From 49ba42651ec1373d218a6dae3a83f2cac8ed0f81 Mon Sep 17 00:00:00 2001
From: Eric Fahlgren <ericfahlgren@gmail.com>
Date: Sun, 20 Apr 2025 12:58:43 -0700
Subject: [PATCH] rq: add a garbage collector to the worker

Implement a maintenance hook on the standard redis queue worker to do garbage
collection on expired builds.  When a result expires from the queue,
its data will be removed from the public/store/ directory at the regular
maintenance interval (default is every 600 seconds).

Signed-off-by: Eric Fahlgren <ericfahlgren@gmail.com>
---
 .github/workflows/publish.yml |  2 +-
 asu/__init__.py               |  2 +
 asu/rq.py                     | 72 +++++++++++++++++++++++++++++++++++
 podman-compose.yml            |  2 +-
 4 files changed, 76 insertions(+), 2 deletions(-)
 create mode 100644 asu/rq.py

diff --git a/.github/workflows/publish.yml b/.github/workflows/publish.yml
index e37b2e2d..18fb0e4f 100644
--- a/.github/workflows/publish.yml
+++ b/.github/workflows/publish.yml
@@ -29,7 +29,7 @@ jobs:
       - name: Set __version__ and poetry version
         run: |
           TAG="$(git describe --tags --always | awk -F"-" '{if (NF>1) {print substr($1, 2)".post"$2} else {print substr($1, 2)}}')"
-          echo "__version__ = \"$TAG\"" > asu/__init__.py
+          sed "s/__version__.*/__version__ = \"$TAG\"/" -i asu/__init__.py
           poetry version "$TAG"
 
       - name: Build and publish PyPi package
diff --git a/asu/__init__.py b/asu/__init__.py
index 6c8e6b97..5d03f98d 100644
--- a/asu/__init__.py
+++ b/asu/__init__.py
@@ -1 +1,3 @@
 __version__ = "0.0.0"
+
+from .rq import GCWorker as GCWorker
diff --git a/asu/rq.py b/asu/rq.py
new file mode 100644
index 00000000..7c842796
--- /dev/null
+++ b/asu/rq.py
@@ -0,0 +1,72 @@
+from re import compile
+from pathlib import Path
+from rq import Queue, Worker
+from rq.job import Job
+from podman import PodmanClient
+from shutil import rmtree
+
+from asu.config import settings
+from asu.util import log, get_podman
+
+REQUEST_HASH_LENGTH = 64
+store: Path = settings.public_path / "store"
+podman: PodmanClient = get_podman()
+
+
+class GCWorker(Worker):
+    """A Worker class that does periodic garbage collection on ASU's
+    public store directory.  We tie into the standard `Worker` maintenance
+    sequence, so the period is controlled by the base class.  You may change
+    the garbage collection frequency in podman-compose.yml by adding a
+    `--maintenance-interval` option to the startup command as follows (the
+    default is 600 seconds).
+
+    >>> command: rqworker ... --maintenance-interval 1800
+    """
+
+    hash_match = compile(f"^[0-9a-f]{{{REQUEST_HASH_LENGTH}}}$")
+
+    def clean_store(self) -> None:
+        """For performance testing, the store directory was mounted on a
+        slow external USB hard drive.  A typical timing result showed ~1000
+        directories deleted per second on that test system.  The synthetic
+        test directories were created containing 10 files in each.
+        File count dominated the timing, with file size being relatively
+        insignificant, likely due to `stat` calls being the bottleneck.
+        (Just for comparison, tests against store mounted on a fast SSD
+        were about twice as fast.)
+
+        >>> Cleaning /mnt/slow/public/store: deleted 5000/5000 builds
+        >>> Timing analysis for clean_store: 5.081s
+        """
+
+        deleted: int = 0
+        total: int = 0
+        dir: Path
+        queue: Queue
+        for dir in store.glob("*"):
+            if not dir.is_dir() or not self.hash_match.match(dir.name):
+                continue
+            total += 1
+            for queue in self.queues:
+                job: Job = queue.fetch_job(dir.name)
+                log.info(f"  Found {dir.name = } {job = }")
+                if job is None:
+                    rmtree(dir)
+                    deleted += 1
+
+        log.info(f"Cleaning {store}: deleted {deleted}/{total} builds")
+
+    def clean_podman(self) -> None:
+        """Reclaim space from the various podman disk entities as they are orphaned."""
+        removed = podman.containers.prune()
+        log.info(f"Reclaimed {removed.get('SpaceReclaimed', 0):,d}B from containers")
+        removed = podman.images.prune()
+        log.info(f"Reclaimed {removed.get('SpaceReclaimed', 0):,d}B from images")
+        removed = podman.volumes.prune()
+        log.info(f"Reclaimed {removed.get('SpaceReclaimed', 0):,d}B from volumes")
+
+    def run_maintenance_tasks(self):
+        super().run_maintenance_tasks()
+        self.clean_store()
+        self.clean_podman()
diff --git a/podman-compose.yml b/podman-compose.yml
index 3c467125..5c6804a6 100644
--- a/podman-compose.yml
+++ b/podman-compose.yml
@@ -22,7 +22,7 @@ services:
       context: .
       dockerfile: Containerfile
     restart: unless-stopped
-    command: rqworker --logging_level INFO
+    command: rqworker --logging_level INFO --with-scheduler --worker-class asu.GCWorker
     env_file: .env
     environment:
       REDIS_URL: "redis://redis:6379/0"