Last week I worked with over 10,000 files of tabular data, each with about 50,000 rows and
10 columns separated by a '|'. This is the kind of problem that falls in the uncanny
valley between small and biggish data. My goal was to make some quick checks and, if
possible, concatenate the files into a single csv file I could load into Python.
ProtonVPN provides a GUI and a console application for Linux, but on 2023, it is nearly
unusable. It consumes up to 2 GB of memory idled, drops connections, and fails to
reconnect unless the kill switch is deleted using nmcli. The command line application has
a bizarre dependency on nm-applet, which of course means it cannot be used on a headless
machine. It appears that ProtonVPN is not going to fix this. The solution is to use
ProtonVPN with Wireguard.