 3
Hi guys,

Strange problem has arisen in the last couple of days with my offline astrometry solver failing with the message
Reading input file 1 of 2: ""...

I'm using kstars-bleeding on ubuntu 15.10 and it *was* working at the weekend. I must have broken something. The command line the 'verbose' output gives me, when transplanted to the commandline by hand, runs fine. The online solver, its own issues aside, seems to work as well.

I've been enabling debug logging on the indi and kstars side but I can't seem to find any output relevent to this to see what's going wrong. It seems the FITS file name is somehow not being passed.

I've tried wiping ~/local/share/kstars, restarting the indi server and purging and re-installing kstars-bleeding. No change.

Has anyone seen a similar problem and knows where I should look, or even tell me whether this falls into the realm of kstars debugging, ekos debugging or indi debugging!

Thanks,

BW.
8 years 4 months ago #6021

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

Try removing my PPA if you had it before and remove astrometry.et then install it from Ubuntu directly.
8 years 4 months ago #6023

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

  • Posts: 18
  • Thank you received: 3
Hi Knro,

Thanks for the response, but I've never used the old astrometry PPA. This is fresh from a new install two weeks ago and it does work when run from the command line. It's as if the actual command being issued is different to the one I'm copying and pasting in.

Do you know which component I should be extracting debug info from in this instance?

Thanks,

BW.
8 years 4 months ago #6024

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

  • Posts: 18
  • Thank you received: 3
Now this has become even stranger.

I've written a small wrapper script to record what's happening at write the output to a file as well as back to EKOS. It now works! I go back to using /usr/bin/solve-field -- fails again.

Is this a possible timing issue of when the FITS file becomes available in the /tmp directory, do you think?

BW
8 years 4 months ago #6025

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

The same error was reported by others before. It appears that it starts and disappears mysteriously. Could this perhaps be some environment variable? I'm not sure, if the reason is found, it would help a lot of users. I don't think it has to do with /tmp at all
8 years 4 months ago #6027

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

  • Posts: 18
  • Thank you received: 3
Happy to look into this a bit more if it's a recurrent problem.

Here's my little wrapper:-

/usr/local/sin/solve:
#!/bin/bash
echo /usr/bin/solve-field $* >> /tmp/solve.log
/usr/bin/solve-field $* | tee -a /tmp/solve.log

Then I changed the 'solver' entry in the Ekos section of the Kstars config to be /usr/local/bin/solve.

This works, and records everything the solver says into the file /tmp/solve.log.

Interestingly the main difference is that solve-field reports opening file 1 of 1 (/tmp/fits*) rather than file 1 of 2 (blank).

Now, this did start occurring after an apt-get update ; apt-get upgrade on kstars-bleeding. The same update broke my kstars install on my virtual ekos box as arabic text started appearing in numeric fields. I changed the locale variables to en_GB.UTF-8 away from the Kuwaiti defaults in the ekos VM. I wonder if that's connected?

Something to experiment with later. Alas cloudy skies last night so solve-field couldn't be used in anger anyway. ;-)
8 years 4 months ago #6032

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

  • Posts: 193
  • Thank you received: 46

This is precisely the reason I wrote the original camera simulator, it allows lots of testing and development on those cloudy nights.
The following user(s) said Thank You: David
8 years 4 months ago #6034

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

  • Posts: 171
  • Thank you received: 41
In case of tests for astrometry you need installed gsc, then you'll get nice images to test :)
8 years 4 months ago #6035

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

  • Posts: 18
  • Thank you received: 3

An excellent point and one well made.

I do have that set up, gsc et al, as I was using it earlier. Slipped my mind!

BW
8 years 4 months ago #6036

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

  • Posts: 14
  • Thank you received: 2
I had the problem with non-solving offline solver in EKOS too. Your wrapper made it work! Thanks a lot!
Is there a root cause or a clean solution available yet?
7 years 8 months ago #9543

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

  • Posts: 18
  • Thank you received: 3
I've not seen a recurrence since my initial work around, to be honest, so I never got to the bottom of it.

It does seem a strange one, doesn't it? Perhaps some kind of timing issue/race condition and the wrapper simply adds a couple of milliseconds delay whilst it fires up bash?

Difficult to see where it could be coming from, though.

BW.
7 years 7 months ago #9868

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

  • Posts: 643
  • Thank you received: 62
Hi!

Suddenly I have the same problem, and your little wrapper seems to have solved it. I wonder what it is all about...... Just did an apt-get dist-upgrade today, after that the problem was there.

Magnus
7 years 7 months ago #9878

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

Time to create page: 1.076 seconds