439ccd4aa3
This is a very useful stack for local development as well, by allowing to easily access a local grafana instance and look at the accessed endpoints, query usage and durations etc. As a nice side effect this also makes sure we have an easy way to actually test any changes to the opentelemetry integration in an actual environment instead of just listening to a raw socket. |
||
|---|---|---|
| .. | ||
| config | ||
| health | ||
| scripts | ||
| ca-entrypoint.sh | ||
| cgit-entrypoint.sh | ||
| cron-entrypoint.sh | ||
| fastapi-entrypoint.sh | ||
| git-entrypoint.sh | ||
| hypercorn.env | ||
| localhost.ext | ||
| logging.conf | ||
| mariadb-entrypoint.sh | ||
| mariadb-init-entrypoint.sh | ||
| nginx-entrypoint.sh | ||
| README.md | ||
| redis-entrypoint.sh | ||
| sharness-entrypoint.sh | ||
| smartgit-entrypoint.sh | ||
| test-mysql-entrypoint.sh | ||
| tests-entrypoint.sh | ||
Aurweb and Docker
The INSTALL document details a manual Aurweb setup, but Docker images are also
provided here to avoid the complications of database configuration (and so
forth).
Setup
Naturally, both docker and docker-compose must be installed, and your Docker
service must be started:
systemctl start docker.service
The main image - aurweb - must be built manually:
docker compose build
Starting and Stopping the Services
With the above steps complete, you can bring up an initial cluster:
docker compose up
Subsequent runs will be done with start instead of up. The cluster can be
stopped with docker compose stop.
Testing
With a running cluster, execute the following in a new terminal:
docker compose run test
Generating Dummy Data
Before you can make meaningful queries to the cluster, it needs some data. Luckily such data can be generated.
docker compose exec fastapi /bin/bash
pacman -S words fortune-mod
./schema/gendummydata.py dummy.sql
mysql aurweb < dummy.sql
The generation script may prompt you to install other Arch packages before it can proceed.
Querying the RPC
The Fast (Python) API runs on Port 8444. You can query one like so:
curl -k "https://localhost:8444/rpc/?v=5&type=search&arg=python"
-k bypasses local certificate issues that curl will otherwise complain about.