fix(db): make the postgres major-version migration safe and correct

The in-place pg_upgrade in the db entrypoint could crash-loop or fail on real
clusters. This reworks it:

- Idempotent, crash-safe: replace the fragile migration_in_progress marker with
  a state-driven guard on the old_data/new_data scratch dirs. An empty leftover
  means a run was interrupted before any data moved (data still intact at
  $PGDATA) so it is discarded and retried; a non-empty one means data may live
  only there, so it stops for manual recovery. Removes both the
  "mkdir: File exists" crash-loop and the silent fresh-initdb-over-live-data
  window.

- Correct install user: pg_upgrade must run as the old cluster's bootstrap
  superuser (oid 10), and the new cluster must be initialised with that same
  user. It is not necessarily $POSTGRES_USER (clusters created with the default
  "postgres" superuser plus a separate app role are common). Detect it from the
  old cluster (briefly start it and read pg_roles where oid = 10) and use it for
  both the new cluster's initdb and the pg_upgrade -U argument.

- Bump DB_ENTRYPOINT_VERSION to v3 so swarm reloads the (immutable) config.

Verified on cctest: clean 13->17, interrupted-then-retried, and prod-like
clusters whose install user is "postgres" with a separate "discourse" app role.

.drone.yml

@@ -0,0 +1,40 @@
+---
+kind: pipeline
+name: deploy to swarm-test.autonomic.zone
+steps:
+  - name: deployment
+    image: git.coopcloud.tech/coop-cloud/stack-ssh-deploy:latest
+    settings:
+      host: swarm-test.autonomic.zone
+      stack: discourse
+      generate_secrets: true
+      purge: true
+      deploy_key:
+        from_secret: drone_ssh_swarm_test
+      networks:
+        - proxy
+    environment:
+      DOMAIN: discourse.swarm-test.autonomic.zone
+      STACK_NAME: discourse
+      LETS_ENCRYPT_ENV: production
+      SECRET_DB_PASSWORD_VERSION: v1
+      DB_ENTRYPOINT_VERSION: v1
+trigger:
+  branch:
+    - main
+---
+kind: pipeline
+name: generate recipe catalogue
+steps:
+  - name: release a new version
+    image: plugins/downstream
+    settings:
+      server: https://build.coopcloud.tech
+      token:
+        from_secret: drone_abra-bot_token
+      fork: true
+      repositories:
+        - toolshed/auto-recipes-catalogue-json
+
+trigger:
+  event: tag

.env.sample

@@ -0,0 +1,21 @@
+TYPE=discourse
+
+DOMAIN=discourse.example.com
+## Domain aliases
+#EXTRA_DOMAINS=', `www.discourse.example.com`'
+LETS_ENCRYPT_ENV=production
+
+# Outgoing email
+#DISCOURSE_SMTP_HOST=
+#DISCOURSE_SMTP_PORT=
+#DISCOURSE_SMTP_USER=
+#DISCOURSE_SMTP_PROTOCOL=
+#DISCOURSE_SMTP_AUTH=
+# Set this if you send e-mails from a different domain than noreply@$DOMAIN
+#DISCOURSE_NOTIFICATION_EMAIL=$SMTP_USER
+
+# SMTP authentication
+#COMPOSE_FILE="compose.yml:compose.smtpauth.yml"
+#SECRET_SMTP_PASSWORD_VERSION=v1
+
+SECRET_DB_PASSWORD_VERSION=v1

.gitignore

@@ -1 +0,0 @@
-.env

README.md

@@ -1,24 +1,72 @@
 # Discourse
 
-Wiki Cafe's configuration for a Discourse deployment. Originally slimmed down from an `abra` [recipe](https://git.coopcloud.tech/coop-cloud/discourse) by [Co-op Cloud](https://coopcloud.tech/).
+[![Build Status](https://build.coopcloud.tech/api/badges/coop-cloud/discourse/status.svg?ref=refs/heads/main)](https://build.coopcloud.tech/coop-cloud/discourse)
 
+A platform for community discussion
 
-## Deploying the app with Docker Swarm
+<!-- metadata -->
+* **Category**: Apps
+* **Status**: 
+* **Image**: [`bitnami/discourse`](https://hub.docker.com/r/bitnami/discourse)
+* **Healthcheck**: yes
+* **Backups**: no
+* **Email**: yes
+* **Tests**: no
+* **SSO**: no
+<!-- endmetadata -->
 
-Set the environment variables from the .env file during the shell session.
+## Basic usage
 
-```
-set -a && source .env && set +a
-```
+1. Set up Docker Swarm and [`abra`]
+2. Deploy [`coop-cloud/traefik`]
+3. `abra app new discourse --secrets` (optionally with `--pass` if you'd like
+   to save secrets in `pass`)
+4. `abra app config YOURAPPDOMAIN` - be sure to change `$DOMAIN` to something that resolves to
+   your Docker swarm box
+5. `abra app deploy YOURAPPDOMAIN`
+6. Open the configured domain in your browser to finish set-up
 
-Set the secrets.
+[`abra`]: https://git.autonomic.zone/autonomic-cooperative/abra
+[`coop-cloud/traefik`]: https://git.autonomic.zone/coop-cloud/traefik
 
-```
-printf "SECRET_HERE" | docker secret create SECRET_NAME -
-```
+## To add a new admin user
 
-Deploy using the `-c` flag to specify one or multiple compose files.
+1. Login to the instance `abra app run APPNAME app sh`
+2. `cd /opt/bitnami/discourse`
+3. `RAILS_ENV=production bundle exec rake admin:create` and follow prompts.
 
-```
-docker stack deploy discourse --detach=true -c compose.yaml -c compose.smtpauth.yaml
-```
+## Install plugins
+
+1. Login to instance `abra app run APPNAME app sh`
+2. `cd /bitnami/discourse/plugins/`
+3. `git clone plugin.git` for example `https://github.com/discourse/discourse-openid-connect.git`
+4. `abra app restart YOURAPPDOMAIN app`
+
+### Events / calendar plugin
+
+We've had some luck running [discourse-events](https://github.com/paviliondev/discourse-events).
+
+## Setup Notes
+
+Until issue #1 is fixed, the default user is `user` and the default password is `bitnami123`
+
+## Postgres major version upgrades
+
+Welcome to hell.
+
+1. `abra app run YOURAPPDOMAIN db pg_dumpall -U discourse | gzip > YOURAPPDOMAIN_db_DATE.sql.gz`
+2. `abra app volume ls YOURAPPDOMAIN`, find the name of the Postgres data volume
+3. `scp` the backup to your VPS
+4. `abra app undeploy YOURAPPDOMAIN`
+5. `abra app volume rm YOURAPPDOMAIN`, choose the Postgres data volume
+6. `abra app deploy YOURAPPDOMAIN`, then `abra app undeploy YOURAPPDOMAIN`
+7. `ssh` to the VPS, run (replacing `13-alpine` with the new Postgres version)
+   `docker run -v YOURDATAVOLUME:/var/lib/postgresql/data -e POSTGRES_HOST_AUTH_METHOD=trust -it postgres:13-alpine`
+8. In another SSH session on the server, run `docker ps` to find the ID of the
+   new Postgres container, then `docker exec -it CONTAINERID bash`
+9. In the shell you just launched, run `dropdb -U discourse discourse`, then
+   `createdb -U discourse discourse`, then Ctrl+D or run `exit`
+10. In the second SSH session, run `zcat YOURAPPDOMAIN_db_DATE.sql.gz | docker exec -it CONTAINERID psql -U discourse`
+11. Exit the second SSH session
+12. Back in the first SSH session, Ctrl+C to shut down the database
+13. `abra app deploy YOURAPPDOMAIN`

abra.sh

@@ -0,0 +1,2 @@
+export DB_ENTRYPOINT_VERSION=v3
+export PG_BACKUP_VERSION=v2

compose.smtpauth.yml

@@ -1,3 +1,6 @@
+---
+version: "3.8"
+
 services:
   app:
     environment:
@@ -14,4 +17,4 @@ services:
 secrets:
   smtp_password:
     external: true
-    name: ${STACK_NAME}_smtp_password
+    name: ${STACK_NAME}_smtp_password_${SECRET_SMTP_PASSWORD_VERSION}

compose.yml

@@ -1,10 +1,13 @@
+---
+version: "3.8"
+
 services:
   app:
-    image: bitnami/discourse:3.2.2
+    image: bitnamilegacy/discourse:3.5.0
     networks:
       - proxy
       - internal
-    #entrypoint: ['tail', '-f', '/dev/null']
+    # entrypoint: ['tail', '-f', '/dev/null']
     environment:
       - ALLOW_EMPTY_PASSWORD=yes
       - DISCOURSE_DATABASE_HOST=${STACK_NAME}_db
@@ -36,18 +39,20 @@ services:
         - "traefik.http.routers.${STACK_NAME}.rule=Host(`${DOMAIN}`${EXTRA_DOMAINS})"
         - "traefik.http.routers.${STACK_NAME}.entrypoints=web-secure"
         - "traefik.http.routers.${STACK_NAME}.tls.certresolver=${LETS_ENCRYPT_ENV}"
-        - "caddy=${DOMAIN}"
-        - "caddy.reverse_proxy={{upstreams 3000}}"
-        - "caddy.tls.on_demand="
-    # healthcheck:
-    #   test: ["CMD", "curl", "-f", "http://localhost:3000"]
-    #   interval: 30s
-    #   timeout: 10s
-    #   retries: 10
-    #   start_period: 1m
+        ## Redirect from EXTRA_DOMAINS to DOMAIN
+        #- "traefik.http.routers.${STACK_NAME}.middlewares=${STACK_NAME}-redirect"
+        #- "traefik.http.middlewares.${STACK_NAME}-redirect.headers.SSLForceHost=true"
+        #- "traefik.http.middlewares.${STACK_NAME}-redirect.headers.SSLHost=${DOMAIN}"
+        - "coop-cloud.${STACK_NAME}.version=0.8.0+3.5.0"
+    healthcheck:
+      test: "ruby -e \"require 'uri'; require 'net/http'; uri = URI('http://localhost:3000/srv/status'); res = Net::HTTP.get_response(uri); if res.is_a?(Net::HTTPSuccess) then exit (0) else exit (1) end\""
+      interval: 30s
+      timeout: 10s
+      retries: 6
+      start_period: 20m
 
   db:
-    image: postgres:13
+    image: pgvector/pgvector:pg17
     networks:
       - internal
     secrets:
@@ -58,6 +63,9 @@ services:
       - source: db_entrypoint
         target: /docker-entrypoint.sh
         mode: 0555
+      - source: pg_backup
+        target: /pg_backup.sh
+        mode: 0555
     entrypoint: /docker-entrypoint.sh
     environment:
       - POSTGRES_HOST_AUTH_METHOD=trust
@@ -67,20 +75,21 @@ services:
     deploy:
       labels:
         backupbot.backup: "true"
-        backupbot.backup.pre-hook: "bash -c 'PGPASSWORD=$$(cat $${POSTGRES_PASSWORD_FILE}) pg_dump -U $${POSTGRES_USER} $${POSTGRES_DB} > /tmp/backup.sql'"
-        backupbot.backup.post-hook: "rm -rf /tmp/backup.sql"
-        backupbot.backup.path: "/tmp/backup.sql"
+        backupbot.backup.pre-hook: "/pg_backup.sh backup"
+        backupbot.backup.volumes.postgresql_data.path: "backup.sql"
+        backupbot.restore.post-hook: "/pg_backup.sh restore"
 
   redis:
-    image: redis:7.2-alpine
+    image: redis:7.4-alpine
     networks:
       - internal
     volumes:
       - 'redis_data:/data'
 
   sidekiq:
-    image: bitnami/discourse:3.2.2
+    image: bitnamilegacy/discourse:3.5.0
     networks:
+      - proxy
       - internal
     depends_on:
       - discourse
@@ -109,7 +118,7 @@ services:
 secrets:
   db_password:
     external: true
-    name: ${STACK_NAME}_db_password
+    name: ${STACK_NAME}_db_password_${SECRET_DB_PASSWORD_VERSION}
 
 volumes:
   postgresql_data:
@@ -123,6 +132,9 @@ networks:
 
 configs:
   db_entrypoint:
-    name: ${STACK_NAME}_db_entrypoint
+    name: ${STACK_NAME}_db_entrypoint_${DB_ENTRYPOINT_VERSION}
     file: entrypoint.postgres.sh.tmpl
     template_driver: golang
+  pg_backup:
+    name: ${STACK_NAME}_pg_backup_${PG_BACKUP_VERSION}
+    file: pg_backup.sh

entrypoint.postgres.sh.tmpl

@@ -2,16 +2,23 @@
 
 set -e
 
-MIGRATION_MARKER=$PGDATA/migration_in_progress
 OLDDATA=$PGDATA/old_data
 NEWDATA=$PGDATA/new_data
 
 echo "Running as $(id)"
 
-if [ -e $MIGRATION_MARKER ]; then
-  echo "FATAL: migration was started but did not complete in a previous run. manual recovery necessary"
-  exit 1
-fi
+# The migration uses $OLDDATA/$NEWDATA as scratch and removes them when it
+# finishes; a leftover *empty* one means a run was interrupted before any data
+# moved (data still intact at $PGDATA) so we clear it and retry, while a
+# *non-empty* one means data may live only there, so we stop for manual recovery.
+for scratch in $OLDDATA $NEWDATA; do
+  if [ -d "$scratch" ] && [ -n "$(ls -A "$scratch")" ]; then
+    echo "FATAL: $scratch exists and is not empty - a previous migration did not"
+    echo "complete and the data may only exist there. manual recovery necessary."
+    exit 1
+  fi
+done
+rm -rf $OLDDATA $NEWDATA
 
 if [ -f $PGDATA/PG_VERSION ]; then
   DATA_VERSION=$(cat $PGDATA/PG_VERSION)
@@ -23,22 +30,33 @@ if [ -f $PGDATA/PG_VERSION ]; then
     apt-get update && apt-get install -y --no-install-recommends \
       postgresql-$DATA_VERSION \
       && rm -rf /var/lib/apt/lists/*
+    # pg_upgrade must run as the old cluster's bootstrap superuser (the "install
+    # user", oid 10), and the new cluster must be initialised with that same
+    # user. It is not necessarily $POSTGRES_USER (e.g. clusters created with the
+    # default "postgres" superuser and a separate app role), so read it from the
+    # old cluster: briefly start it and ask, connecting as the app role we know.
+    PGBIN=/usr/lib/postgresql/$DATA_VERSION/bin
+    gosu postgres $PGBIN/pg_ctl -D $PGDATA -w \
+      -o "-c listen_addresses= -c unix_socket_directories=/tmp" start
+    INSTALL_USER=$(gosu postgres psql -h /tmp -U "$POSTGRES_USER" -d postgres -tAc \
+      "select rolname from pg_roles where oid = 10")
+    gosu postgres $PGBIN/pg_ctl -D $PGDATA -w stop
+    echo "old cluster install user: $INSTALL_USER"
     echo "shuffling around"
     gosu postgres mkdir $OLDDATA $NEWDATA
     chmod 700 $OLDDATA $NEWDATA
     mv $PGDATA/* $OLDDATA/ || true
-    touch $MIGRATION_MARKER
     echo "running initdb"
-    # abuse entrypoint script for initdb by making server error out
-    gosu postgres bash -c "export PGDATA=$NEWDATA ; /usr/local/bin/docker-entrypoint.sh --invalid-arg || true"
+    # abuse entrypoint script for initdb by making server error out; initialise
+    # the new cluster with the same superuser as the old one so pg_upgrade matches
+    gosu postgres bash -c "export PGDATA=$NEWDATA POSTGRES_USER=$INSTALL_USER ; /usr/local/bin/docker-entrypoint.sh --invalid-arg || true"
     echo "running pg_upgrade"
     cd /tmp
-    gosu postgres pg_upgrade --link -b /usr/lib/postgresql/$DATA_VERSION/bin -d $OLDDATA -D $NEWDATA -U $POSTGRES_USER
+    gosu postgres pg_upgrade --link -b /usr/lib/postgresql/$DATA_VERSION/bin -d $OLDDATA -D $NEWDATA -U $INSTALL_USER
     cp $OLDDATA/pg_hba.conf $NEWDATA/
     mv $NEWDATA/* $PGDATA
     rm -rf $OLDDATA
     rmdir $NEWDATA
-    rm $MIGRATION_MARKER
     echo "migration complete"
   fi
 fi

pg_backup.sh

@@ -0,0 +1,44 @@
+#!/bin/bash
+
+# Postgres backup/restore hook for the discourse `db` service.
+
+set -e
+
+BACKUP_FILE='/var/lib/postgresql/data/backup.sql'
+export PGPASSWORD=$(cat "${POSTGRES_PASSWORD_FILE:-/run/secrets/db_password}")
+DB_USER="${POSTGRES_USER:-discourse}"
+DB_NAME="${POSTGRES_DB:-discourse}"
+
+function backup {
+  pg_dump -U "$DB_USER" "$DB_NAME" | gzip > "$BACKUP_FILE"
+}
+
+function restore {
+  cd /var/lib/postgresql/data/
+
+  # Block all non-local connections so the running discourse app + sidekiq cannot reconnect and
+  # interfere with the drop/recreate/reimport. Restored on exit.
+  restore_hba() {
+    cat pg_hba.conf.bak > pg_hba.conf
+    rm -f pg_hba.conf.bak
+    su postgres -c 'pg_ctl reload'
+  }
+  cp pg_hba.conf pg_hba.conf.bak
+  echo 'local all all trust' > pg_hba.conf
+  su postgres -c 'pg_ctl reload'
+  trap restore_hba EXIT INT TERM
+
+  # terminate any lingering local sessions before recreate
+  # see https://stackoverflow.com/questions/5108876/kill-a-postgresql-session-connection
+  psql -U "$DB_USER" -d postgres -c \
+    "SELECT pg_terminate_backend(pid) FROM pg_stat_activity WHERE datname='${DB_NAME}' AND pid<>pg_backend_pid();"
+
+  # drop database and then recreate it
+  psql -U "$DB_USER" -d postgres -c "DROP DATABASE ${DB_NAME} WITH (FORCE);"
+  createdb -U "$DB_USER" "$DB_NAME"
+
+  # reimport data 
+  gunzip -c "$BACKUP_FILE" | psql -U "$DB_USER" -d "$DB_NAME" -1 -v ON_ERROR_STOP=1 -f -
+}
+
+$@