Using SWAP on AWS EC2
By default, all Linux instances of EC2 come without swap. I’m not sure what the reasons for that are, BUT in a few situations it is important to have swap, instead of scaling up to an instance with more main memory. For instance, PHP Composer might break on an EC2 micro instance, due to the lack of available memory.
A very practical way of having swap is to create it in a file. This way, you don’t have to partition your EBS volume or add a new one. If you are using encrypted volumes and/or it is important to keep all data at rest encrypted, then create the swap file inside an encrypted volume.
Before executing this tutorial, back up all your data and/or take a snapshot of the machine.
To configure swap on a file, in a way that is persistent across reboots, do the following steps:
- Log in as root
sudo -i - Create the swap file (preferably inside an encrypted partition), say
/opt/swap
dd if=/dev/zero of=/opt/swap bs=1M count=1024. In this command we are creating a file with 1GB (1024 blocks of 1MB). Change/opt/swapto the desired path - Format the file as a swap partition
mkswap /opt/swap - Fix possible permission problems
chmod 600 /opt/swap - Turn on the swap using the new file
swapon /opt/swap - At this point, if you run top/htop you will be able to see the swap partition
already being used but if you reboot your instance, it will come up again
without the swap. To make it persistent across reboots, edit the
/etc/fstabfile to add the line/opt/swap swap swap defaults 0 0. It is important to notice that the partition where the swap file is located must be configured to mount on boot
Done. Now, you can reboot your instance and your swap will still be working on the next boot.