Line data Source code
1 : //!
2 : //! This provides an abstraction to store PostgreSQL relations and other files
3 : //! in the key-value store that implements the Repository interface.
4 : //!
5 : //! (TODO: The line between PUT-functions here and walingest.rs is a bit blurry, as
6 : //! walingest.rs handles a few things like implicit relation creation and extension.
7 : //! Clarify that)
8 : //!
9 : use super::tenant::{PageReconstructError, Timeline};
10 : use crate::context::RequestContext;
11 : use crate::keyspace::{KeySpace, KeySpaceAccum};
12 : use crate::repository::*;
13 : use crate::span::debug_assert_current_span_has_tenant_and_timeline_id_no_shard_id;
14 : use crate::walrecord::NeonWalRecord;
15 : use anyhow::{ensure, Context};
16 : use bytes::{Buf, Bytes, BytesMut};
17 : use enum_map::Enum;
18 : use pageserver_api::key::{
19 : dbdir_key_range, is_rel_block_key, is_slru_block_key, rel_block_to_key, rel_dir_to_key,
20 : rel_key_range, rel_size_to_key, relmap_file_key, slru_block_to_key, slru_dir_to_key,
21 : slru_segment_key_range, slru_segment_size_to_key, twophase_file_key, twophase_key_range,
22 : AUX_FILES_KEY, CHECKPOINT_KEY, CONTROLFILE_KEY, DBDIR_KEY, TWOPHASEDIR_KEY,
23 : };
24 : use pageserver_api::reltag::{BlockNumber, RelTag, SlruKind};
25 : use postgres_ffi::relfile_utils::{FSM_FORKNUM, VISIBILITYMAP_FORKNUM};
26 : use postgres_ffi::BLCKSZ;
27 : use postgres_ffi::{Oid, TimestampTz, TransactionId};
28 : use serde::{Deserialize, Serialize};
29 : use std::collections::{hash_map, HashMap, HashSet};
30 : use std::ops::ControlFlow;
31 : use std::ops::Range;
32 : use strum::IntoEnumIterator;
33 : use tokio_util::sync::CancellationToken;
34 : use tracing::{debug, trace, warn};
35 : use utils::bin_ser::DeserializeError;
36 : use utils::{bin_ser::BeSer, lsn::Lsn};
37 :
38 : const MAX_AUX_FILE_DELTAS: usize = 1024;
39 :
40 0 : #[derive(Debug)]
41 : pub enum LsnForTimestamp {
42 : /// Found commits both before and after the given timestamp
43 : Present(Lsn),
44 :
45 : /// Found no commits after the given timestamp, this means
46 : /// that the newest data in the branch is older than the given
47 : /// timestamp.
48 : ///
49 : /// All commits <= LSN happened before the given timestamp
50 : Future(Lsn),
51 :
52 : /// The queried timestamp is past our horizon we look back at (PITR)
53 : ///
54 : /// All commits > LSN happened after the given timestamp,
55 : /// but any commits < LSN might have happened before or after
56 : /// the given timestamp. We don't know because no data before
57 : /// the given lsn is available.
58 : Past(Lsn),
59 :
60 : /// We have found no commit with a timestamp,
61 : /// so we can't return anything meaningful.
62 : ///
63 : /// The associated LSN is the lower bound value we can safely
64 : /// create branches on, but no statement is made if it is
65 : /// older or newer than the timestamp.
66 : ///
67 : /// This variant can e.g. be returned right after a
68 : /// cluster import.
69 : NoData(Lsn),
70 : }
71 :
72 0 : #[derive(Debug, thiserror::Error)]
73 : pub enum CalculateLogicalSizeError {
74 : #[error("cancelled")]
75 : Cancelled,
76 : #[error(transparent)]
77 : Other(#[from] anyhow::Error),
78 : }
79 :
80 0 : #[derive(Debug, thiserror::Error)]
81 : pub(crate) enum CollectKeySpaceError {
82 : #[error(transparent)]
83 : Decode(#[from] DeserializeError),
84 : #[error(transparent)]
85 : PageRead(PageReconstructError),
86 : #[error("cancelled")]
87 : Cancelled,
88 : }
89 :
90 : impl From<PageReconstructError> for CollectKeySpaceError {
91 0 : fn from(err: PageReconstructError) -> Self {
92 0 : match err {
93 0 : PageReconstructError::Cancelled => Self::Cancelled,
94 0 : err => Self::PageRead(err),
95 : }
96 0 : }
97 : }
98 :
99 : impl From<PageReconstructError> for CalculateLogicalSizeError {
100 0 : fn from(pre: PageReconstructError) -> Self {
101 0 : match pre {
102 : PageReconstructError::AncestorStopping(_) | PageReconstructError::Cancelled => {
103 0 : Self::Cancelled
104 : }
105 0 : _ => Self::Other(pre.into()),
106 : }
107 0 : }
108 : }
109 :
110 0 : #[derive(Debug, thiserror::Error)]
111 : pub enum RelationError {
112 : #[error("Relation Already Exists")]
113 : AlreadyExists,
114 : #[error("invalid relnode")]
115 : InvalidRelnode,
116 : #[error(transparent)]
117 : Other(#[from] anyhow::Error),
118 : }
119 :
120 : ///
121 : /// This impl provides all the functionality to store PostgreSQL relations, SLRUs,
122 : /// and other special kinds of files, in a versioned key-value store. The
123 : /// Timeline struct provides the key-value store.
124 : ///
125 : /// This is a separate impl, so that we can easily include all these functions in a Timeline
126 : /// implementation, and might be moved into a separate struct later.
127 : impl Timeline {
128 : /// Start ingesting a WAL record, or other atomic modification of
129 : /// the timeline.
130 : ///
131 : /// This provides a transaction-like interface to perform a bunch
132 : /// of modifications atomically.
133 : ///
134 : /// To ingest a WAL record, call begin_modification(lsn) to get a
135 : /// DatadirModification object. Use the functions in the object to
136 : /// modify the repository state, updating all the pages and metadata
137 : /// that the WAL record affects. When you're done, call commit() to
138 : /// commit the changes.
139 : ///
140 : /// Lsn stored in modification is advanced by `ingest_record` and
141 : /// is used by `commit()` to update `last_record_lsn`.
142 : ///
143 : /// Calling commit() will flush all the changes and reset the state,
144 : /// so the `DatadirModification` struct can be reused to perform the next modification.
145 : ///
146 : /// Note that any pending modifications you make through the
147 : /// modification object won't be visible to calls to the 'get' and list
148 : /// functions of the timeline until you finish! And if you update the
149 : /// same page twice, the last update wins.
150 : ///
151 268272 : pub fn begin_modification(&self, lsn: Lsn) -> DatadirModification
152 268272 : where
153 268272 : Self: Sized,
154 268272 : {
155 268272 : DatadirModification {
156 268272 : tline: self,
157 268272 : pending_lsns: Vec::new(),
158 268272 : pending_updates: HashMap::new(),
159 268272 : pending_deletions: Vec::new(),
160 268272 : pending_nblocks: 0,
161 268272 : pending_directory_entries: Vec::new(),
162 268272 : lsn,
163 268272 : }
164 268272 : }
165 :
166 : //------------------------------------------------------------------------------
167 : // Public GET functions
168 : //------------------------------------------------------------------------------
169 :
170 : /// Look up given page version.
171 18384 : pub(crate) async fn get_rel_page_at_lsn(
172 18384 : &self,
173 18384 : tag: RelTag,
174 18384 : blknum: BlockNumber,
175 18384 : version: Version<'_>,
176 18384 : latest: bool,
177 18384 : ctx: &RequestContext,
178 18384 : ) -> Result<Bytes, PageReconstructError> {
179 18384 : if tag.relnode == 0 {
180 0 : return Err(PageReconstructError::Other(
181 0 : RelationError::InvalidRelnode.into(),
182 0 : ));
183 18384 : }
184 :
185 18384 : let nblocks = self.get_rel_size(tag, version, latest, ctx).await?;
186 18384 : if blknum >= nblocks {
187 0 : debug!(
188 0 : "read beyond EOF at {} blk {} at {}, size is {}: returning all-zeros page",
189 0 : tag,
190 0 : blknum,
191 0 : version.get_lsn(),
192 0 : nblocks
193 0 : );
194 0 : return Ok(ZERO_PAGE.clone());
195 18384 : }
196 18384 :
197 18384 : let key = rel_block_to_key(tag, blknum);
198 18384 : version.get(self, key, ctx).await
199 18384 : }
200 :
201 : // Get size of a database in blocks
202 0 : pub(crate) async fn get_db_size(
203 0 : &self,
204 0 : spcnode: Oid,
205 0 : dbnode: Oid,
206 0 : version: Version<'_>,
207 0 : latest: bool,
208 0 : ctx: &RequestContext,
209 0 : ) -> Result<usize, PageReconstructError> {
210 0 : let mut total_blocks = 0;
211 :
212 0 : let rels = self.list_rels(spcnode, dbnode, version, ctx).await?;
213 :
214 0 : for rel in rels {
215 0 : let n_blocks = self.get_rel_size(rel, version, latest, ctx).await?;
216 0 : total_blocks += n_blocks as usize;
217 : }
218 0 : Ok(total_blocks)
219 0 : }
220 :
221 : /// Get size of a relation file
222 24434 : pub(crate) async fn get_rel_size(
223 24434 : &self,
224 24434 : tag: RelTag,
225 24434 : version: Version<'_>,
226 24434 : latest: bool,
227 24434 : ctx: &RequestContext,
228 24434 : ) -> Result<BlockNumber, PageReconstructError> {
229 24434 : if tag.relnode == 0 {
230 0 : return Err(PageReconstructError::Other(
231 0 : RelationError::InvalidRelnode.into(),
232 0 : ));
233 24434 : }
234 :
235 24434 : if let Some(nblocks) = self.get_cached_rel_size(&tag, version.get_lsn()) {
236 19294 : return Ok(nblocks);
237 5140 : }
238 5140 :
239 5140 : if (tag.forknum == FSM_FORKNUM || tag.forknum == VISIBILITYMAP_FORKNUM)
240 0 : && !self.get_rel_exists(tag, version, latest, ctx).await?
241 : {
242 : // FIXME: Postgres sometimes calls smgrcreate() to create
243 : // FSM, and smgrnblocks() on it immediately afterwards,
244 : // without extending it. Tolerate that by claiming that
245 : // any non-existent FSM fork has size 0.
246 0 : return Ok(0);
247 5140 : }
248 5140 :
249 5140 : let key = rel_size_to_key(tag);
250 5140 : let mut buf = version.get(self, key, ctx).await?;
251 5136 : let nblocks = buf.get_u32_le();
252 5136 :
253 5136 : if latest {
254 0 : // Update relation size cache only if "latest" flag is set.
255 0 : // This flag is set by compute when it is working with most recent version of relation.
256 0 : // Typically master compute node always set latest=true.
257 0 : // Please notice, that even if compute node "by mistake" specifies old LSN but set
258 0 : // latest=true, then it can not cause cache corruption, because with latest=true
259 0 : // pageserver choose max(request_lsn, last_written_lsn) and so cached value will be
260 0 : // associated with most recent value of LSN.
261 0 : self.update_cached_rel_size(tag, version.get_lsn(), nblocks);
262 5136 : }
263 5136 : Ok(nblocks)
264 24434 : }
265 :
266 : /// Does relation exist?
267 6050 : pub(crate) async fn get_rel_exists(
268 6050 : &self,
269 6050 : tag: RelTag,
270 6050 : version: Version<'_>,
271 6050 : _latest: bool,
272 6050 : ctx: &RequestContext,
273 6050 : ) -> Result<bool, PageReconstructError> {
274 6050 : if tag.relnode == 0 {
275 0 : return Err(PageReconstructError::Other(
276 0 : RelationError::InvalidRelnode.into(),
277 0 : ));
278 6050 : }
279 :
280 : // first try to lookup relation in cache
281 6050 : if let Some(_nblocks) = self.get_cached_rel_size(&tag, version.get_lsn()) {
282 6032 : return Ok(true);
283 18 : }
284 18 : // fetch directory listing
285 18 : let key = rel_dir_to_key(tag.spcnode, tag.dbnode);
286 18 : let buf = version.get(self, key, ctx).await?;
287 :
288 18 : match RelDirectory::des(&buf).context("deserialization failure") {
289 18 : Ok(dir) => {
290 18 : let exists = dir.rels.get(&(tag.relnode, tag.forknum)).is_some();
291 18 : Ok(exists)
292 : }
293 0 : Err(e) => Err(PageReconstructError::from(e)),
294 : }
295 6050 : }
296 :
297 : /// Get a list of all existing relations in given tablespace and database.
298 : ///
299 : /// # Cancel-Safety
300 : ///
301 : /// This method is cancellation-safe.
302 0 : pub(crate) async fn list_rels(
303 0 : &self,
304 0 : spcnode: Oid,
305 0 : dbnode: Oid,
306 0 : version: Version<'_>,
307 0 : ctx: &RequestContext,
308 0 : ) -> Result<HashSet<RelTag>, PageReconstructError> {
309 0 : // fetch directory listing
310 0 : let key = rel_dir_to_key(spcnode, dbnode);
311 0 : let buf = version.get(self, key, ctx).await?;
312 :
313 0 : match RelDirectory::des(&buf).context("deserialization failure") {
314 0 : Ok(dir) => {
315 0 : let rels: HashSet<RelTag> =
316 0 : HashSet::from_iter(dir.rels.iter().map(|(relnode, forknum)| RelTag {
317 0 : spcnode,
318 0 : dbnode,
319 0 : relnode: *relnode,
320 0 : forknum: *forknum,
321 0 : }));
322 0 :
323 0 : Ok(rels)
324 : }
325 0 : Err(e) => Err(PageReconstructError::from(e)),
326 : }
327 0 : }
328 :
329 : /// Get the whole SLRU segment
330 0 : pub(crate) async fn get_slru_segment(
331 0 : &self,
332 0 : kind: SlruKind,
333 0 : segno: u32,
334 0 : lsn: Lsn,
335 0 : ctx: &RequestContext,
336 0 : ) -> Result<Bytes, PageReconstructError> {
337 0 : let n_blocks = self
338 0 : .get_slru_segment_size(kind, segno, Version::Lsn(lsn), ctx)
339 0 : .await?;
340 0 : let mut segment = BytesMut::with_capacity(n_blocks as usize * BLCKSZ as usize);
341 0 : for blkno in 0..n_blocks {
342 0 : let block = self
343 0 : .get_slru_page_at_lsn(kind, segno, blkno, lsn, ctx)
344 0 : .await?;
345 0 : segment.extend_from_slice(&block[..BLCKSZ as usize]);
346 : }
347 0 : Ok(segment.freeze())
348 0 : }
349 :
350 : /// Look up given SLRU page version.
351 0 : pub(crate) async fn get_slru_page_at_lsn(
352 0 : &self,
353 0 : kind: SlruKind,
354 0 : segno: u32,
355 0 : blknum: BlockNumber,
356 0 : lsn: Lsn,
357 0 : ctx: &RequestContext,
358 0 : ) -> Result<Bytes, PageReconstructError> {
359 0 : let key = slru_block_to_key(kind, segno, blknum);
360 0 : self.get(key, lsn, ctx).await
361 0 : }
362 :
363 : /// Get size of an SLRU segment
364 0 : pub(crate) async fn get_slru_segment_size(
365 0 : &self,
366 0 : kind: SlruKind,
367 0 : segno: u32,
368 0 : version: Version<'_>,
369 0 : ctx: &RequestContext,
370 0 : ) -> Result<BlockNumber, PageReconstructError> {
371 0 : let key = slru_segment_size_to_key(kind, segno);
372 0 : let mut buf = version.get(self, key, ctx).await?;
373 0 : Ok(buf.get_u32_le())
374 0 : }
375 :
376 : /// Get size of an SLRU segment
377 0 : pub(crate) async fn get_slru_segment_exists(
378 0 : &self,
379 0 : kind: SlruKind,
380 0 : segno: u32,
381 0 : version: Version<'_>,
382 0 : ctx: &RequestContext,
383 0 : ) -> Result<bool, PageReconstructError> {
384 0 : // fetch directory listing
385 0 : let key = slru_dir_to_key(kind);
386 0 : let buf = version.get(self, key, ctx).await?;
387 :
388 0 : match SlruSegmentDirectory::des(&buf).context("deserialization failure") {
389 0 : Ok(dir) => {
390 0 : let exists = dir.segments.get(&segno).is_some();
391 0 : Ok(exists)
392 : }
393 0 : Err(e) => Err(PageReconstructError::from(e)),
394 : }
395 0 : }
396 :
397 : /// Locate LSN, such that all transactions that committed before
398 : /// 'search_timestamp' are visible, but nothing newer is.
399 : ///
400 : /// This is not exact. Commit timestamps are not guaranteed to be ordered,
401 : /// so it's not well defined which LSN you get if there were multiple commits
402 : /// "in flight" at that point in time.
403 : ///
404 0 : pub(crate) async fn find_lsn_for_timestamp(
405 0 : &self,
406 0 : search_timestamp: TimestampTz,
407 0 : cancel: &CancellationToken,
408 0 : ctx: &RequestContext,
409 0 : ) -> Result<LsnForTimestamp, PageReconstructError> {
410 0 : let gc_cutoff_lsn_guard = self.get_latest_gc_cutoff_lsn();
411 0 : // We use this method to figure out the branching LSN for the new branch, but the
412 0 : // GC cutoff could be before the branching point and we cannot create a new branch
413 0 : // with LSN < `ancestor_lsn`. Thus, pick the maximum of these two to be
414 0 : // on the safe side.
415 0 : let min_lsn = std::cmp::max(*gc_cutoff_lsn_guard, self.get_ancestor_lsn());
416 0 : let max_lsn = self.get_last_record_lsn();
417 0 :
418 0 : // LSNs are always 8-byte aligned. low/mid/high represent the
419 0 : // LSN divided by 8.
420 0 : let mut low = min_lsn.0 / 8;
421 0 : let mut high = max_lsn.0 / 8 + 1;
422 0 :
423 0 : let mut found_smaller = false;
424 0 : let mut found_larger = false;
425 0 : while low < high {
426 0 : if cancel.is_cancelled() {
427 0 : return Err(PageReconstructError::Cancelled);
428 0 : }
429 0 : // cannot overflow, high and low are both smaller than u64::MAX / 2
430 0 : let mid = (high + low) / 2;
431 :
432 0 : let cmp = self
433 0 : .is_latest_commit_timestamp_ge_than(
434 0 : search_timestamp,
435 0 : Lsn(mid * 8),
436 0 : &mut found_smaller,
437 0 : &mut found_larger,
438 0 : ctx,
439 0 : )
440 0 : .await?;
441 :
442 0 : if cmp {
443 0 : high = mid;
444 0 : } else {
445 0 : low = mid + 1;
446 0 : }
447 : }
448 : // If `found_smaller == true`, `low = t + 1` where `t` is the target LSN,
449 : // so the LSN of the last commit record before or at `search_timestamp`.
450 : // Remove one from `low` to get `t`.
451 : //
452 : // FIXME: it would be better to get the LSN of the previous commit.
453 : // Otherwise, if you restore to the returned LSN, the database will
454 : // include physical changes from later commits that will be marked
455 : // as aborted, and will need to be vacuumed away.
456 0 : let commit_lsn = Lsn((low - 1) * 8);
457 0 : match (found_smaller, found_larger) {
458 : (false, false) => {
459 : // This can happen if no commit records have been processed yet, e.g.
460 : // just after importing a cluster.
461 0 : Ok(LsnForTimestamp::NoData(min_lsn))
462 : }
463 : (false, true) => {
464 : // Didn't find any commit timestamps smaller than the request
465 0 : Ok(LsnForTimestamp::Past(min_lsn))
466 : }
467 : (true, false) => {
468 : // Only found commits with timestamps smaller than the request.
469 : // It's still a valid case for branch creation, return it.
470 : // And `update_gc_info()` ignores LSN for a `LsnForTimestamp::Future`
471 : // case, anyway.
472 0 : Ok(LsnForTimestamp::Future(commit_lsn))
473 : }
474 0 : (true, true) => Ok(LsnForTimestamp::Present(commit_lsn)),
475 : }
476 0 : }
477 :
478 : /// Subroutine of find_lsn_for_timestamp(). Returns true, if there are any
479 : /// commits that committed after 'search_timestamp', at LSN 'probe_lsn'.
480 : ///
481 : /// Additionally, sets 'found_smaller'/'found_Larger, if encounters any commits
482 : /// with a smaller/larger timestamp.
483 : ///
484 0 : pub(crate) async fn is_latest_commit_timestamp_ge_than(
485 0 : &self,
486 0 : search_timestamp: TimestampTz,
487 0 : probe_lsn: Lsn,
488 0 : found_smaller: &mut bool,
489 0 : found_larger: &mut bool,
490 0 : ctx: &RequestContext,
491 0 : ) -> Result<bool, PageReconstructError> {
492 0 : self.map_all_timestamps(probe_lsn, ctx, |timestamp| {
493 0 : if timestamp >= search_timestamp {
494 0 : *found_larger = true;
495 0 : return ControlFlow::Break(true);
496 0 : } else {
497 0 : *found_smaller = true;
498 0 : }
499 0 : ControlFlow::Continue(())
500 0 : })
501 0 : .await
502 0 : }
503 :
504 : /// Obtain the possible timestamp range for the given lsn.
505 : ///
506 : /// If the lsn has no timestamps, returns None. returns `(min, max, median)` if it has timestamps.
507 0 : pub(crate) async fn get_timestamp_for_lsn(
508 0 : &self,
509 0 : probe_lsn: Lsn,
510 0 : ctx: &RequestContext,
511 0 : ) -> Result<Option<TimestampTz>, PageReconstructError> {
512 0 : let mut max: Option<TimestampTz> = None;
513 0 : self.map_all_timestamps(probe_lsn, ctx, |timestamp| {
514 0 : if let Some(max_prev) = max {
515 0 : max = Some(max_prev.max(timestamp));
516 0 : } else {
517 0 : max = Some(timestamp);
518 0 : }
519 0 : ControlFlow::Continue(())
520 0 : })
521 0 : .await?;
522 :
523 0 : Ok(max)
524 0 : }
525 :
526 : /// Runs the given function on all the timestamps for a given lsn
527 : ///
528 : /// The return value is either given by the closure, or set to the `Default`
529 : /// impl's output.
530 0 : async fn map_all_timestamps<T: Default>(
531 0 : &self,
532 0 : probe_lsn: Lsn,
533 0 : ctx: &RequestContext,
534 0 : mut f: impl FnMut(TimestampTz) -> ControlFlow<T>,
535 0 : ) -> Result<T, PageReconstructError> {
536 0 : for segno in self
537 0 : .list_slru_segments(SlruKind::Clog, Version::Lsn(probe_lsn), ctx)
538 0 : .await?
539 : {
540 0 : let nblocks = self
541 0 : .get_slru_segment_size(SlruKind::Clog, segno, Version::Lsn(probe_lsn), ctx)
542 0 : .await?;
543 0 : for blknum in (0..nblocks).rev() {
544 0 : let clog_page = self
545 0 : .get_slru_page_at_lsn(SlruKind::Clog, segno, blknum, probe_lsn, ctx)
546 0 : .await?;
547 :
548 0 : if clog_page.len() == BLCKSZ as usize + 8 {
549 0 : let mut timestamp_bytes = [0u8; 8];
550 0 : timestamp_bytes.copy_from_slice(&clog_page[BLCKSZ as usize..]);
551 0 : let timestamp = TimestampTz::from_be_bytes(timestamp_bytes);
552 0 :
553 0 : match f(timestamp) {
554 0 : ControlFlow::Break(b) => return Ok(b),
555 0 : ControlFlow::Continue(()) => (),
556 : }
557 0 : }
558 : }
559 : }
560 0 : Ok(Default::default())
561 0 : }
562 :
563 0 : pub(crate) async fn get_slru_keyspace(
564 0 : &self,
565 0 : version: Version<'_>,
566 0 : ctx: &RequestContext,
567 0 : ) -> Result<KeySpace, PageReconstructError> {
568 0 : let mut accum = KeySpaceAccum::new();
569 :
570 0 : for kind in SlruKind::iter() {
571 0 : let mut segments: Vec<u32> = self
572 0 : .list_slru_segments(kind, version, ctx)
573 0 : .await?
574 0 : .into_iter()
575 0 : .collect();
576 0 : segments.sort_unstable();
577 :
578 0 : for seg in segments {
579 0 : let block_count = self.get_slru_segment_size(kind, seg, version, ctx).await?;
580 :
581 0 : accum.add_range(
582 0 : slru_block_to_key(kind, seg, 0)..slru_block_to_key(kind, seg, block_count),
583 0 : );
584 : }
585 : }
586 :
587 0 : Ok(accum.to_keyspace())
588 0 : }
589 :
590 : /// Get a list of SLRU segments
591 0 : pub(crate) async fn list_slru_segments(
592 0 : &self,
593 0 : kind: SlruKind,
594 0 : version: Version<'_>,
595 0 : ctx: &RequestContext,
596 0 : ) -> Result<HashSet<u32>, PageReconstructError> {
597 0 : // fetch directory entry
598 0 : let key = slru_dir_to_key(kind);
599 :
600 0 : let buf = version.get(self, key, ctx).await?;
601 0 : match SlruSegmentDirectory::des(&buf).context("deserialization failure") {
602 0 : Ok(dir) => Ok(dir.segments),
603 0 : Err(e) => Err(PageReconstructError::from(e)),
604 : }
605 0 : }
606 :
607 0 : pub(crate) async fn get_relmap_file(
608 0 : &self,
609 0 : spcnode: Oid,
610 0 : dbnode: Oid,
611 0 : version: Version<'_>,
612 0 : ctx: &RequestContext,
613 0 : ) -> Result<Bytes, PageReconstructError> {
614 0 : let key = relmap_file_key(spcnode, dbnode);
615 :
616 0 : let buf = version.get(self, key, ctx).await?;
617 0 : Ok(buf)
618 0 : }
619 :
620 0 : pub(crate) async fn list_dbdirs(
621 0 : &self,
622 0 : lsn: Lsn,
623 0 : ctx: &RequestContext,
624 0 : ) -> Result<HashMap<(Oid, Oid), bool>, PageReconstructError> {
625 : // fetch directory entry
626 0 : let buf = self.get(DBDIR_KEY, lsn, ctx).await?;
627 :
628 0 : match DbDirectory::des(&buf).context("deserialization failure") {
629 0 : Ok(dir) => Ok(dir.dbdirs),
630 0 : Err(e) => Err(PageReconstructError::from(e)),
631 : }
632 0 : }
633 :
634 0 : pub(crate) async fn get_twophase_file(
635 0 : &self,
636 0 : xid: TransactionId,
637 0 : lsn: Lsn,
638 0 : ctx: &RequestContext,
639 0 : ) -> Result<Bytes, PageReconstructError> {
640 0 : let key = twophase_file_key(xid);
641 0 : let buf = self.get(key, lsn, ctx).await?;
642 0 : Ok(buf)
643 0 : }
644 :
645 0 : pub(crate) async fn list_twophase_files(
646 0 : &self,
647 0 : lsn: Lsn,
648 0 : ctx: &RequestContext,
649 0 : ) -> Result<HashSet<TransactionId>, PageReconstructError> {
650 : // fetch directory entry
651 0 : let buf = self.get(TWOPHASEDIR_KEY, lsn, ctx).await?;
652 :
653 0 : match TwoPhaseDirectory::des(&buf).context("deserialization failure") {
654 0 : Ok(dir) => Ok(dir.xids),
655 0 : Err(e) => Err(PageReconstructError::from(e)),
656 : }
657 0 : }
658 :
659 0 : pub(crate) async fn get_control_file(
660 0 : &self,
661 0 : lsn: Lsn,
662 0 : ctx: &RequestContext,
663 0 : ) -> Result<Bytes, PageReconstructError> {
664 0 : self.get(CONTROLFILE_KEY, lsn, ctx).await
665 0 : }
666 :
667 12 : pub(crate) async fn get_checkpoint(
668 12 : &self,
669 12 : lsn: Lsn,
670 12 : ctx: &RequestContext,
671 12 : ) -> Result<Bytes, PageReconstructError> {
672 12 : self.get(CHECKPOINT_KEY, lsn, ctx).await
673 12 : }
674 :
675 6 : pub(crate) async fn list_aux_files(
676 6 : &self,
677 6 : lsn: Lsn,
678 6 : ctx: &RequestContext,
679 6 : ) -> Result<HashMap<String, Bytes>, PageReconstructError> {
680 6 : match self.get(AUX_FILES_KEY, lsn, ctx).await {
681 6 : Ok(buf) => match AuxFilesDirectory::des(&buf).context("deserialization failure") {
682 6 : Ok(dir) => Ok(dir.files),
683 0 : Err(e) => Err(PageReconstructError::from(e)),
684 : },
685 0 : Err(e) => {
686 : // This is expected: historical databases do not have the key.
687 0 : debug!("Failed to get info about AUX files: {}", e);
688 0 : Ok(HashMap::new())
689 : }
690 : }
691 6 : }
692 :
693 : /// Does the same as get_current_logical_size but counted on demand.
694 : /// Used to initialize the logical size tracking on startup.
695 : ///
696 : /// Only relation blocks are counted currently. That excludes metadata,
697 : /// SLRUs, twophase files etc.
698 : ///
699 : /// # Cancel-Safety
700 : ///
701 : /// This method is cancellation-safe.
702 0 : pub async fn get_current_logical_size_non_incremental(
703 0 : &self,
704 0 : lsn: Lsn,
705 0 : ctx: &RequestContext,
706 0 : ) -> Result<u64, CalculateLogicalSizeError> {
707 0 : debug_assert_current_span_has_tenant_and_timeline_id_no_shard_id();
708 :
709 : // Fetch list of database dirs and iterate them
710 0 : let buf = self.get(DBDIR_KEY, lsn, ctx).await?;
711 0 : let dbdir = DbDirectory::des(&buf).context("deserialize db directory")?;
712 :
713 0 : let mut total_size: u64 = 0;
714 0 : for (spcnode, dbnode) in dbdir.dbdirs.keys() {
715 0 : for rel in self
716 0 : .list_rels(*spcnode, *dbnode, Version::Lsn(lsn), ctx)
717 0 : .await?
718 : {
719 0 : if self.cancel.is_cancelled() {
720 0 : return Err(CalculateLogicalSizeError::Cancelled);
721 0 : }
722 0 : let relsize_key = rel_size_to_key(rel);
723 0 : let mut buf = self.get(relsize_key, lsn, ctx).await?;
724 0 : let relsize = buf.get_u32_le();
725 0 :
726 0 : total_size += relsize as u64;
727 : }
728 : }
729 0 : Ok(total_size * BLCKSZ as u64)
730 0 : }
731 :
732 : ///
733 : /// Get a KeySpace that covers all the Keys that are in use at the given LSN.
734 : /// Anything that's not listed maybe removed from the underlying storage (from
735 : /// that LSN forwards).
736 174 : pub(crate) async fn collect_keyspace(
737 174 : &self,
738 174 : lsn: Lsn,
739 174 : ctx: &RequestContext,
740 174 : ) -> Result<KeySpace, CollectKeySpaceError> {
741 174 : // Iterate through key ranges, greedily packing them into partitions
742 174 : let mut result = KeySpaceAccum::new();
743 174 :
744 174 : // The dbdir metadata always exists
745 174 : result.add_key(DBDIR_KEY);
746 :
747 : // Fetch list of database dirs and iterate them
748 2205 : let buf = self.get(DBDIR_KEY, lsn, ctx).await?;
749 174 : let dbdir = DbDirectory::des(&buf)?;
750 :
751 174 : let mut dbs: Vec<(Oid, Oid)> = dbdir.dbdirs.keys().cloned().collect();
752 174 : dbs.sort_unstable();
753 174 : for (spcnode, dbnode) in dbs {
754 0 : result.add_key(relmap_file_key(spcnode, dbnode));
755 0 : result.add_key(rel_dir_to_key(spcnode, dbnode));
756 :
757 0 : let mut rels: Vec<RelTag> = self
758 0 : .list_rels(spcnode, dbnode, Version::Lsn(lsn), ctx)
759 0 : .await?
760 0 : .into_iter()
761 0 : .collect();
762 0 : rels.sort_unstable();
763 0 : for rel in rels {
764 0 : let relsize_key = rel_size_to_key(rel);
765 0 : let mut buf = self.get(relsize_key, lsn, ctx).await?;
766 0 : let relsize = buf.get_u32_le();
767 0 :
768 0 : result.add_range(rel_block_to_key(rel, 0)..rel_block_to_key(rel, relsize));
769 0 : result.add_key(relsize_key);
770 : }
771 : }
772 :
773 : // Iterate SLRUs next
774 522 : for kind in [
775 174 : SlruKind::Clog,
776 174 : SlruKind::MultiXactMembers,
777 174 : SlruKind::MultiXactOffsets,
778 : ] {
779 522 : let slrudir_key = slru_dir_to_key(kind);
780 522 : result.add_key(slrudir_key);
781 7978 : let buf = self.get(slrudir_key, lsn, ctx).await?;
782 522 : let dir = SlruSegmentDirectory::des(&buf)?;
783 522 : let mut segments: Vec<u32> = dir.segments.iter().cloned().collect();
784 522 : segments.sort_unstable();
785 522 : for segno in segments {
786 0 : let segsize_key = slru_segment_size_to_key(kind, segno);
787 0 : let mut buf = self.get(segsize_key, lsn, ctx).await?;
788 0 : let segsize = buf.get_u32_le();
789 0 :
790 0 : result.add_range(
791 0 : slru_block_to_key(kind, segno, 0)..slru_block_to_key(kind, segno, segsize),
792 0 : );
793 0 : result.add_key(segsize_key);
794 : }
795 : }
796 :
797 : // Then pg_twophase
798 174 : result.add_key(TWOPHASEDIR_KEY);
799 3050 : let buf = self.get(TWOPHASEDIR_KEY, lsn, ctx).await?;
800 174 : let twophase_dir = TwoPhaseDirectory::des(&buf)?;
801 174 : let mut xids: Vec<TransactionId> = twophase_dir.xids.iter().cloned().collect();
802 174 : xids.sort_unstable();
803 174 : for xid in xids {
804 0 : result.add_key(twophase_file_key(xid));
805 0 : }
806 :
807 174 : result.add_key(CONTROLFILE_KEY);
808 174 : result.add_key(CHECKPOINT_KEY);
809 174 : if self.get(AUX_FILES_KEY, lsn, ctx).await.is_ok() {
810 74 : result.add_key(AUX_FILES_KEY);
811 100 : }
812 174 : Ok(result.to_keyspace())
813 174 : }
814 :
815 : /// Get cached size of relation if it not updated after specified LSN
816 448540 : pub fn get_cached_rel_size(&self, tag: &RelTag, lsn: Lsn) -> Option<BlockNumber> {
817 448540 : let rel_size_cache = self.rel_size_cache.read().unwrap();
818 448540 : if let Some((cached_lsn, nblocks)) = rel_size_cache.get(tag) {
819 448518 : if lsn >= *cached_lsn {
820 443372 : return Some(*nblocks);
821 5146 : }
822 22 : }
823 5168 : None
824 448540 : }
825 :
826 : /// Update cached relation size if there is no more recent update
827 0 : pub fn update_cached_rel_size(&self, tag: RelTag, lsn: Lsn, nblocks: BlockNumber) {
828 0 : let mut rel_size_cache = self.rel_size_cache.write().unwrap();
829 0 : match rel_size_cache.entry(tag) {
830 0 : hash_map::Entry::Occupied(mut entry) => {
831 0 : let cached_lsn = entry.get_mut();
832 0 : if lsn >= cached_lsn.0 {
833 0 : *cached_lsn = (lsn, nblocks);
834 0 : }
835 : }
836 0 : hash_map::Entry::Vacant(entry) => {
837 0 : entry.insert((lsn, nblocks));
838 0 : }
839 : }
840 0 : }
841 :
842 : /// Store cached relation size
843 288732 : pub fn set_cached_rel_size(&self, tag: RelTag, lsn: Lsn, nblocks: BlockNumber) {
844 288732 : let mut rel_size_cache = self.rel_size_cache.write().unwrap();
845 288732 : rel_size_cache.insert(tag, (lsn, nblocks));
846 288732 : }
847 :
848 : /// Remove cached relation size
849 2 : pub fn remove_cached_rel_size(&self, tag: &RelTag) {
850 2 : let mut rel_size_cache = self.rel_size_cache.write().unwrap();
851 2 : rel_size_cache.remove(tag);
852 2 : }
853 : }
854 :
855 : /// DatadirModification represents an operation to ingest an atomic set of
856 : /// updates to the repository. It is created by the 'begin_record'
857 : /// function. It is called for each WAL record, so that all the modifications
858 : /// by a one WAL record appear atomic.
859 : pub struct DatadirModification<'a> {
860 : /// The timeline this modification applies to. You can access this to
861 : /// read the state, but note that any pending updates are *not* reflected
862 : /// in the state in 'tline' yet.
863 : pub tline: &'a Timeline,
864 :
865 : /// Current LSN of the modification
866 : lsn: Lsn,
867 :
868 : // The modifications are not applied directly to the underlying key-value store.
869 : // The put-functions add the modifications here, and they are flushed to the
870 : // underlying key-value store by the 'finish' function.
871 : pending_lsns: Vec<Lsn>,
872 : pending_updates: HashMap<Key, Vec<(Lsn, Value)>>,
873 : pending_deletions: Vec<(Range<Key>, Lsn)>,
874 : pending_nblocks: i64,
875 :
876 : /// For special "directory" keys that store key-value maps, track the size of the map
877 : /// if it was updated in this modification.
878 : pending_directory_entries: Vec<(DirectoryKind, usize)>,
879 : }
880 :
881 : impl<'a> DatadirModification<'a> {
882 : /// Get the current lsn
883 418056 : pub(crate) fn get_lsn(&self) -> Lsn {
884 418056 : self.lsn
885 418056 : }
886 :
887 : /// Set the current lsn
888 145856 : pub(crate) fn set_lsn(&mut self, lsn: Lsn) -> anyhow::Result<()> {
889 145856 : ensure!(
890 145856 : lsn >= self.lsn,
891 0 : "setting an older lsn {} than {} is not allowed",
892 : lsn,
893 : self.lsn
894 : );
895 145856 : if lsn > self.lsn {
896 145856 : self.pending_lsns.push(self.lsn);
897 145856 : self.lsn = lsn;
898 145856 : }
899 145856 : Ok(())
900 145856 : }
901 :
902 : /// Initialize a completely new repository.
903 : ///
904 : /// This inserts the directory metadata entries that are assumed to
905 : /// always exist.
906 76 : pub fn init_empty(&mut self) -> anyhow::Result<()> {
907 76 : let buf = DbDirectory::ser(&DbDirectory {
908 76 : dbdirs: HashMap::new(),
909 76 : })?;
910 76 : self.pending_directory_entries.push((DirectoryKind::Db, 0));
911 76 : self.put(DBDIR_KEY, Value::Image(buf.into()));
912 76 :
913 76 : // Create AuxFilesDirectory
914 76 : self.init_aux_dir()?;
915 :
916 76 : let buf = TwoPhaseDirectory::ser(&TwoPhaseDirectory {
917 76 : xids: HashSet::new(),
918 76 : })?;
919 76 : self.pending_directory_entries
920 76 : .push((DirectoryKind::TwoPhase, 0));
921 76 : self.put(TWOPHASEDIR_KEY, Value::Image(buf.into()));
922 :
923 76 : let buf: Bytes = SlruSegmentDirectory::ser(&SlruSegmentDirectory::default())?.into();
924 76 : let empty_dir = Value::Image(buf);
925 76 : self.put(slru_dir_to_key(SlruKind::Clog), empty_dir.clone());
926 76 : self.pending_directory_entries
927 76 : .push((DirectoryKind::SlruSegment(SlruKind::Clog), 0));
928 76 : self.put(
929 76 : slru_dir_to_key(SlruKind::MultiXactMembers),
930 76 : empty_dir.clone(),
931 76 : );
932 76 : self.pending_directory_entries
933 76 : .push((DirectoryKind::SlruSegment(SlruKind::Clog), 0));
934 76 : self.put(slru_dir_to_key(SlruKind::MultiXactOffsets), empty_dir);
935 76 : self.pending_directory_entries
936 76 : .push((DirectoryKind::SlruSegment(SlruKind::MultiXactOffsets), 0));
937 76 :
938 76 : Ok(())
939 76 : }
940 :
941 : #[cfg(test)]
942 74 : pub fn init_empty_test_timeline(&mut self) -> anyhow::Result<()> {
943 74 : self.init_empty()?;
944 74 : self.put_control_file(bytes::Bytes::from_static(
945 74 : b"control_file contents do not matter",
946 74 : ))
947 74 : .context("put_control_file")?;
948 74 : self.put_checkpoint(bytes::Bytes::from_static(
949 74 : b"checkpoint_file contents do not matter",
950 74 : ))
951 74 : .context("put_checkpoint_file")?;
952 74 : Ok(())
953 74 : }
954 :
955 : /// Put a new page version that can be constructed from a WAL record
956 : ///
957 : /// NOTE: this will *not* implicitly extend the relation, if the page is beyond the
958 : /// current end-of-file. It's up to the caller to check that the relation size
959 : /// matches the blocks inserted!
960 145630 : pub fn put_rel_wal_record(
961 145630 : &mut self,
962 145630 : rel: RelTag,
963 145630 : blknum: BlockNumber,
964 145630 : rec: NeonWalRecord,
965 145630 : ) -> anyhow::Result<()> {
966 145630 : anyhow::ensure!(rel.relnode != 0, RelationError::InvalidRelnode);
967 145630 : self.put(rel_block_to_key(rel, blknum), Value::WalRecord(rec));
968 145630 : Ok(())
969 145630 : }
970 :
971 : // Same, but for an SLRU.
972 8 : pub fn put_slru_wal_record(
973 8 : &mut self,
974 8 : kind: SlruKind,
975 8 : segno: u32,
976 8 : blknum: BlockNumber,
977 8 : rec: NeonWalRecord,
978 8 : ) -> anyhow::Result<()> {
979 8 : self.put(
980 8 : slru_block_to_key(kind, segno, blknum),
981 8 : Value::WalRecord(rec),
982 8 : );
983 8 : Ok(())
984 8 : }
985 :
986 : /// Like put_wal_record, but with ready-made image of the page.
987 280864 : pub fn put_rel_page_image(
988 280864 : &mut self,
989 280864 : rel: RelTag,
990 280864 : blknum: BlockNumber,
991 280864 : img: Bytes,
992 280864 : ) -> anyhow::Result<()> {
993 280864 : anyhow::ensure!(rel.relnode != 0, RelationError::InvalidRelnode);
994 280864 : self.put(rel_block_to_key(rel, blknum), Value::Image(img));
995 280864 : Ok(())
996 280864 : }
997 :
998 6 : pub fn put_slru_page_image(
999 6 : &mut self,
1000 6 : kind: SlruKind,
1001 6 : segno: u32,
1002 6 : blknum: BlockNumber,
1003 6 : img: Bytes,
1004 6 : ) -> anyhow::Result<()> {
1005 6 : self.put(slru_block_to_key(kind, segno, blknum), Value::Image(img));
1006 6 : Ok(())
1007 6 : }
1008 :
1009 : /// Store a relmapper file (pg_filenode.map) in the repository
1010 16 : pub async fn put_relmap_file(
1011 16 : &mut self,
1012 16 : spcnode: Oid,
1013 16 : dbnode: Oid,
1014 16 : img: Bytes,
1015 16 : ctx: &RequestContext,
1016 16 : ) -> anyhow::Result<()> {
1017 : // Add it to the directory (if it doesn't exist already)
1018 16 : let buf = self.get(DBDIR_KEY, ctx).await?;
1019 16 : let mut dbdir = DbDirectory::des(&buf)?;
1020 :
1021 16 : let r = dbdir.dbdirs.insert((spcnode, dbnode), true);
1022 16 : if r.is_none() || r == Some(false) {
1023 : // The dbdir entry didn't exist, or it contained a
1024 : // 'false'. The 'insert' call already updated it with
1025 : // 'true', now write the updated 'dbdirs' map back.
1026 16 : let buf = DbDirectory::ser(&dbdir)?;
1027 16 : self.put(DBDIR_KEY, Value::Image(buf.into()));
1028 16 :
1029 16 : // Create AuxFilesDirectory as well
1030 16 : self.init_aux_dir()?;
1031 0 : }
1032 16 : if r.is_none() {
1033 8 : // Create RelDirectory
1034 8 : let buf = RelDirectory::ser(&RelDirectory {
1035 8 : rels: HashSet::new(),
1036 8 : })?;
1037 8 : self.pending_directory_entries.push((DirectoryKind::Rel, 0));
1038 8 : self.put(
1039 8 : rel_dir_to_key(spcnode, dbnode),
1040 8 : Value::Image(Bytes::from(buf)),
1041 8 : );
1042 8 : }
1043 :
1044 16 : self.put(relmap_file_key(spcnode, dbnode), Value::Image(img));
1045 16 : Ok(())
1046 16 : }
1047 :
1048 0 : pub async fn put_twophase_file(
1049 0 : &mut self,
1050 0 : xid: TransactionId,
1051 0 : img: Bytes,
1052 0 : ctx: &RequestContext,
1053 0 : ) -> anyhow::Result<()> {
1054 : // Add it to the directory entry
1055 0 : let buf = self.get(TWOPHASEDIR_KEY, ctx).await?;
1056 0 : let mut dir = TwoPhaseDirectory::des(&buf)?;
1057 0 : if !dir.xids.insert(xid) {
1058 0 : anyhow::bail!("twophase file for xid {} already exists", xid);
1059 0 : }
1060 0 : self.pending_directory_entries
1061 0 : .push((DirectoryKind::TwoPhase, dir.xids.len()));
1062 0 : self.put(
1063 0 : TWOPHASEDIR_KEY,
1064 0 : Value::Image(Bytes::from(TwoPhaseDirectory::ser(&dir)?)),
1065 : );
1066 :
1067 0 : self.put(twophase_file_key(xid), Value::Image(img));
1068 0 : Ok(())
1069 0 : }
1070 :
1071 76 : pub fn put_control_file(&mut self, img: Bytes) -> anyhow::Result<()> {
1072 76 : self.put(CONTROLFILE_KEY, Value::Image(img));
1073 76 : Ok(())
1074 76 : }
1075 :
1076 90 : pub fn put_checkpoint(&mut self, img: Bytes) -> anyhow::Result<()> {
1077 90 : self.put(CHECKPOINT_KEY, Value::Image(img));
1078 90 : Ok(())
1079 90 : }
1080 :
1081 0 : pub async fn drop_dbdir(
1082 0 : &mut self,
1083 0 : spcnode: Oid,
1084 0 : dbnode: Oid,
1085 0 : ctx: &RequestContext,
1086 0 : ) -> anyhow::Result<()> {
1087 0 : let total_blocks = self
1088 0 : .tline
1089 0 : .get_db_size(spcnode, dbnode, Version::Modified(self), true, ctx)
1090 0 : .await?;
1091 :
1092 : // Remove entry from dbdir
1093 0 : let buf = self.get(DBDIR_KEY, ctx).await?;
1094 0 : let mut dir = DbDirectory::des(&buf)?;
1095 0 : if dir.dbdirs.remove(&(spcnode, dbnode)).is_some() {
1096 0 : let buf = DbDirectory::ser(&dir)?;
1097 0 : self.pending_directory_entries
1098 0 : .push((DirectoryKind::Db, dir.dbdirs.len()));
1099 0 : self.put(DBDIR_KEY, Value::Image(buf.into()));
1100 : } else {
1101 0 : warn!(
1102 0 : "dropped dbdir for spcnode {} dbnode {} did not exist in db directory",
1103 0 : spcnode, dbnode
1104 0 : );
1105 : }
1106 :
1107 : // Update logical database size.
1108 0 : self.pending_nblocks -= total_blocks as i64;
1109 0 :
1110 0 : // Delete all relations and metadata files for the spcnode/dnode
1111 0 : self.delete(dbdir_key_range(spcnode, dbnode));
1112 0 : Ok(())
1113 0 : }
1114 :
1115 : /// Create a relation fork.
1116 : ///
1117 : /// 'nblocks' is the initial size.
1118 1920 : pub async fn put_rel_creation(
1119 1920 : &mut self,
1120 1920 : rel: RelTag,
1121 1920 : nblocks: BlockNumber,
1122 1920 : ctx: &RequestContext,
1123 1920 : ) -> Result<(), RelationError> {
1124 1920 : if rel.relnode == 0 {
1125 0 : return Err(RelationError::InvalidRelnode);
1126 1920 : }
1127 : // It's possible that this is the first rel for this db in this
1128 : // tablespace. Create the reldir entry for it if so.
1129 1920 : let mut dbdir = DbDirectory::des(&self.get(DBDIR_KEY, ctx).await.context("read db")?)
1130 1920 : .context("deserialize db")?;
1131 1920 : let rel_dir_key = rel_dir_to_key(rel.spcnode, rel.dbnode);
1132 1920 : let mut rel_dir = if dbdir.dbdirs.get(&(rel.spcnode, rel.dbnode)).is_none() {
1133 : // Didn't exist. Update dbdir
1134 8 : dbdir.dbdirs.insert((rel.spcnode, rel.dbnode), false);
1135 8 : let buf = DbDirectory::ser(&dbdir).context("serialize db")?;
1136 8 : self.pending_directory_entries
1137 8 : .push((DirectoryKind::Db, dbdir.dbdirs.len()));
1138 8 : self.put(DBDIR_KEY, Value::Image(buf.into()));
1139 8 :
1140 8 : // and create the RelDirectory
1141 8 : RelDirectory::default()
1142 : } else {
1143 : // reldir already exists, fetch it
1144 1912 : RelDirectory::des(&self.get(rel_dir_key, ctx).await.context("read db")?)
1145 1912 : .context("deserialize db")?
1146 : };
1147 :
1148 : // Add the new relation to the rel directory entry, and write it back
1149 1920 : if !rel_dir.rels.insert((rel.relnode, rel.forknum)) {
1150 0 : return Err(RelationError::AlreadyExists);
1151 1920 : }
1152 1920 :
1153 1920 : self.pending_directory_entries
1154 1920 : .push((DirectoryKind::Rel, rel_dir.rels.len()));
1155 1920 :
1156 1920 : self.put(
1157 1920 : rel_dir_key,
1158 1920 : Value::Image(Bytes::from(
1159 1920 : RelDirectory::ser(&rel_dir).context("serialize")?,
1160 : )),
1161 : );
1162 :
1163 : // Put size
1164 1920 : let size_key = rel_size_to_key(rel);
1165 1920 : let buf = nblocks.to_le_bytes();
1166 1920 : self.put(size_key, Value::Image(Bytes::from(buf.to_vec())));
1167 1920 :
1168 1920 : self.pending_nblocks += nblocks as i64;
1169 1920 :
1170 1920 : // Update relation size cache
1171 1920 : self.tline.set_cached_rel_size(rel, self.lsn, nblocks);
1172 1920 :
1173 1920 : // Even if nblocks > 0, we don't insert any actual blocks here. That's up to the
1174 1920 : // caller.
1175 1920 : Ok(())
1176 1920 : }
1177 :
1178 : /// Truncate relation
1179 6012 : pub async fn put_rel_truncation(
1180 6012 : &mut self,
1181 6012 : rel: RelTag,
1182 6012 : nblocks: BlockNumber,
1183 6012 : ctx: &RequestContext,
1184 6012 : ) -> anyhow::Result<()> {
1185 6012 : anyhow::ensure!(rel.relnode != 0, RelationError::InvalidRelnode);
1186 6012 : if self
1187 6012 : .tline
1188 6012 : .get_rel_exists(rel, Version::Modified(self), true, ctx)
1189 0 : .await?
1190 : {
1191 6012 : let size_key = rel_size_to_key(rel);
1192 : // Fetch the old size first
1193 6012 : let old_size = self.get(size_key, ctx).await?.get_u32_le();
1194 6012 :
1195 6012 : // Update the entry with the new size.
1196 6012 : let buf = nblocks.to_le_bytes();
1197 6012 : self.put(size_key, Value::Image(Bytes::from(buf.to_vec())));
1198 6012 :
1199 6012 : // Update relation size cache
1200 6012 : self.tline.set_cached_rel_size(rel, self.lsn, nblocks);
1201 6012 :
1202 6012 : // Update relation size cache
1203 6012 : self.tline.set_cached_rel_size(rel, self.lsn, nblocks);
1204 6012 :
1205 6012 : // Update logical database size.
1206 6012 : self.pending_nblocks -= old_size as i64 - nblocks as i64;
1207 0 : }
1208 6012 : Ok(())
1209 6012 : }
1210 :
1211 : /// Extend relation
1212 : /// If new size is smaller, do nothing.
1213 276680 : pub async fn put_rel_extend(
1214 276680 : &mut self,
1215 276680 : rel: RelTag,
1216 276680 : nblocks: BlockNumber,
1217 276680 : ctx: &RequestContext,
1218 276680 : ) -> anyhow::Result<()> {
1219 276680 : anyhow::ensure!(rel.relnode != 0, RelationError::InvalidRelnode);
1220 :
1221 : // Put size
1222 276680 : let size_key = rel_size_to_key(rel);
1223 276680 : let old_size = self.get(size_key, ctx).await?.get_u32_le();
1224 276680 :
1225 276680 : // only extend relation here. never decrease the size
1226 276680 : if nblocks > old_size {
1227 274788 : let buf = nblocks.to_le_bytes();
1228 274788 : self.put(size_key, Value::Image(Bytes::from(buf.to_vec())));
1229 274788 :
1230 274788 : // Update relation size cache
1231 274788 : self.tline.set_cached_rel_size(rel, self.lsn, nblocks);
1232 274788 :
1233 274788 : self.pending_nblocks += nblocks as i64 - old_size as i64;
1234 274788 : }
1235 276680 : Ok(())
1236 276680 : }
1237 :
1238 : /// Drop a relation.
1239 2 : pub async fn put_rel_drop(&mut self, rel: RelTag, ctx: &RequestContext) -> anyhow::Result<()> {
1240 2 : anyhow::ensure!(rel.relnode != 0, RelationError::InvalidRelnode);
1241 :
1242 : // Remove it from the directory entry
1243 2 : let dir_key = rel_dir_to_key(rel.spcnode, rel.dbnode);
1244 2 : let buf = self.get(dir_key, ctx).await?;
1245 2 : let mut dir = RelDirectory::des(&buf)?;
1246 :
1247 2 : self.pending_directory_entries
1248 2 : .push((DirectoryKind::Rel, dir.rels.len()));
1249 2 :
1250 2 : if dir.rels.remove(&(rel.relnode, rel.forknum)) {
1251 2 : self.put(dir_key, Value::Image(Bytes::from(RelDirectory::ser(&dir)?)));
1252 : } else {
1253 0 : warn!("dropped rel {} did not exist in rel directory", rel);
1254 : }
1255 :
1256 : // update logical size
1257 2 : let size_key = rel_size_to_key(rel);
1258 2 : let old_size = self.get(size_key, ctx).await?.get_u32_le();
1259 2 : self.pending_nblocks -= old_size as i64;
1260 2 :
1261 2 : // Remove enty from relation size cache
1262 2 : self.tline.remove_cached_rel_size(&rel);
1263 2 :
1264 2 : // Delete size entry, as well as all blocks
1265 2 : self.delete(rel_key_range(rel));
1266 2 :
1267 2 : Ok(())
1268 2 : }
1269 :
1270 6 : pub async fn put_slru_segment_creation(
1271 6 : &mut self,
1272 6 : kind: SlruKind,
1273 6 : segno: u32,
1274 6 : nblocks: BlockNumber,
1275 6 : ctx: &RequestContext,
1276 6 : ) -> anyhow::Result<()> {
1277 6 : // Add it to the directory entry
1278 6 : let dir_key = slru_dir_to_key(kind);
1279 6 : let buf = self.get(dir_key, ctx).await?;
1280 6 : let mut dir = SlruSegmentDirectory::des(&buf)?;
1281 :
1282 6 : if !dir.segments.insert(segno) {
1283 0 : anyhow::bail!("slru segment {kind:?}/{segno} already exists");
1284 6 : }
1285 6 : self.pending_directory_entries
1286 6 : .push((DirectoryKind::SlruSegment(kind), dir.segments.len()));
1287 6 : self.put(
1288 6 : dir_key,
1289 6 : Value::Image(Bytes::from(SlruSegmentDirectory::ser(&dir)?)),
1290 : );
1291 :
1292 : // Put size
1293 6 : let size_key = slru_segment_size_to_key(kind, segno);
1294 6 : let buf = nblocks.to_le_bytes();
1295 6 : self.put(size_key, Value::Image(Bytes::from(buf.to_vec())));
1296 6 :
1297 6 : // even if nblocks > 0, we don't insert any actual blocks here
1298 6 :
1299 6 : Ok(())
1300 6 : }
1301 :
1302 : /// Extend SLRU segment
1303 0 : pub fn put_slru_extend(
1304 0 : &mut self,
1305 0 : kind: SlruKind,
1306 0 : segno: u32,
1307 0 : nblocks: BlockNumber,
1308 0 : ) -> anyhow::Result<()> {
1309 0 : // Put size
1310 0 : let size_key = slru_segment_size_to_key(kind, segno);
1311 0 : let buf = nblocks.to_le_bytes();
1312 0 : self.put(size_key, Value::Image(Bytes::from(buf.to_vec())));
1313 0 : Ok(())
1314 0 : }
1315 :
1316 : /// This method is used for marking truncated SLRU files
1317 0 : pub async fn drop_slru_segment(
1318 0 : &mut self,
1319 0 : kind: SlruKind,
1320 0 : segno: u32,
1321 0 : ctx: &RequestContext,
1322 0 : ) -> anyhow::Result<()> {
1323 0 : // Remove it from the directory entry
1324 0 : let dir_key = slru_dir_to_key(kind);
1325 0 : let buf = self.get(dir_key, ctx).await?;
1326 0 : let mut dir = SlruSegmentDirectory::des(&buf)?;
1327 :
1328 0 : if !dir.segments.remove(&segno) {
1329 0 : warn!("slru segment {:?}/{} does not exist", kind, segno);
1330 0 : }
1331 0 : self.pending_directory_entries
1332 0 : .push((DirectoryKind::SlruSegment(kind), dir.segments.len()));
1333 0 : self.put(
1334 0 : dir_key,
1335 0 : Value::Image(Bytes::from(SlruSegmentDirectory::ser(&dir)?)),
1336 : );
1337 :
1338 : // Delete size entry, as well as all blocks
1339 0 : self.delete(slru_segment_key_range(kind, segno));
1340 0 :
1341 0 : Ok(())
1342 0 : }
1343 :
1344 : /// Drop a relmapper file (pg_filenode.map)
1345 0 : pub fn drop_relmap_file(&mut self, _spcnode: Oid, _dbnode: Oid) -> anyhow::Result<()> {
1346 0 : // TODO
1347 0 : Ok(())
1348 0 : }
1349 :
1350 : /// This method is used for marking truncated SLRU files
1351 0 : pub async fn drop_twophase_file(
1352 0 : &mut self,
1353 0 : xid: TransactionId,
1354 0 : ctx: &RequestContext,
1355 0 : ) -> anyhow::Result<()> {
1356 : // Remove it from the directory entry
1357 0 : let buf = self.get(TWOPHASEDIR_KEY, ctx).await?;
1358 0 : let mut dir = TwoPhaseDirectory::des(&buf)?;
1359 :
1360 0 : if !dir.xids.remove(&xid) {
1361 0 : warn!("twophase file for xid {} does not exist", xid);
1362 0 : }
1363 0 : self.pending_directory_entries
1364 0 : .push((DirectoryKind::TwoPhase, dir.xids.len()));
1365 0 : self.put(
1366 0 : TWOPHASEDIR_KEY,
1367 0 : Value::Image(Bytes::from(TwoPhaseDirectory::ser(&dir)?)),
1368 : );
1369 :
1370 : // Delete it
1371 0 : self.delete(twophase_key_range(xid));
1372 0 :
1373 0 : Ok(())
1374 0 : }
1375 :
1376 92 : pub fn init_aux_dir(&mut self) -> anyhow::Result<()> {
1377 92 : let buf = AuxFilesDirectory::ser(&AuxFilesDirectory {
1378 92 : files: HashMap::new(),
1379 92 : })?;
1380 92 : self.pending_directory_entries
1381 92 : .push((DirectoryKind::AuxFiles, 0));
1382 92 : self.put(AUX_FILES_KEY, Value::Image(Bytes::from(buf)));
1383 92 : Ok(())
1384 92 : }
1385 :
1386 8 : pub async fn put_file(
1387 8 : &mut self,
1388 8 : path: &str,
1389 8 : content: &[u8],
1390 8 : ctx: &RequestContext,
1391 8 : ) -> anyhow::Result<()> {
1392 8 : let file_path = path.to_string();
1393 8 : let content = if content.is_empty() {
1394 2 : None
1395 : } else {
1396 6 : Some(Bytes::copy_from_slice(content))
1397 : };
1398 :
1399 : let n_files;
1400 8 : let mut aux_files = self.tline.aux_files.lock().await;
1401 8 : if let Some(mut dir) = aux_files.dir.take() {
1402 : // We already updated aux files in `self`: emit a delta and update our latest value
1403 6 : dir.upsert(file_path.clone(), content.clone());
1404 6 : n_files = dir.files.len();
1405 6 : if aux_files.n_deltas == MAX_AUX_FILE_DELTAS {
1406 0 : self.put(
1407 0 : AUX_FILES_KEY,
1408 0 : Value::Image(Bytes::from(
1409 0 : AuxFilesDirectory::ser(&dir).context("serialize")?,
1410 : )),
1411 : );
1412 0 : aux_files.n_deltas = 0;
1413 6 : } else {
1414 6 : self.put(
1415 6 : AUX_FILES_KEY,
1416 6 : Value::WalRecord(NeonWalRecord::AuxFile { file_path, content }),
1417 6 : );
1418 6 : aux_files.n_deltas += 1;
1419 6 : }
1420 6 : aux_files.dir = Some(dir);
1421 : } else {
1422 : // Check if the AUX_FILES_KEY is initialized
1423 2 : match self.get(AUX_FILES_KEY, ctx).await {
1424 0 : Ok(dir_bytes) => {
1425 0 : let mut dir = AuxFilesDirectory::des(&dir_bytes)?;
1426 : // Key is already set, we may append a delta
1427 0 : self.put(
1428 0 : AUX_FILES_KEY,
1429 0 : Value::WalRecord(NeonWalRecord::AuxFile {
1430 0 : file_path: file_path.clone(),
1431 0 : content: content.clone(),
1432 0 : }),
1433 0 : );
1434 0 : dir.upsert(file_path, content);
1435 0 : n_files = dir.files.len();
1436 0 : aux_files.dir = Some(dir);
1437 : }
1438 : Err(
1439 0 : e @ (PageReconstructError::AncestorStopping(_)
1440 : | PageReconstructError::Cancelled
1441 : | PageReconstructError::AncestorLsnTimeout(_)),
1442 : ) => {
1443 : // Important that we do not interpret a shutdown error as "not found" and thereby
1444 : // reset the map.
1445 0 : return Err(e.into());
1446 : }
1447 : // FIXME: PageReconstructError doesn't have an explicit variant for key-not-found, so
1448 : // we are assuming that all _other_ possible errors represents a missing key. If some
1449 : // other error occurs, we may incorrectly reset the map of aux files.
1450 : Err(PageReconstructError::Other(_) | PageReconstructError::WalRedo(_)) => {
1451 : // Key is missing, we must insert an image as the basis for subsequent deltas.
1452 :
1453 2 : let mut dir = AuxFilesDirectory {
1454 2 : files: HashMap::new(),
1455 2 : };
1456 2 : dir.upsert(file_path, content);
1457 2 : self.put(
1458 2 : AUX_FILES_KEY,
1459 2 : Value::Image(Bytes::from(
1460 2 : AuxFilesDirectory::ser(&dir).context("serialize")?,
1461 : )),
1462 : );
1463 2 : n_files = 1;
1464 2 : aux_files.dir = Some(dir);
1465 : }
1466 : }
1467 : }
1468 :
1469 8 : self.pending_directory_entries
1470 8 : .push((DirectoryKind::AuxFiles, n_files));
1471 8 :
1472 8 : Ok(())
1473 8 : }
1474 :
1475 : ///
1476 : /// Flush changes accumulated so far to the underlying repository.
1477 : ///
1478 : /// Usually, changes made in DatadirModification are atomic, but this allows
1479 : /// you to flush them to the underlying repository before the final `commit`.
1480 : /// That allows to free up the memory used to hold the pending changes.
1481 : ///
1482 : /// Currently only used during bulk import of a data directory. In that
1483 : /// context, breaking the atomicity is OK. If the import is interrupted, the
1484 : /// whole import fails and the timeline will be deleted anyway.
1485 : /// (Or to be precise, it will be left behind for debugging purposes and
1486 : /// ignored, see <https://github.com/neondatabase/neon/pull/1809>)
1487 : ///
1488 : /// Note: A consequence of flushing the pending operations is that they
1489 : /// won't be visible to subsequent operations until `commit`. The function
1490 : /// retains all the metadata, but data pages are flushed. That's again OK
1491 : /// for bulk import, where you are just loading data pages and won't try to
1492 : /// modify the same pages twice.
1493 1930 : pub async fn flush(&mut self, ctx: &RequestContext) -> anyhow::Result<()> {
1494 1930 : // Unless we have accumulated a decent amount of changes, it's not worth it
1495 1930 : // to scan through the pending_updates list.
1496 1930 : let pending_nblocks = self.pending_nblocks;
1497 1930 : if pending_nblocks < 10000 {
1498 1930 : return Ok(());
1499 0 : }
1500 :
1501 0 : let writer = self.tline.writer().await;
1502 :
1503 : // Flush relation and SLRU data blocks, keep metadata.
1504 0 : let mut retained_pending_updates = HashMap::<_, Vec<_>>::new();
1505 0 : for (key, values) in self.pending_updates.drain() {
1506 0 : for (lsn, value) in values {
1507 0 : if is_rel_block_key(&key) || is_slru_block_key(key) {
1508 : // This bails out on first error without modifying pending_updates.
1509 : // That's Ok, cf this function's doc comment.
1510 0 : writer.put(key, lsn, &value, ctx).await?;
1511 0 : } else {
1512 0 : retained_pending_updates
1513 0 : .entry(key)
1514 0 : .or_default()
1515 0 : .push((lsn, value));
1516 0 : }
1517 : }
1518 : }
1519 :
1520 0 : self.pending_updates = retained_pending_updates;
1521 0 :
1522 0 : if pending_nblocks != 0 {
1523 0 : writer.update_current_logical_size(pending_nblocks * i64::from(BLCKSZ));
1524 0 : self.pending_nblocks = 0;
1525 0 : }
1526 :
1527 0 : for (kind, count) in std::mem::take(&mut self.pending_directory_entries) {
1528 0 : writer.update_directory_entries_count(kind, count as u64);
1529 0 : }
1530 :
1531 0 : Ok(())
1532 1930 : }
1533 :
1534 : ///
1535 : /// Finish this atomic update, writing all the updated keys to the
1536 : /// underlying timeline.
1537 : /// All the modifications in this atomic update are stamped by the specified LSN.
1538 : ///
1539 742956 : pub async fn commit(&mut self, ctx: &RequestContext) -> anyhow::Result<()> {
1540 742956 : let writer = self.tline.writer().await;
1541 :
1542 742956 : let pending_nblocks = self.pending_nblocks;
1543 742956 : self.pending_nblocks = 0;
1544 742956 :
1545 742956 : if !self.pending_updates.is_empty() {
1546 413940 : writer.put_batch(&self.pending_updates, ctx).await?;
1547 413940 : self.pending_updates.clear();
1548 329016 : }
1549 :
1550 742956 : if !self.pending_deletions.is_empty() {
1551 2 : writer.delete_batch(&self.pending_deletions).await?;
1552 2 : self.pending_deletions.clear();
1553 742954 : }
1554 :
1555 742956 : self.pending_lsns.push(self.lsn);
1556 888812 : for pending_lsn in self.pending_lsns.drain(..) {
1557 888812 : // Ideally, we should be able to call writer.finish_write() only once
1558 888812 : // with the highest LSN. However, the last_record_lsn variable in the
1559 888812 : // timeline keeps track of the latest LSN and the immediate previous LSN
1560 888812 : // so we need to record every LSN to not leave a gap between them.
1561 888812 : writer.finish_write(pending_lsn);
1562 888812 : }
1563 :
1564 742956 : if pending_nblocks != 0 {
1565 270570 : writer.update_current_logical_size(pending_nblocks * i64::from(BLCKSZ));
1566 472386 : }
1567 :
1568 742956 : for (kind, count) in std::mem::take(&mut self.pending_directory_entries) {
1569 2424 : writer.update_directory_entries_count(kind, count as u64);
1570 2424 : }
1571 :
1572 742956 : Ok(())
1573 742956 : }
1574 :
1575 291704 : pub(crate) fn len(&self) -> usize {
1576 291704 : self.pending_updates.len() + self.pending_deletions.len()
1577 291704 : }
1578 :
1579 : // Internal helper functions to batch the modifications
1580 :
1581 286562 : async fn get(&self, key: Key, ctx: &RequestContext) -> Result<Bytes, PageReconstructError> {
1582 : // Have we already updated the same key? Read the latest pending updated
1583 : // version in that case.
1584 : //
1585 : // Note: we don't check pending_deletions. It is an error to request a
1586 : // value that has been removed, deletion only avoids leaking storage.
1587 286562 : if let Some(values) = self.pending_updates.get(&key) {
1588 15928 : if let Some((_, value)) = values.last() {
1589 15928 : return if let Value::Image(img) = value {
1590 15928 : Ok(img.clone())
1591 : } else {
1592 : // Currently, we never need to read back a WAL record that we
1593 : // inserted in the same "transaction". All the metadata updates
1594 : // work directly with Images, and we never need to read actual
1595 : // data pages. We could handle this if we had to, by calling
1596 : // the walredo manager, but let's keep it simple for now.
1597 0 : Err(PageReconstructError::from(anyhow::anyhow!(
1598 0 : "unexpected pending WAL record"
1599 0 : )))
1600 : };
1601 0 : }
1602 270634 : }
1603 270634 : let lsn = Lsn::max(self.tline.get_last_record_lsn(), self.lsn);
1604 270634 : self.tline.get(key, lsn, ctx).await
1605 286562 : }
1606 :
1607 711856 : fn put(&mut self, key: Key, val: Value) {
1608 711856 : let values = self.pending_updates.entry(key).or_default();
1609 : // Replace the previous value if it exists at the same lsn
1610 711856 : if let Some((last_lsn, last_value)) = values.last_mut() {
1611 12164 : if *last_lsn == self.lsn {
1612 12160 : *last_value = val;
1613 12160 : return;
1614 4 : }
1615 699692 : }
1616 699696 : values.push((self.lsn, val));
1617 711856 : }
1618 :
1619 2 : fn delete(&mut self, key_range: Range<Key>) {
1620 2 : trace!("DELETE {}-{}", key_range.start, key_range.end);
1621 2 : self.pending_deletions.push((key_range, self.lsn));
1622 2 : }
1623 : }
1624 :
1625 : /// This struct facilitates accessing either a committed key from the timeline at a
1626 : /// specific LSN, or the latest uncommitted key from a pending modification.
1627 : /// During WAL ingestion, the records from multiple LSNs may be batched in the same
1628 : /// modification before being flushed to the timeline. Hence, the routines in WalIngest
1629 : /// need to look up the keys in the modification first before looking them up in the
1630 : /// timeline to not miss the latest updates.
1631 0 : #[derive(Clone, Copy)]
1632 : pub enum Version<'a> {
1633 : Lsn(Lsn),
1634 : Modified(&'a DatadirModification<'a>),
1635 : }
1636 :
1637 : impl<'a> Version<'a> {
1638 23542 : async fn get(
1639 23542 : &self,
1640 23542 : timeline: &Timeline,
1641 23542 : key: Key,
1642 23542 : ctx: &RequestContext,
1643 23542 : ) -> Result<Bytes, PageReconstructError> {
1644 23542 : match self {
1645 23532 : Version::Lsn(lsn) => timeline.get(key, *lsn, ctx).await,
1646 10 : Version::Modified(modification) => modification.get(key, ctx).await,
1647 : }
1648 23542 : }
1649 :
1650 30484 : fn get_lsn(&self) -> Lsn {
1651 30484 : match self {
1652 24438 : Version::Lsn(lsn) => *lsn,
1653 6046 : Version::Modified(modification) => modification.lsn,
1654 : }
1655 30484 : }
1656 : }
1657 :
1658 : //--- Metadata structs stored in key-value pairs in the repository.
1659 :
1660 2110 : #[derive(Debug, Serialize, Deserialize)]
1661 : struct DbDirectory {
1662 : // (spcnode, dbnode) -> (do relmapper and PG_VERSION files exist)
1663 : dbdirs: HashMap<(Oid, Oid), bool>,
1664 : }
1665 :
1666 174 : #[derive(Debug, Serialize, Deserialize)]
1667 : struct TwoPhaseDirectory {
1668 : xids: HashSet<TransactionId>,
1669 : }
1670 :
1671 3860 : #[derive(Debug, Serialize, Deserialize, Default)]
1672 : struct RelDirectory {
1673 : // Set of relations that exist. (relfilenode, forknum)
1674 : //
1675 : // TODO: Store it as a btree or radix tree or something else that spans multiple
1676 : // key-value pairs, if you have a lot of relations
1677 : rels: HashSet<(Oid, u8)>,
1678 : }
1679 :
1680 208 : #[derive(Debug, Serialize, Deserialize, Default)]
1681 : pub(crate) struct AuxFilesDirectory {
1682 : pub(crate) files: HashMap<String, Bytes>,
1683 : }
1684 :
1685 : impl AuxFilesDirectory {
1686 24 : pub(crate) fn upsert(&mut self, key: String, value: Option<Bytes>) {
1687 24 : if let Some(value) = value {
1688 18 : self.files.insert(key, value);
1689 18 : } else {
1690 6 : self.files.remove(&key);
1691 6 : }
1692 24 : }
1693 : }
1694 :
1695 0 : #[derive(Debug, Serialize, Deserialize)]
1696 : struct RelSizeEntry {
1697 : nblocks: u32,
1698 : }
1699 :
1700 528 : #[derive(Debug, Serialize, Deserialize, Default)]
1701 : struct SlruSegmentDirectory {
1702 : // Set of SLRU segments that exist.
1703 : segments: HashSet<u32>,
1704 : }
1705 :
1706 4848 : #[derive(Copy, Clone, PartialEq, Eq, Debug, enum_map::Enum)]
1707 : #[repr(u8)]
1708 : pub(crate) enum DirectoryKind {
1709 : Db,
1710 : TwoPhase,
1711 : Rel,
1712 : AuxFiles,
1713 : SlruSegment(SlruKind),
1714 : }
1715 :
1716 : impl DirectoryKind {
1717 : pub(crate) const KINDS_NUM: usize = <DirectoryKind as Enum>::LENGTH;
1718 4848 : pub(crate) fn offset(&self) -> usize {
1719 4848 : self.into_usize()
1720 4848 : }
1721 : }
1722 :
1723 : static ZERO_PAGE: Bytes = Bytes::from_static(&[0u8; BLCKSZ as usize]);
1724 :
1725 : #[allow(clippy::bool_assert_comparison)]
1726 : #[cfg(test)]
1727 : mod tests {
1728 : use hex_literal::hex;
1729 : use utils::id::TimelineId;
1730 :
1731 : use super::*;
1732 :
1733 : use crate::{tenant::harness::TenantHarness, DEFAULT_PG_VERSION};
1734 :
1735 : /// Test a round trip of aux file updates, from DatadirModification to reading back from the Timeline
1736 2 : #[tokio::test]
1737 2 : async fn aux_files_round_trip() -> anyhow::Result<()> {
1738 2 : let name = "aux_files_round_trip";
1739 2 : let harness = TenantHarness::create(name)?;
1740 2 :
1741 2 : pub const TIMELINE_ID: TimelineId =
1742 2 : TimelineId::from_array(hex!("11223344556677881122334455667788"));
1743 2 :
1744 2 : let (tenant, ctx) = harness.load().await;
1745 2 : let tline = tenant
1746 2 : .create_empty_timeline(TIMELINE_ID, Lsn(0), DEFAULT_PG_VERSION, &ctx)
1747 2 : .await?;
1748 2 : let tline = tline.raw_timeline().unwrap();
1749 2 :
1750 2 : // First modification: insert two keys
1751 2 : let mut modification = tline.begin_modification(Lsn(0x1000));
1752 2 : modification.put_file("foo/bar1", b"content1", &ctx).await?;
1753 2 : modification.set_lsn(Lsn(0x1008))?;
1754 2 : modification.put_file("foo/bar2", b"content2", &ctx).await?;
1755 2 : modification.commit(&ctx).await?;
1756 2 : let expect_1008 = HashMap::from([
1757 2 : ("foo/bar1".to_string(), Bytes::from_static(b"content1")),
1758 2 : ("foo/bar2".to_string(), Bytes::from_static(b"content2")),
1759 2 : ]);
1760 2 :
1761 2 : let readback = tline.list_aux_files(Lsn(0x1008), &ctx).await?;
1762 2 : assert_eq!(readback, expect_1008);
1763 2 :
1764 2 : // Second modification: update one key, remove the other
1765 2 : let mut modification = tline.begin_modification(Lsn(0x2000));
1766 2 : modification.put_file("foo/bar1", b"content3", &ctx).await?;
1767 2 : modification.set_lsn(Lsn(0x2008))?;
1768 2 : modification.put_file("foo/bar2", b"", &ctx).await?;
1769 2 : modification.commit(&ctx).await?;
1770 2 : let expect_2008 =
1771 2 : HashMap::from([("foo/bar1".to_string(), Bytes::from_static(b"content3"))]);
1772 2 :
1773 2 : let readback = tline.list_aux_files(Lsn(0x2008), &ctx).await?;
1774 2 : assert_eq!(readback, expect_2008);
1775 2 :
1776 2 : // Reading back in time works
1777 2 : let readback = tline.list_aux_files(Lsn(0x1008), &ctx).await?;
1778 2 : assert_eq!(readback, expect_1008);
1779 2 :
1780 2 : Ok(())
1781 2 : }
1782 :
1783 : /*
1784 : fn assert_current_logical_size<R: Repository>(timeline: &DatadirTimeline<R>, lsn: Lsn) {
1785 : let incremental = timeline.get_current_logical_size();
1786 : let non_incremental = timeline
1787 : .get_current_logical_size_non_incremental(lsn)
1788 : .unwrap();
1789 : assert_eq!(incremental, non_incremental);
1790 : }
1791 : */
1792 :
1793 : /*
1794 : ///
1795 : /// Test list_rels() function, with branches and dropped relations
1796 : ///
1797 : #[test]
1798 : fn test_list_rels_drop() -> Result<()> {
1799 : let repo = RepoHarness::create("test_list_rels_drop")?.load();
1800 : let tline = create_empty_timeline(repo, TIMELINE_ID)?;
1801 : const TESTDB: u32 = 111;
1802 :
1803 : // Import initial dummy checkpoint record, otherwise the get_timeline() call
1804 : // after branching fails below
1805 : let mut writer = tline.begin_record(Lsn(0x10));
1806 : writer.put_checkpoint(ZERO_CHECKPOINT.clone())?;
1807 : writer.finish()?;
1808 :
1809 : // Create a relation on the timeline
1810 : let mut writer = tline.begin_record(Lsn(0x20));
1811 : writer.put_rel_page_image(TESTREL_A, 0, TEST_IMG("foo blk 0 at 2"))?;
1812 : writer.finish()?;
1813 :
1814 : let writer = tline.begin_record(Lsn(0x00));
1815 : writer.finish()?;
1816 :
1817 : // Check that list_rels() lists it after LSN 2, but no before it
1818 : assert!(!tline.list_rels(0, TESTDB, Lsn(0x10))?.contains(&TESTREL_A));
1819 : assert!(tline.list_rels(0, TESTDB, Lsn(0x20))?.contains(&TESTREL_A));
1820 : assert!(tline.list_rels(0, TESTDB, Lsn(0x30))?.contains(&TESTREL_A));
1821 :
1822 : // Create a branch, check that the relation is visible there
1823 : repo.branch_timeline(&tline, NEW_TIMELINE_ID, Lsn(0x30))?;
1824 : let newtline = match repo.get_timeline(NEW_TIMELINE_ID)?.local_timeline() {
1825 : Some(timeline) => timeline,
1826 : None => panic!("Should have a local timeline"),
1827 : };
1828 : let newtline = DatadirTimelineImpl::new(newtline);
1829 : assert!(newtline
1830 : .list_rels(0, TESTDB, Lsn(0x30))?
1831 : .contains(&TESTREL_A));
1832 :
1833 : // Drop it on the branch
1834 : let mut new_writer = newtline.begin_record(Lsn(0x40));
1835 : new_writer.drop_relation(TESTREL_A)?;
1836 : new_writer.finish()?;
1837 :
1838 : // Check that it's no longer listed on the branch after the point where it was dropped
1839 : assert!(newtline
1840 : .list_rels(0, TESTDB, Lsn(0x30))?
1841 : .contains(&TESTREL_A));
1842 : assert!(!newtline
1843 : .list_rels(0, TESTDB, Lsn(0x40))?
1844 : .contains(&TESTREL_A));
1845 :
1846 : // Run checkpoint and garbage collection and check that it's still not visible
1847 : newtline.checkpoint(CheckpointConfig::Forced)?;
1848 : repo.gc_iteration(Some(NEW_TIMELINE_ID), 0, true)?;
1849 :
1850 : assert!(!newtline
1851 : .list_rels(0, TESTDB, Lsn(0x40))?
1852 : .contains(&TESTREL_A));
1853 :
1854 : Ok(())
1855 : }
1856 : */
1857 :
1858 : /*
1859 : #[test]
1860 : fn test_read_beyond_eof() -> Result<()> {
1861 : let repo = RepoHarness::create("test_read_beyond_eof")?.load();
1862 : let tline = create_test_timeline(repo, TIMELINE_ID)?;
1863 :
1864 : make_some_layers(&tline, Lsn(0x20))?;
1865 : let mut writer = tline.begin_record(Lsn(0x60));
1866 : walingest.put_rel_page_image(
1867 : &mut writer,
1868 : TESTREL_A,
1869 : 0,
1870 : TEST_IMG(&format!("foo blk 0 at {}", Lsn(0x60))),
1871 : )?;
1872 : writer.finish()?;
1873 :
1874 : // Test read before rel creation. Should error out.
1875 : assert!(tline.get_rel_page_at_lsn(TESTREL_A, 1, Lsn(0x10), false).is_err());
1876 :
1877 : // Read block beyond end of relation at different points in time.
1878 : // These reads should fall into different delta, image, and in-memory layers.
1879 : assert_eq!(tline.get_rel_page_at_lsn(TESTREL_A, 1, Lsn(0x20), false)?, ZERO_PAGE);
1880 : assert_eq!(tline.get_rel_page_at_lsn(TESTREL_A, 1, Lsn(0x25), false)?, ZERO_PAGE);
1881 : assert_eq!(tline.get_rel_page_at_lsn(TESTREL_A, 1, Lsn(0x30), false)?, ZERO_PAGE);
1882 : assert_eq!(tline.get_rel_page_at_lsn(TESTREL_A, 1, Lsn(0x35), false)?, ZERO_PAGE);
1883 : assert_eq!(tline.get_rel_page_at_lsn(TESTREL_A, 1, Lsn(0x40), false)?, ZERO_PAGE);
1884 : assert_eq!(tline.get_rel_page_at_lsn(TESTREL_A, 1, Lsn(0x45), false)?, ZERO_PAGE);
1885 : assert_eq!(tline.get_rel_page_at_lsn(TESTREL_A, 1, Lsn(0x50), false)?, ZERO_PAGE);
1886 : assert_eq!(tline.get_rel_page_at_lsn(TESTREL_A, 1, Lsn(0x55), false)?, ZERO_PAGE);
1887 : assert_eq!(tline.get_rel_page_at_lsn(TESTREL_A, 1, Lsn(0x60), false)?, ZERO_PAGE);
1888 :
1889 : // Test on an in-memory layer with no preceding layer
1890 : let mut writer = tline.begin_record(Lsn(0x70));
1891 : walingest.put_rel_page_image(
1892 : &mut writer,
1893 : TESTREL_B,
1894 : 0,
1895 : TEST_IMG(&format!("foo blk 0 at {}", Lsn(0x70))),
1896 : )?;
1897 : writer.finish()?;
1898 :
1899 : assert_eq!(tline.get_rel_page_at_lsn(TESTREL_B, 1, Lsn(0x70), false)?6, ZERO_PAGE);
1900 :
1901 : Ok(())
1902 : }
1903 : */
1904 : }
|
