From bc4e1953f54bd1941ddcbb4712aff9e5e1ad1380 Mon Sep 17 00:00:00 2001
From: Nick Sweeting <git@sweeting.me>
Date: Thu, 27 Aug 2026 17:08:10 -0700
Subject: [PATCH] keep maintenance runner out of regular work

---
 archivebox/services/runner.py    | 49 ++++++++++++++++----------------
 archivebox/tests/test_cli_run.py | 37 ++++++++++++++++++++++++
 2 files changed, 62 insertions(+), 24 deletions(-)

diff --git a/archivebox/services/runner.py b/archivebox/services/runner.py
index ed13c102..5035ef48 100644
--- a/archivebox/services/runner.py
+++ b/archivebox/services/runner.py
@@ -2313,31 +2313,32 @@ def run_pending_crawls(
         ):
             continue
 
-        # Broad final-state maintenance is intentionally a fallback. Specific
-        # queued plugin work above can use ArchiveResult's scheduler indexes;
-        # this branch may need to prove that no due sealed snapshot remains, so
-        # avoid paying that scan while targeted work is already available.
-        sealed_snapshots = Snapshot.objects.filter(
-            retry_at__lte=timezone.now(),
-            status=Snapshot.StatusChoices.SEALED,
-        )
-        if search_plugin_names:
-            queued_search_snapshot_ids = ArchiveResult.objects.filter(
-                status=ArchiveResult.StatusChoices.QUEUED,
-                plugin__in=search_plugin_names,
-            ).values("snapshot_id")
-            sealed_snapshots = sealed_snapshots.exclude(
-                id__in=queued_search_snapshot_ids,
+        if not maintenance_only:
+            # Broad final-state maintenance is intentionally a fallback. Specific
+            # queued plugin work above can use ArchiveResult's scheduler indexes;
+            # this branch may need to prove that no due sealed snapshot remains, so
+            # avoid paying that scan while targeted work is already available.
+            sealed_snapshots = Snapshot.objects.filter(
+                retry_at__lte=timezone.now(),
+                status=Snapshot.StatusChoices.SEALED,
             )
-        if crawl_id:
-            sealed_snapshots = sealed_snapshots.filter(crawl_id=crawl_id)
-        if _run_due_snapshot_query(
-            sealed_snapshots,
-            lock_seconds=60,
-            interactive_interrupts=interactive_interrupts,
-            runtime_config=runtime_config,
-        ):
-            continue
+            if search_plugin_names:
+                queued_search_snapshot_ids = ArchiveResult.objects.filter(
+                    status=ArchiveResult.StatusChoices.QUEUED,
+                    plugin__in=search_plugin_names,
+                ).values("snapshot_id")
+                sealed_snapshots = sealed_snapshots.exclude(
+                    id__in=queued_search_snapshot_ids,
+                )
+            if crawl_id:
+                sealed_snapshots = sealed_snapshots.filter(crawl_id=crawl_id)
+            if _run_due_snapshot_query(
+                sealed_snapshots,
+                lock_seconds=60,
+                interactive_interrupts=interactive_interrupts,
+                runtime_config=runtime_config,
+            ):
+                continue
 
         if not maintenance_only:
             if _run_due_crawl_status(
diff --git a/archivebox/tests/test_cli_run.py b/archivebox/tests/test_cli_run.py
index 81905274..8efd66b2 100644
--- a/archivebox/tests/test_cli_run.py
+++ b/archivebox/tests/test_cli_run.py
@@ -1945,6 +1945,43 @@ class TestRunDueCrawlState:
         assert crawl.retry_at == now
         assert crawl.snapshot_set.count() == 0
 
+    def test_maintenance_only_runner_ignores_disabled_queued_results_on_sealed_snapshots(self):
+        from django.utils import timezone
+
+        from archivebox.base_models.models import get_or_create_system_user_pk
+        from archivebox.core.models import ArchiveResult, Snapshot
+        from archivebox.crawls.models import Crawl
+        from archivebox.services.runner import run_pending_crawls
+
+        now = timezone.now()
+        crawl = Crawl.objects.create(
+            urls="https://example.com/disabled-result",
+            created_by_id=get_or_create_system_user_pk(),
+            status=Crawl.StatusChoices.SEALED,
+            retry_at=None,
+        )
+        snapshot = Snapshot.objects.create(
+            url="https://example.com/disabled-result",
+            crawl=crawl,
+            status=Snapshot.StatusChoices.SEALED,
+            retry_at=now,
+        )
+        result = ArchiveResult.objects.create(
+            snapshot=snapshot,
+            plugin="disabled_plugin",
+            hook_name="on_Snapshot__50_disabled",
+            status=ArchiveResult.StatusChoices.QUEUED,
+        )
+
+        assert run_pending_crawls(daemon=False, maintenance_only=True) == 0
+
+        snapshot.refresh_from_db()
+        result.refresh_from_db()
+        assert snapshot.status == Snapshot.StatusChoices.SEALED
+        assert snapshot.fs_version == Snapshot._fs_current_version()
+        assert snapshot.retry_at is not None
+        assert result.status == ArchiveResult.StatusChoices.QUEUED
+
     def test_snapshot_start_writes_short_future_lease(self):
         from django.utils import timezone