Line data Source code
1 : //! This module contains functions to serve per-tenant background processes,
2 : //! such as compaction and GC
3 :
4 : use std::ops::ControlFlow;
5 : use std::str::FromStr;
6 : use std::sync::Arc;
7 : use std::time::{Duration, Instant};
8 :
9 : use crate::context::{DownloadBehavior, RequestContext};
10 : use crate::metrics::TENANT_TASK_EVENTS;
11 : use crate::task_mgr;
12 : use crate::task_mgr::{TaskKind, BACKGROUND_RUNTIME};
13 : use crate::tenant::throttle::Stats;
14 : use crate::tenant::timeline::CompactionError;
15 : use crate::tenant::{Tenant, TenantState};
16 : use rand::Rng;
17 : use tokio_util::sync::CancellationToken;
18 : use tracing::*;
19 : use utils::{backoff, completion, pausable_failpoint};
20 :
21 : static CONCURRENT_BACKGROUND_TASKS: once_cell::sync::Lazy<tokio::sync::Semaphore> =
22 60 : once_cell::sync::Lazy::new(|| {
23 60 : let total_threads = task_mgr::TOKIO_WORKER_THREADS.get();
24 60 : let permits = usize::max(
25 60 : 1,
26 60 : // while a lot of the work is done on spawn_blocking, we still do
27 60 : // repartitioning in the async context. this should give leave us some workers
28 60 : // unblocked to be blocked on other work, hopefully easing any outside visible
29 60 : // effects of restarts.
30 60 : //
31 60 : // 6/8 is a guess; previously we ran with unlimited 8 and more from
32 60 : // spawn_blocking.
33 60 : (total_threads * 3).checked_div(4).unwrap_or(0),
34 60 : );
35 60 : assert_ne!(permits, 0, "we will not be adding in permits later");
36 60 : assert!(
37 60 : permits < total_threads,
38 0 : "need threads avail for shorter work"
39 : );
40 60 : tokio::sync::Semaphore::new(permits)
41 60 : });
42 :
43 1080 : #[derive(Debug, PartialEq, Eq, Clone, Copy, strum_macros::IntoStaticStr, enum_map::Enum)]
44 : #[strum(serialize_all = "snake_case")]
45 : pub(crate) enum BackgroundLoopKind {
46 : Compaction,
47 : Gc,
48 : Eviction,
49 : IngestHouseKeeping,
50 : ConsumptionMetricsCollectMetrics,
51 : ConsumptionMetricsSyntheticSizeWorker,
52 : InitialLogicalSizeCalculation,
53 : HeatmapUpload,
54 : SecondaryDownload,
55 : }
56 :
57 : impl BackgroundLoopKind {
58 0 : fn as_static_str(&self) -> &'static str {
59 0 : self.into()
60 0 : }
61 : }
62 :
63 : /// Cancellation safe.
64 1092 : pub(crate) async fn concurrent_background_tasks_rate_limit_permit(
65 1092 : loop_kind: BackgroundLoopKind,
66 1092 : _ctx: &RequestContext,
67 1092 : ) -> tokio::sync::SemaphorePermit<'static> {
68 1092 : let _guard = crate::metrics::BACKGROUND_LOOP_SEMAPHORE.measure_acquisition(loop_kind);
69 1092 :
70 1092 : pausable_failpoint!(
71 : "initial-size-calculation-permit-pause",
72 1092 : loop_kind == BackgroundLoopKind::InitialLogicalSizeCalculation
73 : );
74 :
75 : // TODO: assert that we run on BACKGROUND_RUNTIME; requires tokio_unstable Handle::id();
76 1092 : match CONCURRENT_BACKGROUND_TASKS.acquire().await {
77 1092 : Ok(permit) => permit,
78 0 : Err(_closed) => unreachable!("we never close the semaphore"),
79 : }
80 1092 : }
81 :
82 : /// Start per tenant background loops: compaction and gc.
83 0 : pub fn start_background_loops(
84 0 : tenant: &Arc<Tenant>,
85 0 : background_jobs_can_start: Option<&completion::Barrier>,
86 0 : ) {
87 0 : let tenant_shard_id = tenant.tenant_shard_id;
88 0 : task_mgr::spawn(
89 0 : BACKGROUND_RUNTIME.handle(),
90 0 : TaskKind::Compaction,
91 0 : tenant_shard_id,
92 0 : None,
93 0 : &format!("compactor for tenant {tenant_shard_id}"),
94 0 : {
95 0 : let tenant = Arc::clone(tenant);
96 0 : let background_jobs_can_start = background_jobs_can_start.cloned();
97 0 : async move {
98 0 : let cancel = task_mgr::shutdown_token();
99 0 : tokio::select! {
100 0 : _ = cancel.cancelled() => { return Ok(()) },
101 0 : _ = completion::Barrier::maybe_wait(background_jobs_can_start) => {}
102 0 : };
103 0 : compaction_loop(tenant, cancel)
104 0 : // If you rename this span, change the RUST_LOG env variable in test_runner/performance/test_branch_creation.py
105 0 : .instrument(info_span!("compaction_loop", tenant_id = %tenant_shard_id.tenant_id, shard_id = %tenant_shard_id.shard_slug()))
106 0 : .await;
107 0 : Ok(())
108 0 : }
109 0 : },
110 0 : );
111 0 : task_mgr::spawn(
112 0 : BACKGROUND_RUNTIME.handle(),
113 0 : TaskKind::GarbageCollector,
114 0 : tenant_shard_id,
115 0 : None,
116 0 : &format!("garbage collector for tenant {tenant_shard_id}"),
117 0 : {
118 0 : let tenant = Arc::clone(tenant);
119 0 : let background_jobs_can_start = background_jobs_can_start.cloned();
120 0 : async move {
121 0 : let cancel = task_mgr::shutdown_token();
122 0 : tokio::select! {
123 0 : _ = cancel.cancelled() => { return Ok(()) },
124 0 : _ = completion::Barrier::maybe_wait(background_jobs_can_start) => {}
125 0 : };
126 0 : gc_loop(tenant, cancel)
127 0 : .instrument(info_span!("gc_loop", tenant_id = %tenant_shard_id.tenant_id, shard_id = %tenant_shard_id.shard_slug()))
128 0 : .await;
129 0 : Ok(())
130 0 : }
131 0 : },
132 0 : );
133 0 :
134 0 : task_mgr::spawn(
135 0 : BACKGROUND_RUNTIME.handle(),
136 0 : TaskKind::IngestHousekeeping,
137 0 : tenant_shard_id,
138 0 : None,
139 0 : &format!("ingest housekeeping for tenant {tenant_shard_id}"),
140 0 : {
141 0 : let tenant = Arc::clone(tenant);
142 0 : let background_jobs_can_start = background_jobs_can_start.cloned();
143 0 : async move {
144 0 : let cancel = task_mgr::shutdown_token();
145 0 : tokio::select! {
146 0 : _ = cancel.cancelled() => { return Ok(()) },
147 0 : _ = completion::Barrier::maybe_wait(background_jobs_can_start) => {}
148 0 : };
149 0 : ingest_housekeeping_loop(tenant, cancel)
150 0 : .instrument(info_span!("ingest_housekeeping_loop", tenant_id = %tenant_shard_id.tenant_id, shard_id = %tenant_shard_id.shard_slug()))
151 0 : .await;
152 0 : Ok(())
153 0 : }
154 0 : },
155 0 : );
156 0 : }
157 :
158 : ///
159 : /// Compaction task's main loop
160 : ///
161 0 : async fn compaction_loop(tenant: Arc<Tenant>, cancel: CancellationToken) {
162 : const MAX_BACKOFF_SECS: f64 = 300.0;
163 : // How many errors we have seen consequtively
164 0 : let mut error_run_count = 0;
165 0 :
166 0 : let mut last_throttle_flag_reset_at = Instant::now();
167 0 :
168 0 : TENANT_TASK_EVENTS.with_label_values(&["start"]).inc();
169 0 : async {
170 0 : let ctx = RequestContext::todo_child(TaskKind::Compaction, DownloadBehavior::Download);
171 0 : let mut first = true;
172 : loop {
173 0 : tokio::select! {
174 0 : _ = cancel.cancelled() => {
175 0 : return;
176 : },
177 0 : tenant_wait_result = wait_for_active_tenant(&tenant) => match tenant_wait_result {
178 0 : ControlFlow::Break(()) => return,
179 0 : ControlFlow::Continue(()) => (),
180 0 : },
181 0 : }
182 0 :
183 0 : let period = tenant.get_compaction_period();
184 0 :
185 0 : // TODO: we shouldn't need to await to find tenant and this could be moved outside of
186 0 : // loop, #3501. There are also additional "allowed_errors" in tests.
187 0 : if first {
188 0 : first = false;
189 0 : if random_init_delay(period, &cancel).await.is_err() {
190 0 : break;
191 0 : }
192 0 : }
193 :
194 :
195 :
196 : let sleep_duration;
197 0 : if period == Duration::ZERO {
198 : #[cfg(not(feature = "testing"))]
199 : info!("automatic compaction is disabled");
200 : // check again in 10 seconds, in case it's been enabled again.
201 0 : sleep_duration = Duration::from_secs(10)
202 : } else {
203 0 : let iteration = Iteration {
204 0 : started_at: Instant::now(),
205 0 : period,
206 0 : kind: BackgroundLoopKind::Compaction,
207 0 : };
208 :
209 : // Run compaction
210 0 : let IterationResult { output, elapsed } = iteration.run(tenant.compaction_iteration(&cancel, &ctx)).await;
211 0 : match output {
212 0 : Ok(has_pending_task) => {
213 0 : error_run_count = 0;
214 0 : // schedule the next compaction immediately in case there is a pending compaction task
215 0 : sleep_duration = if has_pending_task { Duration::ZERO } else { period };
216 : }
217 0 : Err(e) => {
218 0 : let wait_duration = backoff::exponential_backoff_duration_seconds(
219 0 : error_run_count + 1,
220 0 : 1.0,
221 0 : MAX_BACKOFF_SECS,
222 0 : );
223 0 : error_run_count += 1;
224 0 : let wait_duration = Duration::from_secs_f64(wait_duration);
225 0 : log_compaction_error(
226 0 : &e,
227 0 : error_run_count,
228 0 : &wait_duration,
229 0 : cancel.is_cancelled(),
230 0 : );
231 0 : sleep_duration = wait_duration;
232 0 : }
233 : }
234 :
235 : // the duration is recorded by performance tests by enabling debug in this function
236 0 : tracing::debug!(elapsed_ms=elapsed.as_millis(), "compaction iteration complete");
237 : };
238 :
239 :
240 : // Perhaps we did no work and the walredo process has been idle for some time:
241 : // give it a chance to shut down to avoid leaving walredo process running indefinitely.
242 0 : if let Some(walredo_mgr) = &tenant.walredo_mgr {
243 0 : walredo_mgr.maybe_quiesce(period * 10);
244 0 : }
245 :
246 : // TODO: move this (and walredo quiesce) to a separate task that isn't affected by the back-off,
247 : // so we get some upper bound guarantee on when walredo quiesce / this throttling reporting here happens.
248 0 : info_span!(parent: None, "timeline_get_throttle", tenant_id=%tenant.tenant_shard_id, shard_id=%tenant.tenant_shard_id.shard_slug()).in_scope(|| {
249 0 : let now = Instant::now();
250 0 : let prev = std::mem::replace(&mut last_throttle_flag_reset_at, now);
251 0 : let Stats { count_accounted, count_throttled, sum_throttled_usecs } = tenant.timeline_get_throttle.reset_stats();
252 0 : if count_throttled == 0 {
253 0 : return;
254 0 : }
255 0 : let allowed_rps = tenant.timeline_get_throttle.steady_rps();
256 0 : let delta = now - prev;
257 0 : info!(
258 0 : n_seconds=%format_args!("{:.3}",
259 0 : delta.as_secs_f64()),
260 : count_accounted,
261 : count_throttled,
262 : sum_throttled_usecs,
263 0 : allowed_rps=%format_args!("{allowed_rps:.0}"),
264 0 : "shard was throttled in the last n_seconds"
265 : );
266 0 : });
267 0 :
268 0 : // Sleep
269 0 : if tokio::time::timeout(sleep_duration, cancel.cancelled())
270 0 : .await
271 0 : .is_ok()
272 : {
273 0 : break;
274 0 : }
275 : }
276 0 : }
277 0 : .await;
278 0 : TENANT_TASK_EVENTS.with_label_values(&["stop"]).inc();
279 0 : }
280 :
281 0 : fn log_compaction_error(
282 0 : e: &CompactionError,
283 0 : error_run_count: u32,
284 0 : sleep_duration: &std::time::Duration,
285 0 : task_cancelled: bool,
286 0 : ) {
287 : use crate::tenant::upload_queue::NotInitialized;
288 : use crate::tenant::PageReconstructError;
289 : use CompactionError::*;
290 :
291 : enum LooksLike {
292 : Info,
293 : Error,
294 : }
295 :
296 0 : let decision = match e {
297 0 : ShuttingDown => None,
298 0 : _ if task_cancelled => Some(LooksLike::Info),
299 0 : Other(e) => {
300 0 : let root_cause = e.root_cause();
301 :
302 0 : let is_stopping = {
303 0 : let upload_queue = root_cause
304 0 : .downcast_ref::<NotInitialized>()
305 0 : .is_some_and(|e| e.is_stopping());
306 0 :
307 0 : let timeline = root_cause
308 0 : .downcast_ref::<PageReconstructError>()
309 0 : .is_some_and(|e| e.is_stopping());
310 0 :
311 0 : upload_queue || timeline
312 : };
313 :
314 0 : if is_stopping {
315 0 : Some(LooksLike::Info)
316 : } else {
317 0 : Some(LooksLike::Error)
318 : }
319 : }
320 : };
321 :
322 0 : match decision {
323 0 : Some(LooksLike::Info) => info!(
324 0 : "Compaction failed {error_run_count} times, retrying in {sleep_duration:?}: {e:#}",
325 : ),
326 0 : Some(LooksLike::Error) => error!(
327 0 : "Compaction failed {error_run_count} times, retrying in {sleep_duration:?}: {e:?}",
328 : ),
329 0 : None => {}
330 : }
331 0 : }
332 :
333 : ///
334 : /// GC task's main loop
335 : ///
336 0 : async fn gc_loop(tenant: Arc<Tenant>, cancel: CancellationToken) {
337 : const MAX_BACKOFF_SECS: f64 = 300.0;
338 : // How many errors we have seen consequtively
339 0 : let mut error_run_count = 0;
340 0 :
341 0 : TENANT_TASK_EVENTS.with_label_values(&["start"]).inc();
342 0 : async {
343 0 : // GC might require downloading, to find the cutoff LSN that corresponds to the
344 0 : // cutoff specified as time.
345 0 : let ctx =
346 0 : RequestContext::todo_child(TaskKind::GarbageCollector, DownloadBehavior::Download);
347 0 :
348 0 : let mut first = true;
349 0 : tenant.gc_block.set_lsn_lease_deadline(tenant.get_lsn_lease_length());
350 : loop {
351 0 : tokio::select! {
352 0 : _ = cancel.cancelled() => {
353 0 : return;
354 : },
355 0 : tenant_wait_result = wait_for_active_tenant(&tenant) => match tenant_wait_result {
356 0 : ControlFlow::Break(()) => return,
357 0 : ControlFlow::Continue(()) => (),
358 0 : },
359 0 : }
360 0 :
361 0 : let period = tenant.get_gc_period();
362 0 :
363 0 : if first {
364 0 : first = false;
365 0 :
366 0 : let delays = async {
367 0 : random_init_delay(period, &cancel).await?;
368 0 : Ok::<_, Cancelled>(())
369 0 : };
370 :
371 0 : if delays.await.is_err() {
372 0 : break;
373 0 : }
374 0 : }
375 :
376 0 : let gc_horizon = tenant.get_gc_horizon();
377 0 : let sleep_duration;
378 0 : if period == Duration::ZERO || gc_horizon == 0 {
379 0 : #[cfg(not(feature = "testing"))]
380 0 : info!("automatic GC is disabled");
381 0 : // check again in 10 seconds, in case it's been enabled again.
382 0 : sleep_duration = Duration::from_secs(10);
383 0 : } else {
384 0 : let iteration = Iteration {
385 0 : started_at: Instant::now(),
386 0 : period,
387 0 : kind: BackgroundLoopKind::Gc,
388 0 : };
389 : // Run gc
390 0 : let IterationResult { output, elapsed: _ } =
391 0 : iteration.run(tenant.gc_iteration(None, gc_horizon, tenant.get_pitr_interval(), &cancel, &ctx))
392 0 : .await;
393 0 : match output {
394 0 : Ok(_) => {
395 0 : error_run_count = 0;
396 0 : sleep_duration = period;
397 0 : }
398 : Err(crate::tenant::GcError::TenantCancelled) => {
399 0 : return;
400 : }
401 0 : Err(e) => {
402 0 : let wait_duration = backoff::exponential_backoff_duration_seconds(
403 0 : error_run_count + 1,
404 0 : 1.0,
405 0 : MAX_BACKOFF_SECS,
406 0 : );
407 0 : error_run_count += 1;
408 0 : let wait_duration = Duration::from_secs_f64(wait_duration);
409 :
410 0 : if matches!(e, crate::tenant::GcError::TimelineCancelled) {
411 : // Timeline was cancelled during gc. We might either be in an event
412 : // that affects the entire tenant (tenant deletion, pageserver shutdown),
413 : // or in one that affects the timeline only (timeline deletion).
414 : // Therefore, don't exit the loop.
415 0 : info!("Gc failed {error_run_count} times, retrying in {wait_duration:?}: {e:?}");
416 : } else {
417 0 : error!("Gc failed {error_run_count} times, retrying in {wait_duration:?}: {e:?}");
418 : }
419 :
420 0 : sleep_duration = wait_duration;
421 : }
422 : }
423 : };
424 :
425 0 : if tokio::time::timeout(sleep_duration, cancel.cancelled())
426 0 : .await
427 0 : .is_ok()
428 : {
429 0 : break;
430 0 : }
431 : }
432 0 : }
433 0 : .await;
434 0 : TENANT_TASK_EVENTS.with_label_values(&["stop"]).inc();
435 0 : }
436 :
437 0 : async fn ingest_housekeeping_loop(tenant: Arc<Tenant>, cancel: CancellationToken) {
438 0 : TENANT_TASK_EVENTS.with_label_values(&["start"]).inc();
439 0 : async {
440 : loop {
441 0 : tokio::select! {
442 0 : _ = cancel.cancelled() => {
443 0 : return;
444 : },
445 0 : tenant_wait_result = wait_for_active_tenant(&tenant) => match tenant_wait_result {
446 0 : ControlFlow::Break(()) => return,
447 0 : ControlFlow::Continue(()) => (),
448 0 : },
449 0 : }
450 0 :
451 0 : // We run ingest housekeeping with the same frequency as compaction: it is not worth
452 0 : // having a distinct setting. But we don't run it in the same task, because compaction
453 0 : // blocks on acquiring the background job semaphore.
454 0 : let period = tenant.get_compaction_period();
455 :
456 : // If compaction period is set to zero (to disable it), then we will use a reasonable default
457 0 : let period = if period == Duration::ZERO {
458 0 : humantime::Duration::from_str(
459 0 : pageserver_api::config::tenant_conf_defaults::DEFAULT_COMPACTION_PERIOD,
460 0 : )
461 0 : .unwrap()
462 0 : .into()
463 : } else {
464 0 : period
465 : };
466 :
467 : // Jitter the period by +/- 5%
468 0 : let period =
469 0 : rand::thread_rng().gen_range((period * (95)) / 100..(period * (105)) / 100);
470 0 :
471 0 : // Always sleep first: we do not need to do ingest housekeeping early in the lifetime of
472 0 : // a tenant, since it won't have started writing any ephemeral files yet.
473 0 : if tokio::time::timeout(period, cancel.cancelled())
474 0 : .await
475 0 : .is_ok()
476 : {
477 0 : break;
478 0 : }
479 0 :
480 0 : let iteration = Iteration {
481 0 : started_at: Instant::now(),
482 0 : period,
483 0 : kind: BackgroundLoopKind::IngestHouseKeeping,
484 0 : };
485 0 : iteration.run(tenant.ingest_housekeeping()).await;
486 : }
487 0 : }
488 0 : .await;
489 0 : TENANT_TASK_EVENTS.with_label_values(&["stop"]).inc();
490 0 : }
491 :
492 0 : async fn wait_for_active_tenant(tenant: &Arc<Tenant>) -> ControlFlow<()> {
493 0 : // if the tenant has a proper status already, no need to wait for anything
494 0 : if tenant.current_state() == TenantState::Active {
495 0 : ControlFlow::Continue(())
496 : } else {
497 0 : let mut tenant_state_updates = tenant.subscribe_for_state_updates();
498 : loop {
499 0 : match tenant_state_updates.changed().await {
500 : Ok(()) => {
501 0 : let new_state = &*tenant_state_updates.borrow();
502 0 : match new_state {
503 : TenantState::Active => {
504 0 : debug!("Tenant state changed to active, continuing the task loop");
505 0 : return ControlFlow::Continue(());
506 : }
507 0 : state => {
508 0 : debug!("Not running the task loop, tenant is not active: {state:?}");
509 0 : continue;
510 : }
511 : }
512 : }
513 0 : Err(_sender_dropped_error) => {
514 0 : return ControlFlow::Break(());
515 : }
516 : }
517 : }
518 : }
519 0 : }
520 :
521 0 : #[derive(thiserror::Error, Debug)]
522 : #[error("cancelled")]
523 : pub(crate) struct Cancelled;
524 :
525 : /// Provide a random delay for background task initialization.
526 : ///
527 : /// This delay prevents a thundering herd of background tasks and will likely keep them running on
528 : /// different periods for more stable load.
529 0 : pub(crate) async fn random_init_delay(
530 0 : period: Duration,
531 0 : cancel: &CancellationToken,
532 0 : ) -> Result<(), Cancelled> {
533 0 : if period == Duration::ZERO {
534 0 : return Ok(());
535 0 : }
536 0 :
537 0 : let d = {
538 0 : let mut rng = rand::thread_rng();
539 0 : rng.gen_range(Duration::ZERO..=period)
540 0 : };
541 0 : match tokio::time::timeout(d, cancel.cancelled()).await {
542 0 : Ok(_) => Err(Cancelled),
543 0 : Err(_) => Ok(()),
544 : }
545 0 : }
546 :
547 : struct Iteration {
548 : started_at: Instant,
549 : period: Duration,
550 : kind: BackgroundLoopKind,
551 : }
552 :
553 : struct IterationResult<O> {
554 : output: O,
555 : elapsed: Duration,
556 : }
557 :
558 : impl Iteration {
559 0 : #[instrument(skip_all)]
560 : pub(crate) async fn run<Fut, O>(self, fut: Fut) -> IterationResult<O>
561 : where
562 : Fut: std::future::Future<Output = O>,
563 : {
564 : let Self {
565 : started_at,
566 : period,
567 : kind,
568 : } = self;
569 :
570 : let mut fut = std::pin::pin!(fut);
571 :
572 : // Wrap `fut` into a future that logs a message every `period` so that we get a
573 : // very obvious breadcrumb in the logs _while_ a slow iteration is happening.
574 0 : let liveness_logger = async move {
575 : loop {
576 0 : match tokio::time::timeout(period, &mut fut).await {
577 0 : Ok(x) => return x,
578 : Err(_) => {
579 : // info level as per the same rationale why warn_when_period_overrun is info
580 : // => https://github.com/neondatabase/neon/pull/5724
581 0 : info!("still running");
582 : }
583 : }
584 : }
585 0 : };
586 :
587 : let output = liveness_logger.await;
588 :
589 : let elapsed = started_at.elapsed();
590 : warn_when_period_overrun(elapsed, period, kind);
591 :
592 : IterationResult { output, elapsed }
593 : }
594 : }
595 : /// Attention: the `task` and `period` beocme labels of a pageserver-wide prometheus metric.
596 0 : pub(crate) fn warn_when_period_overrun(
597 0 : elapsed: Duration,
598 0 : period: Duration,
599 0 : task: BackgroundLoopKind,
600 0 : ) {
601 0 : // Duration::ZERO will happen because it's the "disable [bgtask]" value.
602 0 : if elapsed >= period && period != Duration::ZERO {
603 : // humantime does no significant digits clamping whereas Duration's debug is a bit more
604 : // intelligent. however it makes sense to keep the "configuration format" for period, even
605 : // though there's no way to output the actual config value.
606 0 : info!(
607 : ?elapsed,
608 0 : period = %humantime::format_duration(period),
609 0 : ?task,
610 0 : "task iteration took longer than the configured period"
611 : );
612 0 : crate::metrics::BACKGROUND_LOOP_PERIOD_OVERRUN_COUNT
613 0 : .with_label_values(&[task.as_static_str(), &format!("{}", period.as_secs())])
614 0 : .inc();
615 0 : }
616 0 : }
|
