Due to large data, I have to knit my .Rmd files by batching them to a remote server and running rmarkdown::render(). When I encounter an error the partial stack trace given in my .out file is generally uninformative; for example: Calls: <Anonymous> ... .BLOCK_colSums -> .process_grid_by_col -> normarg_grid. Is there a way that I can get a full traceback, i.e. similar to the result of traceback() ? I've looked through the optional arguments to render() and don't see anything. This Stack Exchange post discusses a similar issue, but the solutions only apply to interactive sessions. Thank you for your consideration.