forked from mtusi2026/mtusi2026lab3
79 lines
4.1 KiB
Bash
Executable File
79 lines
4.1 KiB
Bash
Executable File
#!/bin/sh
|
|
# Local three-daemon Swarm for the node-loss experiment. All data are synthetic.
|
|
set -eu
|
|
cd "$(dirname "$0")/.."
|
|
|
|
manager() { docker exec lab3-manager docker "$@"; }
|
|
|
|
case "${1:-}" in
|
|
up)
|
|
test -f .env || { echo "Create .env first" >&2; exit 2; }
|
|
docker network create lab3-nodes >/dev/null
|
|
docker run -d --privileged --network lab3-nodes --name lab3-manager --hostname lab3-manager -p 127.0.0.1:18080:8080 docker:29.3-dind --host=unix:///var/run/docker.sock >/dev/null
|
|
for node in lab3-worker1 lab3-worker2; do
|
|
docker run -d --privileged --network lab3-nodes --name "$node" --hostname "$node" docker:29.3-dind --host=unix:///var/run/docker.sock >/dev/null
|
|
done
|
|
for node in lab3-manager lab3-worker1 lab3-worker2; do
|
|
count=0
|
|
until docker exec "$node" docker info >/dev/null 2>&1; do
|
|
count=$((count + 1))
|
|
test "$count" -lt 60 || { echo "$node daemon unavailable" >&2; exit 1; }
|
|
sleep 1
|
|
done
|
|
done
|
|
manager_ip=$(docker inspect -f '{{range .NetworkSettings.Networks}}{{.IPAddress}}{{end}}' lab3-manager)
|
|
manager swarm init --advertise-addr "$manager_ip" >/dev/null
|
|
join_token=$(manager swarm join-token -q worker)
|
|
for node in lab3-worker1 lab3-worker2; do
|
|
docker exec "$node" docker swarm join --token "$join_token" "$manager_ip:2377" >/dev/null
|
|
manager node update --label-add lab3_compute=true "$node" >/dev/null
|
|
done
|
|
manager node update --label-add lab3_storage=true lab3-manager >/dev/null
|
|
docker build -q -t lab3-api:local . >/dev/null
|
|
for node in lab3-manager lab3-worker1 lab3-worker2; do
|
|
docker save lab3-api:local | docker exec -i "$node" docker load >/dev/null
|
|
done
|
|
python3 -c 'from pathlib import Path; import sys; lines=Path(".env").read_text().splitlines(); values=[line.split("=",1)[1] for line in lines if line.startswith("POSTGRES_PASSWORD=")]; sys.stdout.write(values[-1])' | docker exec -i lab3-manager docker secret create lab3_db_password - >/dev/null
|
|
docker exec -i lab3-manager docker config create lab3_haproxy_conf - < infra/haproxy-swarm.cfg >/dev/null
|
|
docker exec -i lab3-manager docker stack deploy --resolve-image never -c - lab3data < infra/swarm-data.yaml
|
|
count=0
|
|
until manager service ls --format '{{.Name}} {{.Replicas}}' | grep -q '^lab3data_postgres 1/1$'; do
|
|
count=$((count + 1))
|
|
test "$count" -lt 120 || { manager service ps lab3data_postgres; exit 1; }
|
|
sleep 1
|
|
done
|
|
manager service create --detach=true --name lab3-migrate --network lab3data_data --secret lab3_db_password --env DATABASE_HOST=lab3data_postgres --env DATABASE_NAME=messenger --env DATABASE_USER=messenger --env DATABASE_PASSWORD_FILE=/run/secrets/lab3_db_password --restart-condition on-failure --restart-max-attempts 10 --limit-cpu 0.5 --limit-memory 256M lab3-api:local python -m scripts.migrate >/dev/null
|
|
count=0
|
|
until manager service ps --format '{{.CurrentState}}' lab3-migrate | head -1 | grep -q '^Complete'; do
|
|
count=$((count + 1))
|
|
test "$count" -lt 90 || { manager service ps lab3-migrate; exit 1; }
|
|
sleep 1
|
|
done
|
|
docker exec -i lab3-manager docker stack deploy --resolve-image never -c - lab3app < infra/swarm-app.yaml
|
|
echo "Swarm endpoint: http://127.0.0.1:18080 (wait for /health/ready)"
|
|
;;
|
|
status)
|
|
manager node ls
|
|
manager service ls
|
|
manager service ps lab3app_api
|
|
;;
|
|
fail-node)
|
|
case "${2:-}" in lab3-worker1|lab3-worker2) docker stop -t 1 "$2" >/dev/null ;; *) echo "Choose lab3-worker1 or lab3-worker2" >&2; exit 2 ;; esac
|
|
;;
|
|
recover-node)
|
|
case "${2:-}" in lab3-worker1|lab3-worker2) docker start "$2" >/dev/null ;; *) echo "Choose lab3-worker1 or lab3-worker2" >&2; exit 2 ;; esac
|
|
;;
|
|
rolling-update)
|
|
manager service update --force --update-parallelism 1 --update-order start-first lab3app_api
|
|
;;
|
|
down)
|
|
echo "Removing the ephemeral synthetic Swarm cluster and its data."
|
|
docker rm -fv lab3-manager lab3-worker1 lab3-worker2 >/dev/null
|
|
docker network rm lab3-nodes >/dev/null
|
|
;;
|
|
*)
|
|
echo "Usage: $0 up|status|fail-node WORKER|recover-node WORKER|rolling-update|down" >&2
|
|
exit 2
|
|
;;
|
|
esac
|