Merge pull request #6093 from nymtech/bugfix/ns-api-node-custom-http-api-port

ns-api: fix scraping bug when operator specifies custom node HTTP API port in bond
This commit is contained in:
benedetta davico
2025-10-30 16:49:02 +01:00
committed by GitHub
15 changed files with 221 additions and 67 deletions
Generated
+3 -3
View File
@@ -133,9 +133,9 @@ checksum = "683d7910e743518b0e34f1186f92494becacb047c7b6bf616c96772180fef923"
[[package]]
name = "ammonia"
version = "4.1.1"
version = "4.1.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d6b346764dd0814805de8abf899fe03065bcee69bb1a4771c785817e39f3978f"
checksum = "17e913097e1a2124b46746c980134e8c954bc17a6a59bb3fde96f088d126dde6"
dependencies = [
"cssparser",
"html5ever",
@@ -6546,7 +6546,7 @@ dependencies = [
[[package]]
name = "nym-node-status-api"
version = "4.0.10"
version = "4.0.11-rc1"
dependencies = [
"ammonia",
"anyhow",
@@ -1,6 +1,6 @@
{
"db_name": "PostgreSQL",
"query": "SELECT\n node_id,\n ed25519_identity_pubkey,\n total_stake,\n ip_addresses as \"ip_addresses!: serde_json::Value\",\n mix_port,\n x25519_sphinx_pubkey,\n node_role as \"node_role: serde_json::Value\",\n supported_roles as \"supported_roles: serde_json::Value\",\n entry as \"entry: serde_json::Value\",\n performance,\n self_described as \"self_described: serde_json::Value\",\n bond_info as \"bond_info: serde_json::Value\"\n FROM\n nym_nodes\n ORDER BY\n node_id\n ",
"query": "SELECT\n node_id,\n ed25519_identity_pubkey,\n total_stake,\n ip_addresses as \"ip_addresses!: serde_json::Value\",\n mix_port,\n x25519_sphinx_pubkey,\n node_role as \"node_role: serde_json::Value\",\n supported_roles as \"supported_roles: serde_json::Value\",\n entry as \"entry: serde_json::Value\",\n performance,\n self_described as \"self_described: serde_json::Value\",\n bond_info as \"bond_info: serde_json::Value\",\n http_api_port\n FROM\n nym_nodes\n WHERE\n self_described IS NOT NULL\n AND\n bond_info IS NOT NULL\n ",
"describe": {
"columns": [
{
@@ -62,6 +62,11 @@
"ordinal": 11,
"name": "bond_info: serde_json::Value",
"type_info": "Jsonb"
},
{
"ordinal": 12,
"name": "http_api_port",
"type_info": "Int4"
}
],
"parameters": {
@@ -79,8 +84,9 @@
true,
false,
true,
true,
true
]
},
"hash": "c48d04fc3de59dd484f0a63d40336ced54e08785f77e9ef85f3157d004ec85dc"
"hash": "0b51df277ed66c6553f66af9b135342dee177abc1c92e4a89147de3c22d3d1a5"
}
@@ -1,6 +1,6 @@
{
"db_name": "PostgreSQL",
"query": "SELECT\n node_id,\n ed25519_identity_pubkey,\n total_stake,\n ip_addresses as \"ip_addresses!: serde_json::Value\",\n mix_port,\n x25519_sphinx_pubkey,\n node_role as \"node_role: serde_json::Value\",\n supported_roles as \"supported_roles: serde_json::Value\",\n entry as \"entry: serde_json::Value\",\n performance,\n self_described as \"self_described: serde_json::Value\",\n bond_info as \"bond_info: serde_json::Value\"\n FROM\n nym_nodes\n WHERE\n self_described IS NOT NULL\n AND\n bond_info IS NOT NULL\n ",
"query": "SELECT\n node_id,\n ed25519_identity_pubkey,\n total_stake,\n ip_addresses as \"ip_addresses!: serde_json::Value\",\n mix_port,\n x25519_sphinx_pubkey,\n node_role as \"node_role: serde_json::Value\",\n supported_roles as \"supported_roles: serde_json::Value\",\n entry as \"entry: serde_json::Value\",\n performance,\n self_described as \"self_described: serde_json::Value\",\n bond_info as \"bond_info: serde_json::Value\",\n http_api_port\n FROM\n nym_nodes\n ORDER BY\n node_id\n ",
"describe": {
"columns": [
{
@@ -62,6 +62,11 @@
"ordinal": 11,
"name": "bond_info: serde_json::Value",
"type_info": "Jsonb"
},
{
"ordinal": 12,
"name": "http_api_port",
"type_info": "Int4"
}
],
"parameters": {
@@ -79,8 +84,9 @@
true,
false,
true,
true,
true
]
},
"hash": "283f49a65c7d70bf271702ff6a5c7ad6e68c81932d295ff18ed198c54706a57c"
"hash": "3ddc12cc4e1796b787a50c40560d2bd71d1cfe5f5265e6f161b3122d1317a421"
}
@@ -1,26 +0,0 @@
{
"db_name": "PostgreSQL",
"query": "INSERT INTO nym_nodes\n (node_id, ed25519_identity_pubkey,\n total_stake,\n ip_addresses, mix_port,\n x25519_sphinx_pubkey, node_role,\n supported_roles, entry,\n self_described,\n bond_info,\n performance, last_updated_utc\n )\n VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11, $12, $13)\n ON CONFLICT(node_id) DO UPDATE SET\n ed25519_identity_pubkey=excluded.ed25519_identity_pubkey,\n ip_addresses=excluded.ip_addresses,\n mix_port=excluded.mix_port,\n x25519_sphinx_pubkey=excluded.x25519_sphinx_pubkey,\n node_role=excluded.node_role,\n supported_roles=excluded.supported_roles,\n entry=excluded.entry,\n self_described=excluded.self_described,\n bond_info=excluded.bond_info,\n performance=excluded.performance,\n last_updated_utc=excluded.last_updated_utc\n ;",
"describe": {
"columns": [],
"parameters": {
"Left": [
"Int4",
"Varchar",
"Int8",
"Jsonb",
"Int4",
"Varchar",
"Jsonb",
"Jsonb",
"Jsonb",
"Jsonb",
"Jsonb",
"Varchar",
"Int4"
]
},
"nullable": []
},
"hash": "b010fb91828f7e4f0b72bdfe3b58b2abb437cccdb6ebd2e1087cc822ed737b0e"
}
@@ -0,0 +1,27 @@
{
"db_name": "PostgreSQL",
"query": "INSERT INTO nym_nodes\n (node_id, ed25519_identity_pubkey,\n total_stake,\n ip_addresses, mix_port,\n x25519_sphinx_pubkey, node_role,\n supported_roles, entry,\n self_described,\n bond_info,\n performance, last_updated_utc, http_api_port\n )\n VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11, $12, $13, $14)\n ON CONFLICT(node_id) DO UPDATE SET\n ed25519_identity_pubkey=excluded.ed25519_identity_pubkey,\n ip_addresses=excluded.ip_addresses,\n mix_port=excluded.mix_port,\n x25519_sphinx_pubkey=excluded.x25519_sphinx_pubkey,\n node_role=excluded.node_role,\n supported_roles=excluded.supported_roles,\n entry=excluded.entry,\n self_described=excluded.self_described,\n bond_info=excluded.bond_info,\n performance=excluded.performance,\n last_updated_utc=excluded.last_updated_utc,\n http_api_port=excluded.http_api_port\n ;",
"describe": {
"columns": [],
"parameters": {
"Left": [
"Int4",
"Varchar",
"Int8",
"Jsonb",
"Int4",
"Varchar",
"Jsonb",
"Jsonb",
"Jsonb",
"Jsonb",
"Jsonb",
"Varchar",
"Int4",
"Int4"
]
},
"nullable": []
},
"hash": "dde9aff827c34086077927bbe33fa3d5c939e7122ba7c88b78a353f00b271ec2"
}
@@ -3,7 +3,7 @@
[package]
name = "nym-node-status-api"
version = "4.0.10"
version = "4.0.11-rc1"
authors.workspace = true
repository.workspace = true
homepage.workspace = true
@@ -0,0 +1,2 @@
ALTER TABLE nym_nodes
ADD COLUMN IF NOT EXISTS http_api_port INTEGER;
@@ -1,5 +1,5 @@
use crate::ticketbook_manager::TicketbookManagerConfig;
use clap::Parser;
use clap::{Parser, Subcommand};
use nym_bin_common::bin_info;
use nym_credential_proxy_lib::shared_state::ecash_state::TicketType;
use reqwest::Url;
@@ -105,6 +105,19 @@ pub(crate) struct Cli {
#[clap(flatten)]
pub(crate) ticketbook: TicketbookArgs,
#[command(subcommand)]
pub(crate) command: Option<Commands>,
}
#[derive(Subcommand, Debug)]
pub(crate) enum Commands {
/// Scrape a single node and output detailed debug logs
ScrapeNode {
/// The id of the node to scrape
#[arg(long)]
node_id: i64,
},
}
#[derive(Debug, Parser)]
@@ -381,7 +381,7 @@ impl ScrapeNodeKind {
pub(crate) struct ScraperNodeInfo {
pub node_kind: ScrapeNodeKind,
pub hosts: Vec<String>,
pub http_api_port: i64,
pub http_api_port: Option<u16>,
}
impl ScraperNodeInfo {
@@ -395,8 +395,21 @@ impl ScraperNodeInfo {
format!("http://{}", host),
]);
if self.http_api_port != DEFAULT_NYM_NODE_HTTP_PORT as i64 {
urls.insert(0, format!("http://{}:{}", host, self.http_api_port));
if let Some(custom_http_api_port) = self.http_api_port {
urls = Vec::new();
for host in &self.hosts {
urls.append(&mut vec![format!(
"http://{}:{}",
host, custom_http_api_port
)]);
}
// do not fall back to default ports, if the operator sets a custom http api port
// in their bond, use it and error out if it's not available
// this will correctly handle cases where some operators run multiple nodes
// on a single IP address and assign different custom http port apis at bond time
// urls.insert(0, format!("http://{}:{}", host, custom_http_api_port));
}
}
@@ -423,6 +436,7 @@ pub(crate) struct NymNodeDto {
pub performance: String,
pub self_described: Option<serde_json::Value>,
pub bond_info: Option<serde_json::Value>,
pub http_api_port: Option<i32>,
}
#[allow(dead_code)] // it's not dead code but clippy doesn't detect usage in sqlx macros
@@ -440,6 +454,7 @@ pub(crate) struct NymNodeInsertRecord {
pub entry: Option<serde_json::Value>,
pub self_described: Option<serde_json::Value>,
pub bond_info: Option<serde_json::Value>,
pub http_api_port: Option<i32>,
pub last_updated_utc: i64,
}
@@ -456,6 +471,12 @@ impl NymNodeInsertRecord {
.map(|info| decimal_to_i64(info.total_stake()))
.unwrap_or(0);
let entry = serialize_opt_to_value!(skimmed_node.entry)?;
let http_api_port = bond_info.and_then(|bond| {
bond.bond_information
.node
.custom_http_port
.map(|port| port as i32)
});
let bond_info = serialize_opt_to_value!(bond_info)?;
let self_described = serialize_opt_to_value!(self_described)?;
@@ -472,6 +493,7 @@ impl NymNodeInsertRecord {
entry,
self_described,
bond_info,
http_api_port,
last_updated_utc: now,
};
@@ -35,7 +35,8 @@ pub(crate) async fn get_all_nym_nodes(pool: &DbPool) -> anyhow::Result<Vec<NymNo
entry as "entry: serde_json::Value",
performance,
self_described as "self_described: serde_json::Value",
bond_info as "bond_info: serde_json::Value"
bond_info as "bond_info: serde_json::Value",
http_api_port
FROM
nym_nodes
ORDER BY
@@ -72,7 +73,8 @@ pub(crate) async fn get_described_bonded_nym_nodes(
entry as "entry: serde_json::Value",
performance,
self_described as "self_described: serde_json::Value",
bond_info as "bond_info: serde_json::Value"
bond_info as "bond_info: serde_json::Value",
http_api_port
FROM
nym_nodes
WHERE
@@ -115,9 +117,9 @@ pub(crate) async fn update_nym_nodes(
supported_roles, entry,
self_described,
bond_info,
performance, last_updated_utc
performance, last_updated_utc, http_api_port
)
VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11, $12, $13)
VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11, $12, $13, $14)
ON CONFLICT(node_id) DO UPDATE SET
ed25519_identity_pubkey=excluded.ed25519_identity_pubkey,
ip_addresses=excluded.ip_addresses,
@@ -129,7 +131,8 @@ pub(crate) async fn update_nym_nodes(
self_described=excluded.self_described,
bond_info=excluded.bond_info,
performance=excluded.performance,
last_updated_utc=excluded.last_updated_utc
last_updated_utc=excluded.last_updated_utc,
http_api_port=excluded.http_api_port
;",
record.node_id,
record.ed25519_identity_pubkey,
@@ -144,6 +147,7 @@ pub(crate) async fn update_nym_nodes(
record.bond_info,
record.performance,
record.last_updated_utc as i32,
record.http_api_port,
)
.execute(&mut *tx)
.await
@@ -21,10 +21,11 @@ pub(crate) async fn get_nodes_for_scraping(pool: &DbPool) -> Result<Vec<ScraperN
let skimmed_nodes = queries::get_described_bonded_nym_nodes(pool)
.await
.map(|nodes_dto| {
nodes_dto.into_iter().filter_map(|node| {
let node_id = node.node_id;
match SkimmedNode::try_from(node) {
Ok(node) => Some(node),
nodes_dto.into_iter().filter_map(|node_dto| {
let node_id = node_dto.node_id;
let http_api_port = node_dto.http_api_port;
match SkimmedNode::try_from(node_dto) {
Ok(node) => Some((node, http_api_port)),
Err(e) => {
tracing::error!("Failed to decode node_id={}: {}", node_id, e);
None
@@ -33,7 +34,7 @@ pub(crate) async fn get_nodes_for_scraping(pool: &DbPool) -> Result<Vec<ScraperN
})
})?;
skimmed_nodes.for_each(|node| {
skimmed_nodes.for_each(|(node, http_api_port)| {
// TODO: relies on polyfilling: Nym nodes table might contain legacy mixnodes
// as well. Categorize them here.
let node_kind = if gateway_keys.contains(&node.ed25519_identity_pubkey.to_base58_string()) {
@@ -54,7 +55,7 @@ pub(crate) async fn get_nodes_for_scraping(pool: &DbPool) -> Result<Vec<ScraperN
.into_iter()
.map(|ip| ip.to_string())
.collect::<Vec<_>>(),
http_api_port: node.mix_port.into(),
http_api_port: http_api_port.map(|port| port as u16),
})
});
@@ -138,6 +138,7 @@ mod db_tests {
performance: "1.0".to_string(),
self_described: None,
bond_info: None,
http_api_port: None,
};
let skimmed_node: nym_validator_client::nym_api::SkimmedNode =
@@ -362,22 +363,42 @@ fn test_scraper_node_info_contact_addresses() {
let node_info = ScraperNodeInfo {
node_kind: ScrapeNodeKind::MixingNymNode { node_id: 123 },
hosts: vec!["1.1.1.1".to_string(), "example.com".to_string()],
http_api_port: 8080,
http_api_port: None,
};
let addresses = node_info.contact_addresses();
// Should generate multiple URLs for each host
// Custom port (8080) should be inserted at the beginning
// When no custom port is specified only default ports should be used
assert!(addresses.contains(&"http://1.1.1.1:8080".to_string()));
assert!(addresses.contains(&"http://example.com:8080".to_string()));
assert!(addresses.contains(&"http://1.1.1.1:8000".to_string()));
assert!(addresses.contains(&"https://1.1.1.1".to_string()));
assert!(addresses.contains(&"http://1.1.1.1".to_string()));
assert!(addresses.contains(&"http://example.com:8000".to_string()));
// Check that URLs follow the expected pattern
assert!(addresses.len() >= 8); // At least 4 URLs per host
}
#[test]
fn test_scraper_node_info_contact_addresses_with_custom_http_api_port() {
use crate::db::models::{ScrapeNodeKind, ScraperNodeInfo};
let node_info = ScraperNodeInfo {
node_kind: ScrapeNodeKind::MixingNymNode { node_id: 123 },
hosts: vec!["1.1.1.1".to_string(), "example.com".to_string()],
http_api_port: Some(4444),
};
let addresses = node_info.contact_addresses();
// Should generate multiple URLs for each host
// Custom port (4444) should be the only port in the list
assert!(addresses.contains(&"http://1.1.1.1:4444".to_string()));
assert!(addresses.contains(&"http://example.com:4444".to_string()));
// Check that URLs follow the expected pattern
assert!(addresses.len() >= 2); // At least 4 URLs per host
}
#[test]
fn test_scrape_node_kind_node_id() {
use crate::db::models::ScrapeNodeKind;
@@ -414,6 +435,7 @@ fn test_nym_node_dto_with_invalid_keys() {
performance: "1.0".to_string(),
self_described: None,
bond_info: None,
http_api_port: None,
};
let result: Result<nym_validator_client::nym_api::SkimmedNode, _> = nym_node_dto.try_into();
@@ -451,6 +473,7 @@ fn test_nym_node_dto_with_invalid_performance() {
performance: "invalid_percent".to_string(),
self_described: None,
bond_info: None,
http_api_port: None,
};
let result: Result<nym_validator_client::nym_api::SkimmedNode, _> = nym_node_dto.try_into();
@@ -1,4 +1,6 @@
use crate::cli::Commands;
use crate::monitor::DelegationsCache;
use crate::node_scraper::helpers::scrape_and_store_description_by_node_id;
use crate::ticketbook_manager::TicketbookManager;
use crate::ticketbook_manager::state::TicketbookManagerState;
use clap::Parser;
@@ -40,11 +42,49 @@ async fn main() -> anyhow::Result<()> {
tracing::info!("Registered {} agent keys", agent_key_list.len());
let connection_url = args.database_url.clone();
tracing::debug!("Using config:\n{:#?}", args);
if std::env::var("SHOW_CONFIG").ok().is_some() {
tracing::debug!("Using config:\n{:#?}", args);
}
let storage = db::Storage::init(connection_url, args.sqlx_busy_timeout_s).await?;
let db_pool = storage.pool_owned();
// node geocache is shared between node monitor and HTTP server
let geocache = moka::future::Cache::builder()
.time_to_live(args.geodata_ttl)
.build();
let delegations_cache = DelegationsCache::new();
let client_config = nym_validator_client::nyxd::Config::try_from_nym_network_details(
&nym_network_defaults::NymNetworkDetails::new_from_env(),
)?;
let nyxd_client = NyxdClient::connect(client_config.clone(), args.nyxd_addr.as_str())
.map_err(|err| anyhow::anyhow!("Couldn't connect: {}", err))?;
match args.command {
Some(Commands::ScrapeNode { node_id }) => {
if std::env::var("RUN_ONCE_INIT_NODES").ok().is_some() {
let geocache_clone = geocache.clone();
let delegations_cache_clone = Arc::clone(&delegations_cache);
monitor::run_once(
db_pool.clone(),
args.nym_api_client_timeout,
nyxd_client,
args.ipinfo_api_token,
geocache_clone,
delegations_cache_clone,
)
.await?;
}
tracing::info!("Scraping node with id {node_id}...");
scrape_and_store_description_by_node_id(&db_pool, node_id).await?;
return Ok(());
}
None => {
// default behaviour
}
}
// Start the node scraper
let scraper = node_scraper::DescriptionScraper::new(storage.pool_owned());
shutdown_manager.spawn_with_shutdown(async move {
@@ -58,20 +98,9 @@ async fn main() -> anyhow::Result<()> {
scraper.start().await;
});
// node geocache is shared between node monitor and HTTP server
let geocache = moka::future::Cache::builder()
.time_to_live(args.geodata_ttl)
.build();
let delegations_cache = DelegationsCache::new();
// Start the monitor
let geocache_clone = geocache.clone();
let delegations_cache_clone = Arc::clone(&delegations_cache);
let client_config = nym_validator_client::nyxd::Config::try_from_nym_network_details(
&nym_network_defaults::NymNetworkDetails::new_from_env(),
)?;
let nyxd_client = NyxdClient::connect(client_config.clone(), args.nyxd_addr.as_str())
.map_err(|err| anyhow::anyhow!("Couldn't connect: {}", err))?;
shutdown_manager.spawn_with_shutdown(async move {
monitor::run_in_background(
@@ -68,7 +68,7 @@ pub(crate) async fn run_in_background(
loop {
tracing::info!("Refreshing node info...");
if let Err(e) = monitor.run().await {
if let Err(e) = monitor.run(false).await {
tracing::error!(
"Monitor run failed: {e}, retrying in {}s...",
MONITOR_FAILURE_RETRY_DELAY.as_secs()
@@ -84,8 +84,33 @@ pub(crate) async fn run_in_background(
}
}
#[instrument(level = "debug", name = "data_monitor", skip_all)]
pub(crate) async fn run_once(
db_pool: DbPool,
nym_api_client_timeout: Duration,
nyxd_client: nym_validator_client::QueryHttpRpcNyxdClient,
ipinfo_api_token: String,
geocache: NodeGeoCache,
node_delegations: Arc<RwLock<DelegationsCache>>,
) -> anyhow::Result<()> {
let ipinfo = IpInfoClient::new(ipinfo_api_token.clone());
let mut monitor = Monitor {
db_pool,
network_details: nym_network_defaults::NymNetworkDetails::new_from_env(),
nym_api_client_timeout,
nyxd_client,
ipinfo,
geocache,
node_delegations,
};
tracing::info!("Refreshing node info...");
monitor.run(true).await
}
impl Monitor {
async fn run(&mut self) -> anyhow::Result<()> {
async fn run(&mut self, exit_early: bool) -> anyhow::Result<()> {
self.check_ipinfo_bandwidth().await;
let default_api_url = self
@@ -153,6 +178,11 @@ impl Monitor {
tracing::debug!("{} nym nodes written to DB!", inserted);
})?;
// stop here if running once
if exit_early {
return Ok(());
}
// refresh geodata for all nodes
for node_description in described_nodes.values() {
self.location_cached(node_description).await;
@@ -118,6 +118,17 @@ pub fn sanitize_description(
}
}
pub async fn scrape_and_store_description_by_node_id(pool: &DbPool, node_id: i64) -> Result<()> {
let nodes = crate::db::queries::get_nodes_for_scraping(pool).await?;
match nodes.iter().find(|n| *n.node_kind.node_id() == node_id) {
Some(node) => Ok(scrape_and_store_description(pool, node.clone()).await?),
None => {
error!("Could not find node with id {node_id}");
Err(anyhow!("Could not find node with id {node_id}"))
}
}
}
pub async fn scrape_and_store_description(pool: &DbPool, node: ScraperNodeInfo) -> Result<()> {
let client = build_client()?;
let urls = node.contact_addresses();
@@ -152,7 +163,13 @@ pub async fn scrape_and_store_description(pool: &DbPool, node: ScraperNodeInfo)
anyhow::anyhow!("Failed to fetch description from any URL: {}", err_msg)
})?;
let sanitized_description = sanitize_description(description, *node.node_id());
let sanitized_description = sanitize_description(description.clone(), *node.node_id());
trace!("tried_url_list = {tried_url_list:?}");
trace!("ndoe_id = {}", node.node_id());
trace!("description = {:?}", description);
trace!("sanitized_description = {:?}", sanitized_description);
insert_scraped_node_description(pool, &node.node_kind, &sanitized_description).await?;
Ok(())