Chain metadata service not updating metadata on reorgs (#1890)

Merge pull request #1890

- The chain metadata service keeps an internal copy of the chain metadata that it distributes to remote nodes. It should keep this copy uptodate with the blockchain db based on received BlockEvents. It only updated the metadata on block add events and not reorg events, this could have resulted in outdated metadata being shared on the network.
- Fixed an issue with the chain metadata service where it was not updating its copy of the chain metadata on reorg events. This could have resulted in some nodes on the network reporting outdated chain metadata.
- The chain metadata service will look for both block add and reorg events to determine if the blockchain db state was updated and the latest chain metadata needs to be retrieved.
- Also, changed the max send retries used by the Message protocol to 1.

* pull/1890/head:
  - The chain metadata service keeps an internal copy of the chain metadata that it distributes to remote nodes. It should keep this copy uptodate with the blockchain db based on received BlockEvents. It only updated the metadata on block add events and not reorg events, this could have resulted in outdated metadata being shared on the network. - Fixed an issue with the chain metadata service where it was not updating its copy of the chain metadata on reorg events. This could have resulted in some nodes on the network reporting outdated chain metadata. - The chain metadata service will look for both block add and reorg events to determine if the blockchain db state was updated and the latest chain metadata needs to be retrieved. - Also, chained the max send retries used by the Message protocol to 1.

base_layer/core/src/base_node/chain_metadata_service/service.rs

@@ -24,7 +24,7 @@ use super::{error::ChainMetadataSyncError, LOG_TARGET};
 use crate::{
     base_node::{
         chain_metadata_service::handle::{ChainMetadataEvent, PeerChainMetadata},
-        comms_interface::{BlockEvent, Broadcast, LocalNodeCommsInterface},
+        comms_interface::{BlockEvent, LocalNodeCommsInterface},
         proto,
     },
     chain_storage::BlockAddResult,
@@ -109,9 +109,9 @@ impl ChainMetadataService {
 
     /// Handle BlockEvents
     async fn handle_block_event(&mut self, event: &BlockEvent) -> Result<(), ChainMetadataSyncError> {
-        let _broadcast = Broadcast::from(true);
         match event {
-            BlockEvent::Verified((_, BlockAddResult::Ok, _broadcast)) => {
+            BlockEvent::Verified((_, BlockAddResult::Ok, _)) |
+            BlockEvent::Verified((_, BlockAddResult::ChainReorg(_), _)) => {
                 self.update_liveness_chain_metadata().await?;
             },
             BlockEvent::Verified(_) | BlockEvent::Invalid(_) => {},

comms/src/builder/consts.rs

@@ -42,4 +42,4 @@ pub const MESSAGING_EVENTS_BUFFER_SIZE: usize = 100;
 pub const MESSAGING_REQUEST_BUFFER_SIZE: usize = 50;
 /// The default maximum number of times to retry sending a failed message before publishing a SendMessageFailed event.
 /// This can be low because dialing a peer is already attempted a number of times.
-pub const MESSAGING_MAX_SEND_RETRIES: usize = 2;
+pub const MESSAGING_MAX_SEND_RETRIES: usize = 1;