In web applications, you try to decide when to use either JSON Web Tokens (JWTs) or sessions (cookies) for authentication. When you browse the web you use HTTP, which is a stateless protocol. So, the only way to remember the states of your application is using either sessions or tokens.
It happens to the best: some script rockets to the skyline resulting in an instand system administrator headache because some folder – typically, sessions – was stuffed with millions of files. Linux is not quite happy with that, deleting the folder is not an option and the loyal “rm -rf” command decides to call it a day. To make things even worse: you want to remove only files of some days ago… what are the options?
The Linux “find” command is a possible solution, many will go for:
find /yourmagicmap/* -type f -mtime +3 -exec rm -f {} \;
The command above will give a list of files older than 3 days and will pass each found file to the rm command. The rule above has one problem though: it will take some time, since calling the rm command a million times is not exactly what you can call efficient.
A better option is:
find /yourmagicmap/* -type f -mtime +3 -delete
This adds the delete flag to the find command giving it the command to throw it all away. Do the right thing and pass along your command in a cronjob if you need to clean out the folder on a regular basis.
rsync
is without doubt one of the most handy commands when it comes to file
operations. Rsync can do any kind of volume sync – as you may know – but
it also provides a way to empty a folder.
The example below assumes
you have a folder named /tmp/empty/ which contains no files, and a
folder /tmp/session/ that contains too much crap. The rule below allows
you to remove those files:
rsync -a --delete /tmp/empty /tmp/session/
rm: deleting millions of file is a no-can-do!
find -exec: an option, but slower!
find -delete: fast and easy way to remove loads of files.
rsync –delete: without doubt the quickest!