fix: cluster-aware guest deduplication and multi-agent token binding
- Add cluster-aware guest ID generation (clusterName-VMID instead of instanceName-VMID) to prevent duplicate VMs/containers when multiple cluster nodes are monitored - Add cluster deduplication at registration time - when a node is added that belongs to an already-configured cluster, merge as endpoint instead of creating duplicate - Add startup consolidation to automatically merge duplicate cluster instances - Change host agent token binding from agent GUID to hostname, allowing: - Multiple host agents to share a token (each bound by hostname) - Agent reinstalls on same host without token conflicts - Remove 12-character password minimum requirement - Remove emoji from auto-registration success message - Fix grouped view node lookup to support both cluster-aware node IDs (clusterName-nodeName) and legacy guest grouping keys (instance-nodeName) Fixes duplicate guests appearing when agents are installed on multiple cluster nodes. Also improves multi-agent UX by allowing shared tokens.
This commit is contained in:
rcourtman
parent
1b8a310f2e
commit
eb59faaa4a
8 changed files with 358 additions and 91 deletions
|
|
@ -353,10 +353,17 @@ export function Dashboard(props: DashboardProps) {
|
|||
};
|
||||
|
||||
// Create a mapping from node ID to node object
|
||||
// Also maps by instance-nodeName for guest grouping compatibility
|
||||
const nodeByInstance = createMemo(() => {
|
||||
const map: Record<string, Node> = {};
|
||||
props.nodes.forEach((node) => {
|
||||
// Map by node.id (may be clusterName-nodeName or instance-nodeName)
|
||||
map[node.id] = node;
|
||||
// Also map by instance-nodeName for guest grouping (guests use instance-node format)
|
||||
const legacyKey = `${node.instance}-${node.name}`;
|
||||
if (!map[legacyKey]) {
|
||||
map[legacyKey] = node;
|
||||
}
|
||||
});
|
||||
return map;
|
||||
});
|
||||
|
|
|
|||
|
|
@ -149,8 +149,8 @@ export const FirstRunSetup: Component<{ force?: boolean; showLegacyBanner?: bool
|
|||
showError('Passwords do not match');
|
||||
return;
|
||||
}
|
||||
if (password().length < 12) {
|
||||
showError('Password must be at least 12 characters');
|
||||
if (password().length < 1) {
|
||||
showError('Password cannot be empty');
|
||||
return;
|
||||
}
|
||||
}
|
||||
|
|
@ -426,7 +426,7 @@ IMPORTANT: Keep these credentials secure!
|
|||
type="text"
|
||||
value={bootstrapToken()}
|
||||
onInput={(e) => setBootstrapToken(e.currentTarget.value)}
|
||||
onKeyPress={(e) => e.key === 'Enter' && handleUnlock()}
|
||||
onKeyPress={(e) => e.key === 'Enter' && handleUnlock()}
|
||||
class="w-full px-4 py-2 rounded-lg border border-gray-300 dark:border-gray-600 bg-white dark:bg-gray-700 text-gray-900 dark:text-gray-100 focus:ring-2 focus:ring-blue-500 focus:border-transparent font-mono text-sm"
|
||||
placeholder="Paste the token from your host"
|
||||
autofocus
|
||||
|
|
@ -494,22 +494,20 @@ IMPORTANT: Keep these credentials secure!
|
|||
<button
|
||||
type="button"
|
||||
onClick={() => setUseCustomPassword(false)}
|
||||
class={`flex-1 py-2 px-4 rounded-lg text-sm font-medium transition-colors ${
|
||||
!useCustomPassword()
|
||||
? 'bg-blue-600 text-white'
|
||||
: 'bg-gray-200 dark:bg-gray-700 text-gray-700 dark:text-gray-300 hover:bg-gray-300 dark:hover:bg-gray-600'
|
||||
}`}
|
||||
class={`flex-1 py-2 px-4 rounded-lg text-sm font-medium transition-colors ${!useCustomPassword()
|
||||
? 'bg-blue-600 text-white'
|
||||
: 'bg-gray-200 dark:bg-gray-700 text-gray-700 dark:text-gray-300 hover:bg-gray-300 dark:hover:bg-gray-600'
|
||||
}`}
|
||||
>
|
||||
Generate Secure Password
|
||||
</button>
|
||||
<button
|
||||
type="button"
|
||||
onClick={() => setUseCustomPassword(true)}
|
||||
class={`flex-1 py-2 px-4 rounded-lg text-sm font-medium transition-colors ${
|
||||
useCustomPassword()
|
||||
? 'bg-blue-600 text-white'
|
||||
: 'bg-gray-200 dark:bg-gray-700 text-gray-700 dark:text-gray-300 hover:bg-gray-300 dark:hover:bg-gray-600'
|
||||
}`}
|
||||
class={`flex-1 py-2 px-4 rounded-lg text-sm font-medium transition-colors ${useCustomPassword()
|
||||
? 'bg-blue-600 text-white'
|
||||
: 'bg-gray-200 dark:bg-gray-700 text-gray-700 dark:text-gray-300 hover:bg-gray-300 dark:hover:bg-gray-600'
|
||||
}`}
|
||||
>
|
||||
Set Custom Password
|
||||
</button>
|
||||
|
|
@ -522,7 +520,7 @@ IMPORTANT: Keep these credentials secure!
|
|||
value={password()}
|
||||
onInput={(e) => setPassword(e.currentTarget.value)}
|
||||
class="w-full px-4 py-2 rounded-lg border border-gray-300 dark:border-gray-600 bg-white dark:bg-gray-700 text-gray-900 dark:text-gray-100 focus:ring-2 focus:ring-blue-500 focus:border-transparent"
|
||||
placeholder="Enter password (min 12 characters)"
|
||||
placeholder="Enter password"
|
||||
/>
|
||||
<input
|
||||
type="password"
|
||||
|
|
@ -556,11 +554,10 @@ IMPORTANT: Keep these credentials secure!
|
|||
setThemeMode('system');
|
||||
applyTheme('system');
|
||||
}}
|
||||
class={`py-2 px-4 rounded-lg text-sm font-medium transition-colors ${
|
||||
themeMode() === 'system'
|
||||
? 'bg-blue-600 text-white'
|
||||
: 'bg-gray-200 dark:bg-gray-700 text-gray-700 dark:text-gray-300 hover:bg-gray-300 dark:hover:bg-gray-600'
|
||||
}`}
|
||||
class={`py-2 px-4 rounded-lg text-sm font-medium transition-colors ${themeMode() === 'system'
|
||||
? 'bg-blue-600 text-white'
|
||||
: 'bg-gray-200 dark:bg-gray-700 text-gray-700 dark:text-gray-300 hover:bg-gray-300 dark:hover:bg-gray-600'
|
||||
}`}
|
||||
>
|
||||
System
|
||||
</button>
|
||||
|
|
@ -570,11 +567,10 @@ IMPORTANT: Keep these credentials secure!
|
|||
setThemeMode('light');
|
||||
applyTheme('light');
|
||||
}}
|
||||
class={`py-2 px-4 rounded-lg text-sm font-medium transition-colors ${
|
||||
themeMode() === 'light'
|
||||
? 'bg-blue-600 text-white'
|
||||
: 'bg-gray-200 dark:bg-gray-700 text-gray-700 dark:text-gray-300 hover:bg-gray-300 dark:hover:bg-gray-600'
|
||||
}`}
|
||||
class={`py-2 px-4 rounded-lg text-sm font-medium transition-colors ${themeMode() === 'light'
|
||||
? 'bg-blue-600 text-white'
|
||||
: 'bg-gray-200 dark:bg-gray-700 text-gray-700 dark:text-gray-300 hover:bg-gray-300 dark:hover:bg-gray-600'
|
||||
}`}
|
||||
>
|
||||
Light
|
||||
</button>
|
||||
|
|
@ -584,11 +580,10 @@ IMPORTANT: Keep these credentials secure!
|
|||
setThemeMode('dark');
|
||||
applyTheme('dark');
|
||||
}}
|
||||
class={`py-2 px-4 rounded-lg text-sm font-medium transition-colors ${
|
||||
themeMode() === 'dark'
|
||||
? 'bg-blue-600 text-white'
|
||||
: 'bg-gray-200 dark:bg-gray-700 text-gray-700 dark:text-gray-300 hover:bg-gray-300 dark:hover:bg-gray-600'
|
||||
}`}
|
||||
class={`py-2 px-4 rounded-lg text-sm font-medium transition-colors ${themeMode() === 'dark'
|
||||
? 'bg-blue-600 text-white'
|
||||
: 'bg-gray-200 dark:bg-gray-700 text-gray-700 dark:text-gray-300 hover:bg-gray-300 dark:hover:bg-gray-600'
|
||||
}`}
|
||||
>
|
||||
Dark
|
||||
</button>
|
||||
|
|
|
|||
|
|
@ -710,7 +710,7 @@ export function createWebSocketStore(url: string) {
|
|||
const nodeType = node.type === 'pve' ? 'Proxmox VE' : 'Proxmox Backup Server';
|
||||
|
||||
notificationStore.success(
|
||||
`🎉 ${nodeType} node "${nodeName}" was successfully auto-registered and is now being monitored!`,
|
||||
`${nodeType} node "${nodeName}" was successfully auto-registered and is now being monitored!`,
|
||||
8000,
|
||||
);
|
||||
logger.info('Node auto-registered:', node);
|
||||
|
|
|
|||
|
|
@ -1311,11 +1311,70 @@ func (h *ConfigHandlers) HandleAddNode(w http.ResponseWriter, r *http.Request) {
|
|||
isCluster, clusterName, clusterEndpoints = detectPVECluster(clientConfig, req.Name, nil)
|
||||
}
|
||||
|
||||
// CLUSTER DEDUPLICATION: If this node is part of a cluster, check if we already
|
||||
// have that cluster configured. If so, this is a duplicate - we should merge
|
||||
// the node as an endpoint to the existing cluster instead of creating a new instance.
|
||||
// This prevents duplicate VMs/containers when users install agents on multiple cluster nodes.
|
||||
if isCluster && clusterName != "" {
|
||||
for i := range h.config.PVEInstances {
|
||||
existingInstance := &h.config.PVEInstances[i]
|
||||
if existingInstance.IsCluster && existingInstance.ClusterName == clusterName {
|
||||
// Found existing cluster with same name - merge endpoints!
|
||||
log.Info().
|
||||
Str("cluster", clusterName).
|
||||
Str("existingInstance", existingInstance.Name).
|
||||
Str("newNode", req.Name).
|
||||
Msg("New node belongs to already-configured cluster - merging as endpoint instead of creating duplicate")
|
||||
|
||||
// Merge any new endpoints from the detected cluster
|
||||
existingEndpointMap := make(map[string]bool)
|
||||
for _, ep := range existingInstance.ClusterEndpoints {
|
||||
existingEndpointMap[ep.NodeName] = true
|
||||
}
|
||||
for _, newEp := range clusterEndpoints {
|
||||
if !existingEndpointMap[newEp.NodeName] {
|
||||
existingInstance.ClusterEndpoints = append(existingInstance.ClusterEndpoints, newEp)
|
||||
log.Info().
|
||||
Str("cluster", clusterName).
|
||||
Str("endpoint", newEp.NodeName).
|
||||
Msg("Added new endpoint to existing cluster")
|
||||
}
|
||||
}
|
||||
|
||||
// Save the updated configuration
|
||||
if h.persistence != nil {
|
||||
if err := h.persistence.SaveNodesConfig(h.config.PVEInstances, h.config.PBSInstances, h.config.PMGInstances); err != nil {
|
||||
log.Warn().Err(err).Msg("Failed to persist cluster endpoint merge")
|
||||
}
|
||||
}
|
||||
|
||||
// Reload the monitor to pick up the updated endpoints
|
||||
if h.reloadFunc != nil {
|
||||
if err := h.reloadFunc(); err != nil {
|
||||
log.Warn().Err(err).Msg("Failed to reload monitor after cluster merge")
|
||||
}
|
||||
}
|
||||
|
||||
// Return success - the cluster is now updated with new endpoints
|
||||
w.Header().Set("Content-Type", "application/json")
|
||||
json.NewEncoder(w).Encode(map[string]interface{}{
|
||||
"success": true,
|
||||
"merged": true,
|
||||
"cluster": clusterName,
|
||||
"existingNode": existingInstance.Name,
|
||||
"message": fmt.Sprintf("Node merged into existing cluster '%s' (already configured as '%s')", clusterName, existingInstance.Name),
|
||||
"totalEndpoints": len(existingInstance.ClusterEndpoints),
|
||||
})
|
||||
return
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
if isCluster {
|
||||
log.Info().
|
||||
Str("cluster", clusterName).
|
||||
Int("endpoints", len(clusterEndpoints)).
|
||||
Msg("Detected Proxmox cluster, auto-discovering all nodes")
|
||||
Msg("Detected new Proxmox cluster, auto-discovering all nodes")
|
||||
}
|
||||
|
||||
// Use sensible defaults for boolean fields if not provided
|
||||
|
|
@ -5857,6 +5916,61 @@ func (h *ConfigHandlers) HandleAutoRegister(w http.ResponseWriter, r *http.Reque
|
|||
|
||||
isCluster, clusterName, clusterEndpoints := detectPVECluster(clientConfig, nodeConfig.Name, nil)
|
||||
|
||||
// CLUSTER DEDUPLICATION: Check if we already have this cluster configured
|
||||
// If so, merge this node as an endpoint instead of creating a duplicate instance
|
||||
if isCluster && clusterName != "" {
|
||||
for i := range h.config.PVEInstances {
|
||||
existingInstance := &h.config.PVEInstances[i]
|
||||
if existingInstance.IsCluster && existingInstance.ClusterName == clusterName {
|
||||
// Found existing cluster with same name - merge endpoints!
|
||||
log.Info().
|
||||
Str("cluster", clusterName).
|
||||
Str("existingInstance", existingInstance.Name).
|
||||
Str("newNode", nodeConfig.Name).
|
||||
Msg("Auto-registered node belongs to already-configured cluster - merging endpoints")
|
||||
|
||||
// Merge any new endpoints from the detected cluster
|
||||
existingEndpointMap := make(map[string]bool)
|
||||
for _, ep := range existingInstance.ClusterEndpoints {
|
||||
existingEndpointMap[ep.NodeName] = true
|
||||
}
|
||||
for _, newEp := range clusterEndpoints {
|
||||
if !existingEndpointMap[newEp.NodeName] {
|
||||
existingInstance.ClusterEndpoints = append(existingInstance.ClusterEndpoints, newEp)
|
||||
log.Info().
|
||||
Str("cluster", clusterName).
|
||||
Str("endpoint", newEp.NodeName).
|
||||
Msg("Added new endpoint to existing cluster via auto-registration")
|
||||
}
|
||||
}
|
||||
|
||||
// Save and reload
|
||||
if h.persistence != nil {
|
||||
if err := h.persistence.SaveNodesConfig(h.config.PVEInstances, h.config.PBSInstances, h.config.PMGInstances); err != nil {
|
||||
log.Warn().Err(err).Msg("Failed to persist cluster endpoint merge during auto-registration")
|
||||
}
|
||||
}
|
||||
if h.reloadFunc != nil {
|
||||
if err := h.reloadFunc(); err != nil {
|
||||
log.Warn().Err(err).Msg("Failed to reload monitor after cluster merge during auto-registration")
|
||||
}
|
||||
}
|
||||
|
||||
// Return success - merged into existing cluster
|
||||
w.Header().Set("Content-Type", "application/json")
|
||||
json.NewEncoder(w).Encode(map[string]interface{}{
|
||||
"success": true,
|
||||
"merged": true,
|
||||
"cluster": clusterName,
|
||||
"existingNode": existingInstance.Name,
|
||||
"message": fmt.Sprintf("Agent merged into existing cluster '%s'", clusterName),
|
||||
"totalEndpoints": len(existingInstance.ClusterEndpoints),
|
||||
})
|
||||
return
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
monitorVMs := true
|
||||
if nodeConfig.MonitorVMs != nil {
|
||||
monitorVMs = *nodeConfig.MonitorVMs
|
||||
|
|
@ -5895,7 +6009,7 @@ func (h *ConfigHandlers) HandleAutoRegister(w http.ResponseWriter, r *http.Reque
|
|||
log.Info().
|
||||
Str("cluster", clusterName).
|
||||
Int("endpoints", len(clusterEndpoints)).
|
||||
Msg("Added Proxmox cluster via auto-registration")
|
||||
Msg("Added new Proxmox cluster via auto-registration")
|
||||
}
|
||||
} else {
|
||||
verifySSL := false
|
||||
|
|
|
|||
|
|
@ -1267,15 +1267,14 @@ func (s *State) UpdateNodesForInstance(instanceName string, nodes []Node) {
|
|||
s.mu.Lock()
|
||||
defer s.mu.Unlock()
|
||||
|
||||
// Create a map of existing nodes, excluding those from this instance
|
||||
// Build a map of ALL existing nodes by ID (not filtered by instance)
|
||||
// This handles cluster-based IDs where the same node ID comes from multiple instances
|
||||
// Also preserve LinkedHostAgentID for nodes that are being updated
|
||||
existingNodeLinks := make(map[string]string) // nodeID -> linkedHostAgentID
|
||||
nodeMap := make(map[string]Node)
|
||||
for _, node := range s.Nodes {
|
||||
if node.Instance != instanceName {
|
||||
nodeMap[node.ID] = node
|
||||
} else if node.LinkedHostAgentID != "" {
|
||||
// Preserve the link for nodes from this instance
|
||||
nodeMap[node.ID] = node
|
||||
if node.LinkedHostAgentID != "" {
|
||||
existingNodeLinks[node.ID] = node.LinkedHostAgentID
|
||||
}
|
||||
}
|
||||
|
|
|
|||
|
|
@ -205,6 +205,46 @@ func TestSafeFloat(t *testing.T) {
|
|||
}
|
||||
}
|
||||
|
||||
func TestMakeGuestID(t *testing.T) {
|
||||
t.Parallel()
|
||||
|
||||
cases := []struct {
|
||||
name string
|
||||
instanceName string
|
||||
clusterName string
|
||||
isCluster bool
|
||||
vmid int
|
||||
want string
|
||||
}{
|
||||
// Standalone nodes use instance name
|
||||
{name: "standalone node", instanceName: "pve-host1", clusterName: "", isCluster: false, vmid: 100, want: "pve-host1-100"},
|
||||
{name: "standalone with empty cluster name", instanceName: "pve-standalone", clusterName: "", isCluster: false, vmid: 200, want: "pve-standalone-200"},
|
||||
{name: "non-cluster even with cluster name", instanceName: "pve-node", clusterName: "my-cluster", isCluster: false, vmid: 150, want: "pve-node-150"},
|
||||
|
||||
// Cluster nodes use cluster name
|
||||
{name: "cluster node uses cluster name", instanceName: "pve-host1", clusterName: "my-cluster", isCluster: true, vmid: 100, want: "my-cluster-100"},
|
||||
{name: "different cluster node same cluster", instanceName: "pve-host2", clusterName: "my-cluster", isCluster: true, vmid: 100, want: "my-cluster-100"},
|
||||
{name: "cluster with different vmid", instanceName: "pve-host1", clusterName: "production", isCluster: true, vmid: 999, want: "production-999"},
|
||||
|
||||
// Edge case: isCluster true but no cluster name falls back to instance name
|
||||
{name: "cluster flag but no name", instanceName: "pve-node", clusterName: "", isCluster: true, vmid: 300, want: "pve-node-300"},
|
||||
|
||||
// Edge case: special characters and spaces (shouldn't happen but test anyway)
|
||||
{name: "cluster name with hyphen", instanceName: "pve-node1", clusterName: "my-prod-cluster", isCluster: true, vmid: 101, want: "my-prod-cluster-101"},
|
||||
}
|
||||
|
||||
for _, tc := range cases {
|
||||
tc := tc
|
||||
t.Run(tc.name, func(t *testing.T) {
|
||||
t.Parallel()
|
||||
if got := makeGuestID(tc.instanceName, tc.clusterName, tc.isCluster, tc.vmid); got != tc.want {
|
||||
t.Fatalf("makeGuestID(%q, %q, %v, %d) = %q, want %q", tc.instanceName, tc.clusterName, tc.isCluster, tc.vmid, got, tc.want)
|
||||
}
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
func TestConvertPoolInfoToModel(t *testing.T) {
|
||||
t.Parallel()
|
||||
|
||||
|
|
|
|||
|
|
@ -686,6 +686,22 @@ func safeFloat(val float64) float64 {
|
|||
return val
|
||||
}
|
||||
|
||||
// makeGuestID generates a stable guest ID that is cluster-aware.
|
||||
// When the instance is part of a cluster, the cluster name is used as the primary identifier
|
||||
// to prevent duplicate guests when multiple cluster nodes are configured as separate PVE instances.
|
||||
// Format when in cluster: clusterName-VMID (e.g., "my-cluster-100")
|
||||
// Format when standalone: instanceName-VMID (e.g., "pve-host1-100")
|
||||
// This ensures VMs/containers are properly deduplicated across multiple agents in the same cluster.
|
||||
func makeGuestID(instanceName string, clusterName string, isCluster bool, vmid int) string {
|
||||
// Use cluster name as the identifier when the instance is part of a cluster
|
||||
// This ensures guests are identified by their cluster, not by which node reported them
|
||||
if isCluster && clusterName != "" {
|
||||
return fmt.Sprintf("%s-%d", clusterName, vmid)
|
||||
}
|
||||
// For standalone nodes, use the instance name
|
||||
return fmt.Sprintf("%s-%d", instanceName, vmid)
|
||||
}
|
||||
|
||||
// parseDurationEnv parses a duration from an environment variable, returning defaultVal if not set or invalid
|
||||
func parseDurationEnv(key string, defaultVal time.Duration) time.Duration {
|
||||
val := os.Getenv(key)
|
||||
|
|
@ -1925,43 +1941,27 @@ func (m *Monitor) ApplyHostReport(report agentshost.Report, tokenRecord *config.
|
|||
if m.hostTokenBindings == nil {
|
||||
m.hostTokenBindings = make(map[string]string)
|
||||
}
|
||||
if boundID, exists := m.hostTokenBindings[tokenID]; exists && boundID != bindingID {
|
||||
m.mu.Unlock()
|
||||
|
||||
conflictingHost := "unknown"
|
||||
for _, candidate := range existingHosts {
|
||||
if candidate.TokenID == tokenID || candidate.ID == boundID {
|
||||
conflictingHost = candidate.Hostname
|
||||
if candidate.DisplayName != "" {
|
||||
conflictingHost = candidate.DisplayName
|
||||
}
|
||||
break
|
||||
}
|
||||
}
|
||||
|
||||
tokenHint := tokenHintFromRecord(tokenRecord)
|
||||
if tokenHint != "" {
|
||||
tokenHint = " (" + tokenHint + ")"
|
||||
}
|
||||
|
||||
log.Warn().
|
||||
// Bind tokens by hostname rather than agent ID. This allows:
|
||||
// - Same host to reconnect after agent reinstall (agent ID changes but hostname doesn't)
|
||||
// - Multiple hosts to use the same token (each hostname gets its own binding entry)
|
||||
// - Prevents a stolen token from being used on a different hostname
|
||||
bindingKey := fmt.Sprintf("%s:%s", tokenID, hostname)
|
||||
if boundID, exists := m.hostTokenBindings[bindingKey]; exists && boundID != bindingID {
|
||||
// Same token+hostname but different agent ID - this is a reinstall, allow it
|
||||
log.Info().
|
||||
Str("tokenID", tokenID).
|
||||
Str("tokenHint", tokenHint).
|
||||
Str("reportingAgentID", bindingID).
|
||||
Str("boundAgentID", boundID).
|
||||
Str("conflictingHost", conflictingHost).
|
||||
Msg("Rejecting host report: token already bound to different agent")
|
||||
|
||||
return models.Host{}, fmt.Errorf("API token%s is already in use by host %q (agent: %s). Generate a new token or set --agent-id before reusing it", tokenHint, conflictingHost, boundID)
|
||||
}
|
||||
|
||||
if _, exists := m.hostTokenBindings[tokenID]; !exists {
|
||||
m.hostTokenBindings[tokenID] = bindingID
|
||||
Str("hostname", hostname).
|
||||
Str("oldAgentID", boundID).
|
||||
Str("newAgentID", bindingID).
|
||||
Msg("Host agent reinstalled - updating token binding")
|
||||
m.hostTokenBindings[bindingKey] = bindingID
|
||||
} else if !exists {
|
||||
m.hostTokenBindings[bindingKey] = bindingID
|
||||
log.Debug().
|
||||
Str("tokenID", tokenID).
|
||||
Str("agentID", bindingID).
|
||||
Str("hostname", hostname).
|
||||
Msg("Bound host agent token to agent identity")
|
||||
Msg("Bound host agent token to hostname")
|
||||
}
|
||||
m.mu.Unlock()
|
||||
}
|
||||
|
|
@ -3536,8 +3536,109 @@ func (m *Monitor) getConfiguredHostIPs() []string {
|
|||
return ips
|
||||
}
|
||||
|
||||
// consolidateDuplicateClusters detects and merges duplicate cluster instances.
|
||||
// When multiple PVE instances belong to the same Proxmox cluster (determined by ClusterName),
|
||||
// they should be merged into a single instance with all endpoints combined.
|
||||
// This prevents duplicate VMs/containers in the UI.
|
||||
func (m *Monitor) consolidateDuplicateClusters() {
|
||||
if m == nil || m.config == nil || len(m.config.PVEInstances) < 2 {
|
||||
return
|
||||
}
|
||||
|
||||
// Group instances by cluster name
|
||||
clusterGroups := make(map[string][]int) // clusterName -> indices of instances
|
||||
for i, instance := range m.config.PVEInstances {
|
||||
if instance.IsCluster && instance.ClusterName != "" {
|
||||
clusterGroups[instance.ClusterName] = append(clusterGroups[instance.ClusterName], i)
|
||||
}
|
||||
}
|
||||
|
||||
// Find clusters that have duplicates
|
||||
var mergedAny bool
|
||||
for clusterName, indices := range clusterGroups {
|
||||
if len(indices) < 2 {
|
||||
continue // No duplicates for this cluster
|
||||
}
|
||||
|
||||
log.Warn().
|
||||
Str("cluster", clusterName).
|
||||
Int("duplicates", len(indices)).
|
||||
Msg("Detected duplicate cluster instances - consolidating")
|
||||
|
||||
// Keep the first instance and merge all others into it
|
||||
primaryIdx := indices[0]
|
||||
primary := &m.config.PVEInstances[primaryIdx]
|
||||
|
||||
// Build a set of existing endpoint node names
|
||||
existingEndpoints := make(map[string]bool)
|
||||
for _, ep := range primary.ClusterEndpoints {
|
||||
existingEndpoints[ep.NodeName] = true
|
||||
}
|
||||
|
||||
// Merge endpoints from all duplicate instances
|
||||
for _, dupIdx := range indices[1:] {
|
||||
duplicate := m.config.PVEInstances[dupIdx]
|
||||
log.Info().
|
||||
Str("cluster", clusterName).
|
||||
Str("primary", primary.Name).
|
||||
Str("duplicate", duplicate.Name).
|
||||
Msg("Merging duplicate cluster instance")
|
||||
|
||||
for _, ep := range duplicate.ClusterEndpoints {
|
||||
if !existingEndpoints[ep.NodeName] {
|
||||
primary.ClusterEndpoints = append(primary.ClusterEndpoints, ep)
|
||||
existingEndpoints[ep.NodeName] = true
|
||||
log.Info().
|
||||
Str("cluster", clusterName).
|
||||
Str("endpoint", ep.NodeName).
|
||||
Msg("Added endpoint from duplicate instance")
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
mergedAny = true
|
||||
}
|
||||
|
||||
if !mergedAny {
|
||||
return
|
||||
}
|
||||
|
||||
// Remove duplicate instances (keeping only the primary for each cluster)
|
||||
var newInstances []config.PVEInstance
|
||||
seenClusters := make(map[string]bool)
|
||||
|
||||
for _, instance := range m.config.PVEInstances {
|
||||
if instance.IsCluster && instance.ClusterName != "" {
|
||||
if seenClusters[instance.ClusterName] {
|
||||
log.Info().
|
||||
Str("cluster", instance.ClusterName).
|
||||
Str("instance", instance.Name).
|
||||
Msg("Removing duplicate cluster instance")
|
||||
continue // Skip duplicates
|
||||
}
|
||||
seenClusters[instance.ClusterName] = true
|
||||
}
|
||||
newInstances = append(newInstances, instance)
|
||||
}
|
||||
|
||||
m.config.PVEInstances = newInstances
|
||||
|
||||
// Persist the consolidated configuration
|
||||
if m.persistence != nil {
|
||||
if err := m.persistence.SaveNodesConfig(m.config.PVEInstances, m.config.PBSInstances, m.config.PMGInstances); err != nil {
|
||||
log.Error().Err(err).Msg("Failed to persist cluster consolidation")
|
||||
} else {
|
||||
log.Info().Msg("Persisted consolidated cluster configuration")
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// Start begins the monitoring loop
|
||||
func (m *Monitor) Start(ctx context.Context, wsHub *websocket.Hub) {
|
||||
// Consolidate any duplicate cluster instances before starting
|
||||
// This fixes the case where multiple agents registered from the same cluster
|
||||
m.consolidateDuplicateClusters()
|
||||
|
||||
pollingInterval := m.effectivePVEPollingInterval()
|
||||
log.Info().
|
||||
Dur("pollingInterval", pollingInterval).
|
||||
|
|
@ -5177,10 +5278,10 @@ func (m *Monitor) pollPVEInstance(ctx context.Context, instanceName string, clie
|
|||
pollErr = ctx.Err()
|
||||
return
|
||||
default:
|
||||
// Always try the efficient cluster/resources endpoint first
|
||||
// Always try the efficient cluster/resources endpoint first
|
||||
// This endpoint works on both clustered and standalone nodes
|
||||
// Testing confirmed it works on standalone nodes like pimox
|
||||
useClusterEndpoint := m.pollVMsAndContainersEfficient(ctx, instanceName, client, nodeEffectiveStatus)
|
||||
useClusterEndpoint := m.pollVMsAndContainersEfficient(ctx, instanceName, instanceCfg.ClusterName, instanceCfg.IsCluster, client, nodeEffectiveStatus)
|
||||
|
||||
if !useClusterEndpoint {
|
||||
// Fall back to traditional polling only if cluster/resources not available
|
||||
|
|
@ -5202,10 +5303,10 @@ func (m *Monitor) pollPVEInstance(ctx context.Context, instanceName string, clie
|
|||
|
||||
// Use optimized parallel polling for better performance
|
||||
if instanceCfg.MonitorVMs {
|
||||
m.pollVMsWithNodes(ctx, instanceName, client, nodes, nodeEffectiveStatus)
|
||||
m.pollVMsWithNodes(ctx, instanceName, instanceCfg.ClusterName, instanceCfg.IsCluster, client, nodes, nodeEffectiveStatus)
|
||||
}
|
||||
if instanceCfg.MonitorContainers {
|
||||
m.pollContainersWithNodes(ctx, instanceName, client, nodes, nodeEffectiveStatus)
|
||||
m.pollContainersWithNodes(ctx, instanceName, instanceCfg.ClusterName, instanceCfg.IsCluster, client, nodes, nodeEffectiveStatus)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
|
@ -5666,8 +5767,14 @@ func (m *Monitor) pollPVEInstance(ctx context.Context, instanceName string, clie
|
|||
|
||||
// pollVMsAndContainersEfficient uses the cluster/resources endpoint to get all VMs and containers in one call
|
||||
// This works on both clustered and standalone nodes for efficient polling
|
||||
func (m *Monitor) pollVMsAndContainersEfficient(ctx context.Context, instanceName string, client PVEClientInterface, nodeEffectiveStatus map[string]string) bool {
|
||||
log.Info().Str("instance", instanceName).Msg("Polling VMs and containers using efficient cluster/resources endpoint")
|
||||
// When the instance is part of a cluster, the cluster name is used for guest IDs to prevent duplicates
|
||||
// when multiple cluster nodes are configured as separate PVE instances.
|
||||
func (m *Monitor) pollVMsAndContainersEfficient(ctx context.Context, instanceName string, clusterName string, isCluster bool, client PVEClientInterface, nodeEffectiveStatus map[string]string) bool {
|
||||
log.Info().
|
||||
Str("instance", instanceName).
|
||||
Str("clusterName", clusterName).
|
||||
Bool("isCluster", isCluster).
|
||||
Msg("Polling VMs and containers using efficient cluster/resources endpoint")
|
||||
|
||||
// Get all resources in a single API call
|
||||
resources, err := client.GetClusterResources(ctx, "vm")
|
||||
|
|
@ -5696,13 +5803,8 @@ func (m *Monitor) pollVMsAndContainersEfficient(ctx context.Context, instanceNam
|
|||
var allContainers []models.Container
|
||||
|
||||
for _, res := range resources {
|
||||
// Avoid duplicating node name in ID when instance name equals node name
|
||||
var guestID string
|
||||
if instanceName == res.Node {
|
||||
guestID = fmt.Sprintf("%s-%d", res.Node, res.VMID)
|
||||
} else {
|
||||
guestID = fmt.Sprintf("%s-%s-%d", instanceName, res.Node, res.VMID)
|
||||
}
|
||||
// Use cluster-aware guest ID to prevent duplicates when multiple cluster nodes are configured
|
||||
guestID := makeGuestID(instanceName, clusterName, isCluster, res.VMID)
|
||||
|
||||
// Debug log the resource type
|
||||
log.Debug().
|
||||
|
|
|
|||
|
|
@ -185,7 +185,9 @@ func convertPoolInfoToModel(poolInfo *proxmox.ZFSPoolInfo) *models.ZFSPool {
|
|||
}
|
||||
|
||||
// pollVMsWithNodes polls VMs from all nodes in parallel using goroutines
|
||||
func (m *Monitor) pollVMsWithNodes(ctx context.Context, instanceName string, client PVEClientInterface, nodes []proxmox.Node, nodeEffectiveStatus map[string]string) {
|
||||
// When the instance is part of a cluster, the cluster name is used for guest IDs to prevent duplicates
|
||||
// when multiple cluster nodes are configured as separate PVE instances.
|
||||
func (m *Monitor) pollVMsWithNodes(ctx context.Context, instanceName string, clusterName string, isCluster bool, client PVEClientInterface, nodes []proxmox.Node, nodeEffectiveStatus map[string]string) {
|
||||
startTime := time.Now()
|
||||
|
||||
// Channel to collect VM results from each node
|
||||
|
|
@ -253,9 +255,8 @@ func (m *Monitor) pollVMsWithNodes(ctx context.Context, instanceName string, cli
|
|||
tags = strings.Split(vm.Tags, ";")
|
||||
}
|
||||
|
||||
// Create guest ID (stable across node migrations)
|
||||
// Format: instance-VMID
|
||||
guestID := fmt.Sprintf("%s-%d", instanceName, vm.VMID)
|
||||
// Use cluster-aware guest ID to prevent duplicates when multiple cluster nodes are configured
|
||||
guestID := makeGuestID(instanceName, clusterName, isCluster, vm.VMID)
|
||||
|
||||
guestRaw := VMMemoryRaw{
|
||||
ListingMem: vm.Mem,
|
||||
|
|
@ -873,7 +874,9 @@ func (m *Monitor) pollVMsWithNodes(ctx context.Context, instanceName string, cli
|
|||
}
|
||||
|
||||
// pollContainersWithNodes polls containers from all nodes in parallel using goroutines
|
||||
func (m *Monitor) pollContainersWithNodes(ctx context.Context, instanceName string, client PVEClientInterface, nodes []proxmox.Node, nodeEffectiveStatus map[string]string) {
|
||||
// When the instance is part of a cluster, the cluster name is used for guest IDs to prevent duplicates
|
||||
// when multiple cluster nodes are configured as separate PVE instances.
|
||||
func (m *Monitor) pollContainersWithNodes(ctx context.Context, instanceName string, clusterName string, isCluster bool, client PVEClientInterface, nodes []proxmox.Node, nodeEffectiveStatus map[string]string) {
|
||||
startTime := time.Now()
|
||||
|
||||
// Channel to collect container results from each node
|
||||
|
|
@ -967,9 +970,8 @@ func (m *Monitor) pollContainersWithNodes(ctx context.Context, instanceName stri
|
|||
tags = strings.Split(container.Tags, ";")
|
||||
}
|
||||
|
||||
// Create guest ID (stable across node migrations)
|
||||
// Format: instance-VMID
|
||||
guestID := fmt.Sprintf("%s-%d", instanceName, container.VMID)
|
||||
// Use cluster-aware guest ID to prevent duplicates when multiple cluster nodes are configured
|
||||
guestID := makeGuestID(instanceName, clusterName, isCluster, int(container.VMID))
|
||||
|
||||
// Calculate I/O rates
|
||||
currentMetrics := IOMetrics{
|
||||
|
|
@ -1713,7 +1715,15 @@ func (m *Monitor) pollPVENode(
|
|||
}
|
||||
|
||||
// Apply grace period for node status to prevent flapping
|
||||
nodeID := instanceName + "-" + node.Node
|
||||
// For clustered nodes, use clusterName-nodeName as the ID to deduplicate
|
||||
// when the same cluster is registered via multiple entry points
|
||||
// (e.g., agent installed with --enable-proxmox on multiple cluster nodes)
|
||||
var nodeID string
|
||||
if instanceCfg.IsCluster && instanceCfg.ClusterName != "" {
|
||||
nodeID = instanceCfg.ClusterName + "-" + node.Node
|
||||
} else {
|
||||
nodeID = instanceName + "-" + node.Node
|
||||
}
|
||||
effectiveStatus := node.Status
|
||||
now := time.Now()
|
||||
|
||||
|
|
|
|||
Loading…
Reference in a new issue