Move genesis/snapshot archive download into Rust (#5515)

Cargo.lock

@@ -1068,7 +1068,6 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 dependencies = [
  "crc32fast 1.2.0 (registry+https://github.com/rust-lang/crates.io-index)",
  "libc 0.2.61 (registry+https://github.com/rust-lang/crates.io-index)",
- "miniz-sys 0.1.12 (registry+https://github.com/rust-lang/crates.io-index)",
  "miniz_oxide_c_api 0.2.3 (registry+https://github.com/rust-lang/crates.io-index)",
 ]
 
@@ -1813,15 +1812,6 @@ dependencies = [
  "unicase 1.4.2 (registry+https://github.com/rust-lang/crates.io-index)",
 ]
 
-[[package]]
-name = "miniz-sys"
-version = "0.1.12"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-dependencies = [
- "cc 1.0.38 (registry+https://github.com/rust-lang/crates.io-index)",
- "libc 0.2.61 (registry+https://github.com/rust-lang/crates.io-index)",
-]
-
 [[package]]
 name = "miniz_oxide"
 version = "0.3.0"
@@ -3053,12 +3043,12 @@ dependencies = [
  "bincode 1.1.4 (registry+https://github.com/rust-lang/crates.io-index)",
  "bs58 0.2.2 (registry+https://github.com/rust-lang/crates.io-index)",
  "byteorder 1.3.2 (registry+https://github.com/rust-lang/crates.io-index)",
+ "bzip2 0.3.3 (registry+https://github.com/rust-lang/crates.io-index)",
  "chrono 0.4.7 (registry+https://github.com/rust-lang/crates.io-index)",
  "core_affinity 0.5.9 (registry+https://github.com/rust-lang/crates.io-index)",
  "crc 1.8.1 (registry+https://github.com/rust-lang/crates.io-index)",
  "crossbeam-channel 0.3.9 (registry+https://github.com/rust-lang/crates.io-index)",
  "dir-diff 0.3.1 (registry+https://github.com/rust-lang/crates.io-index)",
- "flate2 1.0.9 (registry+https://github.com/rust-lang/crates.io-index)",
  "fs_extra 1.1.0 (registry+https://github.com/rust-lang/crates.io-index)",
  "hex-literal 0.2.0 (registry+https://github.com/rust-lang/crates.io-index)",
  "indexmap 1.0.2 (registry+https://github.com/rust-lang/crates.io-index)",
@@ -3830,8 +3820,10 @@ dependencies = [
 name = "solana-validator"
 version = "0.18.0-pre1"
 dependencies = [
+ "bzip2 0.3.3 (registry+https://github.com/rust-lang/crates.io-index)",
  "clap 2.33.0 (registry+https://github.com/rust-lang/crates.io-index)",
  "log 0.4.8 (registry+https://github.com/rust-lang/crates.io-index)",
+ "reqwest 0.9.19 (registry+https://github.com/rust-lang/crates.io-index)",
  "serde_json 1.0.40 (registry+https://github.com/rust-lang/crates.io-index)",
  "solana 0.18.0-pre1",
  "solana-drone 0.18.0-pre1",
@@ -3842,6 +3834,8 @@ dependencies = [
  "solana-sdk 0.18.0-pre1",
  "solana-vote-api 0.18.0-pre1",
  "solana-vote-signer 0.18.0-pre1",
+ "tar 0.4.26 (registry+https://github.com/rust-lang/crates.io-index)",
+ "tempfile 3.1.0 (registry+https://github.com/rust-lang/crates.io-index)",
 ]
 
 [[package]]
@@ -5430,7 +5424,6 @@ dependencies = [
 "checksum mime 0.2.6 (registry+https://github.com/rust-lang/crates.io-index)" = "ba626b8a6de5da682e1caa06bdb42a335aee5a84db8e5046a3e8ab17ba0a3ae0"
 "checksum mime 0.3.13 (registry+https://github.com/rust-lang/crates.io-index)" = "3e27ca21f40a310bd06d9031785f4801710d566c184a6e15bad4f1d9b65f9425"
 "checksum mime_guess 2.0.0-alpha.6 (registry+https://github.com/rust-lang/crates.io-index)" = "30de2e4613efcba1ec63d8133f344076952090c122992a903359be5a4f99c3ed"
-"checksum miniz-sys 0.1.12 (registry+https://github.com/rust-lang/crates.io-index)" = "1e9e3ae51cea1576ceba0dde3d484d30e6e5b86dee0b2d412fe3a16a15c98202"
 "checksum miniz_oxide 0.3.0 (registry+https://github.com/rust-lang/crates.io-index)" = "c061edee74a88eb35d876ce88b94d77a0448a201de111c244b70d047f5820516"
 "checksum miniz_oxide_c_api 0.2.3 (registry+https://github.com/rust-lang/crates.io-index)" = "6c675792957b0d19933816c4e1d56663c341dd9bfa31cb2140ff2267c1d8ecf4"
 "checksum mio 0.6.19 (registry+https://github.com/rust-lang/crates.io-index)" = "83f51996a3ed004ef184e16818edc51fadffe8e7ca68be67f9dee67d84d0ff23"

ci/testnet-deploy.sh

@@ -90,7 +90,7 @@ Deploys a CD testnet
                         - Attempt to generate a TLS certificate using this DNS name
    --fullnode-additional-disk-size-gb [number]
                         - Size of additional disk in GB for all fullnodes
-   --no-snapshot
+   --no-snapshot-fetch
                         - If set, disables booting validators from a snapshot
 
    Note: the SOLANA_METRICS_CONFIG environment variable is used to configure
@@ -137,7 +137,7 @@ while [[ -n $1 ]]; do
     elif [[ $1 == --machine-type* ]]; then # Bypass quoted long args for GPUs
       shortArgs+=("$1")
       shift
-    elif [[ $1 = --no-snapshot ]]; then
+    elif [[ $1 = --no-snapshot-fetch ]]; then
       maybeNoSnapshot="$1"
       shift 1
     else

core/Cargo.toml

@@ -21,12 +21,12 @@ kvstore = ["solana-kvstore"]
 bincode = "1.1.4"
 bs58 = "0.2.0"
 byteorder = "1.3.2"
+bzip2 = "0.3.3"
 chrono = { version = "0.4.7", features = ["serde"] }
 core_affinity = "0.5.9"
 crc = { version = "1.8.1", optional = true }
 crossbeam-channel = "0.3"
 dir-diff = "0.3.1"
-flate2 = "1.0.9"
 fs_extra = "1.1.0"
 indexmap = "1.0"
 itertools = "0.8.0"

core/src/lib.rs

@@ -102,9 +102,9 @@ extern crate solana_metrics;
 #[macro_use]
 extern crate matches;
 
+extern crate bzip2;
 extern crate crossbeam_channel;
 extern crate dir_diff;
-extern crate flate2;
 extern crate fs_extra;
 extern crate tar;
 extern crate tempfile;

core/src/rpc_service.rs

@@ -70,8 +70,8 @@ impl RequestMiddleware for RpcRequestMiddleware {
     fn on_request(&self, request: hyper::Request<hyper::Body>) -> RequestMiddlewareAction {
         trace!("request uri: {}", request.uri());
         match request.uri().path() {
-            "/snapshot.tgz" => self.get("snapshot.tgz"),
+            "/snapshot.tar.bz2" => self.get("snapshot.tar.bz2"),
-            "/genesis.tgz" => self.get("genesis.tgz"),
+            "/genesis.tar.bz2" => self.get("genesis.tar.bz2"),
             _ => RequestMiddlewareAction::Proceed {
                 should_continue_on_invalid_cors: false,
                 request,

core/src/snapshot_package.rs

@@ -1,7 +1,6 @@
 use crate::result::{Error, Result};
 use crate::service::Service;
-use flate2::write::GzEncoder;
+use bzip2::write::BzEncoder;
-use flate2::Compression;
 use solana_runtime::accounts_db::AccountStorageEntry;
 use std::fs;
 use std::path::Path;
@@ -77,11 +76,11 @@ impl SnapshotPackagerService {
         // Create the tar builder
         let tar_gz = tempfile::Builder::new()
             .prefix("new_state")
-            .suffix(".tgz")
+            .suffix(".tar.bz2")
             .tempfile_in(tar_dir)?;
 
         let temp_tar_path = tar_gz.path();
-        let enc = GzEncoder::new(&tar_gz, Compression::default());
+        let enc = BzEncoder::new(&tar_gz, bzip2::Compression::Default);
         let mut tar = tar::Builder::new(enc);
 
         // Create the list of paths to compress, starting with the snapshots

core/src/snapshot_utils.rs

@@ -3,7 +3,7 @@ use crate::result::{Error, Result};
 use crate::snapshot_package::SnapshotPackage;
 use crate::snapshot_package::{TAR_ACCOUNTS_DIR, TAR_SNAPSHOTS_DIR};
 use bincode::{deserialize_from, serialize_into};
-use flate2::read::GzDecoder;
+use bzip2::bufread::BzDecoder;
 use fs_extra::dir::CopyOptions;
 use solana_runtime::bank::Bank;
 use solana_runtime::status_cache::SlotDelta;
@@ -198,15 +198,15 @@ pub fn bank_from_archive<P: AsRef<Path>>(
 }
 
 pub fn get_snapshot_tar_path<P: AsRef<Path>>(snapshot_output_dir: P) -> PathBuf {
-    snapshot_output_dir.as_ref().join("snapshot.tgz")
+    snapshot_output_dir.as_ref().join("snapshot.tar.bz2")
 }
 
 pub fn untar_snapshot_in<P: AsRef<Path>, Q: AsRef<Path>>(
     snapshot_tar: P,
     unpack_dir: Q,
 ) -> Result<()> {
-    let tar_gz = File::open(snapshot_tar)?;
+    let tar_bz2 = File::open(snapshot_tar)?;
-    let tar = GzDecoder::new(tar_gz);
+    let tar = BzDecoder::new(BufReader::new(tar_bz2));
     let mut archive = Archive::new(tar);
     archive.unpack(&unpack_dir)?;
     Ok(())

multinode-demo/setup.sh

@@ -34,5 +34,5 @@ $solana_genesis "${args[@]}"
 (
   cd "$SOLANA_CONFIG_DIR"/bootstrap-leader
   set -x
-  tar zcvfS genesis.tgz genesis.bin rocksdb
+  tar jcvfS genesis.tar.bz2 genesis.bin rocksdb
 )

multinode-demo/validator.sh

@@ -41,10 +41,6 @@ label=
 identity_keypair_path=
 voting_keypair_path=
 no_restart=0
-# TODO: Enable boot_from_snapshot when snapshots work
-#boot_from_snapshot=1
-boot_from_snapshot=0
-reset_ledger=0
 gossip_entrypoint=
 ledger_dir=
 
@@ -57,9 +53,6 @@ while [[ -n $1 ]]; do
     elif [[ $1 = --no-restart ]]; then
       no_restart=1
       shift
-    elif [[ $1 = --no-snapshot ]]; then
-      boot_from_snapshot=0
-      shift
     elif [[ $1 = --poll-for-new-genesis-block ]]; then
       poll_for_new_genesis_block=1
       shift
@@ -91,6 +84,9 @@ while [[ -n $1 ]]; do
     elif [[ $1 = --node-lamports ]]; then
       node_lamports="$2"
       shift 2
+    elif [[ $1 = --no-snapshot-fetch ]]; then
+      args+=("$1")
+      shift
     elif [[ $1 = --no-voting ]]; then
       args+=("$1")
       shift
@@ -115,9 +111,6 @@ while [[ -n $1 ]]; do
     elif [[ $1 = --no-airdrop ]]; then
       airdrops_enabled=0
       shift
-    elif [[ $1 = --reset-ledger ]]; then
-      reset_ledger=1
-      shift
     elif [[ $1 = --ledger ]]; then
       ledger_dir=$2
       shift 2
@@ -187,14 +180,10 @@ if ((airdrops_enabled)); then
   default_arg --rpc-drone-address "$drone_address"
 fi
 
-accounts_dir="$ledger_dir"/accounts
-snapshot_dir="$ledger_dir"/snapshots
-
 default_arg --identity "$identity_keypair_path"
 default_arg --voting-keypair "$voting_keypair_path"
 default_arg --storage-keypair "$storage_keypair_path"
 default_arg --ledger "$ledger_dir"
-default_arg --accounts "$accounts_dir"
 #default_arg --snapshot-interval-slots 100
 
 if [[ -n $SOLANA_CUDA ]]; then
@@ -213,14 +202,14 @@ new_genesis_block() {
     return
   fi
 
-  rm -f "$ledger_dir"/new-genesis.tgz
+  rm -f "$ledger_dir"/new-genesis.tar.bz2
   (
     set -x
-    curl -f "$rpc_url"/genesis.tgz -o "$ledger_dir"/new-genesis.tgz
+    curl -f "$rpc_url"/genesis.tar.bz2 -o "$ledger_dir"/new-genesis.tar.bz2
   ) || {
     echo "Error: failed to fetch new genesis ledger"
   }
-  ! diff -q "$ledger_dir"/new-genesis.tgz "$ledger_dir"/genesis.tgz >/dev/null 2>&1
+  ! diff -q "$ledger_dir"/new-genesis.tar.bz2 "$ledger_dir"/genesis.tar.bz2 >/dev/null 2>&1
 }
 
 set -e
@@ -245,15 +234,6 @@ kill_node_and_exit() {
 }
 trap 'kill_node_and_exit' INT TERM ERR
 
-if ((reset_ledger)); then
-  echo "Resetting ledger..."
-  (
-    set -x
-    rm -rf "$ledger_dir"
-  )
-fi
-
-
 wallet() {
   (
     set -x
@@ -301,48 +281,6 @@ while true; do
     )
   fi
 
-  if [[ ! -f "$ledger_dir"/.genesis ]]; then
-      echo "Fetching ledger from $rpc_url/genesis.tgz..."
-      SECONDS=
-      mkdir -p "$ledger_dir"
-      while ! curl -f "$rpc_url"/genesis.tgz -o "$ledger_dir"/genesis.tgz; do
-        echo "Genesis ledger fetch failed"
-        sleep 5
-      done
-      echo "Fetched genesis ledger in $SECONDS seconds"
-
-      (
-        set -x
-        cd "$ledger_dir"
-        tar -zxf genesis.tgz
-        touch .genesis
-      )
-
-      (
-        if ((boot_from_snapshot)); then
-          SECONDS=
-
-          echo "Fetching state snapshot $rpc_url/snapshot.tgz..."
-          mkdir -p "$snapshot_dir"
-          if ! curl -f "$rpc_url"/snapshot.tgz -o "$snapshot_dir"/snapshot.tgz; then
-            echo "State snapshot fetch failed"
-            rm -f "$snapshot_dir"/snapshot.tgz
-            exit 0  # None fatal
-          fi
-          echo "Fetched snapshot in $SECONDS seconds"
-
-          SECONDS=
-          (
-            set -x
-            cd "$snapshot_dir"
-            tar -zxf snapshot.tgz
-            rm snapshot.tgz
-          )
-          echo "Extracted snapshot in $SECONDS seconds"
-        fi
-      )
-  fi
-
   [[ -r "$identity_keypair_path" ]] || $solana_keygen new -o "$identity_keypair_path"
   [[ -r "$voting_keypair_path" ]] || $solana_keygen new -o "$voting_keypair_path"
   [[ -r "$storage_keypair_path" ]] || $solana_keygen new -o "$storage_keypair_path"
@@ -360,8 +298,6 @@ identity pubkey: $identity_pubkey
 vote pubkey: $vote_pubkey
 storage pubkey: $storage_pubkey
 ledger: $ledger_dir
-accounts: $accounts_dir
-snapshots: $snapshot_dir
 ========================================================================
 EOF
 

net/net.sh

@@ -66,7 +66,7 @@ Operate a configured testnet
                                       - Amount to fund internal nodes in genesis block.
    --external-accounts-file FILE_PATH
                                       - A YML file with a list of account pubkeys and corresponding lamport balances in genesis block for external nodes
-   --no-snapshot
+   --no-snapshot-fetch
                                       - If set, disables booting validators from a snapshot
    --skip-ledger-verify
                                       - If set, validators will skip verifying
@@ -139,7 +139,7 @@ while [[ -n $1 ]]; do
     elif [[ $1 = --lamports ]]; then
       genesisOptions="$genesisOptions $1 $2"
       shift 2
-    elif [[ $1 = --no-snapshot ]]; then
+    elif [[ $1 = --no-snapshot-fetch ]]; then
       maybeNoSnapshot="$1"
       shift 1
     elif [[ $1 = --no-deploy ]]; then

validator/Cargo.toml

@@ -9,8 +9,10 @@ license = "Apache-2.0"
 homepage = "https://solana.com/"
 
 [dependencies]
+bzip2 = "0.3.3"
 clap = "2.33.0"
 log = "0.4.8"
+reqwest = "0.9.19"
 serde_json = "1.0.40"
 solana = { path = "../core", version = "0.18.0-pre1" }
 solana-drone = { path = "../drone", version = "0.18.0-pre1" }
@@ -21,6 +23,8 @@ solana-runtime = { path = "../runtime", version = "0.18.0-pre1" }
 solana-sdk = { path = "../sdk", version = "0.18.0-pre1" }
 solana-vote-api = { path = "../programs/vote_api", version = "0.18.0-pre1" }
 solana-vote-signer = { path = "../vote-signer", version = "0.18.0-pre1" }
+tempfile = "3.1.0"
+tar = "0.4.26"
 
 [features]
 cuda = ["solana/cuda"]

validator/src/main.rs

@@ -1,8 +1,10 @@
+use bzip2::bufread::BzDecoder;
 use clap::{crate_description, crate_name, crate_version, value_t, App, Arg};
 use log::*;
 use solana::bank_forks::SnapshotConfig;
 use solana::cluster_info::{Node, FULLNODE_PORT_RANGE};
 use solana::contact_info::ContactInfo;
+use solana::gossip_service::discover;
 use solana::ledger_cleanup_service::DEFAULT_MAX_LEDGER_SLOTS;
 use solana::local_vote_signer_service::LocalVoteSignerService;
 use solana::service::Service;
@@ -11,12 +13,12 @@ use solana::validator::{Validator, ValidatorConfig};
 use solana_netutil::parse_port_range;
 use solana_sdk::signature::{read_keypair, Keypair, KeypairUtil};
 use solana_sdk::timing::Slot;
-use std::fs;
+use std::fs::{self, File};
-use std::fs::File;
 use std::net::SocketAddr;
-use std::path::PathBuf;
+use std::path::{Path, PathBuf};
 use std::process::exit;
 use std::sync::Arc;
+use std::time::Instant;
 
 fn port_range_validator(port_range: String) -> Result<(), String> {
     if parse_port_range(&port_range).is_some() {
@@ -26,6 +28,100 @@ fn port_range_validator(port_range: String) -> Result<(), String> {
     }
 }
 
+fn download_archive(
+    rpc_addr: &SocketAddr,
+    archive_name: &str,
+    download_path: &Path,
+    extract: bool,
+) -> Result<(), String> {
+    let archive_path = download_path.join(archive_name);
+    if archive_path.is_file() {
+        return Ok(());
+    }
+    let temp_archive_path = {
+        let mut p = archive_path.clone();
+        p.set_extension(".tmp");
+        p
+    };
+
+    let url = format!("http://{}/{}", rpc_addr, archive_name);
+    println!("Downloading {}...", url);
+    let download_start = Instant::now();
+
+    let mut response = reqwest::get(&url).map_err(|err| format!("Unable to get: {:?}", err))?;
+    let mut file = File::create(&temp_archive_path)
+        .map_err(|err| format!("Unable to create {:?}: {:?}", temp_archive_path, err))?;
+    std::io::copy(&mut response, &mut file)
+        .map_err(|err| format!("Unable to write {:?}: {:?}", temp_archive_path, err))?;
+
+    println!(
+        "Downloaded {} in {:?}",
+        archive_name,
+        Instant::now().duration_since(download_start)
+    );
+
+    if extract {
+        println!("Extracting {:?}...", archive_path);
+        let extract_start = Instant::now();
+        let tar_bz2 = File::open(&temp_archive_path)
+            .map_err(|err| format!("Unable to open {}: {:?}", archive_name, err))?;
+        let tar = BzDecoder::new(std::io::BufReader::new(tar_bz2));
+        let mut archive = tar::Archive::new(tar);
+        archive
+            .unpack(download_path)
+            .map_err(|err| format!("Unable to unpack {}: {:?}", archive_name, err))?;
+        println!(
+            "Extracted {} in {:?}",
+            archive_name,
+            Instant::now().duration_since(extract_start)
+        );
+    }
+    std::fs::rename(temp_archive_path, archive_path)
+        .map_err(|err| format!("Unable to rename: {:?}", err))?;
+
+    Ok(())
+}
+
+fn initialize_ledger_path(
+    entrypoint: &ContactInfo,
+    gossip_addr: &SocketAddr,
+    ledger_path: &Path,
+    no_snapshot_fetch: bool,
+) -> Result<(), String> {
+    let (nodes, _replicators) = discover(
+        &entrypoint.gossip,
+        Some(1),
+        Some(60),
+        None,
+        Some(&gossip_addr),
+    )
+    .map_err(|err| err.to_string())?;
+
+    let rpc_addr = nodes
+        .iter()
+        .filter_map(ContactInfo::valid_client_facing_addr)
+        .map(|addrs| addrs.0)
+        .find(|rpc_addr| rpc_addr.ip() == entrypoint.gossip.ip())
+        .unwrap_or_else(|| {
+            eprintln!(
+                "Entrypoint ({:?}) is not running the RPC service",
+                entrypoint.gossip.ip()
+            );
+            exit(1);
+        });
+
+    fs::create_dir_all(ledger_path).map_err(|err| err.to_string())?;
+
+    download_archive(&rpc_addr, "genesis.tar.bz2", ledger_path, true)?;
+    if !no_snapshot_fetch {
+        let _ = fs::remove_file(ledger_path.join("snapshot.tar.bz2"));
+        download_archive(&rpc_addr, "snapshot.tar.bz2", ledger_path, false)
+            .unwrap_or_else(|err| eprintln!("Warning: Unable to fetch snapshot: {:?}", err));
+    }
+
+    Ok(())
+}
+
 fn main() {
     solana_logger::setup_with_filter("solana=info");
     solana_metrics::set_panic_hook("validator");
@@ -40,7 +136,7 @@ fn main() {
                 .long("blockstream")
                 .takes_value(true)
                 .value_name("UNIX DOMAIN SOCKET")
-                .help("Open blockstream at this unix domain socket path")
+                .help("Stream entries to this unix domain socket path")
         )
         .arg(
             Arg::with_name("identity")
@@ -95,6 +191,13 @@ fn main() {
                 .takes_value(true)
                 .help("Rendezvous with the cluster at this entry point"),
         )
+        .arg(
+            Arg::with_name("no_snapshot_fetch")
+                .long("no-snapshot-fetch")
+                .takes_value(false)
+                .requires("entrypoint")
+                .help("Do not attempt to fetch a snapshot from the cluster entrypoint"),
+        )
         .arg(
             Arg::with_name("no_voting")
                 .long("no-voting")
@@ -142,7 +245,7 @@ fn main() {
                 .help("Rendezvous with the vote signer at this RPC end point"),
         )
         .arg(
-            Arg::with_name("accounts")
+            Arg::with_name("account_paths")
                 .long("accounts")
                 .value_name("PATHS")
                 .takes_value(true)
@@ -243,21 +346,19 @@ fn main() {
         ),
     );
 
-    if let Some(paths) = matches.value_of("accounts") {
+    if let Some(account_paths) = matches.value_of("account_paths") {
-        validator_config.account_paths = Some(paths.to_string());
+        validator_config.account_paths = Some(account_paths.to_string());
+    } else {
+        validator_config.account_paths =
+            Some(ledger_path.join("accounts").to_str().unwrap().to_string());
     }
 
     validator_config.snapshot_config = matches.value_of("snapshot_interval_slots").map(|s| {
         let snapshots_dir = ledger_path.clone().join("snapshot");
-        let snapshots_bank_state_dir = snapshots_dir.join("bank_states");
-        let snapshots_tar_dir = snapshots_dir.join("tar");
         fs::create_dir_all(&snapshots_dir).expect("Failed to create snapshots directory");
-        fs::create_dir_all(&snapshots_bank_state_dir)
-            .expect("Failed to create snapshots bank state directory");
-        fs::create_dir_all(&snapshots_tar_dir).expect("Failed  to create snapshots tar directory");
         SnapshotConfig::new(
-            snapshots_bank_state_dir,
+            snapshots_dir,
-            snapshots_tar_dir,
+            ledger_path.clone(),
             s.parse::<usize>().unwrap(),
         )
     });
@@ -291,7 +392,28 @@ fn main() {
         .value_of("blockstream_unix_socket")
         .map(PathBuf::from);
 
-    let keypair = Arc::new(keypair);
+    if let Some(ref entrypoint_addr) = cluster_entrypoint {
+        initialize_ledger_path(
+            entrypoint_addr,
+            &gossip_addr,
+            &ledger_path,
+            !matches.is_present("no_snapshot_fetch"),
+        )
+        .unwrap_or_else(|err| {
+            eprintln!("Failed to download ledger: {}", err);
+            exit(1);
+        });
+    } else {
+        // Without a cluster entrypoint, ledger_path must already be present
+        if !ledger_path.is_dir() {
+            eprintln!(
+                "Error: ledger directory does not exist or is not accessible: {:?}",
+                ledger_path
+            );
+            exit(1);
+        }
+    }
+
     let mut node = Node::new_with_external_ip(&keypair.pubkey(), &gossip_addr, dynamic_port_range);
     if let Some(port) = matches.value_of("rpc_port") {
         let port_number = port.to_string().parse().expect("integer");
@@ -307,7 +429,7 @@ fn main() {
 
     let validator = Validator::new(
         node,
-        &keypair,
+        &Arc::new(keypair),
         &ledger_path,
         &vote_account,
         &Arc::new(voting_keypair),