Proxy release 2025-05-06 06:03 UTC

Proxy release 2025-03-06

.github/workflows/benchmarking.yml

@@ -64,7 +64,7 @@ jobs:
     env:
       ORG_ID: org-solitary-dew-09443886
       LIMIT: 100
-      SEARCH: "GITHUB_RUN_ID="
+      SEARCH: "Created by actions/neon-project-create; GITHUB_RUN_ID"
       BASE_URL: https://console-stage.neon.build/api/v2
       DRY_RUN: "false"  # Set to "true" to just test out the workflow
 

Cargo.lock

@@ -1284,7 +1284,6 @@ name = "compute_tools"
 version = "0.1.0"
 dependencies = [
  "anyhow",
- "async-compression",
  "aws-config",
  "aws-sdk-kms",
  "aws-sdk-s3",
@@ -1421,7 +1420,6 @@ dependencies = [
  "clap",
  "comfy-table",
  "compute_api",
- "endpoint_storage",
  "futures",
  "http-utils",
  "humantime",

Cargo.toml

@@ -243,7 +243,6 @@ azure_storage_blobs = { git = "https://github.com/neondatabase/azure-sdk-for-rus
 ## Local libraries
 compute_api = { version = "0.1", path = "./libs/compute_api/" }
 consumption_metrics = { version = "0.1", path = "./libs/consumption_metrics/" }
-endpoint_storage = { version = "0.0.1", path = "./endpoint_storage/" }
 http-utils = { version = "0.1", path = "./libs/http-utils/" }
 metrics = { version = "0.1", path = "./libs/metrics/" }
 pageserver = { path = "./pageserver" }

compute/compute-node.Dockerfile

@@ -1084,12 +1084,23 @@ RUN cargo install --locked --version 0.12.9 cargo-pgrx && \
     /bin/bash -c 'cargo pgrx init --pg${PG_VERSION:1}=/usr/local/pgsql/bin/pg_config'
 
 USER root
-
 #########################################################################################
 #
 # Layer "rust extensions pgrx14"
 #
-# Version 14 is now required by a few
+#########################################################################################
+FROM pg-build-nonroot-with-cargo AS rust-extensions-build-pgrx14
+ARG PG_VERSION
+
+RUN cargo install --locked --version 0.14.1 cargo-pgrx && \
+    /bin/bash -c 'cargo pgrx init --pg${PG_VERSION:1}=/usr/local/pgsql/bin/pg_config'
+
+USER root
+#########################################################################################
+#
+# Layer "rust extensions pgrx14"
+#
+# Version 14 is now required by a few 
 # This layer should be used as a base for new pgrx extensions,
 # and eventually get merged with `rust-extensions-build`
 #
@@ -1322,8 +1333,8 @@ ARG PG_VERSION
 # Do not update without approve from proxy team
 # Make sure the version is reflected in proxy/src/serverless/local_conn_pool.rs
 WORKDIR /ext-src
-RUN wget https://github.com/neondatabase/pg_session_jwt/archive/refs/tags/v0.3.1.tar.gz -O pg_session_jwt.tar.gz && \
-    echo "62fec9e472cb805c53ba24a0765afdb8ea2720cfc03ae7813e61687b36d1b0ad pg_session_jwt.tar.gz" | sha256sum --check && \
+RUN wget https://github.com/neondatabase/pg_session_jwt/archive/refs/tags/v0.3.0.tar.gz -O pg_session_jwt.tar.gz && \
+    echo "19be2dc0b3834d643706ed430af998bb4c2cdf24b3c45e7b102bb3a550e8660c pg_session_jwt.tar.gz" | sha256sum --check && \
     mkdir pg_session_jwt-src && cd pg_session_jwt-src && tar xzf ../pg_session_jwt.tar.gz --strip-components=1 -C . && \
     sed -i 's/pgrx = "0.12.6"/pgrx = { version = "0.12.9", features = [ "unsafe-postgres" ] }/g' Cargo.toml && \
     sed -i 's/version = "0.12.6"/version = "0.12.9"/g' pgrx-tests/Cargo.toml && \
@@ -1351,8 +1362,7 @@ COPY compute/patches/anon_v2.patch .
 # This is an experimental extension, never got to real production.
 # !Do not remove! It can be present in shared_preload_libraries and compute will fail to start if library is not found.
 ENV PATH="/usr/local/pgsql/bin/:$PATH"
-RUN wget https://gitlab.com/dalibo/postgresql_anonymizer/-/archive/2.1.0/postgresql_anonymizer-latest.tar.gz -O pg_anon.tar.gz && \
-    echo "48e7f5ae2f1ca516df3da86c5c739d48dd780a4e885705704ccaad0faa89d6c0  pg_anon.tar.gz" | sha256sum --check && \
+RUN wget https://gitlab.com/dalibo/postgresql_anonymizer/-/archive/latest/postgresql_anonymizer-latest.tar.gz -O pg_anon.tar.gz && \
     mkdir pg_anon-src && cd pg_anon-src && tar xzf ../pg_anon.tar.gz --strip-components=1 -C . && \
     find /usr/local/pgsql -type f | sed 's|^/usr/local/pgsql/||' > /before.txt && \
     sed -i 's/pgrx = "0.14.1"/pgrx = { version = "=0.14.1", features = [ "unsafe-postgres" ] }/g' Cargo.toml && \

compute_tools/Cargo.toml

@@ -10,7 +10,6 @@ default = []
 testing = ["fail/failpoints"]
 
 [dependencies]
-async-compression.workspace = true
 base64.workspace = true
 aws-config.workspace = true
 aws-sdk-s3.workspace = true

compute_tools/src/compute.rs

@@ -1,10 +1,17 @@
+use std::collections::HashMap;
+use std::os::unix::fs::{PermissionsExt, symlink};
+use std::path::Path;
+use std::process::{Command, Stdio};
+use std::str::FromStr;
+use std::sync::atomic::{AtomicU32, Ordering};
+use std::sync::{Arc, Condvar, Mutex, RwLock};
+use std::time::{Duration, Instant};
+use std::{env, fs};
+
 use anyhow::{Context, Result};
 use chrono::{DateTime, Utc};
 use compute_api::privilege::Privilege;
-use compute_api::responses::{
-    ComputeConfig, ComputeCtlConfig, ComputeMetrics, ComputeStatus, LfcOffloadState,
-    LfcPrewarmState,
-};
+use compute_api::responses::{ComputeConfig, ComputeCtlConfig, ComputeMetrics, ComputeStatus};
 use compute_api::spec::{
     ComputeAudit, ComputeFeature, ComputeMode, ComputeSpec, ExtVersion, PgIdent,
 };
@@ -18,16 +25,6 @@ use postgres;
 use postgres::NoTls;
 use postgres::error::SqlState;
 use remote_storage::{DownloadError, RemotePath};
-use std::collections::HashMap;
-use std::net::SocketAddr;
-use std::os::unix::fs::{PermissionsExt, symlink};
-use std::path::Path;
-use std::process::{Command, Stdio};
-use std::str::FromStr;
-use std::sync::atomic::{AtomicU32, Ordering};
-use std::sync::{Arc, Condvar, Mutex, RwLock};
-use std::time::{Duration, Instant};
-use std::{env, fs};
 use tokio::spawn;
 use tracing::{Instrument, debug, error, info, instrument, warn};
 use utils::id::{TenantId, TimelineId};
@@ -153,9 +150,6 @@ pub struct ComputeState {
     /// set up the span relationship ourselves.
     pub startup_span: Option<tracing::span::Span>,
 
-    pub lfc_prewarm_state: LfcPrewarmState,
-    pub lfc_offload_state: LfcOffloadState,
-
     pub metrics: ComputeMetrics,
 }
 
@@ -169,8 +163,6 @@ impl ComputeState {
             pspec: None,
             startup_span: None,
             metrics: ComputeMetrics::default(),
-            lfc_prewarm_state: LfcPrewarmState::default(),
-            lfc_offload_state: LfcOffloadState::default(),
         }
     }
 
@@ -206,8 +198,6 @@ pub struct ParsedSpec {
     pub pageserver_connstr: String,
     pub safekeeper_connstrings: Vec<String>,
     pub storage_auth_token: Option<String>,
-    pub endpoint_storage_addr: Option<SocketAddr>,
-    pub endpoint_storage_token: Option<String>,
 }
 
 impl TryFrom<ComputeSpec> for ParsedSpec {
@@ -261,18 +251,6 @@ impl TryFrom<ComputeSpec> for ParsedSpec {
                 .or(Err("invalid timeline id"))?
         };
 
-        let endpoint_storage_addr: Option<SocketAddr> = spec
-            .endpoint_storage_addr
-            .clone()
-            .or_else(|| spec.cluster.settings.find("neon.endpoint_storage_addr"))
-            .unwrap_or_default()
-            .parse()
-            .ok();
-        let endpoint_storage_token = spec
-            .endpoint_storage_token
-            .clone()
-            .or_else(|| spec.cluster.settings.find("neon.endpoint_storage_token"));
-
         Ok(ParsedSpec {
             spec,
             pageserver_connstr,
@@ -280,8 +258,6 @@ impl TryFrom<ComputeSpec> for ParsedSpec {
             storage_auth_token,
             tenant_id,
             timeline_id,
-            endpoint_storage_addr,
-            endpoint_storage_token,
         })
     }
 }
@@ -760,9 +736,6 @@ impl ComputeNode {
         // Log metrics so that we can search for slow operations in logs
         info!(?metrics, postmaster_pid = %postmaster_pid, "compute start finished");
 
-        if pspec.spec.prewarm_lfc_on_startup {
-            self.prewarm_lfc();
-        }
         Ok(())
     }
 

compute_tools/src/compute_prewarm.rs

@@ -1,202 +0,0 @@
-use crate::compute::ComputeNode;
-use anyhow::{Context, Result, bail};
-use async_compression::tokio::bufread::{ZstdDecoder, ZstdEncoder};
-use compute_api::responses::LfcOffloadState;
-use compute_api::responses::LfcPrewarmState;
-use http::StatusCode;
-use reqwest::Client;
-use std::sync::Arc;
-use tokio::{io::AsyncReadExt, spawn};
-use tracing::{error, info};
-
-#[derive(serde::Serialize, Default)]
-pub struct LfcPrewarmStateWithProgress {
-    #[serde(flatten)]
-    base: LfcPrewarmState,
-    total: i32,
-    prewarmed: i32,
-    skipped: i32,
-}
-
-/// A pair of url and a token to query endpoint storage for LFC prewarm-related tasks
-struct EndpointStoragePair {
-    url: String,
-    token: String,
-}
-
-const KEY: &str = "lfc_state";
-impl TryFrom<&crate::compute::ParsedSpec> for EndpointStoragePair {
-    type Error = anyhow::Error;
-    fn try_from(pspec: &crate::compute::ParsedSpec) -> Result<Self, Self::Error> {
-        let Some(ref endpoint_id) = pspec.spec.endpoint_id else {
-            bail!("pspec.endpoint_id missing")
-        };
-        let Some(ref base_uri) = pspec.endpoint_storage_addr else {
-            bail!("pspec.endpoint_storage_addr missing")
-        };
-        let tenant_id = pspec.tenant_id;
-        let timeline_id = pspec.timeline_id;
-
-        let url = format!("http://{base_uri}/{tenant_id}/{timeline_id}/{endpoint_id}/{KEY}");
-        let Some(ref token) = pspec.endpoint_storage_token else {
-            bail!("pspec.endpoint_storage_token missing")
-        };
-        let token = token.clone();
-        Ok(EndpointStoragePair { url, token })
-    }
-}
-
-impl ComputeNode {
-    // If prewarm failed, we want to get overall number of segments as well as done ones.
-    // However, this function should be reliable even if querying postgres failed.
-    pub async fn lfc_prewarm_state(&self) -> LfcPrewarmStateWithProgress {
-        info!("requesting LFC prewarm state from postgres");
-        let mut state = LfcPrewarmStateWithProgress::default();
-        {
-            state.base = self.state.lock().unwrap().lfc_prewarm_state.clone();
-        }
-
-        let client = match ComputeNode::get_maintenance_client(&self.tokio_conn_conf).await {
-            Ok(client) => client,
-            Err(err) => {
-                error!(%err, "connecting to postgres");
-                return state;
-            }
-        };
-        let row = match client
-            .query_one("select * from get_prewarm_info()", &[])
-            .await
-        {
-            Ok(row) => row,
-            Err(err) => {
-                error!(%err, "querying LFC prewarm status");
-                return state;
-            }
-        };
-        state.total = row.try_get(0).unwrap_or_default();
-        state.prewarmed = row.try_get(1).unwrap_or_default();
-        state.skipped = row.try_get(2).unwrap_or_default();
-        state
-    }
-
-    pub fn lfc_offload_state(&self) -> LfcOffloadState {
-        self.state.lock().unwrap().lfc_offload_state.clone()
-    }
-
-    /// Returns false if there is a prewarm request ongoing, true otherwise
-    pub fn prewarm_lfc(self: &Arc<Self>) -> bool {
-        crate::metrics::LFC_PREWARM_REQUESTS.inc();
-        {
-            let state = &mut self.state.lock().unwrap().lfc_prewarm_state;
-            if let LfcPrewarmState::Prewarming =
-                std::mem::replace(state, LfcPrewarmState::Prewarming)
-            {
-                return false;
-            }
-        }
-
-        let cloned = self.clone();
-        spawn(async move {
-            let Err(err) = cloned.prewarm_impl().await else {
-                cloned.state.lock().unwrap().lfc_prewarm_state = LfcPrewarmState::Completed;
-                return;
-            };
-            error!(%err);
-            cloned.state.lock().unwrap().lfc_prewarm_state = LfcPrewarmState::Failed {
-                error: err.to_string(),
-            };
-        });
-        true
-    }
-
-    fn endpoint_storage_pair(&self) -> Result<EndpointStoragePair> {
-        let state = self.state.lock().unwrap();
-        state.pspec.as_ref().unwrap().try_into()
-    }
-
-    async fn prewarm_impl(&self) -> Result<()> {
-        let EndpointStoragePair { url, token } = self.endpoint_storage_pair()?;
-        info!(%url, "requesting LFC state from endpoint storage");
-
-        let request = Client::new().get(&url).bearer_auth(token);
-        let res = request.send().await.context("querying endpoint storage")?;
-        let status = res.status();
-        if status != StatusCode::OK {
-            bail!("{status} querying endpoint storage")
-        }
-
-        let mut uncompressed = Vec::new();
-        let lfc_state = res
-            .bytes()
-            .await
-            .context("getting request body from endpoint storage")?;
-        ZstdDecoder::new(lfc_state.iter().as_slice())
-            .read_to_end(&mut uncompressed)
-            .await
-            .context("decoding LFC state")?;
-        let uncompressed_len = uncompressed.len();
-        info!(%url, "downloaded LFC state, uncompressed size {uncompressed_len}, loading into postgres");
-
-        ComputeNode::get_maintenance_client(&self.tokio_conn_conf)
-            .await
-            .context("connecting to postgres")?
-            .query_one("select prewarm_local_cache($1)", &[&uncompressed])
-            .await
-            .context("loading LFC state into postgres")
-            .map(|_| ())
-    }
-
-    /// Returns false if there is an offload request ongoing, true otherwise
-    pub fn offload_lfc(self: &Arc<Self>) -> bool {
-        crate::metrics::LFC_OFFLOAD_REQUESTS.inc();
-        {
-            let state = &mut self.state.lock().unwrap().lfc_offload_state;
-            if let LfcOffloadState::Offloading =
-                std::mem::replace(state, LfcOffloadState::Offloading)
-            {
-                return false;
-            }
-        }
-
-        let cloned = self.clone();
-        spawn(async move {
-            let Err(err) = cloned.offload_lfc_impl().await else {
-                cloned.state.lock().unwrap().lfc_offload_state = LfcOffloadState::Completed;
-                return;
-            };
-            error!(%err);
-            cloned.state.lock().unwrap().lfc_offload_state = LfcOffloadState::Failed {
-                error: err.to_string(),
-            };
-        });
-        true
-    }
-
-    async fn offload_lfc_impl(&self) -> Result<()> {
-        let EndpointStoragePair { url, token } = self.endpoint_storage_pair()?;
-        info!(%url, "requesting LFC state from postgres");
-
-        let mut compressed = Vec::new();
-        ComputeNode::get_maintenance_client(&self.tokio_conn_conf)
-            .await
-            .context("connecting to postgres")?
-            .query_one("select get_local_cache_state()", &[])
-            .await
-            .context("querying LFC state")?
-            .try_get::<usize, &[u8]>(0)
-            .context("deserializing LFC state")
-            .map(ZstdEncoder::new)?
-            .read_to_end(&mut compressed)
-            .await
-            .context("compressing LFC state")?;
-        let compressed_len = compressed.len();
-        info!(%url, "downloaded LFC state, compressed size {compressed_len}, writing to endpoint storage");
-
-        let request = Client::new().put(url).bearer_auth(token).body(compressed);
-        match request.send().await {
-            Ok(res) if res.status() == StatusCode::OK => Ok(()),
-            Ok(res) => bail!("Error writing to endpoint storage: {}", res.status()),
-            Err(err) => Err(err).context("writing to endpoint storage"),
-        }
-    }
-}

compute_tools/src/config.rs

@@ -223,9 +223,6 @@ pub fn write_postgres_conf(
             // TODO: tune this after performance testing
             writeln!(file, "pgaudit.log_rotation_age=5")?;
 
-            // Enable audit logs for pg_session_jwt extension
-            writeln!(file, "pg_session_jwt.audit_log=on")?;
-
             // Add audit shared_preload_libraries, if they are not present.
             //
             // The caller who sets the flag is responsible for ensuring that the necessary

compute_tools/src/http/middleware/authorize.rs

@@ -1,10 +1,12 @@
+use std::collections::HashSet;
+
 use anyhow::{Result, anyhow};
 use axum::{RequestExt, body::Body};
 use axum_extra::{
     TypedHeader,
     headers::{Authorization, authorization::Bearer},
 };
-use compute_api::requests::{COMPUTE_AUDIENCE, ComputeClaims, ComputeClaimsScope};
+use compute_api::requests::ComputeClaims;
 use futures::future::BoxFuture;
 use http::{Request, Response, StatusCode};
 use jsonwebtoken::{Algorithm, DecodingKey, TokenData, Validation, jwk::JwkSet};
@@ -23,14 +25,13 @@ pub(in crate::http) struct Authorize {
 impl Authorize {
     pub fn new(compute_id: String, jwks: JwkSet) -> Self {
         let mut validation = Validation::new(Algorithm::EdDSA);
+        // Nothing is currently required
+        validation.required_spec_claims = HashSet::new();
         validation.validate_exp = true;
         // Unused by the control plane
-        validation.validate_nbf = false;
-        // Unused by the control plane
         validation.validate_aud = false;
-        validation.set_audience(&[COMPUTE_AUDIENCE]);
-        // Nothing is currently required
-        validation.set_required_spec_claims(&[] as &[&str; 0]);
+        // Unused by the control plane
+        validation.validate_nbf = false;
 
         Self {
             compute_id,
@@ -63,47 +64,11 @@ impl AsyncAuthorizeRequest<Body> for Authorize {
                 Err(e) => return Err(JsonResponse::error(StatusCode::UNAUTHORIZED, e)),
             };
 
-            match data.claims.scope {
-                // TODO: We should validate audience for every token, but
-                // instead of this ad-hoc validation, we should turn
-                // [`Validation::validate_aud`] on. This is merely a stopgap
-                // while we roll out `aud` deployment. We return a 401
-                // Unauthorized because when we eventually do use
-                // [`Validation`], we will hit the above `Err` match arm which
-                // returns 401 Unauthorized.
-                Some(ComputeClaimsScope::Admin) => {
-                    let Some(ref audience) = data.claims.audience else {
-                        return Err(JsonResponse::error(
-                            StatusCode::UNAUTHORIZED,
-                            "missing audience in authorization token claims",
-                        ));
-                    };
-
-                    if !audience.iter().any(|a| a == COMPUTE_AUDIENCE) {
-                        return Err(JsonResponse::error(
-                            StatusCode::UNAUTHORIZED,
-                            "invalid audience in authorization token claims",
-                        ));
-                    }
-                }
-
-                // If the scope is not [`ComputeClaimsScope::Admin`], then we
-                // must validate the compute_id
-                _ => {
-                    let Some(ref claimed_compute_id) = data.claims.compute_id else {
-                        return Err(JsonResponse::error(
-                            StatusCode::FORBIDDEN,
-                            "missing compute_id in authorization token claims",
-                        ));
-                    };
-
-                    if *claimed_compute_id != compute_id {
-                        return Err(JsonResponse::error(
-                            StatusCode::FORBIDDEN,
-                            "invalid compute ID in authorization token claims",
-                        ));
-                    }
-                }
+            if data.claims.compute_id != compute_id {
+                return Err(JsonResponse::error(
+                    StatusCode::UNAUTHORIZED,
+                    "invalid compute ID in authorization token claims",
+                ));
             }
 
             // Make claims available to any subsequent middleware or request

compute_tools/src/http/routes/lfc.rs

@@ -1,39 +0,0 @@
-use crate::compute_prewarm::LfcPrewarmStateWithProgress;
-use crate::http::JsonResponse;
-use axum::response::{IntoResponse, Response};
-use axum::{Json, http::StatusCode};
-use compute_api::responses::LfcOffloadState;
-type Compute = axum::extract::State<std::sync::Arc<crate::compute::ComputeNode>>;
-
-pub(in crate::http) async fn prewarm_state(compute: Compute) -> Json<LfcPrewarmStateWithProgress> {
-    Json(compute.lfc_prewarm_state().await)
-}
-
-// Following functions are marked async for axum, as it's more convenient than wrapping these
-// in async lambdas at call site
-
-pub(in crate::http) async fn offload_state(compute: Compute) -> Json<LfcOffloadState> {
-    Json(compute.lfc_offload_state())
-}
-
-pub(in crate::http) async fn prewarm(compute: Compute) -> Response {
-    if compute.prewarm_lfc() {
-        StatusCode::ACCEPTED.into_response()
-    } else {
-        JsonResponse::error(
-            StatusCode::TOO_MANY_REQUESTS,
-            "Multiple requests for prewarm are not allowed",
-        )
-    }
-}
-
-pub(in crate::http) async fn offload(compute: Compute) -> Response {
-    if compute.offload_lfc() {
-        StatusCode::ACCEPTED.into_response()
-    } else {
-        JsonResponse::error(
-            StatusCode::TOO_MANY_REQUESTS,
-            "Multiple requests for prewarm offload are not allowed",
-        )
-    }
-}

compute_tools/src/http/routes/mod.rs

@@ -11,7 +11,6 @@ pub(in crate::http) mod extensions;
 pub(in crate::http) mod failpoints;
 pub(in crate::http) mod grants;
 pub(in crate::http) mod insights;
-pub(in crate::http) mod lfc;
 pub(in crate::http) mod metrics;
 pub(in crate::http) mod metrics_json;
 pub(in crate::http) mod status;

compute_tools/src/http/server.rs

@@ -23,7 +23,7 @@ use super::{
     middleware::authorize::Authorize,
     routes::{
         check_writability, configure, database_schema, dbs_and_roles, extension_server, extensions,
-        grants, insights, lfc, metrics, metrics_json, status, terminate,
+        grants, insights, metrics, metrics_json, status, terminate,
     },
 };
 use crate::compute::ComputeNode;
@@ -85,8 +85,6 @@ impl From<&Server> for Router<Arc<ComputeNode>> {
                     Router::<Arc<ComputeNode>>::new().route("/metrics", get(metrics::get_metrics));
 
                 let authenticated_router = Router::<Arc<ComputeNode>>::new()
-                    .route("/lfc/prewarm", get(lfc::prewarm_state).post(lfc::prewarm))
-                    .route("/lfc/offload", get(lfc::offload_state).post(lfc::offload))
                     .route("/check_writability", post(check_writability::is_writable))
                     .route("/configure", post(configure::configure))
                     .route("/database_schema", get(database_schema::get_schema_dump))

compute_tools/src/lib.rs

@@ -11,7 +11,6 @@ pub mod http;
 pub mod logger;
 pub mod catalog;
 pub mod compute;
-pub mod compute_prewarm;
 pub mod disk_quota;
 pub mod extension_server;
 pub mod installed_extensions;

compute_tools/src/metrics.rs

@@ -1,7 +1,7 @@
 use metrics::core::{AtomicF64, AtomicU64, Collector, GenericCounter, GenericGauge};
 use metrics::proto::MetricFamily;
 use metrics::{
-    IntCounter, IntCounterVec, IntGaugeVec, UIntGaugeVec, register_gauge, register_int_counter,
+    IntCounterVec, IntGaugeVec, UIntGaugeVec, register_gauge, register_int_counter,
     register_int_counter_vec, register_int_gauge_vec, register_uint_gauge_vec,
 };
 use once_cell::sync::Lazy;
@@ -97,24 +97,6 @@ pub(crate) static PG_TOTAL_DOWNTIME_MS: Lazy<GenericCounter<AtomicU64>> = Lazy::
     .expect("failed to define a metric")
 });
 
-/// Needed as neon.file_cache_prewarm_batch == 0 doesn't mean we never tried to prewarm.
-/// On the other hand, LFC_PREWARMED_PAGES is excessive as we can GET /lfc/prewarm
-pub(crate) static LFC_PREWARM_REQUESTS: Lazy<IntCounter> = Lazy::new(|| {
-    register_int_counter!(
-        "compute_ctl_lfc_prewarm_requests_total",
-        "Total number of LFC prewarm requests made by compute_ctl",
-    )
-    .expect("failed to define a metric")
-});
-
-pub(crate) static LFC_OFFLOAD_REQUESTS: Lazy<IntCounter> = Lazy::new(|| {
-    register_int_counter!(
-        "compute_ctl_lfc_offload_requests_total",
-        "Total number of LFC offload requests made by compute_ctl",
-    )
-    .expect("failed to define a metric")
-});
-
 pub fn collect() -> Vec<MetricFamily> {
     let mut metrics = COMPUTE_CTL_UP.collect();
     metrics.extend(INSTALLED_EXTENSIONS.collect());
@@ -124,7 +106,5 @@ pub fn collect() -> Vec<MetricFamily> {
     metrics.extend(AUDIT_LOG_DIR_SIZE.collect());
     metrics.extend(PG_CURR_DOWNTIME_MS.collect());
     metrics.extend(PG_TOTAL_DOWNTIME_MS.collect());
-    metrics.extend(LFC_PREWARM_REQUESTS.collect());
-    metrics.extend(LFC_OFFLOAD_REQUESTS.collect());
     metrics
 }

compute_tools/tests/pg_helpers_tests.rs

@@ -30,7 +30,6 @@ mod pg_helpers_tests {
             r#"fsync = off
 wal_level = logical
 hot_standby = on
-prewarm_lfc_on_startup = off
 neon.safekeepers = '127.0.0.1:6502,127.0.0.1:6503,127.0.0.1:6501'
 wal_log_hints = on
 log_connections = on

control_plane/Cargo.toml

@@ -41,7 +41,7 @@ storage_broker.workspace = true
 http-utils.workspace = true
 utils.workspace = true
 whoami.workspace = true
-endpoint_storage.workspace = true
+
 compute_api.workspace = true
 workspace_hack.workspace = true
 tracing.workspace = true

control_plane/src/bin/neon_local.rs

@@ -16,11 +16,10 @@ use std::time::Duration;
 
 use anyhow::{Context, Result, anyhow, bail};
 use clap::Parser;
-use compute_api::requests::ComputeClaimsScope;
 use compute_api::spec::ComputeMode;
 use control_plane::broker::StorageBroker;
 use control_plane::endpoint::ComputeControlPlane;
-use control_plane::endpoint_storage::{ENDPOINT_STORAGE_DEFAULT_ADDR, EndpointStorage};
+use control_plane::endpoint_storage::{ENDPOINT_STORAGE_DEFAULT_PORT, EndpointStorage};
 use control_plane::local_env;
 use control_plane::local_env::{
     EndpointStorageConf, InitForceMode, LocalEnv, NeonBroker, NeonLocalInitConf,
@@ -706,9 +705,6 @@ struct EndpointStopCmdArgs {
 struct EndpointGenerateJwtCmdArgs {
     #[clap(help = "Postgres endpoint id")]
     endpoint_id: String,
-
-    #[clap(short = 's', long, help = "Scope to generate the JWT with", value_parser = ComputeClaimsScope::from_str)]
-    scope: Option<ComputeClaimsScope>,
 }
 
 #[derive(clap::Subcommand)]
@@ -1022,7 +1018,7 @@ fn handle_init(args: &InitCmdArgs) -> anyhow::Result<LocalEnv> {
                 })
                 .collect(),
             endpoint_storage: EndpointStorageConf {
-                listen_addr: ENDPOINT_STORAGE_DEFAULT_ADDR,
+                port: ENDPOINT_STORAGE_DEFAULT_PORT,
             },
             pg_distrib_dir: None,
             neon_distrib_dir: None,
@@ -1488,25 +1484,10 @@ async fn handle_endpoint(subcmd: &EndpointCmd, env: &local_env::LocalEnv) -> Res
                 None
             };
 
-            let exp = (std::time::SystemTime::now().duration_since(std::time::UNIX_EPOCH)?
-                + Duration::from_secs(86400))
-            .as_secs();
-            let claims = endpoint_storage::claims::EndpointStorageClaims {
-                tenant_id: endpoint.tenant_id,
-                timeline_id: endpoint.timeline_id,
-                endpoint_id: endpoint_id.to_string(),
-                exp,
-            };
-
-            let endpoint_storage_token = env.generate_auth_token(&claims)?;
-            let endpoint_storage_addr = env.endpoint_storage.listen_addr.to_string();
-
             println!("Starting existing endpoint {endpoint_id}...");
             endpoint
                 .start(
                     &auth_token,
-                    endpoint_storage_token,
-                    endpoint_storage_addr,
                     safekeepers_generation,
                     safekeepers,
                     pageservers,
@@ -1559,16 +1540,12 @@ async fn handle_endpoint(subcmd: &EndpointCmd, env: &local_env::LocalEnv) -> Res
             endpoint.stop(&args.mode, args.destroy)?;
         }
         EndpointCmd::GenerateJwt(args) => {
-            let endpoint = {
-                let endpoint_id = &args.endpoint_id;
-
-                cplane
-                    .endpoints
-                    .get(endpoint_id)
-                    .with_context(|| format!("postgres endpoint {endpoint_id} is not found"))?
-            };
-
-            let jwt = endpoint.generate_jwt(args.scope)?;
+            let endpoint_id = &args.endpoint_id;
+            let endpoint = cplane
+                .endpoints
+                .get(endpoint_id)
+                .with_context(|| format!("postgres endpoint {endpoint_id} is not found"))?;
+            let jwt = endpoint.generate_jwt()?;
 
             print!("{jwt}");
         }

control_plane/src/endpoint.rs

@@ -45,9 +45,7 @@ use std::sync::Arc;
 use std::time::{Duration, Instant};
 
 use anyhow::{Context, Result, anyhow, bail};
-use compute_api::requests::{
-    COMPUTE_AUDIENCE, ComputeClaims, ComputeClaimsScope, ConfigurationRequest,
-};
+use compute_api::requests::{ComputeClaims, ConfigurationRequest};
 use compute_api::responses::{
     ComputeConfig, ComputeCtlConfig, ComputeStatus, ComputeStatusResponse, TlsConfig,
 };
@@ -632,17 +630,9 @@ impl Endpoint {
     }
 
     /// Generate a JWT with the correct claims.
-    pub fn generate_jwt(&self, scope: Option<ComputeClaimsScope>) -> Result<String> {
+    pub fn generate_jwt(&self) -> Result<String> {
         self.env.generate_auth_token(&ComputeClaims {
-            audience: match scope {
-                Some(ComputeClaimsScope::Admin) => Some(vec![COMPUTE_AUDIENCE.to_owned()]),
-                _ => None,
-            },
-            compute_id: match scope {
-                Some(ComputeClaimsScope::Admin) => None,
-                _ => Some(self.endpoint_id.clone()),
-            },
-            scope,
+            compute_id: self.endpoint_id.clone(),
         })
     }
 
@@ -650,8 +640,6 @@ impl Endpoint {
     pub async fn start(
         &self,
         auth_token: &Option<String>,
-        endpoint_storage_token: String,
-        endpoint_storage_addr: String,
         safekeepers_generation: Option<SafekeeperGeneration>,
         safekeepers: Vec<NodeId>,
         pageservers: Vec<(Host, u16)>,
@@ -745,9 +733,6 @@ impl Endpoint {
                 drop_subscriptions_before_start: self.drop_subscriptions_before_start,
                 audit_log_level: ComputeAudit::Disabled,
                 logs_export_host: None::<String>,
-                endpoint_storage_addr: Some(endpoint_storage_addr),
-                endpoint_storage_token: Some(endpoint_storage_token),
-                prewarm_lfc_on_startup: false,
             };
 
             // this strange code is needed to support respec() in tests
@@ -918,7 +903,7 @@ impl Endpoint {
                     self.external_http_address.port()
                 ),
             )
-            .bearer_auth(self.generate_jwt(None::<ComputeClaimsScope>)?)
+            .bearer_auth(self.generate_jwt()?)
             .send()
             .await?;
 
@@ -995,7 +980,7 @@ impl Endpoint {
                 self.external_http_address.port()
             ))
             .header(CONTENT_TYPE.as_str(), "application/json")
-            .bearer_auth(self.generate_jwt(None::<ComputeClaimsScope>)?)
+            .bearer_auth(self.generate_jwt()?)
             .body(
                 serde_json::to_string(&ConfigurationRequest {
                     spec,

control_plane/src/endpoint_storage.rs

@@ -3,19 +3,17 @@ use crate::local_env::LocalEnv;
 use anyhow::{Context, Result};
 use camino::Utf8PathBuf;
 use std::io::Write;
-use std::net::SocketAddr;
 use std::time::Duration;
 
 /// Directory within .neon which will be used by default for LocalFs remote storage.
 pub const ENDPOINT_STORAGE_REMOTE_STORAGE_DIR: &str = "local_fs_remote_storage/endpoint_storage";
-pub const ENDPOINT_STORAGE_DEFAULT_ADDR: SocketAddr =
-    SocketAddr::new(std::net::IpAddr::V4(std::net::Ipv4Addr::LOCALHOST), 9993);
+pub const ENDPOINT_STORAGE_DEFAULT_PORT: u16 = 9993;
 
 pub struct EndpointStorage {
     pub bin: Utf8PathBuf,
     pub data_dir: Utf8PathBuf,
     pub pemfile: Utf8PathBuf,
-    pub addr: SocketAddr,
+    pub port: u16,
 }
 
 impl EndpointStorage {
@@ -24,7 +22,7 @@ impl EndpointStorage {
             bin: Utf8PathBuf::from_path_buf(env.endpoint_storage_bin()).unwrap(),
             data_dir: Utf8PathBuf::from_path_buf(env.endpoint_storage_data_dir()).unwrap(),
             pemfile: Utf8PathBuf::from_path_buf(env.public_key_path.clone()).unwrap(),
-            addr: env.endpoint_storage.listen_addr,
+            port: env.endpoint_storage.port,
         }
     }
 
@@ -33,7 +31,7 @@ impl EndpointStorage {
     }
 
     fn listen_addr(&self) -> Utf8PathBuf {
-        format!("{}:{}", self.addr.ip(), self.addr.port()).into()
+        format!("127.0.0.1:{}", self.port).into()
     }
 
     pub fn init(&self) -> Result<()> {

control_plane/src/local_env.rs

@@ -20,9 +20,7 @@ use utils::auth::encode_from_key_file;
 use utils::id::{NodeId, TenantId, TenantTimelineId, TimelineId};
 
 use crate::broker::StorageBroker;
-use crate::endpoint_storage::{
-    ENDPOINT_STORAGE_DEFAULT_ADDR, ENDPOINT_STORAGE_REMOTE_STORAGE_DIR, EndpointStorage,
-};
+use crate::endpoint_storage::{ENDPOINT_STORAGE_REMOTE_STORAGE_DIR, EndpointStorage};
 use crate::pageserver::{PAGESERVER_REMOTE_STORAGE_DIR, PageServerNode};
 use crate::safekeeper::SafekeeperNode;
 
@@ -153,10 +151,10 @@ pub struct NeonLocalInitConf {
     pub generate_local_ssl_certs: bool,
 }
 
-#[derive(Serialize, Deserialize, PartialEq, Eq, Clone, Debug)]
+#[derive(Serialize, Default, Deserialize, PartialEq, Eq, Clone, Debug)]
 #[serde(default)]
 pub struct EndpointStorageConf {
-    pub listen_addr: SocketAddr,
+    pub port: u16,
 }
 
 /// Broker config for cluster internal communication.
@@ -243,14 +241,6 @@ impl Default for NeonStorageControllerConf {
     }
 }
 
-impl Default for EndpointStorageConf {
-    fn default() -> Self {
-        Self {
-            listen_addr: ENDPOINT_STORAGE_DEFAULT_ADDR,
-        }
-    }
-}
-
 impl NeonBroker {
     pub fn client_url(&self) -> Url {
         let url = if let Some(addr) = self.listen_https_addr {

docker-compose/ext-src/pg_session_jwt-src/expected/basic_functions.out

@@ -12,7 +12,6 @@ ERROR:  invalid JWT encoding
 -- Test creating a session with an expired JWT
 SELECT auth.jwt_session_init('eyJhbGciOiJFZERTQSJ9.eyJleHAiOjE3NDI1NjQ0MzIsImlhdCI6MTc0MjU2NDI1MiwianRpIjo0MjQyNDIsInN1YiI6InVzZXIxMjMifQ.A6FwKuaSduHB9O7Gz37g0uoD_U9qVS0JNtT7YABGVgB7HUD1AMFc9DeyhNntWBqncg8k5brv-hrNTuUh5JYMAw');
 ERROR:  Token used after it has expired
-DETAIL:  exp=1742564432
 -- Test creating a session with a valid JWT
 SELECT auth.jwt_session_init('eyJhbGciOiJFZERTQSJ9.eyJleHAiOjQ4OTYxNjQyNTIsImlhdCI6MTc0MjU2NDI1MiwianRpIjo0MzQzNDMsInN1YiI6InVzZXIxMjMifQ.2TXVgjb6JSUq6_adlvp-m_SdOxZSyGS30RS9TLB0xu2N83dMSs2NybwE1NMU8Fb0tcAZR_ET7M2rSxbTrphfCg');
  jwt_session_init 

endpoint_storage/src/app.rs

@@ -343,7 +343,7 @@ MC4CAQAwBQYDK2VwBCIEID/Drmc1AA6U/znNRWpF3zEGegOATQxfkdWxitcOMsIH
         TimelineId::from_array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 1, 2, 3, 4, 5, 7]);
     const ENDPOINT_ID: &str = "ep-winter-frost-a662z3vg";
     fn token() -> String {
-        let claims = endpoint_storage::claims::EndpointStorageClaims {
+        let claims = endpoint_storage::Claims {
             tenant_id: TENANT_ID,
             timeline_id: TIMELINE_ID,
             endpoint_id: ENDPOINT_ID.into(),
@@ -489,8 +489,16 @@ MC4CAQAwBQYDK2VwBCIEID/Drmc1AA6U/znNRWpF3zEGegOATQxfkdWxitcOMsIH
     }
 
     fn delete_prefix_token(uri: &str) -> String {
+        use serde::Serialize;
         let parts = uri.split("/").collect::<Vec<&str>>();
-        let claims = endpoint_storage::claims::DeletePrefixClaims {
+        #[derive(Serialize)]
+        struct PrefixClaims {
+            tenant_id: TenantId,
+            timeline_id: Option<TimelineId>,
+            endpoint_id: Option<endpoint_storage::EndpointId>,
+            exp: u64,
+        }
+        let claims = PrefixClaims {
             tenant_id: parts.get(1).map(|c| c.parse().unwrap()).unwrap(),
             timeline_id: parts.get(2).map(|c| c.parse().unwrap()),
             endpoint_id: parts.get(3).map(ToString::to_string),

endpoint_storage/src/claims.rs

@@ -1,52 +0,0 @@
-use serde::{Deserialize, Serialize};
-use std::fmt::Display;
-use utils::id::{EndpointId, TenantId, TimelineId};
-
-/// Claims to add, remove, or retrieve endpoint data. Used by compute_ctl
-#[derive(Deserialize, Serialize, PartialEq)]
-pub struct EndpointStorageClaims {
-    pub tenant_id: TenantId,
-    pub timeline_id: TimelineId,
-    pub endpoint_id: EndpointId,
-    pub exp: u64,
-}
-
-/// Claims to remove tenant, timeline, or endpoint data. Used by control plane
-#[derive(Deserialize, Serialize, PartialEq)]
-pub struct DeletePrefixClaims {
-    pub tenant_id: TenantId,
-    /// None when tenant is deleted (endpoint_id is also None in this case)
-    pub timeline_id: Option<TimelineId>,
-    /// None when timeline is deleted
-    pub endpoint_id: Option<EndpointId>,
-    pub exp: u64,
-}
-
-impl Display for EndpointStorageClaims {
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
-        write!(
-            f,
-            "EndpointClaims(tenant_id={} timeline_id={} endpoint_id={} exp={})",
-            self.tenant_id, self.timeline_id, self.endpoint_id, self.exp
-        )
-    }
-}
-
-impl Display for DeletePrefixClaims {
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
-        write!(
-            f,
-            "DeletePrefixClaims(tenant_id={} timeline_id={} endpoint_id={}, exp={})",
-            self.tenant_id,
-            self.timeline_id
-                .as_ref()
-                .map(ToString::to_string)
-                .unwrap_or("".to_string()),
-            self.endpoint_id
-                .as_ref()
-                .map(ToString::to_string)
-                .unwrap_or("".to_string()),
-            self.exp
-        )
-    }
-}

endpoint_storage/src/lib.rs

@@ -1,5 +1,3 @@
-pub mod claims;
-use crate::claims::{DeletePrefixClaims, EndpointStorageClaims};
 use anyhow::Result;
 use axum::extract::{FromRequestParts, Path};
 use axum::response::{IntoResponse, Response};
@@ -15,7 +13,7 @@ use std::result::Result as StdResult;
 use std::sync::Arc;
 use tokio_util::sync::CancellationToken;
 use tracing::{debug, error};
-use utils::id::{EndpointId, TenantId, TimelineId};
+use utils::id::{TenantId, TimelineId};
 
 // simplified version of utils::auth::JwtAuth
 pub struct JwtAuth {
@@ -81,6 +79,26 @@ pub struct Storage {
     pub max_upload_file_limit: usize,
 }
 
+pub type EndpointId = String; // If needed, reuse small string from proxy/src/types.rc
+
+#[derive(Deserialize, Serialize, PartialEq)]
+pub struct Claims {
+    pub tenant_id: TenantId,
+    pub timeline_id: TimelineId,
+    pub endpoint_id: EndpointId,
+    pub exp: u64,
+}
+
+impl Display for Claims {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        write!(
+            f,
+            "Claims(tenant_id {} timeline_id {} endpoint_id {} exp {})",
+            self.tenant_id, self.timeline_id, self.endpoint_id, self.exp
+        )
+    }
+}
+
 #[derive(Deserialize, Serialize)]
 struct KeyRequest {
     tenant_id: TenantId,
@@ -89,13 +107,6 @@ struct KeyRequest {
     path: String,
 }
 
-#[derive(Deserialize, Serialize, PartialEq)]
-struct PrefixKeyRequest {
-    tenant_id: TenantId,
-    timeline_id: Option<TimelineId>,
-    endpoint_id: Option<EndpointId>,
-}
-
 #[derive(Debug, PartialEq)]
 pub struct S3Path {
     pub path: RemotePath,
@@ -154,7 +165,7 @@ impl FromRequestParts<Arc<Storage>> for S3Path {
             .extract::<TypedHeader<Authorization<Bearer>>>()
             .await
             .map_err(|e| bad_request(e, "invalid token"))?;
-        let claims: EndpointStorageClaims = state
+        let claims: Claims = state
             .auth
             .decode(bearer.token())
             .map_err(|e| bad_request(e, "decoding token"))?;
@@ -167,7 +178,7 @@ impl FromRequestParts<Arc<Storage>> for S3Path {
             path.endpoint_id.clone()
         };
 
-        let route = EndpointStorageClaims {
+        let route = Claims {
             tenant_id: path.tenant_id,
             timeline_id: path.timeline_id,
             endpoint_id,
@@ -182,13 +193,38 @@ impl FromRequestParts<Arc<Storage>> for S3Path {
     }
 }
 
+#[derive(Deserialize, Serialize, PartialEq)]
+pub struct PrefixKeyPath {
+    pub tenant_id: TenantId,
+    pub timeline_id: Option<TimelineId>,
+    pub endpoint_id: Option<EndpointId>,
+}
+
+impl Display for PrefixKeyPath {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        write!(
+            f,
+            "PrefixKeyPath(tenant_id {} timeline_id {} endpoint_id {})",
+            self.tenant_id,
+            self.timeline_id
+                .as_ref()
+                .map(ToString::to_string)
+                .unwrap_or("".to_string()),
+            self.endpoint_id
+                .as_ref()
+                .map(ToString::to_string)
+                .unwrap_or("".to_string())
+        )
+    }
+}
+
 #[derive(Debug, PartialEq)]
 pub struct PrefixS3Path {
     pub path: RemotePath,
 }
 
-impl From<&DeletePrefixClaims> for PrefixS3Path {
-    fn from(path: &DeletePrefixClaims) -> Self {
+impl From<&PrefixKeyPath> for PrefixS3Path {
+    fn from(path: &PrefixKeyPath) -> Self {
         let timeline_id = path
             .timeline_id
             .as_ref()
@@ -214,27 +250,21 @@ impl FromRequestParts<Arc<Storage>> for PrefixS3Path {
         state: &Arc<Storage>,
     ) -> Result<Self, Self::Rejection> {
         let Path(path) = parts
-            .extract::<Path<PrefixKeyRequest>>()
+            .extract::<Path<PrefixKeyPath>>()
             .await
             .map_err(|e| bad_request(e, "invalid route"))?;
         let TypedHeader(Authorization(bearer)) = parts
             .extract::<TypedHeader<Authorization<Bearer>>>()
             .await
             .map_err(|e| bad_request(e, "invalid token"))?;
-        let claims: DeletePrefixClaims = state
+        let claims: PrefixKeyPath = state
             .auth
             .decode(bearer.token())
             .map_err(|e| bad_request(e, "invalid token"))?;
-        let route = DeletePrefixClaims {
-            tenant_id: path.tenant_id,
-            timeline_id: path.timeline_id,
-            endpoint_id: path.endpoint_id,
-            exp: claims.exp,
-        };
-        if route != claims {
-            return Err(unauthorized(route, claims));
+        if path != claims {
+            return Err(unauthorized(path, claims));
         }
-        Ok((&route).into())
+        Ok((&path).into())
     }
 }
 
@@ -267,7 +297,7 @@ mod tests {
 
     #[test]
     fn s3_path() {
-        let auth = EndpointStorageClaims {
+        let auth = Claims {
             tenant_id: TENANT_ID,
             timeline_id: TIMELINE_ID,
             endpoint_id: ENDPOINT_ID.into(),
@@ -297,11 +327,10 @@ mod tests {
 
     #[test]
     fn prefix_s3_path() {
-        let mut path = DeletePrefixClaims {
+        let mut path = PrefixKeyPath {
             tenant_id: TENANT_ID,
             timeline_id: None,
             endpoint_id: None,
-            exp: 0,
         };
         let prefix_path = |s: String| RemotePath::from_string(&s).unwrap();
         assert_eq!(

libs/compute_api/src/requests.rs

@@ -1,58 +1,16 @@
 //! Structs representing the JSON formats used in the compute_ctl's HTTP API.
-use std::str::FromStr;
-
 use serde::{Deserialize, Serialize};
 
 use crate::privilege::Privilege;
 use crate::responses::ComputeCtlConfig;
 use crate::spec::{ComputeSpec, ExtVersion, PgIdent};
 
-/// The value to place in the [`ComputeClaims::audience`] claim.
-pub static COMPUTE_AUDIENCE: &str = "compute";
-
-/// Available scopes for a compute's JWT.
-#[derive(Copy, Clone, Debug, Deserialize, Eq, PartialEq, Serialize)]
-#[serde(rename_all = "snake_case")]
-pub enum ComputeClaimsScope {
-    /// An admin-scoped token allows access to all of `compute_ctl`'s authorized
-    /// facilities.
-    Admin,
-}
-
-impl FromStr for ComputeClaimsScope {
-    type Err = anyhow::Error;
-
-    fn from_str(s: &str) -> Result<Self, Self::Err> {
-        match s {
-            "admin" => Ok(ComputeClaimsScope::Admin),
-            _ => Err(anyhow::anyhow!("invalid compute claims scope \"{s}\"")),
-        }
-    }
-}
-
 /// When making requests to the `compute_ctl` external HTTP server, the client
 /// must specify a set of claims in `Authorization` header JWTs such that
 /// `compute_ctl` can authorize the request.
 #[derive(Clone, Debug, Deserialize, Serialize)]
-#[serde(rename = "snake_case")]
 pub struct ComputeClaims {
-    /// The compute ID that will validate the token. The only case in which this
-    /// can be [`None`] is if [`Self::scope`] is
-    /// [`ComputeClaimsScope::Admin`].
-    pub compute_id: Option<String>,
-
-    /// The scope of what the token authorizes.
-    pub scope: Option<ComputeClaimsScope>,
-
-    /// The recipient the token is intended for.
-    ///
-    /// See [RFC 7519](https://www.rfc-editor.org/rfc/rfc7519#section-4.1.3) for
-    /// more information.
-    ///
-    /// TODO: Remove the [`Option`] wrapper when control plane learns to send
-    /// the claim.
-    #[serde(rename = "aud")]
-    pub audience: Option<Vec<String>>,
+    pub compute_id: String,
 }
 
 /// Request of the /configure API

libs/compute_api/src/responses.rs

@@ -46,30 +46,6 @@ pub struct ExtensionInstallResponse {
     pub version: ExtVersion,
 }
 
-#[derive(Serialize, Default, Debug, Clone)]
-#[serde(tag = "status", rename_all = "snake_case")]
-pub enum LfcPrewarmState {
-    #[default]
-    NotPrewarmed,
-    Prewarming,
-    Completed,
-    Failed {
-        error: String,
-    },
-}
-
-#[derive(Serialize, Default, Debug, Clone)]
-#[serde(tag = "status", rename_all = "snake_case")]
-pub enum LfcOffloadState {
-    #[default]
-    NotOffloaded,
-    Offloading,
-    Completed,
-    Failed {
-        error: String,
-    },
-}
-
 /// Response of the /status API
 #[derive(Serialize, Debug, Deserialize)]
 #[serde(rename_all = "snake_case")]

libs/compute_api/src/spec.rs

@@ -172,15 +172,6 @@ pub struct ComputeSpec {
     /// Hostname and the port of the otel collector. Leave empty to disable Postgres logs forwarding.
     /// Example: config-shy-breeze-123-collector-monitoring.neon-telemetry.svc.cluster.local:10514
     pub logs_export_host: Option<String>,
-
-    /// Address of endpoint storage service
-    pub endpoint_storage_addr: Option<String>,
-    /// JWT for authorizing requests to endpoint storage service
-    pub endpoint_storage_token: Option<String>,
-
-    /// If true, download LFC state from endpoint_storage and pass it to Postgres on startup
-    #[serde(default)]
-    pub prewarm_lfc_on_startup: bool,
 }
 
 /// Feature flag to signal `compute_ctl` to enable certain experimental functionality.

libs/compute_api/tests/cluster_spec.json

@@ -84,11 +84,6 @@
                 "value": "on",
                 "vartype": "bool"
             },
-            {
-                "name": "prewarm_lfc_on_startup",
-                "value": "off",
-                "vartype": "bool"
-            },
             {
                 "name": "neon.safekeepers",
                 "value": "127.0.0.1:6502,127.0.0.1:6503,127.0.0.1:6501",

libs/metrics/src/more_process_metrics.rs

@@ -16,7 +16,6 @@ pub struct Collector {
 const NMETRICS: usize = 2;
 
 static CLK_TCK_F64: Lazy<f64> = Lazy::new(|| {
-    // SAFETY: libc::sysconf is safe, it merely returns a value.
     let long = unsafe { libc::sysconf(libc::_SC_CLK_TCK) };
     if long == -1 {
         panic!("sysconf(_SC_CLK_TCK) failed");

libs/utils/src/id.rs

@@ -295,9 +295,6 @@ pub struct TenantId(Id);
 
 id_newtype!(TenantId);
 
-/// If needed, reuse small string from proxy/src/types.rc
-pub type EndpointId = String;
-
 // A pair uniquely identifying Neon instance.
 #[derive(Debug, Clone, Copy, PartialOrd, Ord, PartialEq, Eq, Hash, Serialize, Deserialize)]
 pub struct TenantTimelineId {

pageserver/src/page_service.rs

@@ -1038,23 +1038,21 @@ impl PageServerHandler {
                         tracing::info_span!(
                             parent: &parent_span,
                             "handle_get_page_request",
-                            request_id = %req.hdr.reqid,
                             rel = %req.rel,
                             blkno = %req.blkno,
                             req_lsn = %req.hdr.request_lsn,
-                            not_modified_since_lsn = %req.hdr.not_modified_since,
+                            not_modified_since_lsn = %req.hdr.not_modified_since
                         )
                     }};
                     ($shard_id:expr) => {{
                         tracing::info_span!(
                             parent: &parent_span,
                             "handle_get_page_request",
-                            request_id = %req.hdr.reqid,
                             rel = %req.rel,
                             blkno = %req.blkno,
                             req_lsn = %req.hdr.request_lsn,
                             not_modified_since_lsn = %req.hdr.not_modified_since,
-                            shard_id = %$shard_id,
+                            shard_id = %$shard_id
                         )
                     }};
                 }

pageserver/src/pgdatadir_mapping.rs

@@ -40,7 +40,7 @@ use wal_decoder::serialized_batch::{SerializedValueBatch, ValueMeta};
 
 use super::tenant::{PageReconstructError, Timeline};
 use crate::aux_file;
-use crate::context::{PerfInstrumentFutureExt, RequestContext, RequestContextBuilder};
+use crate::context::{PerfInstrumentFutureExt, RequestContext};
 use crate::keyspace::{KeySpace, KeySpaceAccum};
 use crate::metrics::{
     RELSIZE_CACHE_ENTRIES, RELSIZE_CACHE_HITS, RELSIZE_CACHE_MISSES, RELSIZE_CACHE_MISSES_OLD,
@@ -275,30 +275,24 @@ impl Timeline {
                 continue;
             }
 
-            let nblocks = {
-                let ctx = RequestContextBuilder::from(&ctx)
-                    .perf_span(|crnt_perf_span| {
-                        info_span!(
-                            target: PERF_TRACE_TARGET,
-                            parent: crnt_perf_span,
-                            "GET_REL_SIZE",
-                            reltag=%tag,
-                            lsn=%lsn,
-                        )
-                    })
-                    .attached_child();
-
-                match self
-                    .get_rel_size(*tag, Version::Lsn(lsn), &ctx)
-                    .maybe_perf_instrument(&ctx, |crnt_perf_span| crnt_perf_span.clone())
-                    .await
-                {
-                    Ok(nblocks) => nblocks,
-                    Err(err) => {
-                        result_slots[response_slot_idx].write(Err(err));
-                        slots_filled += 1;
-                        continue;
-                    }
+            let nblocks = match self
+                .get_rel_size(*tag, Version::Lsn(lsn), &ctx)
+                .maybe_perf_instrument(&ctx, |crnt_perf_span| {
+                    info_span!(
+                        target: PERF_TRACE_TARGET,
+                        parent: crnt_perf_span,
+                        "GET_REL_SIZE",
+                        reltag=%tag,
+                        lsn=%lsn,
+                    )
+                })
+                .await
+            {
+                Ok(nblocks) => nblocks,
+                Err(err) => {
+                    result_slots[response_slot_idx].write(Err(err));
+                    slots_filled += 1;
+                    continue;
                 }
             };
 
@@ -314,17 +308,6 @@ impl Timeline {
 
             let key = rel_block_to_key(*tag, *blknum);
 
-            let ctx = RequestContextBuilder::from(&ctx)
-                .perf_span(|crnt_perf_span| {
-                    info_span!(
-                        target: PERF_TRACE_TARGET,
-                        parent: crnt_perf_span,
-                        "GET_BATCH",
-                        batch_size = %page_count,
-                    )
-                })
-                .attached_child();
-
             let key_slots = keys_slots.entry(key).or_default();
             key_slots.push((response_slot_idx, ctx));
 
@@ -340,7 +323,14 @@ impl Timeline {
         let query = VersionedKeySpaceQuery::scattered(query);
         let res = self
             .get_vectored(query, io_concurrency, ctx)
-            .maybe_perf_instrument(ctx, |current_perf_span| current_perf_span.clone())
+            .maybe_perf_instrument(ctx, |current_perf_span| {
+                info_span!(
+                    target: PERF_TRACE_TARGET,
+                    parent: current_perf_span,
+                    "GET_BATCH",
+                    batch_size = %page_count,
+                )
+            })
             .await;
 
         match res {

pageserver/src/tenant/storage_layer/delta_layer.rs

@@ -1441,6 +1441,14 @@ impl DeltaLayerInner {
         offset
     }
 
+    pub fn iter<'a>(&'a self, ctx: &'a RequestContext) -> DeltaLayerIterator<'a> {
+        self.iter_with_options(
+            ctx,
+            1024 * 8192, // The default value. Unit tests might use a different value. 1024 * 8K = 8MB buffer.
+            1024,        // The default value. Unit tests might use a different value
+        )
+    }
+
     pub fn iter_with_options<'a>(
         &'a self,
         ctx: &'a RequestContext,
@@ -1626,6 +1634,7 @@ pub(crate) mod test {
     use crate::tenant::disk_btree::tests::TestDisk;
     use crate::tenant::harness::{TIMELINE_ID, TenantHarness};
     use crate::tenant::storage_layer::{Layer, ResidentLayer};
+    use crate::tenant::vectored_blob_io::StreamingVectoredReadPlanner;
     use crate::tenant::{TenantShard, Timeline};
 
     /// Construct an index for a fictional delta layer and and then
@@ -2302,7 +2311,8 @@ pub(crate) mod test {
             for batch_size in [1, 2, 4, 8, 3, 7, 13] {
                 println!("running with batch_size={batch_size} max_read_size={max_read_size}");
                 // Test if the batch size is correctly determined
-                let mut iter = delta_layer.iter_with_options(&ctx, max_read_size, batch_size);
+                let mut iter = delta_layer.iter(&ctx);
+                iter.planner = StreamingVectoredReadPlanner::new(max_read_size, batch_size);
                 let mut num_items = 0;
                 for _ in 0..3 {
                     iter.next_batch().await.unwrap();
@@ -2319,7 +2329,8 @@ pub(crate) mod test {
                     iter.key_values_batch.clear();
                 }
                 // Test if the result is correct
-                let mut iter = delta_layer.iter_with_options(&ctx, max_read_size, batch_size);
+                let mut iter = delta_layer.iter(&ctx);
+                iter.planner = StreamingVectoredReadPlanner::new(max_read_size, batch_size);
                 assert_delta_iter_equal(&mut iter, &test_deltas).await;
             }
         }

pageserver/src/tenant/storage_layer/filter_iterator.rs

@@ -157,7 +157,7 @@ mod tests {
             .await
             .unwrap();
 
-        let merge_iter = MergeIterator::create_for_testing(
+        let merge_iter = MergeIterator::create(
             &[resident_layer_1.get_as_delta(&ctx).await.unwrap()],
             &[],
             &ctx,
@@ -182,7 +182,7 @@ mod tests {
         result.extend(test_deltas1[90..100].iter().cloned());
         assert_filter_iter_equal(&mut filter_iter, &result).await;
 
-        let merge_iter = MergeIterator::create_for_testing(
+        let merge_iter = MergeIterator::create(
             &[resident_layer_1.get_as_delta(&ctx).await.unwrap()],
             &[],
             &ctx,

pageserver/src/tenant/storage_layer/image_layer.rs

@@ -684,6 +684,14 @@ impl ImageLayerInner {
         }
     }
 
+    pub(crate) fn iter<'a>(&'a self, ctx: &'a RequestContext) -> ImageLayerIterator<'a> {
+        self.iter_with_options(
+            ctx,
+            1024 * 8192, // The default value. Unit tests might use a different value. 1024 * 8K = 8MB buffer.
+            1024,        // The default value. Unit tests might use a different value
+        )
+    }
+
     pub(crate) fn iter_with_options<'a>(
         &'a self,
         ctx: &'a RequestContext,
@@ -1232,6 +1240,7 @@ mod test {
     use crate::context::RequestContext;
     use crate::tenant::harness::{TIMELINE_ID, TenantHarness};
     use crate::tenant::storage_layer::{Layer, ResidentLayer};
+    use crate::tenant::vectored_blob_io::StreamingVectoredReadPlanner;
     use crate::tenant::{TenantShard, Timeline};
 
     #[tokio::test]
@@ -1498,7 +1507,8 @@ mod test {
             for batch_size in [1, 2, 4, 8, 3, 7, 13] {
                 println!("running with batch_size={batch_size} max_read_size={max_read_size}");
                 // Test if the batch size is correctly determined
-                let mut iter = img_layer.iter_with_options(&ctx, max_read_size, batch_size);
+                let mut iter = img_layer.iter(&ctx);
+                iter.planner = StreamingVectoredReadPlanner::new(max_read_size, batch_size);
                 let mut num_items = 0;
                 for _ in 0..3 {
                     iter.next_batch().await.unwrap();
@@ -1515,7 +1525,8 @@ mod test {
                     iter.key_values_batch.clear();
                 }
                 // Test if the result is correct
-                let mut iter = img_layer.iter_with_options(&ctx, max_read_size, batch_size);
+                let mut iter = img_layer.iter(&ctx);
+                iter.planner = StreamingVectoredReadPlanner::new(max_read_size, batch_size);
                 assert_img_iter_equal(&mut iter, &test_imgs, Lsn(0x10)).await;
             }
         }

pageserver/src/tenant/storage_layer/layer.rs

@@ -23,7 +23,7 @@ use super::{
     LayerVisibilityHint, PerfInstrumentFutureExt, PersistentLayerDesc, ValuesReconstructState,
 };
 use crate::config::PageServerConf;
-use crate::context::{RequestContext, RequestContextBuilder};
+use crate::context::{DownloadBehavior, RequestContext, RequestContextBuilder};
 use crate::span::debug_assert_current_span_has_tenant_and_timeline_id;
 use crate::task_mgr::TaskKind;
 use crate::tenant::Timeline;
@@ -1076,17 +1076,24 @@ impl LayerInner {
             return Err(DownloadError::DownloadRequired);
         }
 
-        let ctx = RequestContextBuilder::from(ctx)
-            .perf_span(|crnt_perf_span| {
-                info_span!(
-                    target: PERF_TRACE_TARGET,
-                    parent: crnt_perf_span,
-                    "DOWNLOAD_LAYER",
-                    layer = %self,
-                    reason = %reason,
-                )
-            })
-            .attached_child();
+        let ctx = if ctx.has_perf_span() {
+            let dl_ctx = RequestContextBuilder::from(ctx)
+                .task_kind(TaskKind::LayerDownload)
+                .download_behavior(DownloadBehavior::Download)
+                .root_perf_span(|| {
+                    info_span!(
+                        target: PERF_TRACE_TARGET,
+                        "DOWNLOAD_LAYER",
+                        layer = %self,
+                        reason = %reason
+                    )
+                })
+                .detached_child();
+            ctx.perf_follows_from(&dl_ctx);
+            dl_ctx
+        } else {
+            ctx.attached_child()
+        };
 
         async move {
             tracing::info!(%reason, "downloading on-demand");
@@ -1094,7 +1101,7 @@ impl LayerInner {
             let init_cancelled = scopeguard::guard((), |_| LAYER_IMPL_METRICS.inc_init_cancelled());
             let res = self
                 .download_init_and_wait(timeline, permit, ctx.attached_child())
-                .maybe_perf_instrument(&ctx, |current_perf_span| current_perf_span.clone())
+                .maybe_perf_instrument(&ctx, |crnt_perf_span| crnt_perf_span.clone())
                 .await?;
 
             scopeguard::ScopeGuard::into_inner(init_cancelled);
@@ -1702,7 +1709,7 @@ impl DownloadError {
     }
 }
 
-#[derive(Debug, PartialEq, Copy, Clone)]
+#[derive(Debug, PartialEq)]
 pub(crate) enum NeedsDownload {
     NotFound,
     NotFile(std::fs::FileType),

pageserver/src/tenant/storage_layer/merge_iterator.rs

@@ -19,6 +19,14 @@ pub(crate) enum LayerRef<'a> {
 }
 
 impl<'a> LayerRef<'a> {
+    #[allow(dead_code)]
+    fn iter(self, ctx: &'a RequestContext) -> LayerIterRef<'a> {
+        match self {
+            Self::Image(x) => LayerIterRef::Image(x.iter(ctx)),
+            Self::Delta(x) => LayerIterRef::Delta(x.iter(ctx)),
+        }
+    }
+
     fn iter_with_options(
         self,
         ctx: &'a RequestContext,
@@ -314,28 +322,6 @@ impl MergeIteratorItem for ((Key, Lsn, Value), Arc<PersistentLayerKey>) {
 }
 
 impl<'a> MergeIterator<'a> {
-    #[cfg(test)]
-    pub(crate) fn create_for_testing(
-        deltas: &[&'a DeltaLayerInner],
-        images: &[&'a ImageLayerInner],
-        ctx: &'a RequestContext,
-    ) -> Self {
-        Self::create_with_options(deltas, images, ctx, 1024 * 8192, 1024)
-    }
-
-    /// Create a new merge iterator with custom options.
-    ///
-    /// Adjust `max_read_size` and `max_batch_size` to trade memory usage for performance. The size should scale
-    /// with the number of layers to compact. If there are a lot of layers, consider reducing the values, so that
-    /// the buffer does not take too much memory.
-    ///
-    /// The default options for L0 compactions are:
-    /// - max_read_size: 1024 * 8192 (8MB)
-    /// - max_batch_size: 1024
-    ///
-    /// The default options for gc-compaction are:
-    /// - max_read_size: 128 * 8192 (1MB)
-    /// - max_batch_size: 128
     pub fn create_with_options(
         deltas: &[&'a DeltaLayerInner],
         images: &[&'a ImageLayerInner],
@@ -365,6 +351,14 @@ impl<'a> MergeIterator<'a> {
         }
     }
 
+    pub fn create(
+        deltas: &[&'a DeltaLayerInner],
+        images: &[&'a ImageLayerInner],
+        ctx: &'a RequestContext,
+    ) -> Self {
+        Self::create_with_options(deltas, images, ctx, 1024 * 8192, 1024)
+    }
+
     pub(crate) async fn next_inner<R: MergeIteratorItem>(&mut self) -> anyhow::Result<Option<R>> {
         while let Some(mut iter) = self.heap.peek_mut() {
             if !iter.is_loaded() {
@@ -483,7 +477,7 @@ mod tests {
         let resident_layer_2 = produce_delta_layer(&tenant, &tline, test_deltas2.clone(), &ctx)
             .await
             .unwrap();
-        let mut merge_iter = MergeIterator::create_for_testing(
+        let mut merge_iter = MergeIterator::create(
             &[
                 resident_layer_2.get_as_delta(&ctx).await.unwrap(),
                 resident_layer_1.get_as_delta(&ctx).await.unwrap(),
@@ -555,7 +549,7 @@ mod tests {
         let resident_layer_3 = produce_delta_layer(&tenant, &tline, test_deltas3.clone(), &ctx)
             .await
             .unwrap();
-        let mut merge_iter = MergeIterator::create_for_testing(
+        let mut merge_iter = MergeIterator::create(
             &[
                 resident_layer_1.get_as_delta(&ctx).await.unwrap(),
                 resident_layer_2.get_as_delta(&ctx).await.unwrap(),
@@ -676,7 +670,7 @@ mod tests {
         // Test with different layer order for MergeIterator::create to ensure the order
         // is stable.
 
-        let mut merge_iter = MergeIterator::create_for_testing(
+        let mut merge_iter = MergeIterator::create(
             &[
                 resident_layer_4.get_as_delta(&ctx).await.unwrap(),
                 resident_layer_1.get_as_delta(&ctx).await.unwrap(),
@@ -688,7 +682,7 @@ mod tests {
         );
         assert_merge_iter_equal(&mut merge_iter, &expect).await;
 
-        let mut merge_iter = MergeIterator::create_for_testing(
+        let mut merge_iter = MergeIterator::create(
             &[
                 resident_layer_1.get_as_delta(&ctx).await.unwrap(),
                 resident_layer_4.get_as_delta(&ctx).await.unwrap(),

pageserver/src/tenant/timeline/compaction.rs

@@ -1994,13 +1994,7 @@ impl Timeline {
                 let l = l.get_as_delta(ctx).await.map_err(CompactionError::Other)?;
                 deltas.push(l);
             }
-            MergeIterator::create_with_options(
-                &deltas,
-                &[],
-                ctx,
-                1024 * 8192, /* 8 MiB buffer per layer iterator */
-                1024,
-            )
+            MergeIterator::create(&deltas, &[], ctx)
         };
 
         // This iterator walks through all keys and is needed to calculate size used by each key
@@ -2834,7 +2828,7 @@ impl Timeline {
         Ok(())
     }
 
-    /// Check to bail out of gc compaction early if it would use too much memory.
+    /// Check if the memory usage is within the limit.
     async fn check_memory_usage(
         self: &Arc<Self>,
         layer_selection: &[Layer],
@@ -2847,8 +2841,7 @@ impl Timeline {
             let layer_desc = layer.layer_desc();
             if layer_desc.is_delta() {
                 // Delta layers at most have 1MB buffer; 3x to make it safe (there're deltas as large as 16KB).
-                // Scale it by target_layer_size_bytes so that tests can pass (some tests, e.g., `test_pageserver_gc_compaction_preempt
-                // use 3MB layer size and we need to account for that).
+                // Multiply the layer size so that tests can pass.
                 estimated_memory_usage_mb +=
                     3.0 * (layer_desc.file_size / target_layer_size_bytes) as f64;
                 num_delta_layers += 1;

pgxn/neon/neon_walreader.c

@@ -150,7 +150,7 @@ NeonWALReaderFree(NeonWALReader *state)
  * fetched from timeline 'tli'.
  *
  * Returns NEON_WALREAD_SUCCESS if succeeded, NEON_WALREAD_ERROR if an error
- * occurs, in which case 'err' has the description. Error always closes remote
+ * occurs, in which case 'err' has the desciption. Error always closes remote
  * connection, if there was any, so socket subscription should be removed.
  *
  * NEON_WALREAD_WOULDBLOCK means caller should obtain socket to wait for with

pgxn/neon/pagestore_smgr.c

@@ -1989,14 +1989,8 @@ neon_start_unlogged_build(SMgrRelation reln)
 			neon_log(ERROR, "unknown relpersistence '%c'", reln->smgr_relpersistence);
 	}
 
-#if PG_MAJORVERSION_NUM >= 17
-	/*
-	 * We have to disable this check for pg14-16 because sorted build of GIST index requires
-	 * to perform unlogged build several times
-	 */
 	if (smgrnblocks(reln, MAIN_FORKNUM) != 0)
 		neon_log(ERROR, "cannot perform unlogged index build, index is not empty ");
-#endif
 
 	unlogged_build_rel = reln;
 	unlogged_build_phase = UNLOGGED_BUILD_PHASE_1;

pgxn/neon/walproposer.c

@@ -124,7 +124,6 @@ WalProposerCreate(WalProposerConfig *config, walproposer_api api)
 	}
 	else
 	{
-		wp->safekeepers_generation = INVALID_GENERATION;
 		host = wp->config->safekeepers_list;
 	}
 	wp_log(LOG, "safekeepers_generation=%u", wp->safekeepers_generation);
@@ -757,7 +756,7 @@ UpdateMemberSafekeeperPtr(WalProposer *wp, Safekeeper *sk)
 	{
 		SafekeeperId *sk_id = &wp->mconf.members.m[i];
 
-		if (sk_id->node_id == sk->greetResponse.nodeId)
+		if (wp->mconf.members.m[i].node_id == sk->greetResponse.nodeId)
 		{
 			/*
 			 * If mconf or list of safekeepers to connect to changed (the
@@ -782,7 +781,7 @@ UpdateMemberSafekeeperPtr(WalProposer *wp, Safekeeper *sk)
 	{
 		SafekeeperId *sk_id = &wp->mconf.new_members.m[i];
 
-		if (sk_id->node_id == sk->greetResponse.nodeId)
+		if (wp->mconf.new_members.m[i].node_id == sk->greetResponse.nodeId)
 		{
 			if (wp->new_members_safekeepers[i] != NULL && wp->new_members_safekeepers[i] != sk)
 			{
@@ -1072,6 +1071,7 @@ RecvVoteResponse(Safekeeper *sk)
 	/* ready for elected message */
 	sk->state = SS_WAIT_ELECTED;
 
+	wp->n_votes++;
 	/* Are we already elected? */
 	if (wp->state == WPS_CAMPAIGN)
 	{

pgxn/neon/walproposer.h

@@ -845,6 +845,9 @@ typedef struct WalProposer
 	/* timeline globally starts at this LSN */
 	XLogRecPtr	timelineStartLsn;
 
+	/* number of votes collected from safekeepers */
+	int			n_votes;
+
 	/* number of successful connections over the lifetime of walproposer */
 	int			n_connected;
 

proxy/src/auth/backend/jwt.rs

@@ -409,22 +409,14 @@ impl JwkCacheEntryLock {
 
         if let Some(exp) = payload.expiration {
             if now >= exp + CLOCK_SKEW_LEEWAY {
-                return Err(JwtError::InvalidClaims(JwtClaimsError::JwtTokenHasExpired(
-                    exp.duration_since(SystemTime::UNIX_EPOCH)
-                        .unwrap_or_default()
-                        .as_secs(),
-                )));
+                return Err(JwtError::InvalidClaims(JwtClaimsError::JwtTokenHasExpired));
             }
         }
 
         if let Some(nbf) = payload.not_before {
             if nbf >= now + CLOCK_SKEW_LEEWAY {
                 return Err(JwtError::InvalidClaims(
-                    JwtClaimsError::JwtTokenNotYetReadyToUse(
-                        nbf.duration_since(SystemTime::UNIX_EPOCH)
-                            .unwrap_or_default()
-                            .as_secs(),
-                    ),
+                    JwtClaimsError::JwtTokenNotYetReadyToUse,
                 ));
             }
         }
@@ -542,10 +534,10 @@ struct JwtPayload<'a> {
     #[serde(rename = "aud", default)]
     audience: OneOrMany,
     /// Expiration - Time after which the JWT expires
-    #[serde(rename = "exp", deserialize_with = "numeric_date_opt", default)]
+    #[serde(deserialize_with = "numeric_date_opt", rename = "exp", default)]
     expiration: Option<SystemTime>,
-    /// Not before - Time before which the JWT is not valid
-    #[serde(rename = "nbf", deserialize_with = "numeric_date_opt", default)]
+    /// Not before - Time after which the JWT expires
+    #[serde(deserialize_with = "numeric_date_opt", rename = "nbf", default)]
     not_before: Option<SystemTime>,
 
     // the following entries are only extracted for the sake of debug logging.
@@ -617,15 +609,8 @@ impl<'de> Deserialize<'de> for OneOrMany {
 }
 
 fn numeric_date_opt<'de, D: Deserializer<'de>>(d: D) -> Result<Option<SystemTime>, D::Error> {
-    <Option<u64>>::deserialize(d)?
-        .map(|t| {
-            SystemTime::UNIX_EPOCH
-                .checked_add(Duration::from_secs(t))
-                .ok_or_else(|| {
-                    serde::de::Error::custom(format_args!("timestamp out of bounds: {t}"))
-                })
-        })
-        .transpose()
+    let d = <Option<u64>>::deserialize(d)?;
+    Ok(d.map(|n| SystemTime::UNIX_EPOCH + Duration::from_secs(n)))
 }
 
 struct JwkRenewalPermit<'a> {
@@ -761,11 +746,11 @@ pub enum JwtClaimsError {
     #[error("invalid JWT token audience")]
     InvalidJwtTokenAudience,
 
-    #[error("JWT token has expired (exp={0})")]
-    JwtTokenHasExpired(u64),
+    #[error("JWT token has expired")]
+    JwtTokenHasExpired,
 
-    #[error("JWT token is not yet ready to use (nbf={0})")]
-    JwtTokenNotYetReadyToUse(u64),
+    #[error("JWT token is not yet ready to use")]
+    JwtTokenNotYetReadyToUse,
 }
 
 #[allow(dead_code, reason = "Debug use only")]
@@ -1248,14 +1233,14 @@ X0n5X2/pBLJzxZc62ccvZYVnctBiFs6HbSnxpuMQCfkt/BcR/ttIepBQQIW86wHL
                     "nbf": now + 60,
                     "aud": "neon",
                 }},
-                error: JwtClaimsError::JwtTokenNotYetReadyToUse(now + 60),
+                error: JwtClaimsError::JwtTokenNotYetReadyToUse,
             },
             Test {
                 body: json! {{
                     "exp": now - 60,
                     "aud": ["neon"],
                 }},
-                error: JwtClaimsError::JwtTokenHasExpired(now - 60),
+                error: JwtClaimsError::JwtTokenHasExpired,
             },
             Test {
                 body: json! {{

proxy/src/serverless/local_conn_pool.rs

@@ -41,7 +41,7 @@ use crate::control_plane::messages::{ColdStartInfo, MetricsAuxInfo};
 use crate::metrics::Metrics;
 
 pub(crate) const EXT_NAME: &str = "pg_session_jwt";
-pub(crate) const EXT_VERSION: &str = "0.3.1";
+pub(crate) const EXT_VERSION: &str = "0.3.0";
 pub(crate) const EXT_SCHEMA: &str = "auth";
 
 #[derive(Clone)]

storage_controller/src/service.rs

@@ -5181,8 +5181,7 @@ impl Service {
             }
 
             // We don't expect any new_shard_count shards to exist here, but drop them just in case
-            tenants
-                .retain(|id, s| !(id.tenant_id == *tenant_id && s.shard.count == *new_shard_count));
+            tenants.retain(|_id, s| s.shard.count != *new_shard_count);
 
             detach_locations
         };

storage_scrubber/src/checks.rs

@@ -165,17 +165,16 @@ pub(crate) async fn branch_cleanup_and_check_errors(
                                 .head_object(&path, &CancellationToken::new())
                                 .await;
 
-                            if let Err(e) = response {
+                            if response.is_err() {
                                 // Object is not present.
                                 let is_l0 = LayerMap::is_l0(layer.key_range(), layer.is_delta());
 
                                 let msg = format!(
-                                    "index_part.json contains a layer {}{} (shard {}) that is not present in remote storage (layer_is_l0: {}) with error: {}",
+                                    "index_part.json contains a layer {}{} (shard {}) that is not present in remote storage (layer_is_l0: {})",
                                     layer,
                                     metadata.generation.get_suffix(),
                                     metadata.shard,
                                     is_l0,
-                                    e,
                                 );
 
                                 if is_l0 || ignore_error {

storage_scrubber/src/pageserver_physical_gc.rs

@@ -137,10 +137,11 @@ struct TenantRefAccumulator {
 impl TenantRefAccumulator {
     fn update(&mut self, ttid: TenantShardTimelineId, index_part: &IndexPart) {
         let this_shard_idx = ttid.tenant_shard_id.to_index();
-        self.shards_seen
+        (*self
+            .shards_seen
             .entry(ttid.tenant_shard_id.tenant_id)
-            .or_default()
-            .insert(this_shard_idx);
+            .or_default())
+        .insert(this_shard_idx);
 
         let mut ancestor_refs = Vec::new();
         for (layer_name, layer_metadata) in &index_part.layer_metadata {
@@ -766,13 +767,10 @@ pub async fn pageserver_physical_gc(
                 stream_tenant_timelines(remote_client_ref, target_ref, tenant_shard_id).await?,
             );
             Ok(try_stream! {
-                let mut cnt = 0;
                 while let Some(ttid_res) = timelines.next().await {
                     let ttid = ttid_res?;
-                    cnt += 1;
                     yield (ttid, tenant_manifest_arc.clone());
                 }
-                tracing::info!(%tenant_shard_id, "Found {} timelines", cnt);
             })
         }
     });
@@ -792,7 +790,6 @@ pub async fn pageserver_physical_gc(
                 &accumulator,
                 tenant_manifest_arc,
             )
-            .instrument(info_span!("gc_timeline", %ttid))
         });
         let timelines = timelines.try_buffered(CONCURRENCY);
         let mut timelines = std::pin::pin!(timelines);

storage_scrubber/src/scan_pageserver_metadata.rs

@@ -153,10 +153,7 @@ pub async fn scan_pageserver_metadata(
     const CONCURRENCY: usize = 32;
 
     // Generate a stream of TenantTimelineId
-    let timelines = tenants.map_ok(|t| {
-        tracing::info!("Found tenant: {}", t);
-        stream_tenant_timelines(&remote_client, &target, t)
-    });
+    let timelines = tenants.map_ok(|t| stream_tenant_timelines(&remote_client, &target, t));
     let timelines = timelines.try_buffered(CONCURRENCY);
     let timelines = timelines.try_flatten();
 

test_runner/fixtures/endpoint/http.py

@@ -1,7 +1,6 @@
 from __future__ import annotations
 
 import urllib.parse
-from enum import StrEnum
 from typing import TYPE_CHECKING, final
 
 import requests
@@ -10,23 +9,11 @@ from requests.auth import AuthBase
 from typing_extensions import override
 
 from fixtures.log_helper import log
-from fixtures.utils import wait_until
 
 if TYPE_CHECKING:
     from requests import PreparedRequest
 
 
-COMPUTE_AUDIENCE = "compute"
-"""
-The value to place in the `aud` claim.
-"""
-
-
-@final
-class ComputeClaimsScope(StrEnum):
-    ADMIN = "admin"
-
-
 @final
 class BearerAuth(AuthBase):
     """
@@ -63,35 +50,6 @@ class EndpointHttpClient(requests.Session):
         res.raise_for_status()
         return res.json()
 
-    def prewarm_lfc_status(self) -> dict[str, str]:
-        res = self.get(f"http://localhost:{self.external_port}/lfc/prewarm")
-        res.raise_for_status()
-        json: dict[str, str] = res.json()
-        return json
-
-    def prewarm_lfc(self):
-        self.post(f"http://localhost:{self.external_port}/lfc/prewarm").raise_for_status()
-
-        def prewarmed():
-            json = self.prewarm_lfc_status()
-            status, err = json["status"], json.get("error")
-            assert status == "completed", f"{status}, error {err}"
-
-        wait_until(prewarmed)
-
-    def offload_lfc(self):
-        url = f"http://localhost:{self.external_port}/lfc/offload"
-        self.post(url).raise_for_status()
-
-        def offloaded():
-            res = self.get(url)
-            res.raise_for_status()
-            json = res.json()
-            status, err = json["status"], json.get("error")
-            assert status == "completed", f"{status}, error {err}"
-
-        wait_until(offloaded)
-
     def database_schema(self, database: str):
         res = self.get(
             f"http://localhost:{self.external_port}/database_schema?database={urllib.parse.quote(database, safe='')}",

test_runner/fixtures/neon_cli.py

@@ -21,7 +21,6 @@ if TYPE_CHECKING:
         Any,
     )
 
-    from fixtures.endpoint.http import ComputeClaimsScope
     from fixtures.pg_version import PgVersion
 
 
@@ -536,16 +535,12 @@ class NeonLocalCli(AbstractNeonCli):
         res.check_returncode()
         return res
 
-    def endpoint_generate_jwt(
-        self, endpoint_id: str, scope: ComputeClaimsScope | None = None
-    ) -> str:
+    def endpoint_generate_jwt(self, endpoint_id: str) -> str:
         """
         Generate a JWT for making requests to the endpoint's external HTTP
         server.
         """
         args = ["endpoint", "generate-jwt", endpoint_id]
-        if scope:
-            args += ["--scope", str(scope)]
 
         cmd = self.raw_cli(args)
         cmd.check_returncode()

test_runner/fixtures/neon_fixtures.py

@@ -51,7 +51,7 @@ from fixtures.common_types import (
     TimelineId,
 )
 from fixtures.compute_migrations import NUM_COMPUTE_MIGRATIONS
-from fixtures.endpoint.http import ComputeClaimsScope, EndpointHttpClient
+from fixtures.endpoint.http import EndpointHttpClient
 from fixtures.log_helper import log
 from fixtures.metrics import Metrics, MetricsGetter, parse_metrics
 from fixtures.neon_cli import NeonLocalCli, Pagectl
@@ -1185,9 +1185,7 @@ class NeonEnv:
             "broker": {},
             "safekeepers": [],
             "pageservers": [],
-            "endpoint_storage": {
-                "listen_addr": f"127.0.0.1:{self.port_distributor.get_port()}",
-            },
+            "endpoint_storage": {"port": self.port_distributor.get_port()},
             "generate_local_ssl_certs": self.generate_local_ssl_certs,
         }
 
@@ -4220,7 +4218,7 @@ class Endpoint(PgProtocol, LogUtils):
 
         self.config(config_lines)
 
-        self.__jwt = self.generate_jwt()
+        self.__jwt = self.env.neon_cli.endpoint_generate_jwt(self.endpoint_id)
 
         return self
 
@@ -4267,14 +4265,6 @@ class Endpoint(PgProtocol, LogUtils):
 
         return self
 
-    def generate_jwt(self, scope: ComputeClaimsScope | None = None) -> str:
-        """
-        Generate a JWT for making requests to the endpoint's external HTTP
-        server.
-        """
-        assert self.endpoint_id is not None
-        return self.env.neon_cli.endpoint_generate_jwt(self.endpoint_id, scope)
-
     def endpoint_path(self) -> Path:
         """Path to endpoint directory"""
         assert self.endpoint_id

test_runner/random_ops/test_random_ops.py

@@ -206,7 +206,7 @@ class NeonProject:
         self.neon_api = neon_api
         self.pg_bin = pg_bin
         proj = self.neon_api.create_project(
-            pg_version, f"Automatic random API test GITHUB_RUN_ID={os.getenv('GITHUB_RUN_ID')}"
+            pg_version, f"Automatic random API test {os.getenv('GITHUB_RUN_ID')}"
         )
         self.id: str = proj["project"]["id"]
         self.name: str = proj["project"]["name"]

test_runner/regress/test_compute_http.py

@@ -1,78 +0,0 @@
-from __future__ import annotations
-
-from http.client import FORBIDDEN, UNAUTHORIZED
-from typing import TYPE_CHECKING
-
-import jwt
-import pytest
-from fixtures.endpoint.http import COMPUTE_AUDIENCE, ComputeClaimsScope, EndpointHttpClient
-from fixtures.utils import run_only_on_default_postgres
-from requests import RequestException
-
-if TYPE_CHECKING:
-    from fixtures.neon_fixtures import NeonEnv
-
-
-@run_only_on_default_postgres("The code path being tested is not dependent on Postgres version")
-def test_compute_no_scope_claim(neon_simple_env: NeonEnv):
-    """
-    Test that if the JWT scope is not admin and no compute_id is specified,
-    the external HTTP server returns a 403 Forbidden error.
-    """
-    env = neon_simple_env
-
-    endpoint = env.endpoints.create_start("main")
-
-    # Encode nothing in the token
-    token = jwt.encode({}, env.auth_keys.priv, algorithm="EdDSA")
-
-    # Create an admin-scoped HTTP client
-    client = EndpointHttpClient(
-        external_port=endpoint.external_http_port,
-        internal_port=endpoint.internal_http_port,
-        jwt=token,
-    )
-
-    try:
-        client.status()
-        pytest.fail("Exception should have been raised")
-    except RequestException as e:
-        assert e.response is not None
-        assert e.response.status_code == FORBIDDEN
-
-
-@pytest.mark.parametrize(
-    "audience",
-    (COMPUTE_AUDIENCE, "invalid", None),
-    ids=["with_audience", "with_invalid_audience", "without_audience"],
-)
-@run_only_on_default_postgres("The code path being tested is not dependent on Postgres version")
-def test_compute_admin_scope_claim(neon_simple_env: NeonEnv, audience: str | None):
-    """
-    Test that an admin-scoped JWT can access the compute's external HTTP server
-    without the compute_id being specified in the claims.
-    """
-    env = neon_simple_env
-
-    endpoint = env.endpoints.create_start("main")
-
-    data: dict[str, str | list[str]] = {"scope": str(ComputeClaimsScope.ADMIN)}
-    if audience:
-        data["aud"] = [audience]
-
-    token = jwt.encode(data, env.auth_keys.priv, algorithm="EdDSA")
-
-    # Create an admin-scoped HTTP client
-    client = EndpointHttpClient(
-        external_port=endpoint.external_http_port,
-        internal_port=endpoint.internal_http_port,
-        jwt=token,
-    )
-
-    try:
-        client.status()
-        if audience != COMPUTE_AUDIENCE:
-            pytest.fail("Exception should have been raised")
-    except RequestException as e:
-        assert e.response is not None
-        assert e.response.status_code == UNAUTHORIZED

test_runner/regress/test_endpoint_storage.py

@@ -4,12 +4,10 @@ import pytest
 from aiohttp import ClientSession
 from fixtures.log_helper import log
 from fixtures.neon_fixtures import NeonEnv
-from fixtures.utils import run_only_on_default_postgres
 from jwcrypto import jwk, jwt
 
 
 @pytest.mark.asyncio
-@run_only_on_default_postgres("test doesn't use postgres")
 async def test_endpoint_storage_insert_retrieve_delete(neon_simple_env: NeonEnv):
     """
     Inserts, retrieves, and deletes test file using a JWT token
@@ -37,6 +35,7 @@ async def test_endpoint_storage_insert_retrieve_delete(neon_simple_env: NeonEnv)
     key = f"http://{base_url}/{tenant_id}/{timeline_id}/{endpoint_id}/key"
     headers = {"Authorization": f"Bearer {token}"}
     log.info(f"cache key url {key}")
+    log.info(f"token {token}")
 
     async with ClientSession(headers=headers) as session:
         async with session.get(key) as res:

test_runner/regress/test_gist.py

@@ -1,28 +0,0 @@
-from __future__ import annotations
-
-from typing import TYPE_CHECKING
-
-if TYPE_CHECKING:
-    from fixtures.neon_fixtures import NeonEnv
-
-
-#
-# Test unlogged build for GIST index
-#
-def test_gist(neon_simple_env: NeonEnv):
-    env = neon_simple_env
-    endpoint = env.endpoints.create_start("main")
-    con = endpoint.connect()
-    cur = con.cursor()
-    iterations = 100
-
-    for _ in range(iterations):
-        cur.execute(
-            "CREATE TABLE pvactst (i INT, a INT[], p POINT) with (autovacuum_enabled = off)"
-        )
-        cur.execute(
-            "INSERT INTO pvactst SELECT i, array[1,2,3], point(i, i+1) FROM generate_series(1,1000) i"
-        )
-        cur.execute("CREATE INDEX gist_pvactst ON pvactst USING gist (p)")
-        cur.execute("VACUUM pvactst")
-        cur.execute("DROP TABLE pvactst")

test_runner/regress/test_lfc_prewarm.py

@@ -1,24 +1,11 @@
 import random
 import threading
 import time
-from enum import Enum
 
 import pytest
-from fixtures.endpoint.http import EndpointHttpClient
 from fixtures.log_helper import log
 from fixtures.neon_fixtures import NeonEnv
 from fixtures.utils import USE_LFC
-from prometheus_client.parser import text_string_to_metric_families as prom_parse_impl
-
-
-class LfcQueryMethod(Enum):
-    COMPUTE_CTL = False
-    POSTGRES = True
-
-
-PREWARM_LABEL = "compute_ctl_lfc_prewarm_requests_total"
-OFFLOAD_LABEL = "compute_ctl_lfc_offload_requests_total"
-QUERY_OPTIONS = LfcQueryMethod.POSTGRES, LfcQueryMethod.COMPUTE_CTL
 
 
 def check_pinned_entries(cur):
@@ -32,20 +19,11 @@ def check_pinned_entries(cur):
     assert n_pinned == 0
 
 
-def prom_parse(client: EndpointHttpClient) -> dict[str, float]:
-    return {
-        sample.name: sample.value
-        for family in prom_parse_impl(client.metrics())
-        for sample in family.samples
-        if sample.name in (PREWARM_LABEL, OFFLOAD_LABEL)
-    }
-
-
 @pytest.mark.skipif(not USE_LFC, reason="LFC is disabled, skipping")
-@pytest.mark.parametrize("query", QUERY_OPTIONS, ids=["postgres", "compute-ctl"])
-def test_lfc_prewarm(neon_simple_env: NeonEnv, query: LfcQueryMethod):
+def test_lfc_prewarm(neon_simple_env: NeonEnv):
     env = neon_simple_env
     n_records = 1000000
+
     endpoint = env.endpoints.create_start(
         branch_name="main",
         config_lines=[
@@ -56,57 +34,30 @@ def test_lfc_prewarm(neon_simple_env: NeonEnv, query: LfcQueryMethod):
             "neon.file_cache_prewarm_limit=1000",
         ],
     )
-
-    pg_conn = endpoint.connect()
-    pg_cur = pg_conn.cursor()
-    pg_cur.execute("create extension neon version '1.6'")
-    pg_cur.execute("create database lfc")
-
-    lfc_conn = endpoint.connect(dbname="lfc")
-    lfc_cur = lfc_conn.cursor()
-    log.info(f"Inserting {n_records} rows")
-    lfc_cur.execute("create table t(pk integer primary key, payload text default repeat('?', 128))")
-    lfc_cur.execute(f"insert into t (pk) values (generate_series(1,{n_records}))")
-    log.info(f"Inserted {n_records} rows")
-
-    http_client = endpoint.http_client()
-    if query is LfcQueryMethod.COMPUTE_CTL:
-        status = http_client.prewarm_lfc_status()
-        assert status["status"] == "not_prewarmed"
-        assert "error" not in status
-        http_client.offload_lfc()
-        assert http_client.prewarm_lfc_status()["status"] == "not_prewarmed"
-        assert prom_parse(http_client) == {OFFLOAD_LABEL: 1, PREWARM_LABEL: 0}
-    else:
-        pg_cur.execute("select get_local_cache_state()")
-        lfc_state = pg_cur.fetchall()[0][0]
+    conn = endpoint.connect()
+    cur = conn.cursor()
+    cur.execute("create extension neon version '1.6'")
+    cur.execute("create table t(pk integer primary key, payload text default repeat('?', 128))")
+    cur.execute(f"insert into t (pk) values (generate_series(1,{n_records}))")
+    cur.execute("select get_local_cache_state()")
+    lfc_state = cur.fetchall()[0][0]
 
     endpoint.stop()
     endpoint.start()
 
-    # wait until compute_ctl completes downgrade of extension to default version
-    time.sleep(1)
-    pg_conn = endpoint.connect()
-    pg_cur = pg_conn.cursor()
-    pg_cur.execute("alter extension neon update to '1.6'")
+    conn = endpoint.connect()
+    cur = conn.cursor()
+    time.sleep(1)  # wait until compute_ctl complete downgrade of extension to default version
+    cur.execute("alter extension neon update to '1.6'")
+    cur.execute("select prewarm_local_cache(%s)", (lfc_state,))
 
-    lfc_conn = endpoint.connect(dbname="lfc")
-    lfc_cur = lfc_conn.cursor()
-
-    if query is LfcQueryMethod.COMPUTE_CTL:
-        http_client.prewarm_lfc()
-    else:
-        pg_cur.execute("select prewarm_local_cache(%s)", (lfc_state,))
-
-    pg_cur.execute("select lfc_value from neon_lfc_stats where lfc_key='file_cache_used_pages'")
-    lfc_used_pages = pg_cur.fetchall()[0][0]
+    cur.execute("select lfc_value from neon_lfc_stats where lfc_key='file_cache_used_pages'")
+    lfc_used_pages = cur.fetchall()[0][0]
     log.info(f"Used LFC size: {lfc_used_pages}")
-    pg_cur.execute("select * from get_prewarm_info()")
-    prewarm_info = pg_cur.fetchall()[0]
+    cur.execute("select * from get_prewarm_info()")
+    prewarm_info = cur.fetchall()[0]
     log.info(f"Prewarm info: {prewarm_info}")
-    total, prewarmed, skipped, _ = prewarm_info
-    progress = (prewarmed + skipped) * 100 // total
-    log.info(f"Prewarm progress: {progress}%")
+    log.info(f"Prewarm progress: {(prewarm_info[1] + prewarm_info[2]) * 100 // prewarm_info[0]}%")
 
     assert lfc_used_pages > 10000
     assert (
@@ -115,23 +66,18 @@ def test_lfc_prewarm(neon_simple_env: NeonEnv, query: LfcQueryMethod):
         and prewarm_info[0] == prewarm_info[1] + prewarm_info[2]
     )
 
-    lfc_cur.execute("select sum(pk) from t")
-    assert lfc_cur.fetchall()[0][0] == n_records * (n_records + 1) / 2
+    cur.execute("select sum(pk) from t")
+    assert cur.fetchall()[0][0] == n_records * (n_records + 1) / 2
 
-    check_pinned_entries(pg_cur)
-
-    desired = {"status": "completed", "total": total, "prewarmed": prewarmed, "skipped": skipped}
-    if query is LfcQueryMethod.COMPUTE_CTL:
-        assert http_client.prewarm_lfc_status() == desired
-        assert prom_parse(http_client) == {OFFLOAD_LABEL: 0, PREWARM_LABEL: 1}
+    check_pinned_entries(cur)
 
 
 @pytest.mark.skipif(not USE_LFC, reason="LFC is disabled, skipping")
-@pytest.mark.parametrize("query", QUERY_OPTIONS, ids=["postgres", "compute-ctl"])
-def test_lfc_prewarm_under_workload(neon_simple_env: NeonEnv, query: LfcQueryMethod):
+def test_lfc_prewarm_under_workload(neon_simple_env: NeonEnv):
     env = neon_simple_env
     n_records = 10000
     n_threads = 4
+
     endpoint = env.endpoints.create_start(
         branch_name="main",
         config_lines=[
@@ -141,58 +87,40 @@ def test_lfc_prewarm_under_workload(neon_simple_env: NeonEnv, query: LfcQueryMet
             "neon.file_cache_prewarm_limit=1000000",
         ],
     )
-
-    pg_conn = endpoint.connect()
-    pg_cur = pg_conn.cursor()
-    pg_cur.execute("create extension neon version '1.6'")
-    pg_cur.execute("CREATE DATABASE lfc")
-
-    lfc_conn = endpoint.connect(dbname="lfc")
-    lfc_cur = lfc_conn.cursor()
-    lfc_cur.execute(
+    conn = endpoint.connect()
+    cur = conn.cursor()
+    cur.execute("create extension neon version '1.6'")
+    cur.execute(
         "create table accounts(id integer primary key, balance bigint default 0, payload text default repeat('?', 1000)) with (fillfactor=10)"
     )
-    log.info(f"Inserting {n_records} rows")
-    lfc_cur.execute(f"insert into accounts(id) values (generate_series(1,{n_records}))")
-    log.info(f"Inserted {n_records} rows")
-
-    http_client = endpoint.http_client()
-    if query is LfcQueryMethod.COMPUTE_CTL:
-        http_client.offload_lfc()
-    else:
-        pg_cur.execute("select get_local_cache_state()")
-        lfc_state = pg_cur.fetchall()[0][0]
+    cur.execute(f"insert into accounts(id) values (generate_series(1,{n_records}))")
+    cur.execute("select get_local_cache_state()")
+    lfc_state = cur.fetchall()[0][0]
 
     running = True
-    n_prewarms = 0
 
     def workload():
-        lfc_conn = endpoint.connect(dbname="lfc")
-        lfc_cur = lfc_conn.cursor()
+        conn = endpoint.connect()
+        cur = conn.cursor()
         n_transfers = 0
         while running:
             src = random.randint(1, n_records)
             dst = random.randint(1, n_records)
-            lfc_cur.execute("update accounts set balance=balance-100 where id=%s", (src,))
-            lfc_cur.execute("update accounts set balance=balance+100 where id=%s", (dst,))
+            cur.execute("update accounts set balance=balance-100 where id=%s", (src,))
+            cur.execute("update accounts set balance=balance+100 where id=%s", (dst,))
             n_transfers += 1
         log.info(f"Number of transfers: {n_transfers}")
 
     def prewarm():
-        pg_conn = endpoint.connect()
-        pg_cur = pg_conn.cursor()
+        conn = endpoint.connect()
+        cur = conn.cursor()
+        n_prewarms = 0
         while running:
-            pg_cur.execute("alter system set neon.file_cache_size_limit='1MB'")
-            pg_cur.execute("select pg_reload_conf()")
-            pg_cur.execute("alter system set neon.file_cache_size_limit='1GB'")
-            pg_cur.execute("select pg_reload_conf()")
-
-            if query is LfcQueryMethod.COMPUTE_CTL:
-                http_client.prewarm_lfc()
-            else:
-                pg_cur.execute("select prewarm_local_cache(%s)", (lfc_state,))
-
-            nonlocal n_prewarms
+            cur.execute("alter system set neon.file_cache_size_limit='1MB'")
+            cur.execute("select pg_reload_conf()")
+            cur.execute("alter system set neon.file_cache_size_limit='1GB'")
+            cur.execute("select pg_reload_conf()")
+            cur.execute("select prewarm_local_cache(%s)", (lfc_state,))
             n_prewarms += 1
         log.info(f"Number of prewarms: {n_prewarms}")
 
@@ -212,10 +140,8 @@ def test_lfc_prewarm_under_workload(neon_simple_env: NeonEnv, query: LfcQueryMet
         t.join()
     prewarm_thread.join()
 
-    lfc_cur.execute("select sum(balance) from accounts")
-    total_balance = lfc_cur.fetchall()[0][0]
+    cur.execute("select sum(balance) from accounts")
+    total_balance = cur.fetchall()[0][0]
     assert total_balance == 0
 
-    check_pinned_entries(pg_cur)
-    if query is LfcQueryMethod.COMPUTE_CTL:
-        assert prom_parse(http_client) == {OFFLOAD_LABEL: 1, PREWARM_LABEL: n_prewarms}
+    check_pinned_entries(cur)

test_runner/regress/test_sharding.py

@@ -1334,13 +1334,6 @@ def test_sharding_split_failures(
         tenant_id, timeline_id, shard_count=initial_shard_count, placement_policy='{"Attached":1}'
     )
 
-    # Create bystander tenants with various shard counts. They should not be affected by the aborted
-    # splits. Regression test for https://github.com/neondatabase/cloud/issues/28589.
-    bystanders = {}  # id → shard_count
-    for bystander_shard_count in [1, 2, 4, 8]:
-        id, _ = env.create_tenant(shard_count=bystander_shard_count)
-        bystanders[id] = bystander_shard_count
-
     env.storage_controller.allowed_errors.extend(
         [
             # All split failures log a warning when then enqueue the abort operation
@@ -1401,8 +1394,6 @@ def test_sharding_split_failures(
             locations = ps.http_client().tenant_list_locations()["tenant_shards"]
             for loc in locations:
                 tenant_shard_id = TenantShardId.parse(loc[0])
-                if tenant_shard_id.tenant_id != tenant_id:
-                    continue  # skip bystanders
                 log.info(f"Shard {tenant_shard_id} seen on node {ps.id} in mode {loc[1]['mode']}")
                 assert tenant_shard_id.shard_count == initial_shard_count
                 if loc[1]["mode"] == "Secondary":
@@ -1423,8 +1414,6 @@ def test_sharding_split_failures(
             locations = ps.http_client().tenant_list_locations()["tenant_shards"]
             for loc in locations:
                 tenant_shard_id = TenantShardId.parse(loc[0])
-                if tenant_shard_id.tenant_id != tenant_id:
-                    continue  # skip bystanders
                 log.info(f"Shard {tenant_shard_id} seen on node {ps.id} in mode {loc[1]['mode']}")
                 assert tenant_shard_id.shard_count == split_shard_count
                 if loc[1]["mode"] == "Secondary":
@@ -1507,12 +1496,6 @@ def test_sharding_split_failures(
     # the scheduler reaches an idle state
     env.storage_controller.reconcile_until_idle(timeout_secs=30)
 
-    # Check that all bystanders are still around.
-    for bystander_id, bystander_shard_count in bystanders.items():
-        response = env.storage_controller.tenant_describe(bystander_id)
-        assert TenantId(response["tenant_id"]) == bystander_id
-        assert len(response["shards"]) == bystander_shard_count
-
     env.storage_controller.consistency_check()
 
 

vendor/revisions.json

@@ -5,14 +5,14 @@
   ],
   "v16": [
     "16.8",
-    "05ddf212e2e07b788b5c8b88bdcf98630941f6ae"
+    "37496f87b5324af53c56127e278ee5b1e8435253"
   ],
   "v15": [
     "15.12",
-    "b838c8969b7c63f3e637a769656f5f36793b797c"
+    "8ecb12f21d862dfa39f7204b8f5e1c00a2a225b3"
   ],
   "v14": [
     "14.17",
-    "c8dab02bfc003ae7bd59096919042d7840f3c194"
+    "d3c9d61fb7a362a165dac7060819dd9d6ad68c28"
   ]
 }