Using a file as a swap memory (SWAP)

Author Maikel Llamaret Heredia Posted on the site of GUTL.

Long ago, GNU / Linux, was limited to using a single swap memory partition of a maximum of 128 MB, something highly criticized by critics of the Linus Torvalds core.

Fortunately, today there is no such limit, and it is also possible to use as much swap memory as required to satisfy the needs of any system.

Sometimes, after installing our system we find ourselves in the need to increase the memory SWAP that we had configured during the installation, something that can be done in a simple way by increasing the size of the partition chosen for these purposes. But ... What to do if you don't want to touch the partition table?

What is SWAP and how does it work?

The swap memory space or swap, is what is known as virtual memory. The difference between real and virtual memory is that virtual memory uses hard disk space instead of a memory module.

When real memory is exhausted, the system copies part of its contents directly to this swap memory space in order to perform other tasks.

Using SWAP has the advantage of providing the additional memory needed when real memory has run out and a process needs to be done. The disadvantage is that as a result of using space on the hard disk, the use of the hard disk is slower.

Use a file as a swap memory.

This method does not require making changes to the hard disk partition table. Ideal for inexperienced users, for those who want to avoid taking risks when changing the partition table of their hard disk, or for those who require more than occasional swap memory, or in a circumstantial way.

Considering that the swap file can be placed in any directory on the hard disk, the command is executed dd, specifying that zeros will be written (if = / dev / zero) to create the file / swap (of = / swap), in blocks of 1024 bytes (bs = 1024) until completing a certain quantity in bytes (count = [quantity multiplied by the value of bs]). The following example does the above until 524288000 bytes are completed (divided by 1024 equals 512MB):

To create the file that we will use as SWAP, we will open a console and type the following (as root):

dd if=/dev/zero of=/swap bs=1024 count=512000

The above step may take a few seconds, please be patient. Then, to format the created file as swap memory, we will execute the command mkswap, as follows (always as root):

mkswap /swap

You will be returned an output in the console similar to the following:

Setting up swapspace version 1, size = 511996 KiB no label, UUID = fed2aba5-77c6-4780-9a78-4ae5e19c506b

To activate the partition, and have it used immediately by the operating system, execute the command swapon. In our case we will activate as a swap memory partition the / swap file that we created and formatted SWAP in the previous steps:

swapon /swap

To verify that the new swap file is being used by the operating system, we will execute the command again free and we will see that the capacity of the new file has been added to the initial SWAP memory.

In order for this file to be used as swap memory automatically at the next system boot, we will edit  / etc / fstab (using nano, gedit, kate, kwrite, vim or the plain text editor of your choice), adding the corresponding line, as follows, where instead of the device, the path of the swap file created is put:

We open the file

nano /etc/fstab

And we add:

/swap         swap      swap     defaults               0 0

Ready!!!!

We can reboot the system and test that our swap memory has increased with the use of the new file SWAP. Simply what we have done is create a file in the root of the system, format it SWAP and tell our GNU / Linux to use it as a swap memory, together with the partition that we already had destined for this purpose.

Something simple but that could be of some use to many of us ... Without further ado ...