search

solarmonitor / solarmonitoridl

4 stars 7 forks source link

New Forecasting not appearing on website #24

Closed dpshelio closed 8 years ago

dpshelio commented 10 years ago

Originally reported by: David PS (Bitbucket: dvdgc, GitHub: Unknown)

It seems forecasting values are not being generated. Eg.: http://www.solarmonitor.org/forecast.php?date=20131220

This probably have to do with the last push to deployment which included 52f8b6e

dpshelio commented 10 years ago

Original comment by David PS (Bitbucket: dvdgc, GitHub: Unknown):

The problem seems to have been produced by something failing before the prediction was made. Not confirmed, but we are getting the predictions as expected.

dpshelio commented 10 years ago

Original comment by David PS (Bitbucket: dvdgc, GitHub: Unknown):

It's seems ok now again. We've just missed one day. Probably as you said it, something must have failed before. Sorry for the false alarm :)

dpshelio commented 10 years ago

Original comment by Shaun Bloomfield (Bitbucket: s_bloomfield, GitHub: Unknown):

I cloned the allinone repository to my laptop, copied the example sm_batch.cfg file into the main sm-batch directory and edited it to my local directory settings (i.e., where sm-batch and my $SSW are located). After running ./sm_batch I successfully have a complete copy of arm_forecast_20131220.txt in my expected $OUTPUT_PATH directory (/tmp/data/2013/12/20/meta).

I looked at the sm-batch/idl/arm_forecast.pro differences between commits, and it looks like the format of the 'output_path' variable changed in the same commit that the flarehist.txt file and reading method were changed.

However, I have just noticed that the arm_mmmotd_yyyymmdd.txt file is also missing on SM. It would be best to check that there isn't something else crashing before arm_batch.pro gets to trying to run arm_forecast.pro and then mmmotd2arm.pro as these are the last two procedures to run in arm_batch.pro...