Hive

How to operate the website known as Hive.

Initial setup

Add the deployment site to your ~/.ssh/config:

cat >>~/.ssh/config <<EOF
Host hivesite
Hostname (you know it)
User (you know it)
ForwardX11 no
ForwardAgent no
Compression yes
EOF

Recursively clone the repo:

git clone --recursive [email protected]:gbenson/hivesite.git
cd hivesite

The staging subdirectory now contains a local copy of what's on the live site.

Database setup

The first time the database container starts it will create a local copy of the live site's database using files in the database directory. Run the following script to create those files:

./init-database

To force a database reinitialization, do:

docker compose down
docker volume rm hivesite_database
docker compose up

See the comments in init-database if you need to create a fresh backup to initialize from.

Webserver setup

Some files need fetching from the live site. Run the following script to fetch them:

./init-webserver

Updating the the site

The sources directory contains the tarballs that were used to build the gbenson/mediawiki container image. After testing, the container's document root is extracted into the staging directory and then deployed to the live site using git. To update the site:

1. Download any new files you need into sources.
2. Edit Dockerfile with any changes you need.
3. Either build the container image locally (docker build --tag=gbenson/mediawiki .), or just push a X.Y.Z-R tag to GitHub and let the CI workflow build the image and push it to Docker Hub.
4. Run ./push-to-staging to copy everything into staging.
5. cd into staging, use git diff to make sure everything's as you want it. Go back and make changes if necessary.
6. Commit your changes, then git push them. The post-receive hook on the other end will deploy what you've pushed.
7. cd out of staging, and commit your changes there.