You can use this reference to learn how to update any of the kernel / userspace limits.

Increase max number of open files nofile ubuntu (and other os's)

Fun fact: fs.file-max is probably not what you want to change, you probably want to change ulimit. But why? The deault fs.file-max for the kernel on ubuntu is already pretty high: (9223372036854775807) You can check for yourself by running sysctl -a | grep 'fs.file-max'. But that's set at the kernel level. file-max "sets the maximim number of file-handles that the Linux Kernel will allocate". What you probably want to change is the ulimit for the userspace level (does the u stand for 'user space'? Probably. In operating systems there is a disinction between user-space, and kernel space- note this is userspace not 'a user').

So how to change the ulimit on Ubuntu in user space?

Did you know there's two vaues for each? A 'hard' value and
a 'soft' value. Often I make the mistake of changing the 'hard' value, then checking the 'soft' value which hasn't changed.

Check current value for hard and soft value

Display current soft limit of open files:

ulimit -Sn

Where -S means 'show me the soft value',
and n means 'nofiles' , which means 'number of files'
How do I know that? (See help ulimit):

help ulimit
... there's more...
-l	the maximum size a process may lock into memory
-m	the maximum resident set size
-n	the maximum number of open file descriptors
-p	the pipe buffer size
-q	the maximum number of bytes in POSIX message queues
... there's more...

Display current hard limit of open files:

ulimit -Hn

Example checking both:

$ ulimit -Sn
$ ulimit -Hn

How to increase the limit

vi /etc/security/limits.conf

(read the note at the top of the file):
If root is the user you want to increase limit for, add:

root             hard    nofile          9999

You might want to put both a hard and soft limit:

root             hard    nofile          9999
root             soft    nofile          9999

I would like to draw your attention to read all of the documentation at the top of /etc/security/limits.conf.

   - NOTE: group and wildcard limits are not applied to root.
     To apply a limit to the root user, <domain> must be
    the literal username root.

Do I need to reboot after this?

No. But you do need to kill your current proccess, which means
if you have an ssh connection and shell open, then you need to exit and re-connect. Why? Because only new processes
will get the new limit. Think of the process tree (type pstree), child processes inherit permissions and rules of their parent (usually) , so you won't see your change until you have a brand new parent process.

Someone on the internet said I need to edit /etc/pam.d/commons-session. Do I really need to do that?

No. That information appears to be out of date. This information will also be out of date eventually :)

Don't let poets lie to you.

Why is this hard? How to make it easier

Hopefully this document helped. For me the disinction between 'kernel space' and 'user space' helped a lot. But also:

The manual is hard to find if you're like me and use man all the time

Where is the manual?
I tried man ulimit - does not work
Instead: help ulimit - does work
Why does help ulimit work?

ulimit and disown are Bash built in functions and those don't have a manpage of its own but are described in the Bash manpage. src

The command options, and names are short, so it's impossible to instantly know which is which

But you can find out what all the single letter options mean by typing help ulimit

What is the difference between /etc/sysctl.conf and /etc/security/limits.conf I'm confused they seem the same to me?

You're not the only person to ask this question:
Here you can find posts spanning 5 years with two authors copy-pasting the same views without any reference to the source:

Stephen provides most useful answer because it explains these setting affect the linux kernel, and links to the reference:

In summary, this is the official Documentation for /proc/sys/fs/ and edits to the file /etc/sysctl.conf (once reloaded) will write to /proc/sys/fs/, for example /proc/sys/fs/file-max. Which provides an interface to kernel data structures (such as the maximum number of files the kernel will allow to be open).

/proc (see man proc, is a virtual in-memory filesystem

It's not you it's me

This is a lower level tuning part of the operating system. :) stick at it

From the help ulimit

Not all options are available on all platforms, which makes sense, you might be running on hardware which does not have a network card for example, so /proc/net will probably have less valid options to choose from.

Good luck!

If you missed it, the heading (near the top) if you want to just-get-it-done is "How to increase the limit".