Line data Source code
1 : use std::collections::HashSet;
2 : use std::str::FromStr;
3 : use std::sync::Arc;
4 : use std::time::Duration;
5 :
6 : use super::safekeeper_reconciler::ScheduleRequest;
7 : use crate::heartbeater::SafekeeperState;
8 : use crate::metrics;
9 : use crate::persistence::{
10 : DatabaseError, SafekeeperTimelineOpKind, TimelinePendingOpPersistence, TimelinePersistence,
11 : };
12 : use crate::safekeeper::Safekeeper;
13 : use anyhow::Context;
14 : use http_utils::error::ApiError;
15 : use pageserver_api::controller_api::{
16 : SafekeeperDescribeResponse, SkSchedulingPolicy, TimelineImportRequest,
17 : };
18 : use pageserver_api::models::{SafekeeperInfo, SafekeepersInfo, TimelineInfo};
19 : use safekeeper_api::membership::{MemberSet, SafekeeperId};
20 : use tokio::task::JoinSet;
21 : use tokio_util::sync::CancellationToken;
22 : use utils::id::{NodeId, TenantId, TimelineId};
23 : use utils::logging::SecretString;
24 : use utils::lsn::Lsn;
25 :
26 : use super::Service;
27 :
28 : impl Service {
29 : /// Timeline creation on safekeepers
30 : ///
31 : /// Returns `Ok(left)` if the timeline has been created on a quorum of safekeepers,
32 : /// where `left` contains the list of safekeepers that didn't have a successful response.
33 : /// Assumes tenant lock is held while calling this function.
34 0 : pub(super) async fn tenant_timeline_create_safekeepers_quorum(
35 0 : &self,
36 0 : tenant_id: TenantId,
37 0 : timeline_id: TimelineId,
38 0 : pg_version: u32,
39 0 : timeline_persistence: &TimelinePersistence,
40 0 : ) -> Result<Vec<NodeId>, ApiError> {
41 0 : // If quorum is reached, return if we are outside of a specified timeout
42 0 : let jwt = self
43 0 : .config
44 0 : .safekeeper_jwt_token
45 0 : .clone()
46 0 : .map(SecretString::from);
47 0 : let mut joinset = JoinSet::new();
48 0 :
49 0 : // Prepare membership::Configuration from choosen safekeepers.
50 0 : let safekeepers = {
51 0 : let locked = self.inner.read().unwrap();
52 0 : locked.safekeepers.clone()
53 0 : };
54 0 :
55 0 : let mut members = Vec::new();
56 0 : for sk_id in timeline_persistence.sk_set.iter() {
57 0 : let sk_id = NodeId(*sk_id as u64);
58 0 : let Some(safekeeper) = safekeepers.get(&sk_id) else {
59 0 : return Err(ApiError::InternalServerError(anyhow::anyhow!(
60 0 : "couldn't find entry for safekeeper with id {sk_id}"
61 0 : )))?;
62 : };
63 0 : members.push(SafekeeperId {
64 0 : id: sk_id,
65 0 : host: safekeeper.skp.host.clone(),
66 0 : pg_port: safekeeper.skp.port as u16,
67 0 : });
68 : }
69 0 : let mset = MemberSet::new(members).map_err(ApiError::InternalServerError)?;
70 0 : let mconf = safekeeper_api::membership::Configuration::new(mset);
71 0 :
72 0 : let req = safekeeper_api::models::TimelineCreateRequest {
73 0 : commit_lsn: None,
74 0 : mconf,
75 0 : pg_version,
76 0 : start_lsn: timeline_persistence.start_lsn.0,
77 0 : system_id: None,
78 0 : tenant_id,
79 0 : timeline_id,
80 0 : wal_seg_size: None,
81 0 : };
82 : const SK_CREATE_TIMELINE_RECONCILE_TIMEOUT: Duration = Duration::from_secs(30);
83 0 : for sk in timeline_persistence.sk_set.iter() {
84 0 : let sk_id = NodeId(*sk as u64);
85 0 : let safekeepers = safekeepers.clone();
86 0 : let http_client = self.http_client.clone();
87 0 : let jwt = jwt.clone();
88 0 : let req = req.clone();
89 0 : joinset.spawn(async move {
90 0 : // Unwrap is fine as we already would have returned error above
91 0 : let sk_p = safekeepers.get(&sk_id).unwrap();
92 0 : let res = sk_p
93 0 : .with_client_retries(
94 0 : |client| {
95 0 : let req = req.clone();
96 0 : async move { client.create_timeline(&req).await }
97 0 : },
98 0 : &http_client,
99 0 : &jwt,
100 0 : 3,
101 0 : 3,
102 0 : SK_CREATE_TIMELINE_RECONCILE_TIMEOUT,
103 0 : &CancellationToken::new(),
104 0 : )
105 0 : .await;
106 0 : (sk_id, sk_p.skp.host.clone(), res)
107 0 : });
108 0 : }
109 : // After we have built the joinset, we now wait for the tasks to complete,
110 : // but with a specified timeout to make sure we return swiftly, either with
111 : // a failure or success.
112 0 : let reconcile_deadline = tokio::time::Instant::now() + SK_CREATE_TIMELINE_RECONCILE_TIMEOUT;
113 0 :
114 0 : // Wait until all tasks finish or timeout is hit, whichever occurs
115 0 : // first.
116 0 : let mut reconcile_results = Vec::new();
117 : loop {
118 0 : if let Ok(res) = tokio::time::timeout_at(reconcile_deadline, joinset.join_next()).await
119 : {
120 0 : let Some(res) = res else { break };
121 0 : match res {
122 0 : Ok(res) => {
123 0 : tracing::info!(
124 0 : "response from safekeeper id:{} at {}: {:?}",
125 : res.0,
126 : res.1,
127 : res.2
128 : );
129 0 : reconcile_results.push(res);
130 : }
131 0 : Err(join_err) => {
132 0 : tracing::info!("join_err for task in joinset: {join_err}");
133 : }
134 : }
135 : } else {
136 0 : tracing::info!(
137 0 : "timeout for creation call after {} responses",
138 0 : reconcile_results.len()
139 : );
140 0 : break;
141 : }
142 : }
143 :
144 : // Now check now if quorum was reached in reconcile_results.
145 0 : let total_result_count = reconcile_results.len();
146 0 : let remaining = reconcile_results
147 0 : .into_iter()
148 0 : .filter_map(|res| res.2.is_err().then_some(res.0))
149 0 : .collect::<Vec<_>>();
150 0 : tracing::info!(
151 0 : "Got {} non-successful responses from initial creation request of total {total_result_count} responses",
152 0 : remaining.len()
153 : );
154 0 : let target_sk_count = timeline_persistence.sk_set.len();
155 0 : let quorum_size = match target_sk_count {
156 : 0 => {
157 0 : return Err(ApiError::InternalServerError(anyhow::anyhow!(
158 0 : "timeline configured without any safekeepers",
159 0 : )));
160 : }
161 : 1 | 2 => {
162 : #[cfg(feature = "testing")]
163 : {
164 : // In test settings, it is allowed to have one or two safekeepers
165 0 : target_sk_count
166 : }
167 : #[cfg(not(feature = "testing"))]
168 : {
169 : // The region is misconfigured: we need at least three safekeepers to be configured
170 : // in order to schedule work to them
171 : tracing::warn!(
172 : "couldn't find at least 3 safekeepers for timeline, found: {:?}",
173 : timeline_persistence.sk_set
174 : );
175 : return Err(ApiError::InternalServerError(anyhow::anyhow!(
176 : "couldn't find at least 3 safekeepers to put timeline to"
177 : )));
178 : }
179 : }
180 0 : _ => target_sk_count / 2 + 1,
181 : };
182 0 : let success_count = target_sk_count - remaining.len();
183 0 : if success_count < quorum_size {
184 : // Failure
185 0 : return Err(ApiError::InternalServerError(anyhow::anyhow!(
186 0 : "not enough successful reconciliations to reach quorum size: {success_count} of {quorum_size} of total {target_sk_count}"
187 0 : )));
188 0 : }
189 0 :
190 0 : Ok(remaining)
191 0 : }
192 :
193 : /// Create timeline in controller database and on safekeepers.
194 : /// `timeline_info` is result of timeline creation on pageserver.
195 : ///
196 : /// All actions must be idempotent as the call is retried until success. It
197 : /// tries to create timeline in the db and on at least majority of
198 : /// safekeepers + queue creation for safekeepers which missed it in the db
199 : /// for infinite retries; after that, call returns Ok.
200 : ///
201 : /// The idea is that once this is reached as long as we have alive majority
202 : /// of safekeepers it is expected to get eventually operational as storcon
203 : /// will be able to seed timeline on nodes which missed creation by making
204 : /// pull_timeline from peers. On the other hand we don't want to fail
205 : /// timeline creation if one safekeeper is down.
206 0 : pub(super) async fn tenant_timeline_create_safekeepers(
207 0 : self: &Arc<Self>,
208 0 : tenant_id: TenantId,
209 0 : timeline_info: &TimelineInfo,
210 0 : ) -> Result<SafekeepersInfo, ApiError> {
211 0 : let timeline_id = timeline_info.timeline_id;
212 0 : let pg_version = timeline_info.pg_version * 10000;
213 0 : // Initially start_lsn is determined by last_record_lsn in pageserver
214 0 : // response as it does initdb. However, later we persist it and in sk
215 0 : // creation calls replace with the value from the timeline row if it
216 0 : // previously existed as on retries in theory endpoint might have
217 0 : // already written some data and advanced last_record_lsn, while we want
218 0 : // safekeepers to have consistent start_lsn.
219 0 : let start_lsn = timeline_info.last_record_lsn;
220 :
221 : // Choose initial set of safekeepers respecting affinity
222 0 : let sks = self.safekeepers_for_new_timeline().await?;
223 0 : let sks_persistence = sks.iter().map(|sk| sk.id.0 as i64).collect::<Vec<_>>();
224 0 : // Add timeline to db
225 0 : let mut timeline_persist = TimelinePersistence {
226 0 : tenant_id: tenant_id.to_string(),
227 0 : timeline_id: timeline_id.to_string(),
228 0 : start_lsn: start_lsn.into(),
229 0 : generation: 1,
230 0 : sk_set: sks_persistence.clone(),
231 0 : new_sk_set: None,
232 0 : cplane_notified_generation: 0,
233 0 : deleted_at: None,
234 0 : };
235 0 : let inserted = self
236 0 : .persistence
237 0 : .insert_timeline(timeline_persist.clone())
238 0 : .await?;
239 0 : if !inserted {
240 0 : if let Some(existent_persist) = self
241 0 : .persistence
242 0 : .get_timeline(tenant_id, timeline_id)
243 0 : .await?
244 0 : {
245 0 : // Replace with what we have in the db, to get stuff like the generation right.
246 0 : // We do still repeat the http calls to the safekeepers. After all, we could have
247 0 : // crashed right after the wrote to the DB.
248 0 : timeline_persist = existent_persist;
249 0 : } else {
250 0 : return Err(ApiError::InternalServerError(anyhow::anyhow!(
251 0 : "insertion said timeline already in db, but looking it up, it was gone"
252 0 : )));
253 : }
254 0 : }
255 : // Create the timeline on a quorum of safekeepers
256 0 : let remaining = self
257 0 : .tenant_timeline_create_safekeepers_quorum(
258 0 : tenant_id,
259 0 : timeline_id,
260 0 : pg_version,
261 0 : &timeline_persist,
262 0 : )
263 0 : .await?;
264 :
265 : // For the remaining safekeepers, take care of their reconciliation asynchronously
266 0 : for &remaining_id in remaining.iter() {
267 0 : let pending_op = TimelinePendingOpPersistence {
268 0 : tenant_id: tenant_id.to_string(),
269 0 : timeline_id: timeline_id.to_string(),
270 0 : generation: timeline_persist.generation,
271 0 : op_kind: crate::persistence::SafekeeperTimelineOpKind::Pull,
272 0 : sk_id: remaining_id.0 as i64,
273 0 : };
274 0 : tracing::info!("writing pending op for sk id {remaining_id}");
275 0 : self.persistence.insert_pending_op(pending_op).await?;
276 : }
277 0 : if !remaining.is_empty() {
278 0 : let locked = self.inner.read().unwrap();
279 0 : for remaining_id in remaining {
280 0 : let Some(sk) = locked.safekeepers.get(&remaining_id) else {
281 0 : return Err(ApiError::InternalServerError(anyhow::anyhow!(
282 0 : "Couldn't find safekeeper with id {remaining_id}"
283 0 : )));
284 : };
285 0 : let Ok(host_list) = sks
286 0 : .iter()
287 0 : .map(|sk| {
288 0 : Ok((
289 0 : sk.id,
290 0 : locked
291 0 : .safekeepers
292 0 : .get(&sk.id)
293 0 : .ok_or_else(|| {
294 0 : ApiError::InternalServerError(anyhow::anyhow!(
295 0 : "Couldn't find safekeeper with id {} to pull from",
296 0 : sk.id
297 0 : ))
298 0 : })?
299 0 : .base_url(),
300 : ))
301 0 : })
302 0 : .collect::<Result<_, ApiError>>()
303 : else {
304 0 : continue;
305 : };
306 0 : let req = ScheduleRequest {
307 0 : safekeeper: Box::new(sk.clone()),
308 0 : host_list,
309 0 : tenant_id,
310 0 : timeline_id: Some(timeline_id),
311 0 : generation: timeline_persist.generation as u32,
312 0 : kind: crate::persistence::SafekeeperTimelineOpKind::Pull,
313 0 : };
314 0 : locked.safekeeper_reconcilers.schedule_request(req);
315 : }
316 0 : }
317 :
318 0 : Ok(SafekeepersInfo {
319 0 : generation: timeline_persist.generation as u32,
320 0 : safekeepers: sks,
321 0 : tenant_id,
322 0 : timeline_id,
323 0 : })
324 0 : }
325 :
326 : /// Directly insert the timeline into the database without reconciling it with safekeepers.
327 : ///
328 : /// Useful if the timeline already exists on the specified safekeepers,
329 : /// but we want to make it storage controller managed.
330 0 : pub(crate) async fn timeline_import(&self, req: TimelineImportRequest) -> Result<(), ApiError> {
331 0 : let persistence = TimelinePersistence {
332 0 : tenant_id: req.tenant_id.to_string(),
333 0 : timeline_id: req.timeline_id.to_string(),
334 0 : start_lsn: Lsn::INVALID.into(),
335 0 : generation: 1,
336 0 : sk_set: req.sk_set.iter().map(|sk_id| sk_id.0 as i64).collect(),
337 0 : new_sk_set: None,
338 0 : cplane_notified_generation: 1,
339 0 : deleted_at: None,
340 0 : };
341 0 : let inserted = self.persistence.insert_timeline(persistence).await?;
342 0 : if inserted {
343 0 : tracing::info!("imported timeline into db");
344 : } else {
345 0 : tracing::info!("didn't import timeline into db, as it is already present in db");
346 : }
347 0 : Ok(())
348 0 : }
349 :
350 : /// Perform timeline deletion on safekeepers. Will return success: we persist the deletion into the reconciler.
351 0 : pub(super) async fn tenant_timeline_delete_safekeepers(
352 0 : self: &Arc<Self>,
353 0 : tenant_id: TenantId,
354 0 : timeline_id: TimelineId,
355 0 : ) -> Result<(), ApiError> {
356 0 : let tl = self
357 0 : .persistence
358 0 : .get_timeline(tenant_id, timeline_id)
359 0 : .await?;
360 0 : let Some(tl) = tl else {
361 0 : tracing::info!(
362 0 : "timeline {tenant_id}/{timeline_id} doesn't exist in timelines table, no deletions on safekeepers needed"
363 : );
364 0 : return Ok(());
365 : };
366 0 : self.persistence
367 0 : .timeline_set_deleted_at(tenant_id, timeline_id)
368 0 : .await?;
369 0 : let all_sks = tl
370 0 : .new_sk_set
371 0 : .iter()
372 0 : .flatten()
373 0 : .chain(tl.sk_set.iter())
374 0 : .collect::<HashSet<_>>();
375 :
376 : // Schedule reconciliations
377 0 : for &sk_id in all_sks.iter() {
378 0 : let pending_op = TimelinePendingOpPersistence {
379 0 : tenant_id: tenant_id.to_string(),
380 0 : timeline_id: timeline_id.to_string(),
381 0 : generation: i32::MAX,
382 0 : op_kind: SafekeeperTimelineOpKind::Delete,
383 0 : sk_id: *sk_id,
384 0 : };
385 0 : tracing::info!("writing pending op for sk id {sk_id}");
386 0 : self.persistence.insert_pending_op(pending_op).await?;
387 : }
388 : {
389 0 : let locked = self.inner.read().unwrap();
390 0 : for sk_id in all_sks {
391 0 : let sk_id = NodeId(*sk_id as u64);
392 0 : let Some(sk) = locked.safekeepers.get(&sk_id) else {
393 0 : return Err(ApiError::InternalServerError(anyhow::anyhow!(
394 0 : "Couldn't find safekeeper with id {sk_id}"
395 0 : )));
396 : };
397 :
398 0 : let req = ScheduleRequest {
399 0 : safekeeper: Box::new(sk.clone()),
400 0 : // we don't use this for this kind, put a dummy value
401 0 : host_list: Vec::new(),
402 0 : tenant_id,
403 0 : timeline_id: Some(timeline_id),
404 0 : generation: tl.generation as u32,
405 0 : kind: SafekeeperTimelineOpKind::Delete,
406 0 : };
407 0 : locked.safekeeper_reconcilers.schedule_request(req);
408 : }
409 : }
410 0 : Ok(())
411 0 : }
412 :
413 : /// Perform tenant deletion on safekeepers.
414 0 : pub(super) async fn tenant_delete_safekeepers(
415 0 : self: &Arc<Self>,
416 0 : tenant_id: TenantId,
417 0 : ) -> Result<(), ApiError> {
418 0 : let timeline_list = self
419 0 : .persistence
420 0 : .list_timelines_for_tenant(tenant_id)
421 0 : .await?;
422 :
423 0 : if timeline_list.is_empty() {
424 : // Early exit: the tenant is either empty or not migrated to the storcon yet
425 0 : tracing::info!("Skipping tenant delete as the timeline doesn't exist in db");
426 0 : return Ok(());
427 0 : }
428 :
429 0 : let timeline_list = timeline_list
430 0 : .into_iter()
431 0 : .map(|timeline| {
432 0 : let timeline_id = TimelineId::from_str(&timeline.timeline_id)
433 0 : .context("timeline id loaded from db")
434 0 : .map_err(ApiError::InternalServerError)?;
435 0 : Ok((timeline_id, timeline))
436 0 : })
437 0 : .collect::<Result<Vec<_>, ApiError>>()?;
438 :
439 : // Remove pending ops from db, and set `deleted_at`.
440 : // We cancel them in a later iteration once we hold the state lock.
441 0 : for (timeline_id, _timeline) in timeline_list.iter() {
442 0 : self.persistence
443 0 : .remove_pending_ops_for_timeline(tenant_id, Some(*timeline_id))
444 0 : .await?;
445 0 : self.persistence
446 0 : .timeline_set_deleted_at(tenant_id, *timeline_id)
447 0 : .await?;
448 : }
449 :
450 : // The list of safekeepers that have any of the timelines
451 0 : let mut sk_list = HashSet::new();
452 :
453 : // List all pending ops for all timelines, cancel them
454 0 : for (_timeline_id, timeline) in timeline_list.iter() {
455 0 : let sk_iter = timeline
456 0 : .sk_set
457 0 : .iter()
458 0 : .chain(timeline.new_sk_set.iter().flatten())
459 0 : .map(|id| NodeId(*id as u64));
460 0 : sk_list.extend(sk_iter);
461 0 : }
462 :
463 0 : for &sk_id in sk_list.iter() {
464 0 : let pending_op = TimelinePendingOpPersistence {
465 0 : tenant_id: tenant_id.to_string(),
466 0 : timeline_id: String::new(),
467 0 : generation: i32::MAX,
468 0 : op_kind: SafekeeperTimelineOpKind::Delete,
469 0 : sk_id: sk_id.0 as i64,
470 0 : };
471 0 : tracing::info!("writing pending op for sk id {sk_id}");
472 0 : self.persistence.insert_pending_op(pending_op).await?;
473 : }
474 :
475 0 : let mut locked = self.inner.write().unwrap();
476 :
477 0 : for (timeline_id, _timeline) in timeline_list.iter() {
478 0 : for sk_id in sk_list.iter() {
479 0 : locked
480 0 : .safekeeper_reconcilers
481 0 : .cancel_reconciles_for_timeline(*sk_id, tenant_id, Some(*timeline_id));
482 0 : }
483 : }
484 :
485 : // unwrap is safe: we return above for an empty timeline list
486 0 : let max_generation = timeline_list
487 0 : .iter()
488 0 : .map(|(_tl_id, tl)| tl.generation as u32)
489 0 : .max()
490 0 : .unwrap();
491 :
492 0 : for sk_id in sk_list {
493 0 : let Some(safekeeper) = locked.safekeepers.get(&sk_id) else {
494 0 : tracing::warn!("Couldn't find safekeeper with id {sk_id}");
495 0 : continue;
496 : };
497 : // Add pending op for tenant deletion
498 0 : let req = ScheduleRequest {
499 0 : generation: max_generation,
500 0 : host_list: Vec::new(),
501 0 : kind: SafekeeperTimelineOpKind::Delete,
502 0 : safekeeper: Box::new(safekeeper.clone()),
503 0 : tenant_id,
504 0 : timeline_id: None,
505 0 : };
506 0 : locked.safekeeper_reconcilers.schedule_request(req);
507 : }
508 0 : Ok(())
509 0 : }
510 :
511 : /// Choose safekeepers for the new timeline: 3 in different azs.
512 0 : pub(crate) async fn safekeepers_for_new_timeline(
513 0 : &self,
514 0 : ) -> Result<Vec<SafekeeperInfo>, ApiError> {
515 0 : let mut all_safekeepers = {
516 0 : let locked = self.inner.read().unwrap();
517 0 : locked
518 0 : .safekeepers
519 0 : .iter()
520 0 : .filter_map(|sk| {
521 0 : if sk.1.scheduling_policy() != SkSchedulingPolicy::Active {
522 : // If we don't want to schedule stuff onto the safekeeper, respect that.
523 0 : return None;
524 0 : }
525 0 : let utilization_opt = if let SafekeeperState::Available {
526 : last_seen_at: _,
527 0 : utilization,
528 0 : } = sk.1.availability()
529 : {
530 0 : Some(utilization)
531 : } else {
532 : // non-available safekeepers still get a chance for new timelines,
533 : // but put them last in the list.
534 0 : None
535 : };
536 0 : let info = SafekeeperInfo {
537 0 : hostname: sk.1.skp.host.clone(),
538 0 : id: NodeId(sk.1.skp.id as u64),
539 0 : };
540 0 : Some((utilization_opt, info, sk.1.skp.availability_zone_id.clone()))
541 0 : })
542 0 : .collect::<Vec<_>>()
543 0 : };
544 0 : all_safekeepers.sort_by_key(|sk| {
545 0 : (
546 0 : sk.0.as_ref()
547 0 : .map(|ut| ut.timeline_count)
548 0 : .unwrap_or(u64::MAX),
549 0 : // Use the id to decide on equal scores for reliability
550 0 : sk.1.id.0,
551 0 : )
552 0 : });
553 : // Number of safekeepers in different AZs we are looking for
554 0 : let wanted_count = match all_safekeepers.len() {
555 : 0 => {
556 0 : return Err(ApiError::InternalServerError(anyhow::anyhow!(
557 0 : "couldn't find any active safekeeper for new timeline",
558 0 : )));
559 : }
560 : // Have laxer requirements on testig mode as we don't want to
561 : // spin up three safekeepers for every single test
562 : #[cfg(feature = "testing")]
563 0 : 1 | 2 => all_safekeepers.len(),
564 0 : _ => 3,
565 : };
566 0 : let mut sks = Vec::new();
567 0 : let mut azs = HashSet::new();
568 0 : for (_sk_util, sk_info, az_id) in all_safekeepers.iter() {
569 0 : if !azs.insert(az_id) {
570 0 : continue;
571 0 : }
572 0 : sks.push(sk_info.clone());
573 0 : if sks.len() == wanted_count {
574 0 : break;
575 0 : }
576 : }
577 0 : if sks.len() == wanted_count {
578 0 : Ok(sks)
579 : } else {
580 0 : Err(ApiError::InternalServerError(anyhow::anyhow!(
581 0 : "couldn't find {wanted_count} safekeepers in different AZs for new timeline (found: {}, total active: {})",
582 0 : sks.len(),
583 0 : all_safekeepers.len(),
584 0 : )))
585 : }
586 0 : }
587 :
588 0 : pub(crate) async fn safekeepers_list(
589 0 : &self,
590 0 : ) -> Result<Vec<SafekeeperDescribeResponse>, DatabaseError> {
591 0 : let locked = self.inner.read().unwrap();
592 0 : let mut list = locked
593 0 : .safekeepers
594 0 : .iter()
595 0 : .map(|sk| sk.1.describe_response())
596 0 : .collect::<Result<Vec<_>, _>>()?;
597 0 : list.sort_by_key(|v| v.id);
598 0 : Ok(list)
599 0 : }
600 :
601 0 : pub(crate) async fn get_safekeeper(
602 0 : &self,
603 0 : id: i64,
604 0 : ) -> Result<SafekeeperDescribeResponse, DatabaseError> {
605 0 : let locked = self.inner.read().unwrap();
606 0 : let sk = locked
607 0 : .safekeepers
608 0 : .get(&NodeId(id as u64))
609 0 : .ok_or(diesel::result::Error::NotFound)?;
610 0 : sk.describe_response()
611 0 : }
612 :
613 0 : pub(crate) async fn upsert_safekeeper(
614 0 : self: &Arc<Service>,
615 0 : record: crate::persistence::SafekeeperUpsert,
616 0 : ) -> Result<(), ApiError> {
617 0 : let node_id = NodeId(record.id as u64);
618 0 : let use_https = self.config.use_https_safekeeper_api;
619 0 :
620 0 : if use_https && record.https_port.is_none() {
621 0 : return Err(ApiError::PreconditionFailed(
622 0 : format!(
623 0 : "cannot upsert safekeeper {node_id}: \
624 0 : https is enabled, but https port is not specified"
625 0 : )
626 0 : .into(),
627 0 : ));
628 0 : }
629 0 :
630 0 : self.persistence.safekeeper_upsert(record.clone()).await?;
631 : {
632 0 : let mut locked = self.inner.write().unwrap();
633 0 : let mut safekeepers = (*locked.safekeepers).clone();
634 0 : match safekeepers.entry(node_id) {
635 0 : std::collections::hash_map::Entry::Occupied(mut entry) => entry
636 0 : .get_mut()
637 0 : .update_from_record(record)
638 0 : .expect("all preconditions should be checked before upsert to database"),
639 0 : std::collections::hash_map::Entry::Vacant(entry) => {
640 0 : entry.insert(
641 0 : Safekeeper::from_persistence(
642 0 : crate::persistence::SafekeeperPersistence::from_upsert(
643 0 : record,
644 0 : SkSchedulingPolicy::Pause,
645 0 : ),
646 0 : CancellationToken::new(),
647 0 : use_https,
648 0 : )
649 0 : .expect("all preconditions should be checked before upsert to database"),
650 0 : );
651 0 : }
652 : }
653 0 : locked
654 0 : .safekeeper_reconcilers
655 0 : .start_reconciler(node_id, self);
656 0 : locked.safekeepers = Arc::new(safekeepers);
657 0 : metrics::METRICS_REGISTRY
658 0 : .metrics_group
659 0 : .storage_controller_safekeeper_nodes
660 0 : .set(locked.safekeepers.len() as i64);
661 0 : metrics::METRICS_REGISTRY
662 0 : .metrics_group
663 0 : .storage_controller_https_safekeeper_nodes
664 0 : .set(
665 0 : locked
666 0 : .safekeepers
667 0 : .values()
668 0 : .filter(|s| s.has_https_port())
669 0 : .count() as i64,
670 0 : );
671 0 : }
672 0 : Ok(())
673 0 : }
674 :
675 0 : pub(crate) async fn set_safekeeper_scheduling_policy(
676 0 : self: &Arc<Service>,
677 0 : id: i64,
678 0 : scheduling_policy: SkSchedulingPolicy,
679 0 : ) -> Result<(), DatabaseError> {
680 0 : self.persistence
681 0 : .set_safekeeper_scheduling_policy(id, scheduling_policy)
682 0 : .await?;
683 0 : let node_id = NodeId(id as u64);
684 0 : // After the change has been persisted successfully, update the in-memory state
685 0 : {
686 0 : let mut locked = self.inner.write().unwrap();
687 0 : let mut safekeepers = (*locked.safekeepers).clone();
688 0 : let sk = safekeepers
689 0 : .get_mut(&node_id)
690 0 : .ok_or(DatabaseError::Logical("Not found".to_string()))?;
691 0 : sk.set_scheduling_policy(scheduling_policy);
692 0 :
693 0 : match scheduling_policy {
694 0 : SkSchedulingPolicy::Active => {
695 0 : locked
696 0 : .safekeeper_reconcilers
697 0 : .start_reconciler(node_id, self);
698 0 : }
699 0 : SkSchedulingPolicy::Decomissioned | SkSchedulingPolicy::Pause => {
700 0 : locked.safekeeper_reconcilers.stop_reconciler(node_id);
701 0 : }
702 : }
703 :
704 0 : locked.safekeepers = Arc::new(safekeepers);
705 0 : }
706 0 : Ok(())
707 0 : }
708 : }
|
