This topic describes how to use a custom runtime to write functions in a Function Compute.

Background

Custom runtimes allow you to define runtime environments. In a custom runtime, you can define the runtime environment based on your business requirements.
  • Use a specified programming language, such as Rust.
  • Customize a runtime environment of a specified version for a programming language, such as Node.js 16.

Container environment

The following table shows the runtime environments that are supported by the custom runtime.
ParameterIdentifierOperating SystemArchitecture
Custom runtime (Debian 10) (in public preview)custom.debian10Debian 10x86_64
Custom RuntimecustomDebian 9x86_64

Custom runtime (Debian 10) (in public preview)

  • User permission: The function must be executed by the root user.
  • Directory permissions: All directories are writable.
  • The location of the code in the container: /code.

Custom Runtime

  • User permissions
    • If functions are created at and after 00:00:00 December 1, 2021, the functions must be executed by the root user.
    • If functions are created before 00:00:00 December 1, 2021, the functions must be executed by a non-root user.
  • Directory permissions
    • If functions are created at and after 00:00:00 December 1, 2021, all the directories are writable.
    • For the functions that are created before 00:00:00 December 1, 2021, data can be written to only the /tmp directory.
  • The location of the code in the container: /code.

Information about the runtime environment

Custom runtime (Debian10)

The following programming languages are built into a custom runtime of the Debian 10 type. You can create custom runtimes of the following languages without the need to install third-party interpreters or configure layers:

  • Python 2.7 (default Python version)
  • Python 3.10.9

Custom Runtime

A custom runtime supports the following built-in programming languages of the specified versions. You can create custom runtimes of the following languages without the need to install third-party interpreters or configure layers:
  • Python 3.7.4 (default Python version)
  • Node.js 10.16.2
  • OpenJDK 1.8.0
  • Ruby 2.7
  • Nginx 1.10.3
  • PHP 7.4.12
    Expand to see a list of PHP 7.4.12 built-in extensions.
    bcmathcalendarCore
    ctypecurldate
    domexifFFI
    fileinfofilterftp
    gdgettexthash
    iconvimagickimap
    intljsonlibxml
    mbstringmcryptmemcached
    mysqlimysqlndopenSSL
    pcntlpcrePDO
    pdo_mysqlpdo_pgsqlpdo_sqlite
    pgsqlPharposix
    protobufreadlineredis
    Reflectionsessionshmop
    SimpleXMLsoapsockets
    sodiumSPLsqlite3
    standardswoole sysvmsg
    sysvsemsysvshmtokenizer
    xmlxmlreaderxmlrpc
    xmlwriterxsl Zend OPcache
    zipzlibN/A

Use a non-built-in programming language

If you want to create a custom runtime by using a non-built-in language of the custom runtime, you must compress the parser or runtime of the language and your code file into a package and deploy the package in Function Compute. For example, if the runtime environment is Node.js 16, you must download the interpreter required for Node.js 16, compress the interpreter and your code file into a package, and then deploy the package in Function Compute. Perform the following steps:

  1. Download the Linux-x64 edition of node to the directory of the code package.
    wget http://mirrors.nju.edu.cn/nodejs/v16.14.2/node-v16.14.2-linux-x64.tar.gz -O node-v16.14.2-linux-x64.tar.gz && tar -zxvf node-v16.14.2-linux-x64.tar.gz && rm -rf node-v16.14.2-linux-x64.tar.gz
  2. Configure the startup command by which the custom runtime uses node.
    customRuntimeConfig:
      command:
        - /code/node-v16.14.2-linux-x64/bin/node
      args:
        - 'server.js'
    
    # You can also configure environment variables for the function: PATH=/code/node-v16.14.2-linux-x64/bin:/usr/local/bin/apache-maven/bin:/usr/local/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/local/ruby/bin. 
    # After the configuration is complete, you can directly use the node to start the HTTP server. 

Install dependencies

For more information about how to install the dependencies of each language in a custom runtime, see the following documentation.
You can also use the layer feature of Function Compute to install dependencies. We recommend that you use an official common layer or build a dependency layer online. For more information, see the following topics: