fixes

Author: mohammeds1992

cmd/push-validator/root_cobra.go

@@ -781,12 +781,14 @@ func handlePostStartFlow(cfg config.Config, p *ui.Printer) bool {
 		}
 
 		// Create reconfigure function to get fresh trust parameters
-		reconfigFunc := func() error {
+		// Takes attempt number to rotate through fullnode RPCs on each retry
+		reconfigFunc := func(attempt int) error {
 			fmt.Println(p.Colors.Info("    Refreshing state sync parameters..."))
 			svc := bootstrap.New()
 			return svc.ReconfigureStateSync(context.Background(), bootstrap.Options{
 				HomeDir:     cfg.HomeDir,
-				SnapshotRPC: "", // Use default fullnode RPC
+				SnapshotRPC: "", // Empty = rotate through fullnode RPCs based on attempt
+				Attempt:     attempt,
 			})
 		}
 

internal/bootstrap/bootstrap.go

@@ -40,6 +40,7 @@ type Options struct {
 	BinPath              string       // pchaind path
 	SnapshotRPC string // e.g., https://donut.rpc.push.org
 	Progress             func(string) // optional callback for progress messages
+	Attempt              int          // Retry attempt number (0-based), used to rotate fullnode RPCs
 }
 
 type Service interface {
@@ -199,7 +200,7 @@ func (s *svc) Init(ctx context.Context, opts Options) error {
 		TrustHash:           tp.Hash,
 		RPCServers:          rpcServers,
 		TrustPeriod:         "336h0m0s",
-		ChunkFetchers:       12,      // Aggressive: 3x parallel downloads for faster sync
+		ChunkFetchers:       6,       // Balanced: 3 per peer when 2 peers share snapshot
 		ChunkRequestTimeout: "60m0s", // Extended timeout for slow/congested networks
 		DiscoveryTime:       "90s",   // More time to discover all available snapshots
 	}); err != nil {
@@ -217,14 +218,19 @@ func (s *svc) Init(ctx context.Context, opts Options) error {
 
 // ReconfigureStateSync updates state sync configuration with fresh trust parameters.
 // This is used during retry logic when the previous sync attempt failed.
+// The Attempt field in opts is used to rotate through fullnode RPCs, so each retry
+// uses a different peer's snapshot (which may be available from multiple peers).
 func (s *svc) ReconfigureStateSync(ctx context.Context, opts Options) error {
 	if opts.HomeDir == "" {
 		return errors.New("HomeDir required")
 	}
 
 	snapshotRPC := opts.SnapshotRPC
 	if snapshotRPC == "" {
-		snapshotRPC = fullnodeRPCs[0]
+		// Rotate through fullnode RPCs based on attempt number
+		// This ensures each retry uses a different peer's trust height/snapshot
+		rpcIndex := opts.Attempt % len(fullnodeRPCs)
+		snapshotRPC = fullnodeRPCs[rpcIndex]
 	}
 
 	// Compute fresh trust parameters
@@ -245,7 +251,7 @@ func (s *svc) ReconfigureStateSync(ctx context.Context, opts Options) error {
 		TrustHash:           tp.Hash,
 		RPCServers:          rpcServers,
 		TrustPeriod:         "336h0m0s",
-		ChunkFetchers:       12,
+		ChunkFetchers:       6,
 		ChunkRequestTimeout: "60m0s",
 		DiscoveryTime:       "90s",
 	})

internal/sync/monitor.go

@@ -674,9 +674,9 @@ func Run(ctx context.Context, opts Options) error {
 // RetryOptions extends Options with retry configuration
 type RetryOptions struct {
 	Options
-	MaxRetries   int          // Max retry attempts (default: 3)
-	ResetFunc    func() error // Function to reset data before retry
-	ReconfigFunc func() error // Function to reconfigure state sync params
+	MaxRetries   int               // Max retry attempts (default: 3)
+	ResetFunc    func() error      // Function to reset data before retry
+	ReconfigFunc func(int) error   // Function to reconfigure state sync params (receives attempt number)
 }
 
 // RunWithRetry runs sync monitoring with automatic retry on failure
@@ -702,8 +702,9 @@ func RunWithRetry(ctx context.Context, opts RetryOptions) error {
 			}
 
 			// Optionally reconfigure (e.g., get fresh trust height)
+			// Pass attempt number to allow RPC rotation on each retry
 			if opts.ReconfigFunc != nil {
-				_ = opts.ReconfigFunc() // Best-effort reconfigure
+				_ = opts.ReconfigFunc(attempt) // Best-effort reconfigure
 			}
 
 			// Wait before retry (exponential backoff: 10s, 20s, 30s)