feat: wait for +2 blocks after previous fork tip to reset

Author: hstove

stacks-signer/src/chainstate.rs

@@ -143,6 +143,9 @@ pub struct ProposalEvalConfig {
     pub reorg_attempts_activity_timeout: Duration,
     /// Time to wait before submitting a block proposal to the stacks-node
     pub proposal_wait_for_parent_time: Duration,
+    /// How many blocks after a fork should we reset the replay set,
+    /// as a failsafe mechanism?
+    pub reset_replay_set_after_fork_blocks: u64,
 }
 
 impl From<&SignerConfig> for ProposalEvalConfig {
@@ -155,6 +158,7 @@ impl From<&SignerConfig> for ProposalEvalConfig {
             reorg_attempts_activity_timeout: value.reorg_attempts_activity_timeout,
             tenure_idle_timeout_buffer: value.tenure_idle_timeout_buffer,
             proposal_wait_for_parent_time: value.proposal_wait_for_parent_time,
+            reset_replay_set_after_fork_blocks: value.reset_replay_set_after_fork_blocks,
         }
     }
 }

stacks-signer/src/client/mod.rs

@@ -432,6 +432,7 @@ pub(crate) mod tests {
             reorg_attempts_activity_timeout: config.reorg_attempts_activity_timeout,
             proposal_wait_for_parent_time: config.proposal_wait_for_parent_time,
             validate_with_replay_tx: config.validate_with_replay_tx,
+            reset_replay_set_after_fork_blocks: config.reset_replay_set_after_fork_blocks,
         }
     }
 

stacks-signer/src/config.rs

@@ -49,6 +49,9 @@ const DEFAULT_TENURE_IDLE_TIMEOUT_BUFFER_SECS: u64 = 2;
 ///  cannot determine that our stacks-node has processed the parent
 ///  block
 const DEFAULT_PROPOSAL_WAIT_TIME_FOR_PARENT_SECS: u64 = 15;
+/// Default number of blocks after a fork to reset the replay set,
+/// as a failsafe mechanism
+const DEFAULT_RESET_REPLAY_SET_AFTER_FORK_BLOCKS: u64 = 2;
 
 #[derive(thiserror::Error, Debug)]
 /// An error occurred parsing the provided configuration
@@ -184,6 +187,9 @@ pub struct SignerConfig {
     pub proposal_wait_for_parent_time: Duration,
     /// Whether or not to validate blocks with replay transactions
     pub validate_with_replay_tx: bool,
+    /// How many blocks after a fork should we reset the replay set,
+    /// as a failsafe mechanism?
+    pub reset_replay_set_after_fork_blocks: u64,
 }
 
 /// The parsed configuration for the signer
@@ -237,6 +243,9 @@ pub struct GlobalConfig {
     pub dry_run: bool,
     /// Whether or not to validate blocks with replay transactions
     pub validate_with_replay_tx: bool,
+    /// How many blocks after a fork should we reset the replay set,
+    /// as a failsafe mechanism?
+    pub reset_replay_set_after_fork_blocks: u64,
 }
 
 /// Internal struct for loading up the config file
@@ -288,6 +297,9 @@ struct RawConfigFile {
     pub dry_run: Option<bool>,
     /// Whether or not to validate blocks with replay transactions
     pub validate_with_replay_tx: Option<bool>,
+    /// How many blocks after a fork should we reset the replay set,
+    /// as a failsafe mechanism?
+    pub reset_replay_set_after_fork_blocks: Option<u64>,
 }
 
 impl RawConfigFile {
@@ -413,6 +425,10 @@ impl TryFrom<RawConfigFile> for GlobalConfig {
         // https://github.com/stacks-network/stacks-core/issues/6087
         let validate_with_replay_tx = raw_data.validate_with_replay_tx.unwrap_or(false);
 
+        let reset_replay_set_after_fork_blocks = raw_data
+            .reset_replay_set_after_fork_blocks
+            .unwrap_or(DEFAULT_RESET_REPLAY_SET_AFTER_FORK_BLOCKS);
+
         Ok(Self {
             node_host: raw_data.node_host,
             endpoint,
@@ -435,6 +451,7 @@ impl TryFrom<RawConfigFile> for GlobalConfig {
             tenure_idle_timeout_buffer,
             proposal_wait_for_parent_time,
             validate_with_replay_tx,
+            reset_replay_set_after_fork_blocks,
         })
     }
 }
@@ -714,12 +731,14 @@ network = "mainnet"
 auth_password = "abcd"
 db_path = ":memory:"
 validate_with_replay_tx = true
+reset_replay_set_after_fork_blocks = 100
             "#
         );
         let config = GlobalConfig::load_from_str(&config_toml).unwrap();
         assert_eq!(config.stacks_address.to_string(), expected_addr);
         assert_eq!(config.to_chain_id(), CHAIN_ID_MAINNET);
         assert!(config.validate_with_replay_tx);
+        assert_eq!(config.reset_replay_set_after_fork_blocks, 100);
     }
 
     #[test]

stacks-signer/src/runloop.rs

@@ -329,6 +329,7 @@ impl<Signer: SignerTrait<T>, T: StacksMessageCodec + Clone + Send + Debug> RunLo
             reorg_attempts_activity_timeout: self.config.reorg_attempts_activity_timeout,
             proposal_wait_for_parent_time: self.config.proposal_wait_for_parent_time,
             validate_with_replay_tx: self.config.validate_with_replay_tx,
+            reset_replay_set_after_fork_blocks: self.config.reset_replay_set_after_fork_blocks,
         }))
     }
 

stacks-signer/src/tests/chainstate.rs

@@ -92,6 +92,7 @@ fn setup_test_environment(
             tenure_idle_timeout_buffer: Duration::from_secs(2),
             reorg_attempts_activity_timeout: Duration::from_secs(3),
             proposal_wait_for_parent_time: Duration::from_secs(0),
+            reset_replay_set_after_fork_blocks: 2,
         },
     };
 

stacks-signer/src/v0/signer.rs

@@ -127,6 +127,8 @@ pub struct Signer {
     pub validate_with_replay_tx: bool,
     /// Scope of Tx Replay in terms of Burn block boundaries
     pub tx_replay_scope: ReplayScopeOpt,
+    /// The number of blocks after the past tip to reset the replay set
+    pub reset_replay_set_after_fork_blocks: u64,
 }
 
 impl std::fmt::Display for SignerMode {
@@ -244,6 +246,7 @@ impl SignerTrait<SignerMessage> for Signer {
             global_state_evaluator,
             validate_with_replay_tx: signer_config.validate_with_replay_tx,
             tx_replay_scope: None,
+            reset_replay_set_after_fork_blocks: signer_config.reset_replay_set_after_fork_blocks,
         }
     }
 

stacks-signer/src/v0/signer_state.rs

@@ -636,7 +636,7 @@ impl LocalStateMachine {
             } else if Self::handle_possible_replay_failsafe(
                 &replay_state,
                 &expected_burn_block,
-                client,
+                proposal_config.reset_replay_set_after_fork_blocks,
             )? {
                 info!(
                     "Signer state: replay set is stalled after 2 tenures. Clearing the replay set."
@@ -1226,60 +1226,16 @@ impl LocalStateMachine {
     fn handle_possible_replay_failsafe(
         replay_state: &ReplayState,
         new_burn_block: &NewBurnBlock,
-        client: &StacksClient,
+        reset_replay_set_after_fork_blocks: u64,
     ) -> Result<bool, SignerChainstateError> {
         let ReplayState::InProgress(_, replay_scope) = replay_state else {
             // Not in replay - skip
             return Ok(false);
         };
 
-        // if replay_scope.fork_origin.burn_block_height + 2 >= new_burn_block.burn_block_height {
-        if new_burn_block.burn_block_height < replay_scope.fork_origin.burn_block_height + 2 {
-            // We havent' had two burn blocks yet - skip
-            return Ok(false);
-        }
-
-        info!("Signer state: checking for replay set failsafe";
-            "replay_scope.fork_origin.burn_block_height" => replay_scope.fork_origin.burn_block_height,
-            "new_burn_block.burn_block_height" => new_burn_block.burn_block_height,
-        );
-        let Ok(fork_info) = client.get_tenure_forking_info(
-            &replay_scope.fork_origin.consensus_hash,
-            &new_burn_block.consensus_hash,
-        ) else {
-            warn!("Signer state: failed to get fork info");
-            return Ok(false);
-        };
-
-        let tenures_with_sortition = fork_info
-            .iter()
-            .filter(|fork_info| {
-                fork_info.was_sortition
-                    && fork_info
-                        .nakamoto_blocks
-                        .as_ref()
-                        .map(|b| b.len())
-                        .unwrap_or(0)
-                        > 0
-            })
-            .count();
-
-        info!("Signer state: fork info in failsafe check";
-            "tenures_with_sortition" => tenures_with_sortition,
-            "fork_info" => ?fork_info,
-        );
-
-        if tenures_with_sortition < 2 {
-            // We might have had 2 burn blocks, but not 2 tenures.
-            return Ok(false);
-        }
-
-        let forked_txs = Self::get_forked_txs_from_fork_info(&fork_info);
-
-        info!("Signer state: forked txs in failsafe check";
-            "forked_txs_len" => forked_txs.len(),
-        );
+        let failsafe_height =
+            replay_scope.past_tip.burn_block_height + reset_replay_set_after_fork_blocks;
 
-        Ok(forked_txs.is_empty())
+        Ok(new_burn_block.burn_block_height > failsafe_height)
     }
 }

testnet/stacks-node/src/tests/nakamoto_integrations.rs

@@ -6589,6 +6589,7 @@ fn signer_chainstate() {
             tenure_idle_timeout: Duration::from_secs(300),
             tenure_idle_timeout_buffer: Duration::from_secs(2),
             reorg_attempts_activity_timeout: Duration::from_secs(30),
+            reset_replay_set_after_fork_blocks: 2,
         };
         let mut sortitions_view =
             SortitionsView::fetch_view(proposal_conf, &signer_client).unwrap();
@@ -6716,6 +6717,7 @@ fn signer_chainstate() {
             tenure_idle_timeout: Duration::from_secs(300),
             tenure_idle_timeout_buffer: Duration::from_secs(2),
             reorg_attempts_activity_timeout: Duration::from_secs(30),
+            reset_replay_set_after_fork_blocks: 2,
         };
         let burn_block_height = SortitionDB::get_canonical_burn_chain_tip(sortdb.conn())
             .unwrap()
@@ -6794,6 +6796,7 @@ fn signer_chainstate() {
         tenure_idle_timeout: Duration::from_secs(300),
         tenure_idle_timeout_buffer: Duration::from_secs(2),
         reorg_attempts_activity_timeout: Duration::from_secs(30),
+        reset_replay_set_after_fork_blocks: 2,
     };
     let mut sortitions_view = SortitionsView::fetch_view(proposal_conf, &signer_client).unwrap();
     sortitions_view