WRF doesn't halt when beginning LBC time is not in wrfbdy_d01 file

[mgduda]

When restarting a model simulation at a time beyond the last valid boundary data in the wrfbdy_d01 file, the model doesn't stop, but instead appears to go into an infinite loop, generating printout like the following in all rsl files:

WRF restart, LBC starts at 2019-01-25_00:00:00 and restart starts at 2019-01-26_00:00:00
 LBC for restart: Starting valid date = 2019-01-25_00:00:00, Ending valid date = 2019-01-25_03:00:00
 LBC for restart: Restart time            = 2019-01-26_00:00:00
 LBC for restart: Looking for a bounding time
 LBC for restart: Starting valid date = 2019-01-25_03:00:00, Ending valid date = 2019-01-25_06:00:00
 LBC for restart: Starting valid date = 2019-01-25_06:00:00, Ending valid date = 2019-01-25_09:00:00
 LBC for restart: Starting valid date = 2019-01-25_09:00:00, Ending valid date = 2019-01-25_12:00:00
 LBC for restart: Starting valid date = 2019-01-25_12:00:00, Ending valid date = 2019-01-25_15:00:00
 LBC for restart: Starting valid date = 2019-01-25_15:00:00, Ending valid date = 2019-01-25_18:00:00
 LBC for restart: Starting valid date = 2019-01-25_18:00:00, Ending valid date = 2019-01-25_21:00:00
 LBC for restart: Starting valid date = 2019-01-25_21:00:00, Ending valid date = 2019-01-26_00:00:00
---- ERROR: Cannot find a valid time to start the LBC during this restart, likely ran out of time periods to test
 LBC for restart: Starting valid date = 2019-01-25_21:00:00, Ending valid date = 2019-01-26_00:00:00
---- ERROR: Cannot find a valid time to start the LBC during this restart, likely ran out of time periods to test
 LBC for restart: Starting valid date = 2019-01-25_21:00:00, Ending valid date = 2019-01-26_00:00:00
---- ERROR: Cannot find a valid time to start the LBC during this restart, likely ran out of time periods to test
 LBC for restart: Starting valid date = 2019-01-25_21:00:00, Ending valid date = 2019-01-26_00:00:00
---- ERROR: Cannot find a valid time to start the LBC during this restart, likely ran out of time periods to test
 LBC for restart: Starting valid date = 2019-01-25_21:00:00, Ending valid date = 2019-01-26_00:00:00
---- ERROR: Cannot find a valid time to start the LBC during this restart, likely ran out of time periods to test

In a small test run, I was able to generate almost 26 GB of rsl files before manually terminating WRF.

[mgduda]

The shortest route to reproducing this would probably be:

1. Run real.exe to generate a wrfinput_d01 and wrfbdy_d01 file, where the wrfbdy_d01 file has, say, 24 hours of data
2. Run the model for 24 hours, writing a restart file at the end of the simulation
3. Try to restart the model

[mgduda]

Although the code in input_wrf.F limits the number of messages to 10000, the input_boundary routine (which calls input_wrf) appears to be called in what becomes an infinite loop in mediation_integrate.F.

[weiwangncar]

@davegill Can you take a look at this? Thanks.