Clearing Process Logs
Written on October 29, 2014
I’m writing a big hairy database import from a legacy system and not all of the records contain valid data. It’s important to me to see when any of the rows from the old datastore fail model validations so I can either fix my model logic, or insert corrected data later.
My workflow is basically to run the import, view logged exceptions in a tmux split, fix them, and repeat. I found it painful to have to clear the log/screen every time I wanted to rerun the import, so I automated it.
I’ve structured my import into a collection of Rake tasks so each table can be imported individually (after importing its respective dependencies). I have an additional task for running all of the import tasks, and this additional task depends on a task I use for truncating the log file.
When I open the log file in the
clear_import_log task, I use the
w+ which truncates the existing file to zero length, or creates a new file if one doesn’t exist.
To actually clear the screen when I tail the log, I need to employ a touch more trickery. I wrote a bash script which clears the screen when it encounters the line “file truncated”, or prints the line to the screen in any other case.
In my tmux split, I tail the log file and pipe the output to my bash script, so now whenever I restart the database import, the screen clears and I see fresh information.