2
0
mirror of https://github.com/frappe/frappe_docker.git synced 2024-12-23 18:48:58 +00:00
frappe_docker/docs/site-operations.md
Lev a9b6b755ce
Global refactoring (#617)
* Rename `bench-build` target to `bench` in bake file

* Update bake file and break everything

* Rename docker-compose.yml to compose.yml to avoid conflicting on `docker buildx bake`

* Fix groups in bake file

* Update frappe-worker

* Update frappe-nginx, erpnext-nginx

* Remove old erpnext images

* Update frappe-socketio

* Fix develop frappe-nginx build on linux/arm64

* Update dockerignore

* Update gitignore

* Update gitignore

* Update .env files

* Update installation (overrides)

* Update tests

* Fix image names

* Update compose

* Update get-latest-tags

* Update CI

* Setup and remove .env on tests

* Add build bench workflow

* Add triggers to main workflow

* Add release helm job

* Use reusable workflows

* Rollback

* Print configuration before running tests

* Show tests/.env

* Revert "Show tests/.env"

This reverts commit 4bc3bdebaf.

* Fix ci image versions

* Remove `frappe-` prefix in build directories

* Move requirements-dev.txt

* Fix image name in CI

* Update gitignore

* Update pre-commit config

* Drop `version:` in compose files

* Add push-backup

* Fix postgres CI test

* Change .yml to .yaml in compose file to follow compose-spec

* Remove prettierignore

* Fix dockerignore

* Change .yml to .yaml in compose file to follow compose-spec

* Don't depend on boto3 while testing (do it in backend)

* Update erpnext example version

* Don't fail ping on URLError

* Move assets volume to main compose file

* Fix type annotations for v12

* Fix postgres ci override in tests

* Fix spaces in socketio

* Reorder stages in nginx image, improve perfomance

* Remove unused todo

* Optimize worker build

* Install Node in worker image

* Add 502 error page

* Remove unused quiet-pull in tests

* Add configurator service to dynamically set common config

* Remove unused compose.ci-postgres.yml

* Use Python for configurator service: faster and more robust

* Add TODO.md

* Use python script to get latest tags in CI

* Clean up nginx dockerfile

* Remove VOLUME declaration

https://stackoverflow.com/a/55052682

* Add custom app example

* Remove pwd for now

* Remove pwd for now

* Use jq for parsing config in healthcheck

* Take advantage of yaml lang: add defaults in compose file. Also require env vars

* Fix CI

* Use resusable workflow

* Update

* Move release_helm job to main.yml

* Rename docker-build to docker-build-push

* Rename main to build_stable

* Rename bench targets

* Remove quotes from docker-build-push inputs

* Update build develop

* Remove HELM_DEPLOY_KEY secret from docker-build-push

* Add job names

* Remove build_bench workflow

* Update version input description in docker-build-push

* Print .env in tests, if version is develop, change to latest (for tag)

* Fix env setup

* Uncomment tests

* Parse and set short tags from git tag in bake file

* Move devcontainer settings to devcontainer.json

* Add db command notice

* Fix CI?

* Fix inconsistencies in development readme

* Remove pwd for now

* Remove custom apps for production instruction

* Update todos

* Add docs for images and compose files

* Add variables docs and allow custom frappe site name header

* Add notice about internal environment variables

* Update site-operations docs

* Update todos

* Add Overrides header in images-and-compose-files

* Update todos

* Remove extra docs

* Don't log requests in worker image (nginx already does that)

* Remove default value of FRAPPE_SITE_NAME_HEADER in example.env

* Use file that consistent in v12, v13 and develop to check /assets

* Fix paths in CI

* Update todos

* Remove TODO.md

* Update tests/_check_backup_files.py

Co-authored-by: Revant Nandgaonkar <revant.one@gmail.com>

* Change variables MINIO_ACCESS_KEY and MINIO_SECRET_KEY to S3_ACCESS_KEY, S3_SECRET_KEY in tests

* Fix S3 test

* Use `nginxinc/nginx-unprivileged` instead of `nginx` image

Also use Ngnix 1.20 instead of unstable 1.21

* Fix https override

* Update Dockerfile

* Mount assets to backend service in read only mode

* Touch .build (#307), use scripts from nginx image to generate config and touch .build

* Update example env after building stable images

* Touch `.build` on develop image (untill https://github.com/frappe/frappe/issues/15396 is resolved)

* Add `make` to worker build deps for linux/arm64

* Fix update example.env job

* Fix .build creation on develop branch

* Move bench CI to different file

This way workflow runs only on PRs that relevant to bench build

* Fix app name in custom app example

* Update erpnext and frappe versions in example.env

* Don't install `svg-sprite` and `sass` node modules in nginx image on linux/arm64 (https://github.com/frappe/frappe/pull/15275)

* docs: README and docs

* docs: add link to site operations from docker swarm

* ci: fix tests as per changes to compose.yaml

* docs: move wiki articles to docs

* docs: fix add custom domain

* docs: fix patch code from images

* fix: do not expose port 80 for old images

* fix: custom domain labels to frontend container/service

* Add missing descriptions to envs in example.env

* Fix redis depends_on

* Fix docker compose in tests when not running on TTY

* Set -T flag in `docker compose exec` only if not tty

* Run pre-commit on docs

* Remove postgres healthcheck (it gets overriden by mariadb)

* Refactor test

* Update workflow names

* Add pip to dependabot config

* docs: backup and push (#19)

* Beautify changes by @revant (#20)

* feat: add gevent to worker image

* feat: real_ip configuration for nginx

* Return `healthcheck.sh` just for tests

Co-authored-by: Lev Vereshchagin <mail@vrslev.com>

* Make pretend bench catch unknown commands (closes #666)

* Remove debug print in push-backup

* Fix typing issues in push-backup

* Update file keys in push-backups: from abs path to <site>/<file>

* Refactor push-backup

* Move gevent installation in Frappe step

* Don't pin boto stubs requirement

* Cache pip deps on build

* Update example env versions

* Refactor check backup files

* Fix backup test

* Fix backup test

* Rename build/ dir to images/

* Rename build/ dir to images/

* Fix /build -> /images in docs

* Update example.env

* Use reusable workflow in frappe user instead of vrslev

* Fix compose`s `project` option in docs (https://github.com/frappe/frappe_docker/pull/617#issuecomment-1065178792)

* Add note about project option in site-operations doc

* Update example env

* Rename build arg `USERNAME` to `REGISTRY_USER`

* Allow https proxy to access Docker socket

* Revert "Use reusable workflow in frappe user instead of vrslev"

This reverts commit 6062500d0d.

* Revert "Revert "Use reusable workflow in frappe user instead of vrslev""

This reverts commit 4680d18ff8.

Co-authored-by: Revant Nandgaonkar <revant.one@gmail.com>
2022-03-14 11:23:03 +05:30

2.9 KiB

Site operations

💡 You should setup --project-name option in docker-compose commands if you have non-standard project name.

Setup new site

Note:

  • Wait for the db service to start and configurator to exit before trying to create a new site. Usually this takes up to 10 seconds.
docker-compose exec backend bench new-site <site-name> --mariadb-root-password <db-password> --admin-password <admin-password>

If you need to install some app, specify --install-app. To see all options, just run bench new-site --help.

To create Postgres site (assuming you already use Postgres compose override) you need have to do set root_login and root_password in common config before that:

docker-compose exec backend bench set-config -g root_login <root-login>
docker-compose exec backend bench set-config -g root_password <root-password>

Also command is slightly different:

docker-compose exec backend bench new-site <site-name> --db-type postgres --admin-password <admin-password>

Push backup to S3 storage

We have the script that helps to push latest backup to S3.

docker-compose exec backend push_backup.py --site-name <site-name> --bucket <bucket> --region-name <region> --endpoint-url <endpoint-url> --aws-access-key-id <access-key> --aws-secret-access-key <secret-key>

Note that you can restore backup only manually.

Edit configs

Editing config manually might be required in some cases, one such case is to use Amazon RDS (or any other DBaaS). For full instructions, refer to the wiki. Common question can be found in Issues and on forum.

common_site_config.json or site_config.json from sites volume has to be edited using following command:

docker run --rm -it \
    -v <project-name>_sites:/sites \
    alpine vi /sites/common_site_config.json

Instead of alpine use any image of your choice.

Health check

For socketio and gunicorn service ping the hostname:port and that will be sufficient. For workers and scheduler, there is a command that needs to be executed.

docker-compose exec backend healthcheck.sh --ping-service mongodb:27017

Additional services can be pinged as part of health check with option -p or --ping-service.

This check ensures that given service should be connected along with services in common_site_config.json. If connection to service(s) fails, the command fails with exit code 1.


For reference of commands like backup, drop-site or migrate check official guide or run:

docker-compose exec backend bench --help

Migrate site

Note:

  • Wait for the db service to start and configurator to exit before trying to migrate a site. Usually this takes up to 10 seconds.
docker-compose exec backend bench --site <site-name> migrate