fix(backup): block reconnections via pg_hba during pg restore (app reconnect broke reimport)

The db service dumped the DB on backup (pg_dump pre-hook) but shipped no
restore hook, and a file-level restore does not reload into the running
postgres, so a restored backup silently kept the live (un-restored) state.
Add pg_backup.sh (backup=pg_dump|gzip into the postgresql_data volume;
restore=terminate conns + DROP DATABASE WITH FORCE + createdb + reimport),
mount it via a config, and wire the backupbot backup/restore hooks. Same
fix as the immich / mattermost-lts / ghost recipes.

.env.sample

@@ -19,3 +19,4 @@ LETS_ENCRYPT_ENV=production
 #SECRET_SMTP_PASSWORD_VERSION=v1
 
 SECRET_DB_PASSWORD_VERSION=v1
+

compose.yml

@@ -3,7 +3,7 @@ version: "3.8"
 
 services:
   app:
-    image: bitnamilegacy/discourse:3.5.0
+    image: bitnamilegacy/discourse:3.3.1
     networks:
       - proxy
       - internal
@@ -43,7 +43,7 @@ services:
         #- "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"
+        - "coop-cloud.${STACK_NAME}.version=0.8.0+3.3.1"
     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
@@ -52,7 +52,7 @@ services:
       start_period: 20m
 
   db:
-    image: postgres:16
+    image: postgres:13
     networks:
       - internal
     secrets:
@@ -87,7 +87,7 @@ services:
       - 'redis_data:/data'
 
   sidekiq:
-    image: bitnamilegacy/discourse:3.5.0
+    image: bitnamilegacy/discourse:3.3.1
     networks:
       - proxy
       - internal