Line data Source code
1 : use std::collections::{HashMap, HashSet, VecDeque};
2 : use std::future::Future;
3 : use std::num::NonZeroUsize;
4 : use std::pin::Pin;
5 : use std::sync::atomic::{AtomicU64, Ordering};
6 : use std::sync::{Arc, Mutex};
7 : use std::time::{Duration, Instant};
8 :
9 : use anyhow::Context;
10 : use async_trait::async_trait;
11 : use bytes::Bytes;
12 : use camino::Utf8PathBuf;
13 : use futures::stream::FuturesUnordered;
14 : use futures::{Stream, StreamExt as _};
15 : use pageserver_api::key::Key;
16 : use pageserver_api::keyspace::KeySpaceAccum;
17 : use pageserver_api::pagestream_api::{PagestreamGetPageRequest, PagestreamRequest};
18 : use pageserver_api::reltag::RelTag;
19 : use pageserver_api::shard::TenantShardId;
20 : use pageserver_client_grpc::{self as client_grpc, ShardSpec};
21 : use pageserver_page_api as page_api;
22 : use rand::prelude::*;
23 : use tokio::task::JoinSet;
24 : use tokio_util::sync::CancellationToken;
25 : use tracing::info;
26 : use url::Url;
27 : use utils::id::TenantTimelineId;
28 : use utils::lsn::Lsn;
29 : use utils::shard::ShardIndex;
30 :
31 : use crate::util::tokio_thread_local_stats::AllThreadLocalStats;
32 : use crate::util::{request_stats, tokio_thread_local_stats};
33 :
34 : /// GetPage@LatestLSN, uniformly distributed across the compute-accessible keyspace.
35 : #[derive(clap::Parser)]
36 : pub(crate) struct Args {
37 : #[clap(long, default_value = "http://localhost:9898")]
38 : mgmt_api_endpoint: String,
39 : /// Pageserver connection string. Supports postgresql:// and grpc:// protocols.
40 : #[clap(long, default_value = "postgres://postgres@localhost:64000")]
41 : page_service_connstring: String,
42 : /// Use the rich gRPC Pageserver client `client_grpc::PageserverClient`, rather than the basic
43 : /// no-frills `page_api::Client`. Only valid with grpc:// connstrings.
44 : #[clap(long)]
45 : rich_client: bool,
46 : #[clap(long)]
47 : pageserver_jwt: Option<String>,
48 : #[clap(long, default_value = "1")]
49 : num_clients: NonZeroUsize,
50 : #[clap(long)]
51 : runtime: Option<humantime::Duration>,
52 : /// If true, enable compression (only for gRPC).
53 : #[clap(long)]
54 : compression: bool,
55 : /// Each client sends requests at the given rate.
56 : ///
57 : /// If a request takes too long and we should be issuing a new request already,
58 : /// we skip that request and account it as `MISSED`.
59 : #[clap(long)]
60 : per_client_rate: Option<usize>,
61 : /// Probability for sending `latest=true` in the request (uniform distribution).
62 : #[clap(long, default_value = "1")]
63 : req_latest_probability: f64,
64 : #[clap(long)]
65 : limit_to_first_n_targets: Option<usize>,
66 : /// For large pageserver installations, enumerating the keyspace takes a lot of time.
67 : /// If specified, the specified path is used to maintain a cache of the keyspace enumeration result.
68 : /// The cache is tagged and auto-invalided by the tenant/timeline ids only.
69 : /// It doesn't get invalidated if the keyspace changes under the hood, e.g., due to new ingested data or compaction.
70 : #[clap(long)]
71 : keyspace_cache: Option<Utf8PathBuf>,
72 : /// Before starting the benchmark, live-reconfigure the pageserver to use the given
73 : /// [`pageserver_api::models::virtual_file::IoEngineKind`].
74 : #[clap(long)]
75 : set_io_engine: Option<pageserver_api::models::virtual_file::IoEngineKind>,
76 :
77 : /// Before starting the benchmark, live-reconfigure the pageserver to use specified io mode (buffered vs. direct).
78 : #[clap(long)]
79 : set_io_mode: Option<pageserver_api::models::virtual_file::IoMode>,
80 :
81 : /// Queue depth generated in each client.
82 : #[clap(long, default_value = "1")]
83 : queue_depth: NonZeroUsize,
84 :
85 : /// Batch size of contiguous pages generated by each client. This is equivalent to how Postgres
86 : /// will request page batches (e.g. prefetches or vectored reads). A batch counts as 1 RPS and
87 : /// 1 queue depth.
88 : ///
89 : /// The libpq protocol does not support client-side batching, and will submit batches as many
90 : /// individual requests, in the hope that the server will batch them. Each batch still counts as
91 : /// 1 RPS and 1 queue depth.
92 : #[clap(long, default_value = "1")]
93 : batch_size: NonZeroUsize,
94 :
95 : #[clap(long)]
96 : only_relnode: Option<u32>,
97 :
98 : targets: Option<Vec<TenantTimelineId>>,
99 : }
100 :
101 : /// State shared by all clients
102 : #[derive(Debug)]
103 : struct SharedState {
104 : start_work_barrier: tokio::sync::Barrier,
105 : live_stats: LiveStats,
106 : }
107 :
108 : #[derive(Debug, Default)]
109 : struct LiveStats {
110 : completed_requests: AtomicU64,
111 : missed: AtomicU64,
112 : }
113 :
114 : impl LiveStats {
115 0 : fn request_done(&self) {
116 0 : self.completed_requests.fetch_add(1, Ordering::Relaxed);
117 0 : }
118 0 : fn missed(&self, n: u64) {
119 0 : self.missed.fetch_add(n, Ordering::Relaxed);
120 0 : }
121 : }
122 :
123 0 : #[derive(Clone, serde::Serialize, serde::Deserialize)]
124 : struct KeyRange {
125 : timeline: TenantTimelineId,
126 : timeline_lsn: Lsn,
127 : start: i128,
128 : end: i128,
129 : }
130 :
131 : impl KeyRange {
132 0 : fn len(&self) -> i128 {
133 0 : self.end - self.start
134 0 : }
135 : }
136 :
137 : #[derive(PartialEq, Eq, Hash, Copy, Clone)]
138 : struct WorkerId {
139 : timeline: TenantTimelineId,
140 : num_client: usize, // from 0..args.num_clients
141 : }
142 :
143 : #[derive(serde::Serialize)]
144 : struct Output {
145 : total: request_stats::Output,
146 : }
147 :
148 : tokio_thread_local_stats::declare!(STATS: request_stats::Stats);
149 :
150 0 : pub(crate) fn main(args: Args) -> anyhow::Result<()> {
151 0 : tokio_thread_local_stats::main!(STATS, move |thread_local_stats| {
152 0 : main_impl(args, thread_local_stats)
153 0 : })
154 0 : }
155 :
156 0 : async fn main_impl(
157 0 : args: Args,
158 0 : all_thread_local_stats: AllThreadLocalStats<request_stats::Stats>,
159 0 : ) -> anyhow::Result<()> {
160 0 : let args: &'static Args = Box::leak(Box::new(args));
161 :
162 0 : let mgmt_api_client = Arc::new(pageserver_client::mgmt_api::Client::new(
163 0 : reqwest::Client::new(), // TODO: support ssl_ca_file for https APIs in pagebench.
164 0 : args.mgmt_api_endpoint.clone(),
165 0 : args.pageserver_jwt.as_deref(),
166 : ));
167 :
168 0 : if let Some(engine_str) = &args.set_io_engine {
169 0 : mgmt_api_client.put_io_engine(engine_str).await?;
170 0 : }
171 :
172 0 : if let Some(mode) = &args.set_io_mode {
173 0 : mgmt_api_client.put_io_mode(mode).await?;
174 0 : }
175 :
176 : // discover targets
177 0 : let timelines: Vec<TenantTimelineId> = crate::util::cli::targets::discover(
178 0 : &mgmt_api_client,
179 0 : crate::util::cli::targets::Spec {
180 0 : limit_to_first_n_targets: args.limit_to_first_n_targets,
181 0 : targets: args.targets.clone(),
182 0 : },
183 0 : )
184 0 : .await?;
185 :
186 0 : #[derive(serde::Deserialize)]
187 : struct KeyspaceCacheDe {
188 : tag: Vec<TenantTimelineId>,
189 : data: Vec<KeyRange>,
190 : }
191 : #[derive(serde::Serialize)]
192 : struct KeyspaceCacheSer<'a> {
193 : tag: &'a [TenantTimelineId],
194 : data: &'a [KeyRange],
195 : }
196 0 : let cache = args
197 0 : .keyspace_cache
198 0 : .as_ref()
199 0 : .map(|keyspace_cache_file| {
200 0 : let contents = match std::fs::read(keyspace_cache_file) {
201 0 : Err(e) if e.kind() == std::io::ErrorKind::NotFound => {
202 0 : return anyhow::Ok(None);
203 : }
204 0 : x => x.context("read keyspace cache file")?,
205 : };
206 0 : let cache: KeyspaceCacheDe =
207 0 : serde_json::from_slice(&contents).context("deserialize cache file")?;
208 0 : let tag_ok = HashSet::<TenantTimelineId>::from_iter(cache.tag.into_iter())
209 0 : == HashSet::from_iter(timelines.iter().cloned());
210 0 : info!("keyspace cache file matches tag: {tag_ok}");
211 0 : anyhow::Ok(if tag_ok { Some(cache.data) } else { None })
212 0 : })
213 0 : .transpose()?
214 0 : .flatten();
215 0 : let all_ranges: Vec<KeyRange> = if let Some(cached) = cache {
216 0 : info!("using keyspace cache file");
217 0 : cached
218 : } else {
219 0 : let mut js = JoinSet::new();
220 0 : for timeline in &timelines {
221 0 : js.spawn({
222 0 : let mgmt_api_client = Arc::clone(&mgmt_api_client);
223 0 : let timeline = *timeline;
224 0 : async move {
225 0 : let partitioning = mgmt_api_client
226 0 : .keyspace(
227 0 : TenantShardId::unsharded(timeline.tenant_id),
228 0 : timeline.timeline_id,
229 0 : )
230 0 : .await?;
231 0 : let lsn = partitioning.at_lsn;
232 0 : let start = Instant::now();
233 0 : let mut filtered = KeySpaceAccum::new();
234 : // let's hope this is inlined and vectorized...
235 : // TODO: turn this loop into a is_rel_block_range() function.
236 0 : for r in partitioning.keys.ranges.iter() {
237 0 : let mut i = r.start;
238 0 : while i != r.end {
239 0 : let mut include = true;
240 0 : include &= i.is_rel_block_key();
241 0 : if let Some(only_relnode) = args.only_relnode {
242 0 : include &= i.is_rel_block_of_rel(only_relnode);
243 0 : }
244 0 : if include {
245 0 : filtered.add_key(i);
246 0 : }
247 0 : i = i.next();
248 : }
249 : }
250 0 : let filtered = filtered.to_keyspace();
251 0 : let filter_duration = start.elapsed();
252 :
253 0 : anyhow::Ok((
254 0 : filter_duration,
255 0 : filtered.ranges.into_iter().map(move |r| KeyRange {
256 0 : timeline,
257 0 : timeline_lsn: lsn,
258 0 : start: r.start.to_i128(),
259 0 : end: r.end.to_i128(),
260 0 : }),
261 : ))
262 0 : }
263 : });
264 : }
265 0 : let mut total_filter_duration = Duration::from_secs(0);
266 0 : let mut all_ranges: Vec<KeyRange> = Vec::new();
267 0 : while let Some(res) = js.join_next().await {
268 0 : let (filter_duration, range) = res.unwrap().unwrap();
269 0 : all_ranges.extend(range);
270 0 : total_filter_duration += filter_duration;
271 0 : }
272 0 : info!("filter duration: {}", total_filter_duration.as_secs_f64());
273 0 : if let Some(cachefile) = args.keyspace_cache.as_ref() {
274 0 : let cache = KeyspaceCacheSer {
275 0 : tag: &timelines,
276 0 : data: &all_ranges,
277 0 : };
278 0 : let bytes = serde_json::to_vec(&cache).context("serialize keyspace for cache file")?;
279 0 : std::fs::write(cachefile, bytes).context("write keyspace cache file to disk")?;
280 0 : info!("successfully wrote keyspace cache file");
281 0 : }
282 0 : all_ranges
283 : };
284 :
285 0 : let num_live_stats_dump = 1;
286 0 : let num_work_sender_tasks = args.num_clients.get() * timelines.len();
287 0 : let num_main_impl = 1;
288 :
289 0 : let shared_state = Arc::new(SharedState {
290 0 : start_work_barrier: tokio::sync::Barrier::new(
291 0 : num_live_stats_dump + num_work_sender_tasks + num_main_impl,
292 0 : ),
293 0 : live_stats: LiveStats::default(),
294 0 : });
295 0 : let cancel = CancellationToken::new();
296 :
297 0 : let ss = shared_state.clone();
298 0 : tokio::spawn({
299 0 : async move {
300 0 : ss.start_work_barrier.wait().await;
301 : loop {
302 0 : let start = std::time::Instant::now();
303 0 : tokio::time::sleep(std::time::Duration::from_secs(1)).await;
304 0 : let stats = &ss.live_stats;
305 0 : let completed_requests = stats.completed_requests.swap(0, Ordering::Relaxed);
306 0 : let missed = stats.missed.swap(0, Ordering::Relaxed);
307 0 : let elapsed = start.elapsed();
308 0 : info!(
309 0 : "RPS: {:.0} MISSED: {:.0}",
310 0 : completed_requests as f64 / elapsed.as_secs_f64(),
311 0 : missed as f64 / elapsed.as_secs_f64()
312 : );
313 : }
314 : }
315 : });
316 :
317 0 : let rps_period = args
318 0 : .per_client_rate
319 0 : .map(|rps_limit| Duration::from_secs_f64(1.0 / (rps_limit as f64)));
320 0 : let make_worker: &dyn Fn(WorkerId) -> Pin<Box<dyn Send + Future<Output = ()>>> = &|worker_id| {
321 0 : let ss = shared_state.clone();
322 0 : let cancel = cancel.clone();
323 0 : let ranges: Vec<KeyRange> = all_ranges
324 0 : .iter()
325 0 : .filter(|r| r.timeline == worker_id.timeline)
326 0 : .cloned()
327 0 : .collect();
328 0 : let weights =
329 0 : rand::distr::weighted::WeightedIndex::new(ranges.iter().map(|v| v.len())).unwrap();
330 :
331 0 : Box::pin(async move {
332 0 : let scheme = match Url::parse(&args.page_service_connstring) {
333 0 : Ok(url) => url.scheme().to_lowercase().to_string(),
334 0 : Err(url::ParseError::RelativeUrlWithoutBase) => "postgresql".to_string(),
335 0 : Err(err) => panic!("invalid connstring: {err}"),
336 : };
337 0 : let client: Box<dyn Client> = match scheme.as_str() {
338 0 : "postgresql" | "postgres" => {
339 0 : assert!(!args.compression, "libpq does not support compression");
340 0 : assert!(!args.rich_client, "rich client requires grpc://");
341 0 : Box::new(
342 0 : LibpqClient::new(&args.page_service_connstring, worker_id.timeline)
343 0 : .await
344 0 : .unwrap(),
345 : )
346 : }
347 :
348 0 : "grpc" if args.rich_client => Box::new(
349 0 : RichGrpcClient::new(
350 0 : &args.page_service_connstring,
351 0 : worker_id.timeline,
352 0 : args.compression,
353 0 : )
354 0 : .await
355 0 : .unwrap(),
356 : ),
357 :
358 0 : "grpc" => Box::new(
359 0 : GrpcClient::new(
360 0 : &args.page_service_connstring,
361 0 : worker_id.timeline,
362 0 : args.compression,
363 0 : )
364 0 : .await
365 0 : .unwrap(),
366 : ),
367 :
368 0 : scheme => panic!("unsupported scheme {scheme}"),
369 : };
370 0 : run_worker(args, client, ss, cancel, rps_period, ranges, weights).await
371 0 : })
372 0 : };
373 :
374 0 : info!("spawning workers");
375 0 : let mut workers = JoinSet::new();
376 0 : for timeline in timelines.iter().cloned() {
377 0 : for num_client in 0..args.num_clients.get() {
378 0 : let worker_id = WorkerId {
379 0 : timeline,
380 0 : num_client,
381 0 : };
382 0 : workers.spawn(make_worker(worker_id));
383 0 : }
384 : }
385 0 : let workers = async move {
386 0 : while let Some(res) = workers.join_next().await {
387 0 : res.unwrap();
388 0 : }
389 0 : };
390 :
391 0 : info!("waiting for everything to become ready");
392 0 : shared_state.start_work_barrier.wait().await;
393 0 : info!("work started");
394 0 : if let Some(runtime) = args.runtime {
395 0 : tokio::time::sleep(runtime.into()).await;
396 0 : info!("runtime over, signalling cancellation");
397 0 : cancel.cancel();
398 0 : workers.await;
399 0 : info!("work sender exited");
400 : } else {
401 0 : workers.await;
402 0 : unreachable!("work sender never terminates");
403 : }
404 :
405 0 : let output = Output {
406 : total: {
407 0 : let mut agg_stats = request_stats::Stats::new();
408 0 : for stats in all_thread_local_stats.lock().unwrap().iter() {
409 0 : let stats = stats.lock().unwrap();
410 0 : agg_stats.add(&stats);
411 0 : }
412 0 : agg_stats.output()
413 : },
414 : };
415 :
416 0 : let output = serde_json::to_string_pretty(&output).unwrap();
417 0 : println!("{output}");
418 :
419 0 : anyhow::Ok(())
420 0 : }
421 :
422 0 : async fn run_worker(
423 0 : args: &Args,
424 0 : mut client: Box<dyn Client>,
425 0 : shared_state: Arc<SharedState>,
426 0 : cancel: CancellationToken,
427 0 : rps_period: Option<Duration>,
428 0 : ranges: Vec<KeyRange>,
429 0 : weights: rand::distr::weighted::WeightedIndex<i128>,
430 0 : ) {
431 0 : shared_state.start_work_barrier.wait().await;
432 0 : let client_start = Instant::now();
433 0 : let mut ticks_processed = 0;
434 0 : let mut req_id = 0;
435 0 : let batch_size: usize = args.batch_size.into();
436 :
437 : // Track inflight requests by request ID and start time. This times the request duration, and
438 : // ensures responses match requests. We don't expect responses back in any particular order.
439 : //
440 : // NB: this does not check that all requests received a response, because we don't wait for the
441 : // inflight requests to complete when the duration elapses.
442 0 : let mut inflight: HashMap<u64, Instant> = HashMap::new();
443 :
444 0 : while !cancel.is_cancelled() {
445 : // Detect if a request took longer than the RPS rate
446 0 : if let Some(period) = &rps_period {
447 0 : let periods_passed_until_now =
448 0 : usize::try_from(client_start.elapsed().as_micros() / period.as_micros()).unwrap();
449 :
450 0 : if periods_passed_until_now > ticks_processed {
451 0 : shared_state
452 0 : .live_stats
453 0 : .missed((periods_passed_until_now - ticks_processed) as u64);
454 0 : }
455 0 : ticks_processed = periods_passed_until_now;
456 0 : }
457 :
458 0 : while inflight.len() < args.queue_depth.get() {
459 0 : req_id += 1;
460 0 : let start = Instant::now();
461 0 : let (req_lsn, mod_lsn, rel, blks) = {
462 : /// Converts a compact i128 key to a relation tag and block number.
463 0 : fn key_to_block(key: i128) -> (RelTag, u32) {
464 0 : let key = Key::from_i128(key);
465 0 : assert!(key.is_rel_block_key());
466 0 : key.to_rel_block()
467 0 : .expect("we filter non-rel-block keys out above")
468 0 : }
469 :
470 : // Pick a random page from a random relation.
471 0 : let mut rng = rand::rng();
472 0 : let r = &ranges[weights.sample(&mut rng)];
473 0 : let key: i128 = rng.random_range(r.start..r.end);
474 0 : let (rel_tag, block_no) = key_to_block(key);
475 :
476 0 : let mut blks = VecDeque::with_capacity(batch_size);
477 0 : blks.push_back(block_no);
478 :
479 : // If requested, populate a batch of sequential pages. This is how Postgres will
480 : // request page batches (e.g. prefetches). If we hit the end of the relation, we
481 : // grow the batch towards the start too.
482 0 : for i in 1..batch_size {
483 0 : let (r, b) = key_to_block(key + i as i128);
484 0 : if r != rel_tag {
485 0 : break; // went outside relation
486 0 : }
487 0 : blks.push_back(b)
488 : }
489 :
490 0 : if blks.len() < batch_size {
491 : // Grow batch backwards if needed.
492 0 : for i in 1..batch_size {
493 0 : let (r, b) = key_to_block(key - i as i128);
494 0 : if r != rel_tag {
495 0 : break; // went outside relation
496 0 : }
497 0 : blks.push_front(b)
498 : }
499 0 : }
500 :
501 : // We assume that the entire batch can fit within the relation.
502 0 : assert_eq!(blks.len(), batch_size, "incomplete batch");
503 :
504 0 : let req_lsn = if rng.random_bool(args.req_latest_probability) {
505 0 : Lsn::MAX
506 : } else {
507 0 : r.timeline_lsn
508 : };
509 0 : (req_lsn, r.timeline_lsn, rel_tag, blks.into())
510 : };
511 0 : client
512 0 : .send_get_page(req_id, req_lsn, mod_lsn, rel, blks)
513 0 : .await
514 0 : .unwrap();
515 0 : let old = inflight.insert(req_id, start);
516 0 : assert!(old.is_none(), "duplicate request ID {req_id}");
517 : }
518 :
519 0 : let (req_id, pages) = client.recv_get_page().await.unwrap();
520 0 : assert_eq!(pages.len(), batch_size, "unexpected page count");
521 0 : assert!(pages.iter().all(|p| !p.is_empty()), "empty page");
522 0 : let start = inflight
523 0 : .remove(&req_id)
524 0 : .expect("response for unknown request ID");
525 0 : let end = Instant::now();
526 0 : shared_state.live_stats.request_done();
527 0 : ticks_processed += 1;
528 0 : STATS.with(|stats| {
529 0 : stats
530 0 : .borrow()
531 0 : .lock()
532 0 : .unwrap()
533 0 : .observe(end.duration_since(start))
534 0 : .unwrap();
535 0 : });
536 :
537 0 : if let Some(period) = &rps_period {
538 0 : let next_at = client_start
539 0 : + Duration::from_micros(
540 0 : (ticks_processed) as u64 * u64::try_from(period.as_micros()).unwrap(),
541 0 : );
542 0 : tokio::time::sleep_until(next_at.into()).await;
543 0 : }
544 : }
545 0 : }
546 :
547 : /// A benchmark client, to allow switching out the transport protocol.
548 : ///
549 : /// For simplicity, this just uses separate asynchronous send/recv methods. The send method could
550 : /// return a future that resolves when the response is received, but we don't really need it.
551 : #[async_trait]
552 : trait Client: Send {
553 : /// Sends an asynchronous GetPage request to the pageserver.
554 : async fn send_get_page(
555 : &mut self,
556 : req_id: u64,
557 : req_lsn: Lsn,
558 : mod_lsn: Lsn,
559 : rel: RelTag,
560 : blks: Vec<u32>,
561 : ) -> anyhow::Result<()>;
562 :
563 : /// Receives the next GetPage response from the pageserver.
564 : async fn recv_get_page(&mut self) -> anyhow::Result<(u64, Vec<Bytes>)>;
565 : }
566 :
567 : /// A libpq-based Pageserver client.
568 : struct LibpqClient {
569 : inner: pageserver_client::page_service::PagestreamClient,
570 : // Track sent batches, so we know how many responses to expect.
571 : batch_sizes: VecDeque<usize>,
572 : }
573 :
574 : impl LibpqClient {
575 0 : async fn new(connstring: &str, ttid: TenantTimelineId) -> anyhow::Result<Self> {
576 0 : let inner = pageserver_client::page_service::Client::new(connstring.to_string())
577 0 : .await?
578 0 : .pagestream(ttid.tenant_id, ttid.timeline_id)
579 0 : .await?;
580 0 : Ok(Self {
581 0 : inner,
582 0 : batch_sizes: VecDeque::new(),
583 0 : })
584 0 : }
585 : }
586 :
587 : #[async_trait]
588 : impl Client for LibpqClient {
589 0 : async fn send_get_page(
590 : &mut self,
591 : req_id: u64,
592 : req_lsn: Lsn,
593 : mod_lsn: Lsn,
594 : rel: RelTag,
595 : blks: Vec<u32>,
596 0 : ) -> anyhow::Result<()> {
597 : // libpq doesn't support client-side batches, so we send a bunch of individual requests
598 : // instead in the hope that the server will batch them for us. We use the same request ID
599 : // for all, because we'll return a single batch response.
600 0 : self.batch_sizes.push_back(blks.len());
601 0 : for blkno in blks {
602 0 : let req = PagestreamGetPageRequest {
603 0 : hdr: PagestreamRequest {
604 0 : reqid: req_id,
605 0 : request_lsn: req_lsn,
606 0 : not_modified_since: mod_lsn,
607 0 : },
608 0 : rel,
609 0 : blkno,
610 0 : };
611 0 : self.inner.getpage_send(req).await?;
612 : }
613 0 : Ok(())
614 0 : }
615 :
616 0 : async fn recv_get_page(&mut self) -> anyhow::Result<(u64, Vec<Bytes>)> {
617 0 : let batch_size = self.batch_sizes.pop_front().unwrap();
618 0 : let mut batch = Vec::with_capacity(batch_size);
619 0 : let mut req_id = None;
620 0 : for _ in 0..batch_size {
621 0 : let resp = self.inner.getpage_recv().await?;
622 0 : if req_id.is_none() {
623 0 : req_id = Some(resp.req.hdr.reqid);
624 0 : }
625 0 : assert_eq!(req_id, Some(resp.req.hdr.reqid), "request ID mismatch");
626 0 : batch.push(resp.page);
627 : }
628 0 : Ok((req_id.unwrap(), batch))
629 0 : }
630 : }
631 :
632 : /// A gRPC Pageserver client.
633 : struct GrpcClient {
634 : req_tx: tokio::sync::mpsc::Sender<page_api::GetPageRequest>,
635 : resp_rx: Pin<Box<dyn Stream<Item = Result<page_api::GetPageResponse, tonic::Status>> + Send>>,
636 : }
637 :
638 : impl GrpcClient {
639 0 : async fn new(
640 0 : connstring: &str,
641 0 : ttid: TenantTimelineId,
642 0 : compression: bool,
643 0 : ) -> anyhow::Result<Self> {
644 0 : let mut client = page_api::Client::connect(
645 0 : connstring.to_string(),
646 0 : ttid.tenant_id,
647 0 : ttid.timeline_id,
648 0 : ShardIndex::unsharded(),
649 0 : None,
650 0 : compression.then_some(tonic::codec::CompressionEncoding::Zstd),
651 0 : )
652 0 : .await?;
653 :
654 : // The channel has a buffer size of 1, since 0 is not allowed. It does not matter, since the
655 : // benchmark will control the queue depth (i.e. in-flight requests) anyway, and requests are
656 : // buffered by Tonic and the OS too.
657 0 : let (req_tx, req_rx) = tokio::sync::mpsc::channel(1);
658 0 : let req_stream = tokio_stream::wrappers::ReceiverStream::new(req_rx);
659 0 : let resp_rx = Box::pin(client.get_pages(req_stream).await?);
660 :
661 0 : Ok(Self { req_tx, resp_rx })
662 0 : }
663 : }
664 :
665 : #[async_trait]
666 : impl Client for GrpcClient {
667 0 : async fn send_get_page(
668 : &mut self,
669 : req_id: u64,
670 : req_lsn: Lsn,
671 : mod_lsn: Lsn,
672 : rel: RelTag,
673 : blks: Vec<u32>,
674 0 : ) -> anyhow::Result<()> {
675 0 : let req = page_api::GetPageRequest {
676 0 : request_id: req_id.into(),
677 0 : request_class: page_api::GetPageClass::Normal,
678 0 : read_lsn: page_api::ReadLsn {
679 0 : request_lsn: req_lsn,
680 0 : not_modified_since_lsn: Some(mod_lsn),
681 0 : },
682 0 : rel,
683 0 : block_numbers: blks,
684 0 : };
685 0 : self.req_tx.send(req).await?;
686 0 : Ok(())
687 0 : }
688 :
689 0 : async fn recv_get_page(&mut self) -> anyhow::Result<(u64, Vec<Bytes>)> {
690 0 : let resp = self.resp_rx.next().await.unwrap().unwrap();
691 0 : anyhow::ensure!(
692 0 : resp.status_code == page_api::GetPageStatusCode::Ok,
693 0 : "unexpected status code: {}",
694 : resp.status_code,
695 : );
696 : Ok((
697 0 : resp.request_id.id,
698 0 : resp.pages.into_iter().map(|p| p.image).collect(),
699 : ))
700 0 : }
701 : }
702 :
703 : /// A rich gRPC Pageserver client.
704 : struct RichGrpcClient {
705 : inner: Arc<client_grpc::PageserverClient>,
706 : requests: FuturesUnordered<
707 : Pin<Box<dyn Future<Output = anyhow::Result<page_api::GetPageResponse>> + Send>>,
708 : >,
709 : }
710 :
711 : impl RichGrpcClient {
712 0 : async fn new(
713 0 : connstring: &str,
714 0 : ttid: TenantTimelineId,
715 0 : compression: bool,
716 0 : ) -> anyhow::Result<Self> {
717 0 : let inner = Arc::new(client_grpc::PageserverClient::new(
718 0 : ttid.tenant_id,
719 0 : ttid.timeline_id,
720 0 : ShardSpec::new(
721 0 : [(ShardIndex::unsharded(), connstring.to_string())].into(),
722 0 : None,
723 0 : )?,
724 0 : None,
725 0 : compression.then_some(tonic::codec::CompressionEncoding::Zstd),
726 0 : )?);
727 0 : Ok(Self {
728 0 : inner,
729 0 : requests: FuturesUnordered::new(),
730 0 : })
731 0 : }
732 : }
733 :
734 : #[async_trait]
735 : impl Client for RichGrpcClient {
736 0 : async fn send_get_page(
737 : &mut self,
738 : req_id: u64,
739 : req_lsn: Lsn,
740 : mod_lsn: Lsn,
741 : rel: RelTag,
742 : blks: Vec<u32>,
743 0 : ) -> anyhow::Result<()> {
744 0 : let req = page_api::GetPageRequest {
745 0 : request_id: req_id.into(),
746 0 : request_class: page_api::GetPageClass::Normal,
747 0 : read_lsn: page_api::ReadLsn {
748 0 : request_lsn: req_lsn,
749 0 : not_modified_since_lsn: Some(mod_lsn),
750 0 : },
751 0 : rel,
752 0 : block_numbers: blks,
753 0 : };
754 0 : let inner = self.inner.clone();
755 0 : self.requests.push(Box::pin(async move {
756 0 : inner
757 0 : .get_page(req)
758 0 : .await
759 0 : .map_err(|err| anyhow::anyhow!("{err}"))
760 0 : }));
761 0 : Ok(())
762 0 : }
763 :
764 0 : async fn recv_get_page(&mut self) -> anyhow::Result<(u64, Vec<Bytes>)> {
765 0 : let resp = self.requests.next().await.unwrap()?;
766 : Ok((
767 0 : resp.request_id.id,
768 0 : resp.pages.into_iter().map(|p| p.image).collect(),
769 : ))
770 0 : }
771 : }
|
