honeymoose/angular-cn
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
angular-cn/aio/aio-builds-setup/docs/vm-setup--update-docker-container.md
George Kalpakas 2d065a32ec docs(docs-infra): improve docs on creating/updating the preview server Docker image (#37015) 
This commit includes the following improvements:

- Document that the `create-image.sh` script (and by extention the
  `update-preview-server.sh` script) need to have access to a `yarn`
  executable.

- Add a note on cron jobs running in non-interactive, non-login shells
  (which affects their execution context and have different behavior vs
  running the same commands in an interactive, login shell).

- Change the Node.js and `yarn` installation instructions to ensure the
  `yarn` executable will be available on the `PATH` and not require an
  interactive, login shell (as happens, for example, when installing it
  via [nvm](https://github.com/nvm-sh/nvm)). This makes it easier to set
  up a cron job that runs the `update-preview-server.sh` script.

NOTE: The equivalent updates have been made on the GCE VM that hosts the
      PR preview server.

PR Close #37015
2020-05-18 10:24:20 -07:00

2.6 KiB
Raw Blame History

VM setup - Update docker container

Overview

Assuming you have cloned the repository containing the preview server code (as described here), you can use the update-preview-server.sh script on the VM host to update the preview server based on changes in the source code.

The script will pull the latest changes from the origin's master branch and examine if there have been any changes in files inside the preview server source code directory (see below). If there are, it will create a new image and verify that it works as expected. Finally, it will stop and remove the old docker container and image, create a new container based on the new image and start it.

The script assumes that the preview server source code is in the repository's aio/aio-builds-setup/ directory and expects the following inputs:

  • $1: HOST_REPO_DIR
  • $2: HOST_SECRETS_DIR
  • $3: HOST_BUILDS_DIR
  • $4: HOST_LOCALCERTS_DIR
  • $5: HOST_LOGS_DIR

See here for more info on what each input directory is used for.

Note 1: The script has to execute docker commands with sudo.

Note 2: The script has to execute yarn commands, so make sure yarn is on the PATH when invoking the script.

Note 3: Make sure the user that executes the script has access to update the repository.

Run the script manually

You may choose to manually run the script, when necessary. Example:

update-preview-server.sh \
    /path/to/repo \
    /path/to/secrets \
    /path/to/builds \
    /path/to/localcerts \
    /path/to/logs

Run the script automatically

You may choose to automatically trigger the script, e.g. using a cronjob. For example, the following cronjob entry would run the script every 30 minutes, update the preview server (if necessary) and log its output to update-preview-server.log (assuming the user has the necessary permissions):

# Periodically check for changes and update the preview server (if necessary)
*/30 *  * * *   /path/to/update-preview-server.sh /path/to/repo /path/to/secrets /path/to/builds /path/to/localcerts /path/to/logs >> /path/to/update-preview-server.log 2>&1

Note: Keep in mind that cron jobs run in non-interactive, non-login shells. This means that the execution context might be different compared to when running the same commands from an interactive, login shell. For example, .bashrc files are normally not sourced automatically in cron jobs. See here for more info.