solana/core/src/replicator.rs

797 lines
29 KiB
Rust
Raw Normal View History

2018-12-07 19:16:27 -08:00
use crate::blob_fetch_stage::BlobFetchStage;
2019-02-07 20:52:39 -08:00
use crate::blocktree::Blocktree;
use crate::chacha::{chacha_cbc_encrypt_ledger, CHACHA_BLOCK_SIZE};
use crate::cluster_info::{ClusterInfo, Node, FULLNODE_PORT_RANGE};
2019-03-08 17:23:07 -08:00
use crate::contact_info::ContactInfo;
2018-12-07 19:16:27 -08:00
use crate::gossip_service::GossipService;
use crate::packet::to_shared_blob;
use crate::repair_service::{RepairService, RepairSlotRange, RepairStrategy};
use crate::result::{Error, Result};
2018-12-07 19:16:27 -08:00
use crate::service::Service;
use crate::streamer::{blob_receiver, receiver, responder};
use crate::window_service::WindowService;
use crate::{repair_service, window_service};
use bincode::deserialize;
use rand::thread_rng;
use rand::Rng;
use solana_client::rpc_client::RpcClient;
use solana_client::rpc_request::RpcRequest;
use solana_client::thin_client::ThinClient;
2019-05-10 14:26:46 -07:00
use solana_ed25519_dalek as ed25519_dalek;
use solana_netutil::bind_in_range;
use solana_sdk::account_utils::State;
use solana_sdk::client::{AsyncClient, SyncClient};
2018-11-16 08:04:46 -08:00
use solana_sdk::hash::{Hash, Hasher};
use solana_sdk::message::Message;
use solana_sdk::signature::{Keypair, KeypairUtil, Signature};
use solana_sdk::timing::timestamp;
2019-03-22 21:02:00 -07:00
use solana_sdk::transaction::Transaction;
use solana_sdk::transport::TransportError;
use solana_storage_api::storage_contract::StorageContract;
use solana_storage_api::{get_segment_from_slot, storage_instruction, SLOTS_PER_SEGMENT};
use std::fs::File;
use std::io::{self, BufReader, ErrorKind, Read, Seek, SeekFrom};
use std::mem::size_of;
use std::net::{SocketAddr, UdpSocket};
2019-05-20 09:58:27 -07:00
use std::path::{Path, PathBuf};
use std::result;
use std::sync::atomic::{AtomicBool, Ordering};
use std::sync::mpsc::channel;
use std::sync::{Arc, RwLock};
2019-05-20 09:58:27 -07:00
use std::thread::{sleep, spawn, JoinHandle};
use std::time::Duration;
#[derive(Serialize, Deserialize)]
pub enum ReplicatorRequest {
GetSlotHeight(SocketAddr),
}
pub struct Replicator {
gossip_service: GossipService,
fetch_stage: BlobFetchStage,
window_service: WindowService,
thread_handles: Vec<JoinHandle<()>>,
exit: Arc<AtomicBool>,
slot: u64,
ledger_path: String,
keypair: Arc<Keypair>,
storage_keypair: Arc<Keypair>,
blockhash: Hash,
signature: ed25519_dalek::Signature,
cluster_info: Arc<RwLock<ClusterInfo>>,
ledger_data_file_encrypted: PathBuf,
sampling_offsets: Vec<u64>,
sha_state: Hash,
num_chacha_blocks: usize,
blocktree: Arc<Blocktree>,
}
2019-05-20 09:58:27 -07:00
pub(crate) fn sample_file(in_path: &Path, sample_offsets: &[u64]) -> io::Result<Hash> {
let in_file = File::open(in_path)?;
let metadata = in_file.metadata()?;
let mut buffer_file = BufReader::new(in_file);
let mut hasher = Hasher::default();
let sample_size = size_of::<Hash>();
let sample_size64 = sample_size as u64;
let mut buf = vec![0; sample_size];
let file_len = metadata.len();
if file_len < sample_size64 {
return Err(io::Error::new(ErrorKind::Other, "file too short!"));
}
for offset in sample_offsets {
if *offset > (file_len - sample_size64) / sample_size64 {
return Err(io::Error::new(ErrorKind::Other, "offset too large"));
}
buffer_file.seek(SeekFrom::Start(*offset * sample_size64))?;
trace!("sampling @ {} ", *offset);
match buffer_file.read(&mut buf) {
Ok(size) => {
assert_eq!(size, buf.len());
hasher.hash(&buf);
}
Err(e) => {
warn!("Error sampling file");
return Err(e);
}
}
}
Ok(hasher.result())
}
fn get_slot_from_blockhash(signature: &ed25519_dalek::Signature, storage_slot: u64) -> u64 {
let signature_vec = signature.to_bytes();
let mut segment_index = u64::from(signature_vec[0])
| (u64::from(signature_vec[1]) << 8)
| (u64::from(signature_vec[1]) << 16)
| (u64::from(signature_vec[2]) << 24);
let max_segment_index = get_segment_from_slot(storage_slot);
segment_index %= max_segment_index as u64;
segment_index * SLOTS_PER_SEGMENT
}
fn create_request_processor(
socket: UdpSocket,
exit: &Arc<AtomicBool>,
slot: u64,
) -> Vec<JoinHandle<()>> {
let mut thread_handles = vec![];
let (s_reader, r_reader) = channel();
let (s_responder, r_responder) = channel();
let storage_socket = Arc::new(socket);
let t_receiver = receiver(storage_socket.clone(), exit, s_reader);
thread_handles.push(t_receiver);
let t_responder = responder("replicator-responder", storage_socket.clone(), r_responder);
thread_handles.push(t_responder);
let exit = exit.clone();
let t_processor = spawn(move || loop {
let packets = r_reader.recv_timeout(Duration::from_secs(1));
if let Ok(packets) = packets {
for packet in &packets.packets {
let req: result::Result<ReplicatorRequest, Box<bincode::ErrorKind>> =
deserialize(&packet.data[..packet.meta.size]);
match req {
Ok(ReplicatorRequest::GetSlotHeight(from)) => {
if let Ok(blob) = to_shared_blob(slot, from) {
let _ = s_responder.send(vec![blob]);
}
}
Err(e) => {
info!("invalid request: {:?}", e);
}
}
}
}
if exit.load(Ordering::Relaxed) {
break;
}
});
thread_handles.push(t_processor);
thread_handles
}
impl Replicator {
/// Returns a Result that contains a replicator on success
///
/// # Arguments
2019-01-31 12:44:09 -08:00
/// * `ledger_path` - path to where the ledger will be stored.
/// Causes panic if none
/// * `node` - The replicator node
/// * `cluster_entrypoint` - ContactInfo representing an entry into the network
/// * `keypair` - Keypair for this replicator
#[allow(clippy::new_ret_no_self)]
pub fn new(
2019-01-31 12:44:09 -08:00
ledger_path: &str,
node: Node,
cluster_entrypoint: ContactInfo,
keypair: Arc<Keypair>,
storage_keypair: Arc<Keypair>,
) -> Result<Self> {
let exit = Arc::new(AtomicBool::new(false));
info!("Replicator: id: {}", keypair.pubkey());
info!("Creating cluster info....");
let mut cluster_info = ClusterInfo::new(node.info.clone(), keypair.clone());
cluster_info.set_entrypoint(cluster_entrypoint.clone());
let cluster_info = Arc::new(RwLock::new(cluster_info));
// Note for now, this ledger will not contain any of the existing entries
// in the ledger located at ledger_path, and will only append on newly received
// entries after being passed to window_service
let blocktree = Arc::new(
Blocktree::open(ledger_path).expect("Expected to be able to open database ledger"),
);
let gossip_service = GossipService::new(
&cluster_info,
2019-02-07 20:52:39 -08:00
Some(blocktree.clone()),
None,
node.sockets.gossip,
&exit,
);
info!("Connecting to the cluster via {:?}", cluster_entrypoint);
let (nodes, _) =
match crate::gossip_service::discover_cluster(&cluster_entrypoint.gossip, 1) {
Ok(nodes_and_replicators) => nodes_and_replicators,
Err(e) => {
//shutdown services before exiting
exit.store(true, Ordering::Relaxed);
gossip_service.join()?;
return Err(Error::from(e));
}
};
let client = crate::gossip_service::get_client(&nodes);
let (storage_blockhash, storage_slot) =
match Self::poll_for_blockhash_and_slot(&cluster_info, &Hash::default()) {
Ok(blockhash_and_slot) => blockhash_and_slot,
Err(e) => {
//shutdown services before exiting
exit.store(true, Ordering::Relaxed);
gossip_service.join()?;
return Err(e);
}
};
let signature = storage_keypair.sign(storage_blockhash.as_ref());
let slot = get_slot_from_blockhash(&signature, storage_slot);
info!("replicating slot: {}", slot);
let mut repair_slot_range = RepairSlotRange::default();
repair_slot_range.end = slot + SLOTS_PER_SEGMENT;
repair_slot_range.start = slot;
let repair_socket = Arc::new(node.sockets.repair);
let mut blob_sockets: Vec<Arc<UdpSocket>> =
node.sockets.tvu.into_iter().map(Arc::new).collect();
blob_sockets.push(repair_socket.clone());
let (blob_fetch_sender, blob_fetch_receiver) = channel();
2019-03-04 20:50:02 -08:00
let fetch_stage = BlobFetchStage::new_multi_socket(blob_sockets, &blob_fetch_sender, &exit);
let (retransmit_sender, retransmit_receiver) = channel();
let window_service = WindowService::new(
2019-02-07 20:52:39 -08:00
blocktree.clone(),
2018-10-08 19:55:54 -07:00
cluster_info.clone(),
blob_fetch_receiver,
retransmit_sender,
repair_socket,
2019-03-04 20:50:02 -08:00
&exit,
RepairStrategy::RepairRange(repair_slot_range),
|_, _, _| true,
);
if let Err(e) = Self::setup_mining_account(&client, &keypair, &storage_keypair) {
//shutdown services before exiting
exit.store(true, Ordering::Relaxed);
gossip_service.join()?;
window_service.join()?;
fetch_stage.join()?;
return Err(e);
};
let mut thread_handles =
create_request_processor(node.sockets.storage.unwrap(), &exit, slot);
// receive blobs from retransmit and drop them.
let t_retransmit = {
let exit = exit.clone();
spawn(move || loop {
let _ = retransmit_receiver.recv_timeout(Duration::from_secs(1));
if exit.load(Ordering::Relaxed) {
break;
}
})
};
thread_handles.push(t_retransmit);
let t_replicate = {
let exit = exit.clone();
let blocktree = blocktree.clone();
let cluster_info = cluster_info.clone();
2019-05-20 09:58:27 -07:00
let node_info = node.info.clone();
spawn(move || {
Self::wait_for_ledger_download(slot, &blocktree, &exit, &node_info, cluster_info)
})
};
//always push this last
thread_handles.push(t_replicate);
Ok(Self {
gossip_service,
fetch_stage,
window_service,
thread_handles,
exit,
slot,
ledger_path: ledger_path.to_string(),
keypair,
storage_keypair,
blockhash: storage_blockhash,
signature,
cluster_info,
ledger_data_file_encrypted: PathBuf::default(),
sampling_offsets: vec![],
sha_state: Hash::default(),
num_chacha_blocks: 0,
blocktree,
})
}
2019-06-22 17:18:35 -07:00
pub fn run(&mut self) {
info!("waiting for ledger download");
self.thread_handles.pop().unwrap().join().unwrap();
self.encrypt_ledger()
.expect("ledger encrypt not successful");
loop {
self.create_sampling_offsets();
if let Err(err) = self.sample_file_to_create_mining_hash() {
warn!("Error sampling file, exiting: {:?}", err);
break;
}
self.submit_mining_proof();
// Todo make this a lot more frequent by picking a "new" blockhash instead of picking a storage blockhash
// prep the next proof
let (storage_blockhash, _) =
match Self::poll_for_blockhash_and_slot(&self.cluster_info, &self.blockhash) {
Ok(blockhash_and_slot) => blockhash_and_slot,
Err(e) => {
warn!(
"Error couldn't get a newer blockhash than {:?}. {:?}",
self.blockhash, e
);
break;
}
};
self.blockhash = storage_blockhash;
2019-06-22 17:18:35 -07:00
self.redeem_rewards();
}
}
2019-06-22 17:18:35 -07:00
fn redeem_rewards(&self) {
let nodes = self.cluster_info.read().unwrap().tvu_peers();
let client = crate::gossip_service::get_client(&nodes);
if let Ok(Some(account)) = client.get_account(&self.storage_keypair.pubkey()) {
if let Ok(StorageContract::ReplicatorStorage { validations, .. }) = account.state() {
if !validations.is_empty() {
let ix = storage_instruction::claim_reward(
&self.keypair.pubkey(),
&self.storage_keypair.pubkey(),
);
let message = Message::new_with_payer(vec![ix], Some(&self.keypair.pubkey()));
if let Err(e) = client.send_message(&[&self.keypair], message) {
error!("unable to redeem reward, tx failed: {:?}", e);
} else {
info!(
"collected mining rewards: Account balance {:?}",
client.get_balance(&self.keypair.pubkey())
);
}
}
}
} else {
info!("Redeem mining reward: No account data found");
}
}
fn wait_for_ledger_download(
start_slot: u64,
blocktree: &Arc<Blocktree>,
exit: &Arc<AtomicBool>,
node_info: &ContactInfo,
cluster_info: Arc<RwLock<ClusterInfo>>,
) {
info!(
"window created, waiting for ledger download starting at slot {:?}",
start_slot
);
let mut current_slot = start_slot;
'outer: loop {
while blocktree.is_full(current_slot) {
current_slot += 1;
info!("current slot: {}", current_slot);
if current_slot >= start_slot + SLOTS_PER_SEGMENT {
break 'outer;
}
}
if exit.load(Ordering::Relaxed) {
break;
}
sleep(Duration::from_secs(1));
}
info!("Done receiving entries from window_service");
// Remove replicator from the data plane
let mut contact_info = node_info.clone();
2019-03-08 17:23:07 -08:00
contact_info.tvu = "0.0.0.0:0".parse().unwrap();
contact_info.wallclock = timestamp();
{
let mut cluster_info_w = cluster_info.write().unwrap();
2019-03-08 17:23:07 -08:00
cluster_info_w.insert_self(contact_info);
}
}
fn encrypt_ledger(&mut self) -> Result<()> {
let ledger_path = Path::new(&self.ledger_path);
self.ledger_data_file_encrypted = ledger_path.join("ledger.enc");
{
let mut ivec = [0u8; 64];
ivec.copy_from_slice(&self.signature.to_bytes());
let num_encrypted_bytes = chacha_cbc_encrypt_ledger(
&self.blocktree,
self.slot,
&self.ledger_data_file_encrypted,
&mut ivec,
)?;
self.num_chacha_blocks = num_encrypted_bytes / CHACHA_BLOCK_SIZE;
}
info!(
"Done encrypting the ledger: {:?}",
self.ledger_data_file_encrypted
);
Ok(())
}
fn create_sampling_offsets(&mut self) {
self.sampling_offsets.clear();
{
use crate::storage_stage::NUM_STORAGE_SAMPLES;
use rand::SeedableRng;
use rand_chacha::ChaChaRng;
let mut rng_seed = [0u8; 32];
rng_seed.copy_from_slice(&self.blockhash.as_ref());
let mut rng = ChaChaRng::from_seed(rng_seed);
for _ in 0..NUM_STORAGE_SAMPLES {
self.sampling_offsets
.push(rng.gen_range(0, self.num_chacha_blocks) as u64);
}
}
}
fn sample_file_to_create_mining_hash(&mut self) -> Result<()> {
self.sha_state = sample_file(&self.ledger_data_file_encrypted, &self.sampling_offsets)?;
info!("sampled sha_state: {}", self.sha_state);
Ok(())
}
fn setup_mining_account(
client: &ThinClient,
keypair: &Keypair,
storage_keypair: &Keypair,
) -> Result<()> {
// make sure replicator has some balance
if client.poll_get_balance(&keypair.pubkey())? == 0 {
Err(io::Error::new(
io::ErrorKind::Other,
"keypair account has no balance",
))?
}
// check if the storage account exists
let balance = client.poll_get_balance(&storage_keypair.pubkey());
if balance.is_err() || balance.unwrap() == 0 {
let (blockhash, _fee_calculator) = client.get_recent_blockhash().expect("blockhash");
let ix = storage_instruction::create_replicator_storage_account(
&keypair.pubkey(),
&keypair.pubkey(),
&storage_keypair.pubkey(),
1,
);
let tx = Transaction::new_signed_instructions(&[keypair], ix, blockhash);
2019-04-05 20:59:27 -07:00
let signature = client.async_send_transaction(tx)?;
client
.poll_for_signature(&signature)
.map_err(|err| match err {
TransportError::IoError(e) => e,
TransportError::TransactionError(_) => io::Error::new(
ErrorKind::Other,
"setup_mining_account: signature not found",
),
})?;
}
Ok(())
}
fn submit_mining_proof(&self) {
// No point if we've got no storage account...
let nodes = self.cluster_info.read().unwrap().tvu_peers();
let client = crate::gossip_service::get_client(&nodes);
assert!(
client
.poll_get_balance(&self.storage_keypair.pubkey())
.unwrap()
> 0
);
// ...or no lamports for fees
let balance = client.poll_get_balance(&self.keypair.pubkey()).unwrap();
if balance == 0 {
error!("Unable to submit mining proof, insufficient Replicator Account balance");
return;
}
let (blockhash, _) = client.get_recent_blockhash().expect("No recent blockhash");
let instruction = storage_instruction::mining_proof(
&self.storage_keypair.pubkey(),
self.sha_state,
get_segment_from_slot(self.slot),
Signature::new(&self.signature.to_bytes()),
self.blockhash,
);
let message = Message::new_with_payer(vec![instruction], Some(&self.keypair.pubkey()));
let mut transaction = Transaction::new(
&[self.keypair.as_ref(), self.storage_keypair.as_ref()],
message,
blockhash,
);
client
.send_and_confirm_transaction(
&[&self.keypair, &self.storage_keypair],
&mut transaction,
10,
0,
)
.expect("transfer didn't work!");
}
pub fn close(self) {
self.exit.store(true, Ordering::Relaxed);
self.join()
}
pub fn join(self) {
self.gossip_service.join().unwrap();
self.fetch_stage.join().unwrap();
self.window_service.join().unwrap();
for handle in self.thread_handles {
handle.join().unwrap();
}
}
/// Poll for a different blockhash and associated max_slot than `previous_blockhash`
fn poll_for_blockhash_and_slot(
cluster_info: &Arc<RwLock<ClusterInfo>>,
previous_blockhash: &Hash,
) -> result::Result<(Hash, u64), Error> {
for _ in 0..10 {
2019-06-17 15:44:19 -07:00
let rpc_peers = {
let cluster_info = cluster_info.read().unwrap();
2019-06-17 15:44:19 -07:00
cluster_info.rpc_peers()
};
2019-06-17 15:44:19 -07:00
debug!("rpc peers: {:?}", rpc_peers);
if !rpc_peers.is_empty() {
let rpc_client = {
let node_index = thread_rng().gen_range(0, rpc_peers.len());
RpcClient::new_socket(rpc_peers[node_index].rpc)
};
let response = rpc_client
.retry_make_rpc_request(&RpcRequest::GetStorageBlockhash, None, 0)
.map_err(|err| {
warn!("Error while making rpc request {:?}", err);
Error::IO(io::Error::new(ErrorKind::Other, "rpc error"))
})?;
let storage_blockhash =
serde_json::from_value::<(String)>(response).map_err(|err| {
io::Error::new(
io::ErrorKind::Other,
format!("Couldn't parse response: {:?}", err),
)
})?;
let storage_blockhash = storage_blockhash.parse().map_err(|err| {
io::Error::new(
io::ErrorKind::Other,
2019-06-17 15:44:19 -07:00
format!(
"Blockhash parse failure: {:?} on {:?}",
err, storage_blockhash
),
)
})?;
2019-06-17 15:44:19 -07:00
if storage_blockhash != *previous_blockhash {
let storage_slot = rpc_client
.retry_make_rpc_request(&RpcRequest::GetStorageSlot, None, 0)
.map_err(|err| {
warn!("Error while making rpc request {:?}", err);
Error::IO(io::Error::new(ErrorKind::Other, "rpc error"))
})?
.as_u64()
.unwrap();
info!("storage slot: {}", storage_slot);
if get_segment_from_slot(storage_slot) != 0 {
return Ok((storage_blockhash, storage_slot));
}
}
}
info!("waiting for segment...");
sleep(Duration::from_secs(5));
}
Err(io::Error::new(
ErrorKind::Other,
"Couldn't get blockhash or slot",
))?
}
/// Ask a replicator to populate a given blocktree with its segment.
/// Return the slot at the start of the replicator's segment
///
/// It is recommended to use a temporary blocktree for this since the download will not verify
/// blobs received and might impact the chaining of blobs across slots
pub fn download_from_replicator(
cluster_info: &Arc<RwLock<ClusterInfo>>,
replicator_info: &ContactInfo,
blocktree: &Arc<Blocktree>,
) -> Result<(u64)> {
// Create a client which downloads from the replicator and see that it
// can respond with blobs.
let start_slot = Self::get_replicator_segment_slot(replicator_info.storage_addr);
info!("Replicator download: start at {}", start_slot);
let exit = Arc::new(AtomicBool::new(false));
let (s_reader, r_reader) = channel();
let repair_socket = Arc::new(bind_in_range(FULLNODE_PORT_RANGE).unwrap().1);
let t_receiver = blob_receiver(repair_socket.clone(), &exit, s_reader);
let id = cluster_info.read().unwrap().id();
info!(
"Sending repair requests from: {} to: {}",
cluster_info.read().unwrap().my_data().id,
replicator_info.gossip
);
let repair_slot_range = RepairSlotRange {
start: start_slot,
end: start_slot + SLOTS_PER_SEGMENT,
};
// try for upto 180 seconds //TODO needs tuning if segments are huge
for _ in 0..120 {
// Strategy used by replicators
let repairs = RepairService::generate_repairs_in_range(
blocktree,
repair_service::MAX_REPAIR_LENGTH,
&repair_slot_range,
);
//iter over the repairs and send them
if let Ok(repairs) = repairs {
let reqs: Vec<_> = repairs
.into_iter()
.filter_map(|repair_request| {
cluster_info
.read()
.unwrap()
.map_repair_request(&repair_request)
.map(|result| ((replicator_info.gossip, result), repair_request))
.ok()
})
.collect();
for ((to, req), repair_request) in reqs {
if let Ok(local_addr) = repair_socket.local_addr() {
datapoint_info!(
"replicator_download",
("repair_request", format!("{:?}", repair_request), String),
("to", to.to_string(), String),
("from", local_addr.to_string(), String),
("id", id.to_string(), String)
);
}
repair_socket
.send_to(&req, replicator_info.gossip)
.unwrap_or_else(|e| {
error!("{} repair req send_to({}) error {:?}", id, to, e);
0
});
}
}
let res = r_reader.recv_timeout(Duration::new(1, 0));
if let Ok(blobs) = res {
window_service::process_blobs(&blobs, blocktree)?;
}
// check if all the slots in the segment are complete
if Self::segment_complete(start_slot, blocktree) {
break;
}
sleep(Duration::from_millis(500));
}
exit.store(true, Ordering::Relaxed);
t_receiver.join().unwrap();
// check if all the slots in the segment are complete
if !Self::segment_complete(start_slot, blocktree) {
Err(io::Error::new(
ErrorKind::Other,
"Unable to download the full segment",
))?
}
Ok(start_slot)
}
fn segment_complete(start_slot: u64, blocktree: &Arc<Blocktree>) -> bool {
for slot in start_slot..(start_slot + SLOTS_PER_SEGMENT) {
if !blocktree.is_full(slot) {
return false;
}
}
true
}
fn get_replicator_segment_slot(to: SocketAddr) -> u64 {
let (_port, socket) = bind_in_range(FULLNODE_PORT_RANGE).unwrap();
socket
.set_read_timeout(Some(Duration::from_secs(5)))
.unwrap();
let req = ReplicatorRequest::GetSlotHeight(socket.local_addr().unwrap());
let serialized_req = bincode::serialize(&req).unwrap();
for _ in 0..10 {
socket.send_to(&serialized_req, to).unwrap();
let mut buf = [0; 1024];
if let Ok((size, _addr)) = socket.recv_from(&mut buf) {
return deserialize(&buf[..size]).unwrap();
}
sleep(Duration::from_millis(500));
}
panic!("Couldn't get slot height!");
}
}
#[cfg(test)]
mod tests {
2019-05-20 09:58:27 -07:00
use super::*;
use std::fs::{create_dir_all, remove_file};
use std::io::Write;
fn tmp_file_path(name: &str) -> PathBuf {
use std::env;
let out_dir = env::var("OUT_DIR").unwrap_or_else(|_| "target".to_string());
let keypair = Keypair::new();
let mut path = PathBuf::new();
path.push(out_dir);
path.push("tmp");
create_dir_all(&path).unwrap();
path.push(format!("{}-{}", name, keypair.pubkey()));
path
}
#[test]
fn test_sample_file() {
solana_logger::setup();
let in_path = tmp_file_path("test_sample_file_input.txt");
let num_strings = 4096;
let string = "12foobar";
{
let mut in_file = File::create(&in_path).unwrap();
for _ in 0..num_strings {
in_file.write(string.as_bytes()).unwrap();
}
}
let num_samples = (string.len() * num_strings / size_of::<Hash>()) as u64;
let samples: Vec<_> = (0..num_samples).collect();
let res = sample_file(&in_path, samples.as_slice());
let ref_hash: Hash = Hash::new(&[
173, 251, 182, 165, 10, 54, 33, 150, 133, 226, 106, 150, 99, 192, 179, 1, 230, 144,
151, 126, 18, 191, 54, 67, 249, 140, 230, 160, 56, 30, 170, 52,
]);
let res = res.unwrap();
assert_eq!(res, ref_hash);
// Sample just past the end
assert!(sample_file(&in_path, &[num_samples]).is_err());
remove_file(&in_path).unwrap();
}
#[test]
fn test_sample_file_invalid_offset() {
let in_path = tmp_file_path("test_sample_file_invalid_offset_input.txt");
{
let mut in_file = File::create(&in_path).unwrap();
for _ in 0..4096 {
in_file.write("123456foobar".as_bytes()).unwrap();
}
}
let samples = [0, 200000];
let res = sample_file(&in_path, &samples);
assert!(res.is_err());
remove_file(in_path).unwrap();
}
#[test]
fn test_sample_file_missing_file() {
let in_path = tmp_file_path("test_sample_file_that_doesnt_exist.txt");
let samples = [0, 5];
let res = sample_file(&in_path, &samples);
assert!(res.is_err());
}
}