Create a swapfile for your Linux server

Documentation Self-Hosting
1

:bulb: For servers with <= 2GB of RAM, running ./discourse-setup will prompt for and automatically create a 2GB swapfile.

Most cloud virtual machine providers do not set up swapfiles as part of their server provisioning.

In particular, upgrading Discourse produces a lot of memory pressure. With a swap file, rather than randomly terminating processes with an out of memory error, things will slow down instead. Having a swap file is a cheap insurance policy that protects you from many other load related failures.

This can be done at any time from the command line on your server.

Set up a 2GB swap file

We recommend a 2GB swap file for Discourse, unless your server has 4GB or more of memory.

In the shell you have opened to your droplet, do the following:

:warning: You will need to be root for much of this, so either be root or sudo -s before running these commands! Don’t forget to exit root afterwards.

  1. Create an empty swapfile

     install -o root -g root -m 0600 /dev/null /swapfile

  2. write out a 2GB file named ‘swapfile’

     dd if=/dev/zero of=/swapfile bs=1k count=2048k

  3. tell linux this is the swap file:

     mkswap /swapfile

  4. Activate it

     swapon /swapfile

  5. Add it to the file system table so its there after reboot:

     echo "/swapfile       swap    swap    auto      0       0" | tee -a /etc/fstab

  6. Set the swappiness to 10 so its only uses as an emergency buffer

     sysctl -w vm.swappiness=10
 echo vm.swappiness = 10 | tee -a /etc/sysctl.conf

The whole thing as a single copy and pastable script:

install -o root -g root -m 0600 /dev/null /swapfile
dd if=/dev/zero of=/swapfile bs=1k count=2048k
mkswap /swapfile
swapon /swapfile
echo "/swapfile       swap    swap    auto      0       0" | tee -a /etc/fstab
sysctl -w vm.swappiness=10
echo vm.swappiness = 10 | tee -a /etc/sysctl.conf

Last edited by @JammyDodger 2022-05-23T18:45:59Z

Check documentPerform check on document:
94 Likes
77

I’ve been trying to set up a larger swapfile (4GB instead of the default 2GB), but have run into this when running the script above:

rm: cannot remove `swapfile’: Operation not permitted

To get around this, I’ve had to do this first:

  1. Reboot the server (it wouldn’t let me swapoff, presumably due to heavy swapfile use)
  2. swapoff -a -v
3 Likes
78

For similar reasons, I run with two swapfiles:

# swapon
NAME                       TYPE  SIZE  USED PRIO
/var/local/swap/swapfile.0 file 1024M 44.9M   -3
/var/local/swap/swapfile.1 file 1024M 1024M   -2
79

A second swap file could be a clever way to avoid the reboot. Just create a second swap file, enable it, and then remove the first.

4 Likes
80

So we won’t be needing to follow this guide to make the 2GB swap file since the installation already does that if we have less that 2GB of memory?

81

Yes - but if you change your server config, or perhaps if you do a migration, you might need to look after this yourself. (At the same time there are two important kernel configs to change which the installation does not look after - see MKJ’s Opinionated Discourse Deployment Configuration

82

It looks like 2+2 may not be enough any more… I’m managing a fairly unassuming (no big/fancy plugins, etc) Discourse instance that, as of today, is failing to bootstrap because ember is chewing all the RAM it can find, and all the swap, and grinding the machine into unresponsiveness. Adding another 2GB of swap allowed the bootstrap to complete, with a peak swap usage of around 2.5GB.

7 Likes
83

Yikes, this is on @david’s list to investigate.

4 Likes
84

@david has been investigating, we do confirm that as it stands 2gb is enough for docker rebuilds, but not enough for the web upgrader to work.

One idea I have tossed around is just shutting down all ruby processes during the web upgrader to save an extra 300-500mb which would leave enough for asset precompilation.

A long term approach we are likely going to need to go down for self hosters is shipping bootstrapped containers which is a pandoras box cause how would a web upgrader be able to pull that one off. We don’t want to mount docker sockets …

it sure is a pickle.

2 Likes
85

Well, it wasn’t for me.

Is this compared between basic pure install against real world situations?

86

Indeed, it’s not perfectly consistent. Even with everything else shut down, it can still fail.

Unfortunately we’re fighting a losing battle against modern JS build tooling here. It’s all designed to run on 8GB+ of Ram on modern developer machines, not 1GB VPSs :cry:

We do have some solutions in mind. For example: providing pre-built assets which can be automatically downloaded. The big challenge we have is plugins, because they vary on everyone’s site, and right now they’re integrated tightly into the core build.

But for now, doing a full CLI rebuild should have a higher success rate than a web-ui update.

6 Likes