wnm 0.0.9__py3-none-any.whl → 0.0.10__py3-none-any.whl

wnm/executor.py

@@ -0,0 +1,1292 @@
+"""Action executor for performing node lifecycle operations.
+
+This module contains the ActionExecutor class which takes planned actions
+from the DecisionEngine and executes them using ProcessManager abstractions.
+"""
+
+import logging
+import os
+import shutil
+import subprocess
+import time
+from typing import Any, Dict, List, Optional
+
+from packaging.version import Version
+from sqlalchemy import func, insert, select, text
+from sqlalchemy.orm import scoped_session
+
+from wnm.actions import Action, ActionType
+from wnm.common import (
+    DEAD,
+    DISABLED,
+    METRICS_PORT_BASE,
+    PORT_MULTIPLIER,
+    REMOVING,
+    RESTARTING,
+    RUNNING,
+    STOPPED,
+    UPGRADING,
+)
+from wnm.config import LOG_DIR
+from wnm.models import Machine, Node
+from wnm.process_managers.factory import get_default_manager_type, get_process_manager
+from wnm.utils import (
+    get_antnode_version,
+    parse_service_names,
+    update_nodes,
+)
+
+
+class ActionExecutor:
+    """Executes planned actions on nodes.
+
+    The ActionExecutor takes Action objects from the DecisionEngine and
+    performs the actual operations by calling utility functions and
+    managing database state.
+    """
+
+    def __init__(self, session_factory: scoped_session):
+        """Initialize the action executor.
+
+        Args:
+            session_factory: SQLAlchemy session factory for database operations
+        """
+        self.S = session_factory
+
+    def _get_process_manager(self, node: Node):
+        """Get the appropriate process manager for a node.
+
+        Args:
+            node: Node database record
+
+        Returns:
+            ProcessManager instance for the node's manager type
+        """
+        # Get manager type from node, or use machine config default
+        manager_type = getattr(node, "manager_type", None)
+        return get_process_manager(manager_type)
+
+    def _set_node_status(self, node_id: int, status: str) -> bool:
+        """Update node status in database.
+
+        Args:
+            node_id: ID of the node
+            status: New status to set
+
+        Returns:
+            True if status was updated successfully
+        """
+        try:
+            with self.S() as session:
+                session.query(Node).filter(Node.id == node_id).update(
+                    {"status": status, "timestamp": int(time.time())}
+                )
+                session.commit()
+            return True
+        except Exception as e:
+            logging.error(f"Failed to set node status for {node_id}: {e}")
+            return False
+
+    def _upgrade_node_binary(self, node: Node, new_version: str) -> bool:
+        """Upgrade a node's binary and restart it.
+
+        Args:
+            node: Node to upgrade
+            new_version: Version string for the new binary
+
+        Returns:
+            True if upgrade succeeded
+        """
+        # Source binary location
+        source_binary = os.path.expanduser("~/.local/bin/antnode")
+
+        # Copy new binary to node directory
+        try:
+            shutil.copy2(source_binary, node.binary)
+            os.chmod(node.binary, 0o755)
+            logging.info(f"Copied new binary from {source_binary} to {node.binary}")
+        except (OSError, shutil.Error) as err:
+            logging.error(f"Failed to copy binary for upgrade: {err}")
+            return False
+
+        # Restart the node with new binary
+        manager = self._get_process_manager(node)
+        if not manager.restart_node(node):
+            logging.error(f"Failed to restart node {node.id} during upgrade")
+            return False
+
+        # Update status to UPGRADING
+        with self.S() as session:
+            session.query(Node).filter(Node.id == node.id).update(
+                {
+                    "status": UPGRADING,
+                    "timestamp": int(time.time()),
+                    "version": new_version,
+                }
+            )
+            session.commit()
+
+        return True
+
+    def execute(
+        self,
+        actions: List[Action],
+        machine_config: Dict[str, Any],
+        metrics: Dict[str, Any],
+        dry_run: bool = False,
+    ) -> Dict[str, Any]:
+        """Execute a list of actions.
+
+        Args:
+            actions: List of Action objects to execute
+            machine_config: Machine configuration dictionary
+            metrics: Current system metrics
+            dry_run: If True, log actions without executing them
+
+        Returns:
+            Dictionary with execution status and results
+        """
+        if not actions:
+            return {"status": "no-actions", "results": []}
+
+        results = []
+
+        for action in actions:
+            logging.info(
+                f"Executing: {action.type.value} (priority={action.priority}, reason={action.reason})"
+            )
+
+            try:
+                result = self._execute_action(action, machine_config, metrics, dry_run)
+                results.append(result)
+            except Exception as e:
+                logging.error(f"Failed to execute {action.type.value}: {e}")
+                results.append(
+                    {"action": action.type.value, "success": False, "error": str(e)}
+                )
+
+        # Return status from the first (highest priority) action
+        if results:
+            return results[0]
+        return {"status": "no-results"}
+
+    def _execute_action(
+        self,
+        action: Action,
+        machine_config: Dict[str, Any],
+        metrics: Dict[str, Any],
+        dry_run: bool,
+    ) -> Dict[str, Any]:
+        """Execute a single action.
+
+        Args:
+            action: The action to execute
+            machine_config: Machine configuration
+            metrics: Current metrics
+            dry_run: If True, log without executing
+
+        Returns:
+            Dictionary with execution result
+        """
+        if action.type == ActionType.RESURVEY_NODES:
+            return self._execute_resurvey(machine_config, dry_run)
+
+        elif action.type == ActionType.REMOVE_NODE:
+            return self._execute_remove_node(action, dry_run)
+
+        elif action.type == ActionType.STOP_NODE:
+            return self._execute_stop_node(machine_config, dry_run)
+
+        elif action.type == ActionType.UPGRADE_NODE:
+            return self._execute_upgrade_node(metrics, dry_run)
+
+        elif action.type == ActionType.START_NODE:
+            return self._execute_start_node(metrics, dry_run)
+
+        elif action.type == ActionType.ADD_NODE:
+            return self._execute_add_node(machine_config, metrics, dry_run)
+
+        elif action.type == ActionType.SURVEY_NODES:
+            return self._execute_survey(dry_run)
+
+        else:
+            logging.warning(f"Unknown action type: {action.type}")
+            return {"status": "unknown-action", "action": action.type.value}
+
+    def _execute_resurvey(
+        self, machine_config: Dict[str, Any], dry_run: bool
+    ) -> Dict[str, Any]:
+        """Execute node resurvey after system reboot."""
+        if dry_run:
+            logging.warning("DRYRUN: System rebooted, survey nodes")
+        else:
+            update_nodes(self.S)
+            # Update the last stopped time
+            with self.S() as session:
+                session.query(Machine).filter(Machine.id == 1).update(
+                    {"last_stopped_at": int(time.time())}
+                )
+                session.commit()
+
+        return {"status": "system-rebooted"}
+
+    def _execute_remove_node(self, action: Action, dry_run: bool) -> Dict[str, Any]:
+        """Execute node removal.
+
+        If reason contains 'dead', remove all dead nodes.
+        Otherwise, remove youngest stopped or running node based on reason.
+        """
+        if "dead" in action.reason.lower():
+            # Remove all dead nodes
+            if dry_run:
+                logging.warning("DRYRUN: Remove Dead Nodes")
+            else:
+                with self.S() as session:
+                    broken = session.execute(
+                        select(Node)
+                        .where(Node.status == DEAD)
+                        .order_by(Node.timestamp.asc())
+                    ).all()
+
+                for row in broken:
+                    node = row[0]
+                    logging.info(f"Removing dead node {node.id}")
+                    manager = self._get_process_manager(node)
+                    manager.remove_node(node)
+                    # Delete from database immediately (no delay for dead nodes)
+                    with self.S() as session:
+                        session.delete(node)
+                        session.commit()
+
+            return {"status": "removed-dead-nodes"}
+
+        elif "stopped" in action.reason.lower():
+            # Remove youngest stopped node
+            with self.S() as session:
+                youngest = session.execute(
+                    select(Node).where(Node.status == STOPPED).order_by(Node.age.desc())
+                ).first()
+
+            if youngest:
+                if dry_run:
+                    logging.warning("DRYRUN: Remove youngest stopped node")
+                else:
+                    node = youngest[0]
+                    manager = self._get_process_manager(node)
+                    manager.remove_node(node)
+                    # Delete from database immediately (no delay for stopped nodes)
+                    with self.S() as session:
+                        session.delete(node)
+                        session.commit()
+                return {"status": "removed-stopped-node"}
+            else:
+                return {"status": "no-stopped-nodes-to-remove"}
+
+        else:
+            # Remove youngest running node (with delay)
+            with self.S() as session:
+                youngest = session.execute(
+                    select(Node).where(Node.status == RUNNING).order_by(Node.age.desc())
+                ).first()
+
+            if youngest:
+                if dry_run:
+                    logging.warning("DRYRUN: Remove youngest running node")
+                else:
+                    node = youngest[0]
+                    manager = self._get_process_manager(node)
+                    manager.stop_node(node)
+                    # Mark as REMOVING (will be deleted later after delay)
+                    self._set_node_status(node.id, REMOVING)
+                return {"status": "removed-running-node"}
+            else:
+                return {"status": "no-running-nodes-to-remove"}
+
+    def _execute_stop_node(
+        self, machine_config: Dict[str, Any], dry_run: bool
+    ) -> Dict[str, Any]:
+        """Execute node stop (to reduce resource usage)."""
+        with self.S() as session:
+            youngest = session.execute(
+                select(Node).where(Node.status == RUNNING).order_by(Node.age.desc())
+            ).first()
+
+        if youngest:
+            if dry_run:
+                logging.warning("DRYRUN: Stopping youngest node")
+            else:
+                node = youngest[0]
+                manager = self._get_process_manager(node)
+                manager.stop_node(node)
+                self._set_node_status(node.id, STOPPED)
+                # Update the last stopped time
+                with self.S() as session:
+                    session.query(Machine).filter(Machine.id == 1).update(
+                        {"last_stopped_at": int(time.time())}
+                    )
+                    session.commit()
+            return {"status": "stopped-node"}
+        else:
+            return {"status": "no-nodes-to-stop"}
+
+    def _execute_upgrade_node(
+        self, metrics: Dict[str, Any], dry_run: bool
+    ) -> Dict[str, Any]:
+        """Execute node upgrade (oldest running node with outdated version)."""
+        with self.S() as session:
+            oldest = session.execute(
+                select(Node)
+                .where(Node.status == RUNNING)
+                .where(Node.version != metrics["antnode_version"])
+                .order_by(Node.age.asc())
+            ).first()
+
+        if oldest:
+            if dry_run:
+                logging.warning("DRYRUN: Upgrade oldest node")
+            else:
+                node = oldest[0]
+                # If we don't have a version number from metadata, grab from binary
+                if not node.version:
+                    node.version = get_antnode_version(node.binary)
+
+                # Perform the upgrade (copies binary, restarts, sets UPGRADING status)
+                if not self._upgrade_node_binary(node, metrics["antnode_version"]):
+                    return {"status": "upgrade-failed"}
+
+            return {"status": "upgrading-node"}
+        else:
+            return {"status": "no-nodes-to-upgrade"}
+
+    def _execute_start_node(
+        self, metrics: Dict[str, Any], dry_run: bool
+    ) -> Dict[str, Any]:
+        """Execute starting a stopped node (may upgrade first if needed)."""
+        with self.S() as session:
+            oldest = session.execute(
+                select(Node).where(Node.status == STOPPED).order_by(Node.age.asc())
+            ).first()
+
+        if oldest:
+            node = oldest[0]
+            # If we don't have a version number from metadata, grab from binary
+            if not node.version:
+                node.version = get_antnode_version(node.binary)
+
+            # If the stopped version is old, upgrade it (which also starts it)
+            if Version(metrics["antnode_version"]) > Version(node.version):
+                if dry_run:
+                    logging.warning("DRYRUN: Upgrade and start stopped node")
+                else:
+                    # Perform the upgrade (copies binary, restarts, sets UPGRADING status)
+                    if not self._upgrade_node_binary(node, metrics["antnode_version"]):
+                        return {"status": "failed-upgrade"}
+                return {"status": "upgrading-stopped-node"}
+            else:
+                if dry_run:
+                    logging.warning("DRYRUN: Start stopped node")
+                    return {"status": "starting-node"}
+                else:
+                    manager = self._get_process_manager(node)
+                    if manager.start_node(node):
+                        self._set_node_status(node.id, RESTARTING)
+                        return {"status": "started-node"}
+                    else:
+                        return {"status": "failed-start-node"}
+        else:
+            return {"status": "no-stopped-nodes"}
+
+    def _execute_add_node(
+        self, machine_config: Dict[str, Any], metrics: Dict[str, Any], dry_run: bool
+    ) -> Dict[str, Any]:
+        """Execute adding a new node."""
+        if dry_run:
+            logging.warning("DRYRUN: Add a node")
+            return {"status": "add-node"}
+
+        # Find next available node ID (look for holes first)
+        # First check if node 1 exists
+        with self.S() as session:
+            node_1_exists = session.execute(
+                select(Node.id).where(Node.id == 1)
+            ).first()
+
+        if not node_1_exists:
+            # Node 1 is available, use it
+            node_id = 1
+        else:
+            # Look for holes in the sequence
+            sql = text(
+                "select n1.id + 1 as id from node n1 "
+                + "left join node n2 on n2.id = n1.id + 1 "
+                + "where n2.id is null "
+                + "and n1.id <> (select max(id) from node) "
+                + "order by n1.id;"
+            )
+            with self.S() as session:
+                result = session.execute(sql).first()
+
+            if result:
+                node_id = result[0]
+            else:
+                # No holes, use max + 1
+                with self.S() as session:
+                    result = session.execute(
+                        select(Node.id).order_by(Node.id.desc())
+                    ).first()
+                node_id = result[0] + 1 if result else 1
+
+        # Determine the appropriate manager type for this system
+        manager_type = get_default_manager_type()
+
+        # Create node object
+        node = Node(
+            id=node_id,
+            node_name=f"{node_id:04}",
+            service=f"antnode{node_id:04}.service",
+            user=machine_config.get("user", "ant"),
+            version=metrics["antnode_version"],
+            root_dir=f"{machine_config['node_storage']}/antnode{node_id:04}",
+            binary=f"{machine_config['node_storage']}/antnode{node_id:04}/antnode",
+            port=machine_config["port_start"] * PORT_MULTIPLIER + node_id,
+            metrics_port=METRICS_PORT_BASE + node_id,
+            network="evm-arbitrum-one",
+            wallet=machine_config["rewards_address"],
+            peer_id="",
+            status=STOPPED,
+            timestamp=int(time.time()),
+            records=0,
+            uptime=0,
+            shunned=0,
+            age=int(time.time()),
+            host=machine_config["host"],
+            method=manager_type,
+            layout="1",
+            environment=machine_config.get("environment", ""),
+            manager_type=manager_type,
+        )
+
+        # Insert into database
+        with self.S() as session:
+            session.add(node)
+            session.commit()
+            session.refresh(node)  # Get the persisted node
+
+        # Create the node using process manager
+        source_binary = os.path.expanduser("~/.local/bin/antnode")
+        manager = self._get_process_manager(node)
+
+        if not manager.create_node(node, source_binary):
+            logging.error(f"Failed to create node {node.id}")
+            return {"status": "failed-create-node"}
+
+        # Update status to RESTARTING (node is starting up)
+        self._set_node_status(node.id, RESTARTING)
+
+        return {"status": "added-node"}
+
+    def _execute_survey(self, dry_run: bool) -> Dict[str, Any]:
+        """Execute node survey (idle monitoring)."""
+        if dry_run:
+            logging.warning("DRYRUN: Update nodes")
+        else:
+            update_nodes(self.S)
+        return {"status": "idle"}
+
+    def _parse_node_name(self, service_name: str) -> Optional[int]:
+        """Parse node ID from service name like 'antnode0001'.
+
+        Args:
+            service_name: Node name (e.g., 'antnode0001')
+
+        Returns:
+            Node ID as integer, or None if parsing fails
+        """
+        import re
+        match = re.match(r"antnode(\d+)", service_name)
+        if match:
+            return int(match.group(1))
+        return None
+
+    def _get_node_by_name(self, service_name: str) -> Optional[Node]:
+        """Get node by service name.
+
+        Args:
+            service_name: Node name (e.g., 'antnode0001')
+
+        Returns:
+            Node object or None if not found
+        """
+        node_id = self._parse_node_name(service_name)
+        if node_id is None:
+            logging.error(f"Invalid node name format: {service_name}")
+            return None
+
+        with self.S() as session:
+            result = session.execute(
+                select(Node).where(Node.id == node_id)
+            ).first()
+
+        if result:
+            return result[0]
+        else:
+            logging.error(f"Node not found: {service_name} (id={node_id})")
+            return None
+
+    def execute_forced_action(
+        self,
+        action_type: str,
+        machine_config: Dict[str, Any],
+        metrics: Dict[str, Any],
+        service_name: Optional[str] = None,
+        dry_run: bool = False,
+        count: int = 1,
+    ) -> Dict[str, Any]:
+        """Execute a forced action bypassing the decision engine.
+
+        Args:
+            action_type: Type of action ('add', 'remove', 'upgrade', 'start', 'stop', 'disable', 'teardown')
+            machine_config: Machine configuration
+            metrics: Current system metrics
+            service_name: Optional node name for targeted operations
+            dry_run: If True, log without executing
+            count: Number of nodes to affect (for add, remove, start, stop, upgrade actions)
+
+        Returns:
+            Dictionary with execution result
+        """
+        if action_type == "add":
+            return self._force_add_node(machine_config, metrics, dry_run, count)
+        elif action_type == "remove":
+            return self._force_remove_node(service_name, dry_run, count)
+        elif action_type == "upgrade":
+            return self._force_upgrade_node(service_name, metrics, dry_run, count)
+        elif action_type == "start":
+            return self._force_start_node(service_name, metrics, dry_run, count)
+        elif action_type == "stop":
+            return self._force_stop_node(service_name, dry_run, count)
+        elif action_type == "disable":
+            return self._force_disable_node(service_name, dry_run)
+        elif action_type == "teardown":
+            return self._force_teardown_cluster(machine_config, dry_run)
+        elif action_type == "survey":
+            return self._force_survey_nodes(service_name, dry_run)
+        else:
+            return {"status": "error", "message": f"Unknown action type: {action_type}"}
+
+    def _force_add_node(
+        self, machine_config: Dict[str, Any], metrics: Dict[str, Any], dry_run: bool, count: int = 1
+    ) -> Dict[str, Any]:
+        """Force add new nodes.
+
+        Args:
+            machine_config: Machine configuration
+            metrics: Current system metrics
+            dry_run: If True, log without executing
+            count: Number of nodes to add (default: 1)
+
+        Returns:
+            Dictionary with execution result
+        """
+        logging.info(f"Forced action: Adding {count} node(s)")
+
+        if count < 1:
+            return {"status": "error", "message": "count must be at least 1"}
+
+        added_nodes = []
+        failed_nodes = []
+
+        # Track the start time to identify newly created nodes
+        start_time = int(time.time())
+
+        for i in range(count):
+            result = self._execute_add_node(machine_config, metrics, dry_run)
+            if result["status"] in ["added-node", "add-node"]:
+                # Get the node that was just added (youngest by age >= start_time)
+                if not dry_run:
+                    with self.S() as session:
+                        newest = session.execute(
+                            select(Node).where(Node.age >= start_time).order_by(Node.age.desc())
+                        ).first()
+                        if newest:
+                            added_nodes.append(newest[0].service.replace(".service", ""))
+                else:
+                    added_nodes.append(f"node-{i+1}")
+            else:
+                failed_nodes.append({"index": i+1, "error": result.get("status", "unknown error")})
+
+        if count == 1:
+            # Keep backward compatibility for single node
+            return result
+
+        return {
+            "status": "added-nodes" if not dry_run else "add-nodes-dryrun",
+            "added_count": len(added_nodes),
+            "added_nodes": added_nodes if added_nodes else None,
+            "failed_count": len(failed_nodes),
+            "failed_nodes": failed_nodes if failed_nodes else None,
+        }
+
+    def _force_remove_node(
+        self, service_name: Optional[str], dry_run: bool, count: int = 1
+    ) -> Dict[str, Any]:
+        """Force remove nodes (specific or youngest by age).
+
+        Args:
+            service_name: Optional comma-separated list of service names
+            dry_run: If True, log without executing
+            count: Number of nodes to remove when service_name is not specified (default: 1)
+
+        Returns:
+            Dictionary with execution result
+        """
+        # Parse comma-separated service names
+        service_names = parse_service_names(service_name)
+
+        if service_names:
+            # Remove specific nodes
+            removed_nodes = []
+            failed_nodes = []
+
+            for name in service_names:
+                node = self._get_node_by_name(name)
+                if not node:
+                    failed_nodes.append({"service": name, "error": "not found"})
+                    continue
+
+                logging.info(f"Forced action: Removing node {name}")
+                if dry_run:
+                    logging.warning(f"DRYRUN: Remove node {name}")
+                    removed_nodes.append(name)
+                else:
+                    try:
+                        manager = self._get_process_manager(node)
+                        manager.remove_node(node)
+                        # Remove from database immediately
+                        with self.S() as session:
+                            session.delete(node)
+                            session.commit()
+                        removed_nodes.append(name)
+                    except Exception as e:
+                        logging.error(f"Failed to remove node {name}: {e}")
+                        failed_nodes.append({"service": name, "error": str(e)})
+
+            return {
+                "status": "removed-nodes" if not dry_run else "remove-dryrun",
+                "removed_count": len(removed_nodes),
+                "removed_nodes": removed_nodes,
+                "failed_count": len(failed_nodes),
+                "failed_nodes": failed_nodes if failed_nodes else None,
+            }
+        else:
+            # Remove youngest nodes (default behavior - highest age value)
+            if count < 1:
+                return {"status": "error", "message": "count must be at least 1"}
+
+            logging.info(f"Forced action: Removing {count} youngest node(s)")
+
+            # Get youngest nodes (highest age values)
+            with self.S() as session:
+                youngest_nodes = session.execute(
+                    select(Node).order_by(Node.age.desc()).limit(count)
+                ).all()
+
+            if not youngest_nodes:
+                return {"status": "error", "message": "No nodes to remove"}
+
+            if len(youngest_nodes) < count:
+                logging.warning(f"Only {len(youngest_nodes)} nodes available, removing all of them")
+
+            removed_nodes = []
+            failed_nodes = []
+
+            for row in youngest_nodes:
+                node = row[0]
+                if dry_run:
+                    logging.warning(f"DRYRUN: Remove youngest node {node.node_name}")
+                    removed_nodes.append(node.service.replace(".service", ""))
+                else:
+                    try:
+                        manager = self._get_process_manager(node)
+                        manager.remove_node(node)
+                        # Remove from database immediately
+                        with self.S() as session:
+                            session.delete(node)
+                            session.commit()
+                        removed_nodes.append(node.service.replace(".service", ""))
+                    except Exception as e:
+                        logging.error(f"Failed to remove node {node.node_name}: {e}")
+                        failed_nodes.append({"service": node.service.replace(".service", ""), "error": str(e)})
+
+            if count == 1 and len(removed_nodes) == 1:
+                # Keep backward compatibility for single node
+                # Extract node name from service name (e.g., "antnode0001" -> "0001")
+                node_name = removed_nodes[0].replace("antnode", "")
+                return {"status": "removed-node", "node": node_name}
+
+            return {
+                "status": "removed-nodes" if not dry_run else "remove-dryrun",
+                "removed_count": len(removed_nodes),
+                "removed_nodes": removed_nodes if removed_nodes else None,
+                "failed_count": len(failed_nodes),
+                "failed_nodes": failed_nodes if failed_nodes else None,
+            }
+
+    def _force_upgrade_node(
+        self, service_name: Optional[str], metrics: Dict[str, Any], dry_run: bool, count: int = 1
+    ) -> Dict[str, Any]:
+        """Force upgrade nodes (specific or oldest running nodes by age).
+
+        Args:
+            service_name: Optional comma-separated list of service names
+            metrics: Current system metrics
+            dry_run: If True, log without executing
+            count: Number of nodes to upgrade when service_name is not specified (default: 1)
+
+        Returns:
+            Dictionary with execution result
+        """
+        # Parse comma-separated service names
+        service_names = parse_service_names(service_name)
+
+        if service_names:
+            # Upgrade specific nodes
+            upgraded_nodes = []
+            failed_nodes = []
+
+            for name in service_names:
+                node = self._get_node_by_name(name)
+                if not node:
+                    failed_nodes.append({"service": name, "error": "not found"})
+                    continue
+
+                logging.info(f"Forced action: Upgrading node {name}")
+                if dry_run:
+                    logging.warning(f"DRYRUN: Upgrade node {name}")
+                    upgraded_nodes.append(name)
+                else:
+                    try:
+                        if not self._upgrade_node_binary(node, metrics["antnode_version"]):
+                            failed_nodes.append({"service": name, "error": "upgrade failed"})
+                        else:
+                            upgraded_nodes.append(name)
+                    except Exception as e:
+                        logging.error(f"Failed to upgrade node {name}: {e}")
+                        failed_nodes.append({"service": name, "error": str(e)})
+
+            return {
+                "status": "upgraded-nodes" if not dry_run else "upgrade-dryrun",
+                "upgraded_count": len(upgraded_nodes),
+                "upgraded_nodes": upgraded_nodes,
+                "failed_count": len(failed_nodes),
+                "failed_nodes": failed_nodes if failed_nodes else None,
+            }
+        else:
+            # Upgrade oldest running nodes (default behavior - lowest age values)
+            if count < 1:
+                return {"status": "error", "message": "count must be at least 1"}
+
+            logging.info(f"Forced action: Upgrading {count} oldest running node(s)")
+
+            # Get oldest running nodes (lowest age values)
+            with self.S() as session:
+                oldest_nodes = session.execute(
+                    select(Node)
+                    .where(Node.status == RUNNING)
+                    .order_by(Node.age.asc())
+                    .limit(count)
+                ).all()
+
+            if not oldest_nodes:
+                return {"status": "error", "message": "No running nodes to upgrade"}
+
+            if len(oldest_nodes) < count:
+                logging.warning(f"Only {len(oldest_nodes)} running nodes available, upgrading all of them")
+
+            upgraded_nodes = []
+            failed_nodes = []
+
+            for row in oldest_nodes:
+                node = row[0]
+                if dry_run:
+                    logging.warning(f"DRYRUN: Upgrade oldest node {node.node_name}")
+                    upgraded_nodes.append(node.service.replace(".service", ""))
+                else:
+                    try:
+                        if not self._upgrade_node_binary(node, metrics["antnode_version"]):
+                            failed_nodes.append({"service": node.service.replace(".service", ""), "error": "upgrade failed"})
+                        else:
+                            upgraded_nodes.append(node.service.replace(".service", ""))
+                    except Exception as e:
+                        logging.error(f"Failed to upgrade node {node.node_name}: {e}")
+                        failed_nodes.append({"service": node.service.replace(".service", ""), "error": str(e)})
+
+            if count == 1 and len(upgraded_nodes) == 1:
+                # Keep backward compatibility for single node
+                # Extract node name from service name (e.g., "antnode0001" -> "0001")
+                node_name = upgraded_nodes[0].replace("antnode", "")
+                return {"status": "upgraded-node", "node": node_name}
+
+            return {
+                "status": "upgraded-nodes" if not dry_run else "upgrade-dryrun",
+                "upgraded_count": len(upgraded_nodes),
+                "upgraded_nodes": upgraded_nodes if upgraded_nodes else None,
+                "failed_count": len(failed_nodes),
+                "failed_nodes": failed_nodes if failed_nodes else None,
+            }
+
+    def _force_stop_node(
+        self, service_name: Optional[str], dry_run: bool, count: int = 1
+    ) -> Dict[str, Any]:
+        """Force stop nodes (specific or youngest running nodes by age).
+
+        Args:
+            service_name: Optional comma-separated list of service names
+            dry_run: If True, log without executing
+            count: Number of nodes to stop when service_name is not specified (default: 1)
+
+        Returns:
+            Dictionary with execution result
+        """
+        # Parse comma-separated service names
+        service_names = parse_service_names(service_name)
+
+        if service_names:
+            # Stop specific nodes
+            stopped_nodes = []
+            failed_nodes = []
+
+            for name in service_names:
+                node = self._get_node_by_name(name)
+                if not node:
+                    failed_nodes.append({"service": name, "error": "not found"})
+                    continue
+
+                logging.info(f"Forced action: Stopping node {name}")
+                if dry_run:
+                    logging.warning(f"DRYRUN: Stop node {name}")
+                    stopped_nodes.append(name)
+                else:
+                    try:
+                        manager = self._get_process_manager(node)
+                        manager.stop_node(node)
+                        self._set_node_status(node.id, STOPPED)
+                        stopped_nodes.append(name)
+                    except Exception as e:
+                        logging.error(f"Failed to stop node {name}: {e}")
+                        failed_nodes.append({"service": name, "error": str(e)})
+
+            return {
+                "status": "stopped-nodes" if not dry_run else "stop-dryrun",
+                "stopped_count": len(stopped_nodes),
+                "stopped_nodes": stopped_nodes,
+                "failed_count": len(failed_nodes),
+                "failed_nodes": failed_nodes if failed_nodes else None,
+            }
+        else:
+            # Stop youngest running nodes (default behavior - highest age values)
+            if count < 1:
+                return {"status": "error", "message": "count must be at least 1"}
+
+            logging.info(f"Forced action: Stopping {count} youngest running node(s)")
+
+            # Get youngest running nodes (highest age values)
+            with self.S() as session:
+                youngest_nodes = session.execute(
+                    select(Node)
+                    .where(Node.status == RUNNING)
+                    .order_by(Node.age.desc())
+                    .limit(count)
+                ).all()
+
+            if not youngest_nodes:
+                return {"status": "error", "message": "No running nodes to stop"}
+
+            if len(youngest_nodes) < count:
+                logging.warning(f"Only {len(youngest_nodes)} running nodes available, stopping all of them")
+
+            stopped_nodes = []
+            failed_nodes = []
+
+            for row in youngest_nodes:
+                node = row[0]
+                if dry_run:
+                    logging.warning(f"DRYRUN: Stop youngest node {node.node_name}")
+                    stopped_nodes.append(node.service.replace(".service", ""))
+                else:
+                    try:
+                        manager = self._get_process_manager(node)
+                        manager.stop_node(node)
+                        self._set_node_status(node.id, STOPPED)
+                        stopped_nodes.append(node.service.replace(".service", ""))
+                    except Exception as e:
+                        logging.error(f"Failed to stop node {node.node_name}: {e}")
+                        failed_nodes.append({"service": node.service.replace(".service", ""), "error": str(e)})
+
+            if count == 1 and len(stopped_nodes) == 1:
+                # Keep backward compatibility for single node
+                # Extract node name from service name (e.g., "antnode0001" -> "0001")
+                node_name = stopped_nodes[0].replace("antnode", "")
+                return {"status": "stopped-node", "node": node_name}
+
+            return {
+                "status": "stopped-nodes" if not dry_run else "stop-dryrun",
+                "stopped_count": len(stopped_nodes),
+                "stopped_nodes": stopped_nodes if stopped_nodes else None,
+                "failed_count": len(failed_nodes),
+                "failed_nodes": failed_nodes if failed_nodes else None,
+            }
+
+    def _force_start_node(
+        self, service_name: Optional[str], metrics: Dict[str, Any], dry_run: bool, count: int = 1
+    ) -> Dict[str, Any]:
+        """Force start nodes (specific or oldest stopped nodes by age).
+
+        Args:
+            service_name: Optional comma-separated list of service names
+            metrics: Current system metrics
+            dry_run: If True, log without executing
+            count: Number of nodes to start when service_name is not specified (default: 1)
+
+        Returns:
+            Dictionary with execution result
+        """
+        # Parse comma-separated service names
+        service_names = parse_service_names(service_name)
+
+        if service_names:
+            # Start specific nodes
+            started_nodes = []
+            upgraded_nodes = []
+            failed_nodes = []
+
+            for name in service_names:
+                node = self._get_node_by_name(name)
+                if not node:
+                    failed_nodes.append({"service": name, "error": "not found"})
+                    continue
+
+                if node.status == RUNNING:
+                    failed_nodes.append({"service": name, "error": "already running"})
+                    continue
+
+                logging.info(f"Forced action: Starting node {name}")
+                if dry_run:
+                    logging.warning(f"DRYRUN: Start node {name}")
+                    started_nodes.append(name)
+                else:
+                    try:
+                        # Check if node needs upgrade
+                        if not node.version:
+                            node.version = get_antnode_version(node.binary)
+
+                        # If the stopped version is old, upgrade it (which also starts it)
+                        if Version(metrics["antnode_version"]) > Version(node.version):
+                            if not self._upgrade_node_binary(node, metrics["antnode_version"]):
+                                failed_nodes.append({"service": name, "error": "upgrade failed"})
+                            else:
+                                upgraded_nodes.append(name)
+                        else:
+                            manager = self._get_process_manager(node)
+                            if manager.start_node(node):
+                                self._set_node_status(node.id, RESTARTING)
+                                started_nodes.append(name)
+                            else:
+                                failed_nodes.append({"service": name, "error": "start failed"})
+                    except Exception as e:
+                        logging.error(f"Failed to start node {name}: {e}")
+                        failed_nodes.append({"service": name, "error": str(e)})
+
+            return {
+                "status": "started-nodes" if not dry_run else "start-dryrun",
+                "started_count": len(started_nodes),
+                "started_nodes": started_nodes,
+                "upgraded_count": len(upgraded_nodes),
+                "upgraded_nodes": upgraded_nodes if upgraded_nodes else None,
+                "failed_count": len(failed_nodes),
+                "failed_nodes": failed_nodes if failed_nodes else None,
+            }
+        else:
+            # Start oldest stopped nodes (default behavior - lowest age values)
+            if count < 1:
+                return {"status": "error", "message": "count must be at least 1"}
+
+            logging.info(f"Forced action: Starting {count} oldest stopped node(s)")
+
+            # Get oldest stopped nodes (lowest age values)
+            with self.S() as session:
+                oldest_nodes = session.execute(
+                    select(Node)
+                    .where(Node.status == STOPPED)
+                    .order_by(Node.age.asc())
+                    .limit(count)
+                ).all()
+
+            if not oldest_nodes:
+                return {"status": "error", "message": "No stopped nodes to start"}
+
+            if len(oldest_nodes) < count:
+                logging.warning(f"Only {len(oldest_nodes)} stopped nodes available, starting all of them")
+
+            started_nodes = []
+            upgraded_nodes = []
+            failed_nodes = []
+
+            for row in oldest_nodes:
+                node = row[0]
+                if dry_run:
+                    logging.warning(f"DRYRUN: Start oldest stopped node {node.node_name}")
+                    started_nodes.append(node.service.replace(".service", ""))
+                else:
+                    try:
+                        # Check if node needs upgrade
+                        if not node.version:
+                            node.version = get_antnode_version(node.binary)
+
+                        # If the stopped version is old, upgrade it (which also starts it)
+                        if Version(metrics["antnode_version"]) > Version(node.version):
+                            if not self._upgrade_node_binary(node, metrics["antnode_version"]):
+                                failed_nodes.append({"service": node.service.replace(".service", ""), "error": "upgrade failed"})
+                            else:
+                                upgraded_nodes.append(node.service.replace(".service", ""))
+                        else:
+                            manager = self._get_process_manager(node)
+                            if manager.start_node(node):
+                                self._set_node_status(node.id, RESTARTING)
+                                started_nodes.append(node.service.replace(".service", ""))
+                            else:
+                                failed_nodes.append({"service": node.service.replace(".service", ""), "error": "start failed"})
+                    except Exception as e:
+                        logging.error(f"Failed to start node {node.node_name}: {e}")
+                        failed_nodes.append({"service": node.service.replace(".service", ""), "error": str(e)})
+
+            if count == 1 and len(started_nodes) == 1:
+                # Keep backward compatibility for single node
+                # Extract node name from service name (e.g., "antnode0001" -> "0001")
+                node_name = started_nodes[0].replace("antnode", "")
+                return {"status": "started-node", "node": node_name}
+            elif count == 1 and len(upgraded_nodes) == 1:
+                # Keep backward compatibility for single node upgrade
+                # Extract node name from service name (e.g., "antnode0001" -> "0001")
+                node_name = upgraded_nodes[0].replace("antnode", "")
+                return {"status": "upgrading-node", "node": node_name}
+
+            return {
+                "status": "started-nodes" if not dry_run else "start-dryrun",
+                "started_count": len(started_nodes),
+                "started_nodes": started_nodes if started_nodes else None,
+                "upgraded_count": len(upgraded_nodes),
+                "upgraded_nodes": upgraded_nodes if upgraded_nodes else None,
+                "failed_count": len(failed_nodes),
+                "failed_nodes": failed_nodes if failed_nodes else None,
+            }
+
+    def _force_disable_node(
+        self, service_name: Optional[str], dry_run: bool
+    ) -> Dict[str, Any]:
+        """Force disable a specific node (service_name required)."""
+        if not service_name:
+            return {"status": "error", "message": "service_name required for disable action"}
+
+        # Parse comma-separated service names
+        service_names = parse_service_names(service_name)
+
+        disabled_nodes = []
+        failed_nodes = []
+
+        for name in service_names:
+            node = self._get_node_by_name(name)
+            if not node:
+                failed_nodes.append({"service": name, "error": "not found"})
+                continue
+
+            logging.info(f"Forced action: Disabling node {name}")
+            if dry_run:
+                logging.warning(f"DRYRUN: Disable node {name}")
+                disabled_nodes.append(name)
+            else:
+                try:
+                    # Stop the node if it's running
+                    if node.status == RUNNING:
+                        manager = self._get_process_manager(node)
+                        manager.stop_node(node)
+                    self._set_node_status(node.id, DISABLED)
+                    disabled_nodes.append(name)
+                except Exception as e:
+                    logging.error(f"Failed to disable node {name}: {e}")
+                    failed_nodes.append({"service": name, "error": str(e)})
+
+        return {
+            "status": "disabled-nodes" if not dry_run else "disable-dryrun",
+            "disabled_count": len(disabled_nodes),
+            "disabled_nodes": disabled_nodes,
+            "failed_count": len(failed_nodes),
+            "failed_nodes": failed_nodes if failed_nodes else None,
+        }
+
+    def _force_teardown_cluster(
+        self, machine_config: Dict[str, Any], dry_run: bool
+    ) -> Dict[str, Any]:
+        """Force teardown the entire cluster."""
+        logging.info("Forced action: Tearing down cluster")
+
+        # Get all nodes
+        with self.S() as session:
+            all_nodes = session.execute(
+                select(Node).order_by(Node.id.asc())
+            ).all()
+
+        if not all_nodes:
+            return {"status": "no-nodes", "message": "No nodes to teardown"}
+
+        # Get the process manager (use the first node's manager or default)
+        if all_nodes:
+            sample_node = all_nodes[0][0]
+            manager = self._get_process_manager(sample_node)
+        else:
+            manager = get_process_manager()
+
+        # Try manager-specific teardown first
+        if hasattr(manager, 'teardown_cluster'):
+            logging.info(f"Using {manager.__class__.__name__} teardown_cluster method")
+            if dry_run:
+                logging.warning("DRYRUN: Teardown cluster via manager")
+            else:
+                if manager.teardown_cluster():
+                    # Remove all nodes from database
+                    with self.S() as session:
+                        session.query(Node).delete()
+                        session.commit()
+                    return {"status": "cluster-teardown", "method": "manager-specific"}
+
+        # Fall back to removing each node individually (without delay)
+        logging.info("Using default teardown (remove all nodes)")
+        removed_count = 0
+        for row in all_nodes:
+            node = row[0]
+            if dry_run:
+                logging.warning(f"DRYRUN: Remove node {node.node_name}")
+                removed_count += 1
+            else:
+                try:
+                    manager = self._get_process_manager(node)
+                    manager.remove_node(node)
+                    with self.S() as session:
+                        session.delete(node)
+                        session.commit()
+                    removed_count += 1
+                    logging.info(f"Removed node {node.node_name}")
+                except Exception as e:
+                    logging.error(f"Failed to remove node {node.node_name}: {e}")
+
+        return {
+            "status": "cluster-teardown",
+            "method": "individual-remove",
+            "removed_count": removed_count,
+        }
+
+    def _survey_specific_nodes(self, service_names: List[str], dry_run: bool) -> Dict[str, Any]:
+        """Survey specific nodes by service name.
+
+        Args:
+            service_names: List of service names to survey
+            dry_run: If True, log without executing
+
+        Returns:
+            Dictionary with survey results
+        """
+        from wnm.utils import read_node_metrics, read_node_metadata, update_node_from_metrics
+
+        surveyed_nodes = []
+        failed_nodes = []
+
+        for service_name in service_names:
+            node = self._get_node_by_name(service_name)
+            if not node:
+                failed_nodes.append({"service": service_name, "error": "not found"})
+                continue
+
+            if node.status == DISABLED:
+                failed_nodes.append({"service": service_name, "error": "disabled"})
+                continue
+
+            if dry_run:
+                logging.warning(f"DRYRUN: Survey node {service_name}")
+                surveyed_nodes.append(service_name)
+            else:
+                logging.info(f"Surveying node {service_name}")
+
+                # Check metadata first
+                node_metadata = read_node_metadata(node.host, node.metrics_port)
+
+                # If metadata fails, fake metrics with 0's
+                if node_metadata["status"] == STOPPED:
+                    node_metrics = {
+                        "status": STOPPED,
+                        "uptime": 0,
+                        "records": 0,
+                        "shunned": 0,
+                        "connected_peers": 0
+                    }
+                else:
+                    # Metadata succeeded, now get metrics
+                    node_metrics = read_node_metrics(node.host, node.metrics_port)
+
+                # Skip update if node is stopped and already marked as stopped
+                if node_metadata["status"] == STOPPED and node.status == STOPPED:
+                    surveyed_nodes.append(service_name)
+                    continue
+
+                update_node_from_metrics(self.S, node.id, node_metrics, node_metadata)
+                surveyed_nodes.append(service_name)
+
+        return {
+            "status": "survey-complete" if not dry_run else "survey-dryrun",
+            "surveyed_count": len(surveyed_nodes),
+            "surveyed_nodes": surveyed_nodes,
+            "failed_count": len(failed_nodes),
+            "failed_nodes": failed_nodes if failed_nodes else None,
+        }
+
+    def _force_survey_nodes(self, service_name: Optional[str] = None, dry_run: bool = False) -> Dict[str, Any]:
+        """Force a survey of all nodes or specific nodes to update their status and metrics.
+
+        Args:
+            service_name: Optional comma-separated list of service names to survey
+            dry_run: If True, log without executing
+
+        Returns:
+            Dictionary with survey results
+        """
+        # Parse service names if provided
+        service_names = parse_service_names(service_name)
+
+        if service_names:
+            # Survey specific nodes
+            logging.info(f"Forced action: Surveying {len(service_names)} specific nodes")
+            return self._survey_specific_nodes(service_names, dry_run)
+        else:
+            # Survey all nodes
+            logging.info("Forced action: Surveying all nodes")
+
+            if dry_run:
+                logging.warning("DRYRUN: Survey all nodes")
+                # Get count of non-disabled nodes
+                with self.S() as session:
+                    from wnm.common import DISABLED
+                    node_count = session.execute(
+                        select(func.count(Node.id)).where(Node.status != DISABLED)
+                    ).scalar()
+                return {"status": "survey-dryrun", "node_count": node_count}
+
+            # Update all nodes
+            update_nodes(self.S)
+
+            # Get updated count
+            with self.S() as session:
+                from wnm.common import DISABLED
+                node_count = session.execute(
+                    select(func.count(Node.id)).where(Node.status != DISABLED)
+                ).scalar()
+
+            return {"status": "survey-complete", "node_count": node_count}