metrumresearchgroup / bbi

Next generation modeling platform
11 stars 2 forks source link

Exit Code 107 when submitting model #313

Closed poonv closed 9 months ago

poonv commented 9 months ago

Had an issue submitting a nonmem model on windows 10 (R: 4.3.1, Rstudio: 2023.09, bbi version: 3.2.2) where the execution would stop due an exit code 107. Did also try running Sys.unsetenv("GFORTRAN_STDOUT_UNIT") first but didn't seem to work.

Error in check_status_code(p$get_exit_status(), output, .cmd_args) : bbi nonmem run local C:/Users/user1/nm/run004.ctl --overwrite returned status code 1 -- STDOUT and STDERR: time="2023-10-11T14:38:20-07:00" level=info msg="Successfully loaded default configuration from C:\Users\user1\nm\bbi.yaml" time="2023-10-11T14:38:20-07:00" level=info msg="Beginning Local Path" time="2023-10-11T14:38:20-07:00" level=info msg="A total of 1 models have completed the initial preparation phase" time="2023-10-11T14:38:20-07:00" level=info msg="[run004] Beginning local work phase" time="2023-10-11T14:38:21-07:00" level=error msg="[run004] Exit code was 107, details were exit status 107" time="2023-10-11T14:38:21-07:00" level=error msg="[run004] output details were: Starting NMTRAN\r\n \r\n WARNINGS AND ERRORS (IF ANY) FOR PROBLEM 1\r\n \r\n (WARNING 2) NM-TRAN INFERS THAT THE DATA ARE POPULATION.\r\n \r\n (WARNING 13) WITH USER-WRITTEN PRED OR $PRED,

kyleam commented 9 months ago

Hi @poonv, I'm going to transfer this to the bbi repo because it's more likely to be an issue there than in bbr. Are you able to come up with a minimal reproducer for this error (data, control stream, driver R script)? That will make it more likely that we'll be able to help.

With a quick search, I haven't been able to determine what an NM-TRAN exit code of 107 signals.

seth127 commented 9 months ago

Assuming this model is proprietary, it may be tough to get reproducer, though I agree with @kyleam that would greatly increase our chances of figuring out what's going on here. @poonv if you are able to share a model, much appreciated, but if not that's fine.

It looks to me like the error message is truncated here (ends in a comma, no closed quote)

level=error msg="[run004] output details were: Starting NMTRAN\r\n \r\n WARNINGS AND ERRORS (IF ANY) FOR PROBLEM 1\r\n \r\n (WARNING 2) NM-TRAN INFERS THAT THE DATA ARE POPULATION.\r\n \r\n (WARNING 13) WITH USER-WRITTEN PRED OR $PRED,

My bet is that the useful part of the error is below. @poonv could you take a look in the <model>.ctl.out file in your model output directory directory? You should be able to get the path with bbr::build_path_from_model(..., ".ctl.out"). If you can find the full error in there, please paste that into the issue. Thanks.

poonv commented 9 months ago

@seth127 Thanks for the help! I was able to find the output file with the full error description there. Turns out it was just simple dataset pointer issue.

kyleam commented 9 months ago

@poonv Thanks for reporting back. I've opened a dedicated issue about the truncated output (gh-314).