Postprocessing a rerun

[sbauer]

I did a historical run with CESM-WACCM a few months back and post processed the timeseries. However, there are a few months missing in the first 2.5 years of that simulation and so I wanted to rerun them.
I want them to smoothly merge to my existing 25year run afterwards so I reran the 2.5years in the same case directory as the 25years (though I removed the 25 years from /run and /archive). The case is here: /glade/derecho/scratch/sbauer/cesm2.1.5_output/histSST/f.e21.FWHIST.f09_f09_mg17.atmos-scale_fixedSST.001/

I moved the successfully postprocessed 25 years from postprocessed_output to postprocessed_output_1 so that I wouldn't run into issues when postprocessing the 2.5years. However, I keep running into problems when wanting to postprocess the 2.5years. I keep getting an error in the timeseries.log file that I cannot seem to fix. Extra data: line 1 column 3752983 (char 3752982) in here: /glade/derecho/scratch/sbauer/cesm2.1.5_output/histSST/f.e21.FWHIST.f09_f09_mg17.atmos-scale_fixedSST.001/case_scripts/postprocess/logs

Can someone guide me to where the issue arises and what I can do about it?

 

[dobbins]

Hi Susanne (sorry for the slow response!) -- this is another lovely quirk of the old time-series tool.

If you have removed all the data so you don't overwrite what you've already done, the simplest option is just to remove the 'logs/ts_status.log' file. The process reads that to, in theory, know what it's done, but it always causes errors. Removing it and just staging precisely the data you want to process should work.

(If you don't want to 'move' your data, I also just give it a new TIMESERIES_OUTPUT_ROOTDIR directory to write new things to.)

I'll email you too, in case this doesn't give you a ping.

Cheers,
- Brian