Merge remote-tracking branch 'origin/master'

scripts/src/gc-remote-memory-pressure.test.ts

@@ -137,7 +137,7 @@ MEMORY_CONFIG = {
     "minimumSwapFreeBytesAfterReclaim": 402653184,
     "targets": [
       {"id": "file-pages", "path": "/sys/fs/cgroup/system.slice/k3s.service", "reclaimBytes": 402653184, "swappiness": 0, "priority": 10},
-      {"id": "inactive-anon", "path": "/sys/fs/cgroup/system.slice/k3s.service", "reclaimBytes": 1342177280, "swappiness": 100, "priority": 20},
+      {"id": "inactive-anon", "path": "/sys/fs/cgroup/system.slice/k3s.service", "reclaimBytes": 1342177280, "swappiness": 100, "minimumInactiveAnonBytes": 1342177280, "priority": 20},
     ],
   },
 }

scripts/src/gc-remote-web-observe.py

@@ -361,6 +361,7 @@ def configured_cgroup_reclaim_targets():
         path = item.get("path")
         reclaim_bytes = item.get("reclaimBytes")
         swappiness = item.get("swappiness")
+        minimum_inactive_anon = item.get("minimumInactiveAnonBytes", 0)
         if not isinstance(target_id, str) or not re.match(r"^[A-Za-z0-9._-]{1,80}$", target_id) or target_id in seen_ids:
             _MEMORY_RECLAIM_CONFIG_ERROR = "yaml-cgroup-reclaim-target-id-invalid:%s" % index
             return []
@@ -373,6 +374,12 @@ def configured_cgroup_reclaim_targets():
         if isinstance(swappiness, bool) or not isinstance(swappiness, int) or swappiness < 0 or swappiness > 200:
             _MEMORY_RECLAIM_CONFIG_ERROR = "yaml-cgroup-reclaim-swappiness-invalid:%s" % target_id
             return []
+        if isinstance(minimum_inactive_anon, bool) or not isinstance(minimum_inactive_anon, int) or minimum_inactive_anon < 0:
+            _MEMORY_RECLAIM_CONFIG_ERROR = "yaml-cgroup-reclaim-minimum-inactive-anon-invalid:%s" % target_id
+            return []
+        if swappiness > 0 and minimum_inactive_anon < reclaim_bytes:
+            _MEMORY_RECLAIM_CONFIG_ERROR = "yaml-cgroup-reclaim-minimum-inactive-anon-too-small:%s" % target_id
+            return []
         priority = item.get("priority", index + 1)
         if isinstance(priority, bool) or not isinstance(priority, int) or priority < 1 or priority > 1000:
             _MEMORY_RECLAIM_CONFIG_ERROR = "yaml-cgroup-reclaim-priority-invalid:%s" % target_id
@@ -383,6 +390,7 @@ def configured_cgroup_reclaim_targets():
             "path": path,
             "reclaimBytes": reclaim_bytes,
             "swappiness": swappiness,
+            "minimumInactiveAnonBytes": minimum_inactive_anon,
             "priority": priority,
             "targetMemAvailableBytes": target_available,
             "minimumSwapFreeBytesAfterReclaim": swap_reserve,
@@ -416,15 +424,19 @@ def collect_cgroup_reclaim_candidates():
         elif safe_int(item.get("swappiness")) > 0 and swap_free - safe_int(item.get("reclaimBytes")) < safe_int(item.get("minimumSwapFreeBytesAfterReclaim")):
             protected_reason = "minimum-swap-free-reserve"
         current = read_cgroup_integer(path, "memory.current")
+        memory_stat = read_cgroup_memory_stat(path)
+        inactive_anon = safe_int(memory_stat.get("inactive_anon"))
+        if protected_reason is None and safe_int(item.get("swappiness")) > 0 and inactive_anon < safe_int(item.get("minimumInactiveAnonBytes")):
+            protected_reason = "minimum-inactive-anon-not-met"
         if protected_reason is not None:
             diagnostics["protected"].append({
                 "id": item.get("id"),
                 "path": path,
                 "reason": protected_reason,
                 "memoryCurrentBytes": current,
+                "inactiveAnonBytes": inactive_anon,
             })
             continue
-        memory_stat = read_cgroup_memory_stat(path)
         reclaim_bytes = safe_int(item.get("reclaimBytes"))
         candidates.append({
             "id": "cgroup-memory-reclaim:%s" % item.get("id"),
@@ -437,6 +449,7 @@ def collect_cgroup_reclaim_candidates():
             "priority": safe_int(item.get("priority")),
             "reclaimBytes": reclaim_bytes,
             "swappiness": safe_int(item.get("swappiness")),
+            "minimumInactiveAnonBytes": safe_int(item.get("minimumInactiveAnonBytes")),
             "targetMemAvailableBytes": safe_int(item.get("targetMemAvailableBytes")),
             "minimumSwapFreeBytesAfterReclaim": safe_int(item.get("minimumSwapFreeBytesAfterReclaim")),
             "memoryCurrentBytes": current,
@@ -667,7 +680,7 @@ def execute_cgroup_reclaim_candidate(candidate):
     configured = next((item for item in configured_cgroup_reclaim_targets() if item.get("id") == config_id), None)
     if configured is None:
         raise RuntimeError("cgroup reclaim candidate is no longer present in YAML allowlist")
-    for key in ["path", "reclaimBytes", "swappiness", "targetMemAvailableBytes", "minimumSwapFreeBytesAfterReclaim"]:
+    for key in ["path", "reclaimBytes", "swappiness", "minimumInactiveAnonBytes", "targetMemAvailableBytes", "minimumSwapFreeBytesAfterReclaim"]:
         if configured.get(key) != candidate.get(key):
             raise RuntimeError("cgroup reclaim candidate no longer matches YAML: %s" % key)
     path = configured.get("path")
@@ -698,6 +711,9 @@ def execute_cgroup_reclaim_candidate(candidate):
         required = safe_int(configured.get("reclaimBytes")) + safe_int(configured.get("minimumSwapFreeBytesAfterReclaim"))
         if swap_free < required:
             raise RuntimeError("cgroup reclaim would violate YAML minimum SwapFree reserve")
+        inactive_anon = safe_int(read_cgroup_memory_stat(path).get("inactive_anon"))
+        if inactive_anon < safe_int(configured.get("minimumInactiveAnonBytes")):
+            raise RuntimeError("minimum inactive anonymous memory is no longer available")
     cgroup_before = read_cgroup_integer(path, "memory.current")
     write_error = None
     try:

scripts/src/gc-remote.ts

@@ -411,6 +411,7 @@ function compactMemoryPressureCandidate(value: unknown): Record<string, unknown>
       path: candidate.path,
       reclaimBytes: candidate.reclaimBytes,
       swappiness: candidate.swappiness,
+      minimumInactiveAnonBytes: candidate.minimumInactiveAnonBytes,
       targetMemAvailableBytes: candidate.targetMemAvailableBytes,
       minimumSwapFreeBytesAfterReclaim: candidate.minimumSwapFreeBytesAfterReclaim,
       memoryCurrentBytes: candidate.memoryCurrentBytes,
@@ -446,9 +447,14 @@ function memoryPressureCandidateAllowed(value: unknown): boolean {
   const action = recordOrEmpty(candidate.action);
   if (typeof candidate.id !== "string" || candidate.id.length === 0 || positiveInteger(candidate.expectedMemoryRssBytes) === null) return false;
   if (candidate.kind === "cgroup-memory-reclaim") {
+    const swappiness = typeof candidate.swappiness === "number" && Number.isInteger(candidate.swappiness)
+      ? candidate.swappiness
+      : null;
     return typeof candidate.configId === "string" && candidate.configId.length > 0
       && typeof candidate.path === "string" && candidate.path.startsWith("/")
       && positiveInteger(candidate.reclaimBytes) !== null
+      && swappiness !== null && swappiness >= 0 && swappiness <= 200
+      && (swappiness === 0 || positiveInteger(candidate.minimumInactiveAnonBytes) !== null)
       && action.op === "cgroup-v2-memory-reclaim"
       && action.allowlist === "yaml-exact-cgroup-path";
   }