From 21a8ae3897a928e689514ef8dcbef0196a258989 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Sebastian=20Kr=C3=BCger?= <valknar@pivoine.art>
Date: Mon, 24 Aug 2026 12:02:47 +0200
Subject: [PATCH] Revert "fix(gitea): revert cache workarounds, rely on
 CONFIG_FILE fix instead"

This reverts commit b62b9b1871599ad05c0d292d30b2e253613a153c.
---
 gitea/compose.yml        | 39 +++++++++++++++++++++++++++++++++++++++
 gitea/runner-config.yaml | 18 ++++++++++++++++--
 2 files changed, 55 insertions(+), 2 deletions(-)

diff --git a/gitea/compose.yml b/gitea/compose.yml
index 6b66f4a..750a644 100644
--- a/gitea/compose.yml
+++ b/gitea/compose.yml
@@ -80,10 +80,49 @@ services:
       # effect despite being correctly mounted: without CONFIG_FILE set, the
       # daemon never even loaded that file.
       CONFIG_FILE: /data/config.yaml
+    # Lets this container itself resolve host.docker.internal too (used by
+    # cache.external_server in runner-config.yaml) - the daemon needs this
+    # for its own pre-registration calls to runner-cache-server, separately
+    # from job containers (which get it via container.options's --add-host).
+    extra_hosts:
+      - "host.docker.internal:host-gateway"
     volumes:
       - ../.data/gitea/runner:/data
       - /var/run/docker.sock:/var/run/docker.sock
       - ./runner-config.yaml:/data/config.yaml:ro
+      - ../.data/gitea/cache_secret:/data/cache_secret:ro
+    restart: always
+    networks:
+      - compose_network
+  runner-cache-server:
+    image: gitea/runner:latest
+    container_name: gitea_runner_cache
+    # The image's default entrypoint (run.sh) ignores any command args and
+    # always registers-then-execs `gitea-runner daemon` regardless - it has no
+    # cache-server mode. Bypass it entirely and invoke the binary directly.
+    entrypoint: ["/sbin/tini", "--"]
+    command: ["gitea-runner", "cache-server", "--config", "/data/config.yaml", "--host", "0.0.0.0", "--port", "8088", "--dir", "/data/cache"]
+    environment:
+      TZ: ${TIMEZONE:-Europe/Amsterdam}
+    ports:
+      # Published so job containers - isolated on their own per-task Docker
+      # network, unreachable via container name regardless of network
+      # settings we tried - can still reach it via host.docker.internal,
+      # which resolves to their own network's gateway and routes back to
+      # this published port. Exposed on the VPS's public interface too;
+      # requests without a valid job's bearer token are rejected, but
+      # consider a firewall rule restricting this to Docker's private
+      # bridge ranges (172.16.0.0/12) if that's a concern.
+      - "8088:8088"
+    volumes:
+      - ../.data/gitea/runner-cache:/data/cache
+      # Unlike the daemon (runner-config.yaml, external_secret_file works
+      # fine there), the cache-server subcommand only checks the literal
+      # external_secret field and errors out ("must be set") if only
+      # external_secret_file is given. So this file has to hold the actual
+      # secret value - kept out of git entirely, generated straight into
+      # .data on the VPS (see the cache_secret file it reads from).
+      - ../.data/gitea/cache-server-config.yaml:/data/config.yaml:ro
     restart: always
     networks:
       - compose_network
diff --git a/gitea/runner-config.yaml b/gitea/runner-config.yaml
index 59c0007..891c2e1 100644
--- a/gitea/runner-config.yaml
+++ b/gitea/runner-config.yaml
@@ -17,12 +17,26 @@ cache:
   dir: ""
   host: ""
   port: 0
-  external_server: ""
+  # Both the shared network approaches we tried (a custom network name, then
+  # Docker's real "bridge" network) failed to make job containers reachable
+  # from/to this daemon's own embedded cache server - its host/port config
+  # is also unreliable in this act_runner version (confirmed by others: a
+  # fixed port still gets silently ignored in favor of a random one, see
+  # https://codeberg.org/forgejo/docs/issues/996). Using a dedicated
+  # `cache-server` process instead (see runner-cache-server in compose.yml),
+  # whose --host/--port CLI flags are documented and confirmed to actually
+  # work, reached via host.docker.internal + a published fixed port (proven
+  # reachable end-to-end from an isolated per-task-style container).
+  external_server: "http://host.docker.internal:8088/"
+  external_secret_file: "/data/cache_secret"
 
 container:
   network: ""
   privileged: false
-  options: "-v /var/run/docker.sock:/var/run/docker.sock"
+  # --add-host is what makes host.docker.internal (used for
+  # cache.external_server above) resolve inside job containers on Linux -
+  # Docker only wires it up automatically on Docker Desktop (Mac/Windows).
+  options: "-v /var/run/docker.sock:/var/run/docker.sock --add-host=host.docker.internal:host-gateway"
   workdir_parent: ""
   valid_volumes: []
   docker_host: ""