oci-arm-host-capacity

Resolving Oracle Cloud "Out of Capacity" issue and getting free VPS with 4 ARM cores / 24GB of memory

Update 2024: The script is still functional, but many Reddit users now recommend upgrading to Pay As You Go (PAYG) for the best experience. With PAYG, you'll continue to enjoy all the free benefits without any additional cost, but you'll also receive priority for launching instances and are less likely to face "Out of host capacity" errors. Additionally, PAYG unlocks more types of OCI resources, including free Kubernetes-related infrastructure if that's something you're interested in. It's important to set up budget alerts as a safety net and be mindful of the resources you deploy and their associated costs. This way, you can take full advantage of PAYG while keeping your spending in check.

Very neat and useful configuration was recently announced at Oracle Cloud Infrastructure (OCI) blog as a part of Always Free tier. Sometimes it's complicated to launch an instance due to the "Out of Capacity" error. Here we're solving that issue as Oracle constantly adds capacity from time to time.

Each tenancy gets the first 3,000 OCPU hours and 18,000 GB hours per month for free to create Ampere A1 Compute instances using the VM.Standard.A1.Flex shape (equivalent to 4 OCPUs and 24 GB of memory).

This approach requires PHP 7.x or 8.x and composer installed and will call "LaunchInstance" OCI API endpoint. We'll utilise the package which I've written (and published) some time ago, here's the article.

If you prefer article style, here's a link to Medium

YouTube video instruction https://youtu.be/uzAqgjElc64 is a bit outdated regarding Configuration but still can be useful for the rest.

• Generating API key
• Installation
• Configuration
  • Create/copy .env file
  • General
  • Private key
  • Instance parameters
    • Mandatory
      • OCI_SUBNET_ID and OCI_IMAGE_ID
      • OCI_SSH_PUBLIC_KEY (SSH access)
    • Optional
• Running the script
• Periodic job setup (cron)
  • Linux / WSL
  • GitHub actions (workflows)
    • Setup
    • Read This Carefully
• How it works
• Assigning public IP address
• Troubleshooting
  • Private key issues
  • SSH key issues
• Multiple configuration support
• Conclusion

Generating API key

After logging in to OCI Console, click profile icon and then "User Settings"

Go to Resources -> API keys, click "Add API Key" button

Make sure "Generate API Key Pair" radio button is selected, click "Download Private Key" and then "Add".

Copy the contents from textarea and save it to file with a name "config". I put it together with *.pem file in newly created directory /home/ubuntu/.oci

Installation

Clone this repository

git clone https://github.com/hitrov/oci-arm-host-capacity.git

run

cd oci-arm-host-capacity/
composer install

Configuration

Create/copy .env file

Copy .env.example as .env

cp .env.example .env

You must modify .env file below. Don't push/share it as it possibly contains sensitive information.

All parameters except OCI_AVAILABILITY_DOMAIN are mandatory to be set. Please read the comments in .env file as well.

General

Region, user, tenancy, fingerprint should be taken from textarea during API key generation step. Adjust these values in .env file accordingly:

• OCI_REGION
• OCI_USER_ID
• OCI_TENANCY_ID
• OCI_KEY_FINGERPRINT

Private key

OCI_PRIVATE_KEY_FILENAME is an absolute path (including directories) or direct public accessible URL to your *.pem private key file.

Instance parameters

Mandatory

OCI_SUBNET_ID and OCI_IMAGE_ID

You must start instance creation process from the OCI Console in the browser (Menu -> Compute -> Instances -> Create Instance)

Change image and shape. For Always free AMD x64 - make sure that "Always Free Eligible" availabilityDomain label is there:

ARMs can be created anywhere within your home region.

Adjust Networking section, set "Do not assign a public IPv4 address" checkbox. If you don't have existing VNIC/subnet, please create VM.Standard.E2.1.Micro instance before doing everything.

"Add SSH keys" section does not matter for us right now. Before clicking "Create"…

…open browser's dev tools -> network tab. Click "Create" and wait a bit most probably you'll get "Out of capacity" error. Now find /instances API call (red one)…

…and right click on it -> copy as curl. Paste the clipboard contents in any text editor and review the data-binary parameter. Find subnetId, imageId and set OCI_SUBNET_ID, OCI_IMAGE_ID, respectively.

Note availabilityDomain for yourself, then read the corresponding comment in .env file regarding OCI_AVAILABILITY_DOMAIN.

OCI_SSH_PUBLIC_KEY (SSH access)

In order to have secure shell (SSH) access to the instance you need to have a keypair, besically 2 files:

• ~/.ssh/id_rsa
• ~/.ssh/id_rsa.pub

Second one (public key) contents (string) should be provided to a command below. The are plenty of tutorials on how to generate them (if you don't have them yet), we won't cover this part here.

cat ~/.ssh/id_rsa.pub

Output should be similar to

ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIFwZVQa+F41Jrb4X+p9gFMrrcAqh9ks8ATrcGRitK+R/ [email protected]

Change OCI_SSH_PUBLIC_KEY inside double quotes - paste the contents above (or you won't be able to login into the newly created instance). NB! No new lines allowed!

Optional

OCI_OCPUS and OCI_MEMORY_IN_GBS are set 4 and 24 by default. Of course, you can safely adjust them. Possible values are 1/6, 2/12, 3/18 and 4/24, respectively. Please notice that "Oracle Linux Cloud Developer" image can be created with at least 8GB of RAM (OCI_MEMORY_IN_GBS).

If for some reason your home region is running out of Always free AMD x64 (1/8 OPCU + 1GB RAM), replace values below. NB! Setting the OCI_AVAILABILITY_DOMAIN to Always Free Eligible is mandatory for non-ARM architecture!

OCI_SHAPE=VM.Standard.E2.1.Micro
OCI_OCPUS=1
OCI_MEMORY_IN_GBS=1
OCI_AVAILABILITY_DOMAIN=FeVO:EU-FRANKFURT-1-AD-2

If you don't have instances of selected shape at all, and need only one, leave the value of OCI_MAX_INSTANCES=1. When you managed to launch one and need more (or 2 from scratch), set to OCI_MAX_INSTANCES=2.

Running the script

php ./index.php

I bet that the output (error) will be similar to the one in a browser a few minutes ago

{
    "code": "InternalError",
    "message": "Out of host capacity."
}

or if you already have instances:

{
    "code": "LimitExceeded",
    "message": "The following service limits were exceeded: standard-a1-memory-count, standard-a1-core-count. Request a service limit increase from the service limits page in the console. "
}

Periodic job setup (cron)

Linux / WSL

You can now setup periodic job to run the command

Create log file:

touch /path/to/oci-arm-host-capacity/oci.log

Set permissions for PHP script to modify it:

chmod 777 /path/to/oci-arm-host-capacity/oci.log

Get full path to PHP binary

which php

Usually that's /usr/bin/php

Setup itself:

EDITOR=nano crontab -e

Add new line to execute the script every minute and append log the output:

* * * * * /usr/bin/php /path/to/oci-arm-host-capacity/index.php >> /path/to/oci-arm-host-capacity/oci.log

NB! Use absolute paths wherever possible

...and save the file (F2, press Y to confirm overwrite, then Enter).

There could be cases when cron user won't have some permissions, there're ways to solve it:

1. Setup job for root user by executing EDITOR=nano sudo crontab -e
2. Move this directory (oci-arm-host-capacity) into web server's one e.g. /usr/share/nginx/html and setup cron this way:

* * * * * curl http://server.add.re.ss/oci-arm-host-capacity/index.php >> /path/to/oci-arm-host-capacity/oci.log

You can also visit the URL above and see the same command output as by running from the shell.

GitHub actions (workflows)

In order to test the script using GitHub runners (their virtual machines) please complete Setup. NB! To avoid the ban of your Github account Read This Carefully !!!

Setup

1. Fork this repository
2. Never push .env file, it's in .gitignore for a reason
3. Instead of copying/modifying .env file, use Secrets in your own repository Settings:

https://github.com/{your-username}/oci-arm-host-capacity/settings/secrets/actions

1. Click New repository secret and set all the values (one by one) that you'd set in .env file e.g.

NB! No need to double quote any value here!

1. As for the private key, you have 2 options. Either:

• upload to any web server accessible from the Internet by using just URL or...
• upload in the bucket and Create Pre-Authenticated Request.

1. Copy and save the URL from (5) as OCI_PRIVATE_KEY_FILENAME GitHub secret.
2. Go to any other directory e.g. cd /Users/hitrov
3. git clone https://github.com/{your-username}/oci-arm-host-capacity
4. Adjust the file .github/workflows/tests.yml according to this commit, just execute:

git checkout 67fe41ebfb9f385ae1614c97b74195ea318c8db7 -- .github/workflows/tests.yml

1. Commit and push this file

git commit -m "Modify workflow to test out periodic job" .github/workflows/tests.yml
git push origin main

1. Go to https://github.com/{your-username}/oci-arm-host-capacity/actions and check how Run script job.

Here's the example https://github.com/hitrov/oci-arm-host-capacity/runs/4727904401?check_suite_focus=true

Read This Carefully

Specific GitHub Workflows commit used in the Setup take an advantage of Scheduled events
and will endlessly run the script every 5-20 minutes (how exactly often - depends on runners' availability).

NB! After you're done with testing, immediately delete .github/workflows/tests.yml (because you don't need integration tests - they're written taking into account instances that I have) and push to the main branch because infinite run actually violates the Terms of Use:

Actions should not be used for:
...
- if using GitHub-hosted runners, any ot