gc.rs 93.0%

147 of 158 lines run, 7 of 14 functions · run #339 on proposal/124 · run missed

//! Garbage collection (spec §21): delete objects
//! nothing reachable refers to.
//!
//! Roots: the current view, the views of every operation newer than
//! `core.op-retention` (their working copies only within the shorter
//! `core.snapshot-retention`), reference attestations, and the policy; receipts
//! live as long as the changes of their state. Operation
//! objects are always kept so the log stays walkable, but an operation older
//! than the retention window no longer keeps its states alive, so undoing
//! that far back may fail once its objects are gone.

use super::*;
use crate::store::{PackReport, PruneReport, RetiredReport};
use std::collections::HashSet;

#[derive(Debug, Clone, Default, PartialEq, Eq)]
pub struct GcReport {
    pub packed: PackReport,
    pub pruned: PruneReport,
    pub reachable: usize,
    /// Packs folded into the one a full repack writes.
    pub merged_packs: usize,
    /// Packs and loose objects earlier repacks retired, deleted now that
    /// [`crate::store::RETIRED_FOR`] has passed (oxygenhub#249).
    pub retired: RetiredReport,
}

impl Repo {
    /// Delete what earlier repacks retired more than
    /// [`crate::store::RETIRED_FOR`] ago, prune unreachable objects older
    /// than the grace period (`now` ignores it, but not the retired
    /// files' hour), then repack everything into one pack (spec §13.4).
    /// The packs and loose objects this repack replaces are retired in
    /// turn, for the next collection.
    pub fn gc(&self, now: bool) -> Result<GcReport> {
        self.gc_with(now, &mut |_| {})
    }

    /// [`Repo::gc`], reporting each phase and the packing's progress
    /// (oxygen#66), for a spinner or a log.
    pub fn gc_with(&self, now: bool, progress: crate::store::Progress) -> Result<GcReport> {
        self.gc_keeping(now, &[], progress)
    }

    /// [`Repo::gc_with`], keeping `extra` blobs and their chunks as well:
    /// objects a host stored for itself and names from outside the
    /// repository's history, such as a forge's job artifacts
    /// (oxygenhub#126).
    pub fn gc_keeping(
        &self,
        now: bool,
        extra: &[ObjectRef],
        progress: crate::store::Progress,
    ) -> Result<GcReport> {
        // Retired files first: the readers that used them have had their
        // hour, and what replaced them is in the packs read from here on.
        let retired = self
            .store
            .expire_retired(crate::store::RETIRED_FOR, progress)?;
        progress("finding what is reachable");
        let mut keep = self.reachable()?;
        for r in extra {
            if !self.store.has(r) || !keep.insert(*r) {
                continue;
            }
            if let Ok(body) = self.store.get_typed(r, ObjectType::Blob) {
                if let Ok(blob) = Blob::decode(&body) {
                    keep.extend(blob.chunks);
                }
            }
        }
        let grace = if now {
            std::time::Duration::ZERO
        } else {
            std::time::Duration::from_secs(self.config.gc_grace_days * 86_400)
        };
        let pruned = self.store.prune_with(&keep, grace, progress)?;
        // Then everything left in one pack, compressed harder than a
        // commit's (spec §13.4), as `git gc` repacks.
        let (packed, merged_packs) = self
            .store
            .repack_all_with(crate::store::REPACK_ZSTD_LEVEL, progress)?;
        Ok(GcReport {
            packed,
            pruned,
            reachable: keep.len(),
            merged_packs,
            retired,
        })
    }

    /// Every object that must survive garbage collection.
    pub fn reachable(&self) -> Result<HashSet<ObjectRef>> {
        let mut keep: HashSet<ObjectRef> = HashSet::new();
        let mut states: Vec<ObjectRef> = Vec::new();
        let mut changes: Vec<ChangeHash> = Vec::new();
        let mut attests: Vec<ObjectRef> = Vec::new();
        let mut policies: Vec<ObjectRef> = Vec::new();

        // Operations: always kept; views within retention are roots. Past
        // working copies only within the shorter `core.snapshot-retention`
        // (oxygen#98): a file only ever snapshotted, such as a build output
        // a status took in, goes once they expire.
        let now_ms = Hlc::now().wall_ms;
        let cutoff = now_ms.saturating_sub(self.config.op_retention_days * 86_400_000);
        let wc_cutoff = now_ms
            .saturating_sub(self.config.snapshot_retention_days * 86_400_000)
            .max(cutoff);
        for (i, (r, op)) in self.op_log()?.into_iter().enumerate() {
            keep.insert(r);
            if i == 0 || op.hlc.wall_ms >= cutoff {
                states.extend(op.view.bookmarks.values().copied());
                states.extend(op.view.remotes.values().flat_map(|b| b.values().copied()));
                policies.extend(op.view.policy);
            }
            if i == 0 || (self.config.snapshot_retention_days > 0 && op.hlc.wall_ms >= wc_cutoff) {
                changes.extend(op.view.workspaces.values().copied());
            }
        }
        // Checkpoints: the object, its payload blob, and the blob's chunks.
        for r in self.checkpoint_refs()? {
            if let Ok(bytes) = self.store.get_typed(&r, ObjectType::Checkpoint) {
                keep.insert(r);
                let cp = crate::objects::Checkpoint::decode(&bytes)?;
                keep.insert(cp.payload);
                if let Ok(b) = self.store.get_typed(&cp.payload, ObjectType::Blob) {
                    keep.extend(Blob::decode(&b)?.chunks);
                }
                // A version 2 payload's graph blobs and their chunks.
                for g in self.checkpoint_graph_blobs(&cp).unwrap_or_default() {
                    keep.insert(g);
                    if let Ok(b) = self.store.get_typed(&g, ObjectType::Blob) {
                        keep.extend(Blob::decode(&b)?.chunks);
                    }
                }
            }
        }
        // Tags: their states and attestations.
        let view_now = self.view()?;
        states.extend(view_now.tags.values().copied());
        for name in view_now.tags.keys() {
            if let Some((r, _)) = self.tag_attest(name)? {
                attests.push(r);
            }
        }
        // Reference attestations.
        for name in self.view()?.bookmarks.keys() {
            if let Some((r, _)) = self.latest_ref_attest(name)? {
                // With a host's acceptance of it (spec §14.7).
                if let Some(acc) = self.acceptance_of(&r)? {
                    attests.push(acc.attest.hash());
                }
                attests.push(r);
            }
        }
        // A historical view may name objects pruned under an earlier, shorter
        // retention; there is nothing to keep for those.
        for st in states {
            if keep.insert(st) {
                match self.store.get_state(&st) {
                    Ok(s) => changes.extend(s.heads),
                    Err(StoreError::NotFound(_)) => continue,
                    Err(e) => return Err(e.into()),
                }
            }
        }
        for p in policies {
            if keep.insert(p) {
                let bytes = match self.store.get_typed(&p, ObjectType::Policy) {
                    Ok(b) => b,
                    Err(StoreError::NotFound(_)) => continue,
                    Err(e) => return Err(e.into()),
                };
                let policy = crate::policy::Policy::decode(&bytes)?;
                for id in &policy.identities {
                    keep.insert(id.hash());
                }
                for sig in self.store.signatures(&p)? {
                    keep.insert(sig.key);
                }
            }
        }
        // Changes: their closure, identities, blobs, chunks, seals.
        let mut stack = changes;
        while let Some(h) = stack.pop() {
            if !keep.insert(h) {
                continue;
            }
            // Beyond the shallow boundary, or pruned under an earlier
            // retention, there is nothing to keep.
            let c = match self.store.get_change_parts(&h) {
                Ok((c, texts)) => {
                    keep.extend(texts);
                    c
                }
                Err(StoreError::NotFound(_)) => continue,
                Err(e) => return Err(e.into()),
            };
            stack.extend(c.deps.iter().copied());
            if !c.author.is_zero() {
                keep.insert(c.author);
            }
            for sig in self.store.signatures(&h)? {
                self.keep_signer(&mut keep, sig.key)?;
            }
            for e in &c.edits {
                if let Edit::Replace { old, new, .. } = e {
                    for b in [old, new].into_iter().flatten() {
                        if keep.insert(*b) {
                            if let Ok(bytes) = self.store.get_typed(b, ObjectType::Blob) {
                                keep.extend(Blob::decode(&bytes)?.chunks);
                            }
                        }
                    }
                }
            }
            attests.extend(self.store.seals_for(&h)?);
        }
        self.reachable_receipts(&mut keep)?;
        self.reachable_approvals(&mut keep)?;
        for a in attests {
            if keep.insert(a) {
                for sig in self.store.signatures(&a)? {
                    self.keep_signer(&mut keep, sig.key)?;
                }
            }
        }
        Ok(keep)
    }

    /// Keep a signer's identity and, for a short-lived key, its
    /// certificate and the certificate's issuer (spec §14.7).
    pub(crate) fn keep_signer(&self, keep: &mut HashSet<ObjectRef>, key: ObjectRef) -> Result<()> {
        keep.insert(key);
        if let Some(c) = self.store.certificate_of(&key)? {
            keep.insert(c);
            if let Ok(cert) = self.store.get_certificate(&c) {
                keep.insert(cert.cert.issuer);
            }
        }
        Ok(())
    }
}

`/`	Search
`g` `c`	Code
`g` `h`	History
`g` `b`	Bookmarks
`g` `t`	Trust
`y`	Permalink to this state (files and trees)
`?`	This sheet

1		//! Garbage collection (spec §21): delete objects
2		//! nothing reachable refers to.
3		//!
4		//! Roots: the current view, the views of every operation newer than
5		//! `core.op-retention` (their working copies only within the shorter
6		//! `core.snapshot-retention`), reference attestations, and the policy; receipts
7		//! live as long as the changes of their state. Operation
8		//! objects are always kept so the log stays walkable, but an operation older
9		//! than the retention window no longer keeps its states alive, so undoing
10		//! that far back may fail once its objects are gone.
11
12		use super::*;
13		use crate::store::{PackReport, PruneReport, RetiredReport};
14		use std::collections::HashSet;
15
16		#[derive(Debug, Clone, Default, PartialEq, Eq)]
17		pub struct GcReport {
18		pub packed: PackReport,
19		pub pruned: PruneReport,
20		pub reachable: usize,
21		/// Packs folded into the one a full repack writes.
22		pub merged_packs: usize,
23		/// Packs and loose objects earlier repacks retired, deleted now that
24		/// [`crate::store::RETIRED_FOR`] has passed (oxygenhub#249).
25		pub retired: RetiredReport,
26		}
27
28		impl Repo {
29		/// Delete what earlier repacks retired more than
30		/// [`crate::store::RETIRED_FOR`] ago, prune unreachable objects older
31		/// than the grace period (`now` ignores it, but not the retired
32		/// files' hour), then repack everything into one pack (spec §13.4).
33		/// The packs and loose objects this repack replaces are retired in
34		/// turn, for the next collection.
35	21×	pub fn gc(&self, now: bool) -> Result<GcReport> {
36	114×	self.gc_with(now, &mut \|_\| {})
37	21×	}
38
39		/// [`Repo::gc`], reporting each phase and the packing's progress
40		/// (oxygen#66), for a spinner or a log.
41	22×	pub fn gc_with(&self, now: bool, progress: crate::store::Progress) -> Result<GcReport> {
42	22×	self.gc_keeping(now, &[], progress)
43	22×	}
44
45		/// [`Repo::gc_with`], keeping `extra` blobs and their chunks as well:
46		/// objects a host stored for itself and names from outside the
47		/// repository's history, such as a forge's job artifacts
48		/// (oxygenhub#126).
49	23×	pub fn gc_keeping(
50	23×	&self,
51	23×	now: bool,
52	23×	extra: &[ObjectRef],
53	23×	progress: crate::store::Progress,
54	23×	) -> Result<GcReport> {
55		// Retired files first: the readers that used them have had their
56		// hour, and what replaced them is in the packs read from here on.
57	23×	let retired = self
58	23×	.store
59	23×	.expire_retired(crate::store::RETIRED_FOR, progress)?;
60	23×	progress("finding what is reachable");
61	23×	let mut keep = self.reachable()?;
62	23×	for r in extra {
63	1×	if !self.store.has(r) \|\| !keep.insert(*r) {
64	0×	continue;
65	1×	}
66	1×	if let Ok(body) = self.store.get_typed(r, ObjectType::Blob) {
67	1×	if let Ok(blob) = Blob::decode(&body) {
68	1×	keep.extend(blob.chunks);
69	1×	}
70	0×	}
71		}
72	23×	let grace = if now {
73	21×	std::time::Duration::ZERO
74		} else {
75	2×	std::time::Duration::from_secs(self.config.gc_grace_days * 86_400)
76		};
77	23×	let pruned = self.store.prune_with(&keep, grace, progress)?;
78		// Then everything left in one pack, compressed harder than a
79		// commit's (spec §13.4), as `git gc` repacks.
80	23×	let (packed, merged_packs) = self
81	23×	.store
82	23×	.repack_all_with(crate::store::REPACK_ZSTD_LEVEL, progress)?;
83	23×	Ok(GcReport {
84	23×	packed,
85	23×	pruned,
86	23×	reachable: keep.len(),
87	23×	merged_packs,
88	23×	retired,
89	23×	})
90	23×	}
91
92		/// Every object that must survive garbage collection.
93	23×	pub fn reachable(&self) -> Result<HashSet<ObjectRef>> {
94	23×	let mut keep: HashSet<ObjectRef> = HashSet::new();
95	23×	let mut states: Vec<ObjectRef> = Vec::new();
96	23×	let mut changes: Vec<ChangeHash> = Vec::new();
97	23×	let mut attests: Vec<ObjectRef> = Vec::new();
98	23×	let mut policies: Vec<ObjectRef> = Vec::new();
99
100		// Operations: always kept; views within retention are roots. Past
101		// working copies only within the shorter `core.snapshot-retention`
102		// (oxygen#98): a file only ever snapshotted, such as a build output
103		// a status took in, goes once they expire.
104	23×	let now_ms = Hlc::now().wall_ms;
105	23×	let cutoff = now_ms.saturating_sub(self.config.op_retention_days * 86_400_000);
106	23×	let wc_cutoff = now_ms
107	23×	.saturating_sub(self.config.snapshot_retention_days * 86_400_000)
108	23×	.max(cutoff);
109	148×	for (i, (r, op)) in self.op_log()?.into_iter().enumerate() {
110	148×	keep.insert(r);
111	148×	if i == 0 \|\| op.hlc.wall_ms >= cutoff {
112	110×	states.extend(op.view.bookmarks.values().copied());
113	110×	states.extend(op.view.remotes.values().flat_map(\|b\| b.values().copied()));
114	110×	policies.extend(op.view.policy);
115	38×	}
116	148×	if i == 0 \|\| (self.config.snapshot_retention_days > 0 && op.hlc.wall_ms >= wc_cutoff) {
117	107×	changes.extend(op.view.workspaces.values().copied());
118	107×	}
119		}
120		// Checkpoints: the object, its payload blob, and the blob's chunks.
121	23×	for r in self.checkpoint_refs()? {
122	2×	if let Ok(bytes) = self.store.get_typed(&r, ObjectType::Checkpoint) {
123	2×	keep.insert(r);
124	2×	let cp = crate::objects::Checkpoint::decode(&bytes)?;
125	2×	keep.insert(cp.payload);
126	2×	if let Ok(b) = self.store.get_typed(&cp.payload, ObjectType::Blob) {
127	2×	keep.extend(Blob::decode(&b)?.chunks);
128	0×	}
129		// A version 2 payload's graph blobs and their chunks.
130	2×	for g in self.checkpoint_graph_blobs(&cp).unwrap_or_default() {
131	1×	keep.insert(g);
132	1×	if let Ok(b) = self.store.get_typed(&g, ObjectType::Blob) {
133	1×	keep.extend(Blob::decode(&b)?.chunks);
134	0×	}
135		}
136	0×	}
137		}
138		// Tags: their states and attestations.
139	23×	let view_now = self.view()?;
140	23×	states.extend(view_now.tags.values().copied());
141	23×	for name in view_now.tags.keys() {
142	3×	if let Some((r, _)) = self.tag_attest(name)? {
143	3×	attests.push(r);
144	3×	}
145		}
146		// Reference attestations.
147	23×	for name in self.view()?.bookmarks.keys() {
148	23×	if let Some((r, _)) = self.latest_ref_attest(name)? {
149		// With a host's acceptance of it (spec §14.7).
150	1×	if let Some(acc) = self.acceptance_of(&r)? {
151	1×	attests.push(acc.attest.hash());
152	1×	}
153	1×	attests.push(r);
154	22×	}
155		}
156		// A historical view may name objects pruned under an earlier, shorter
157		// retention; there is nothing to keep for those.
158	129×	for st in states {
159	129×	if keep.insert(st) {
160	76×	match self.store.get_state(&st) {
161	73×	Ok(s) => changes.extend(s.heads),
162	3×	Err(StoreError::NotFound(_)) => continue,
163	0×	Err(e) => return Err(e.into()),
164		}
165	53×	}
166		}
167	23×	for p in policies {
168	9×	if keep.insert(p) {
169	3×	let bytes = match self.store.get_typed(&p, ObjectType::Policy) {
170	3×	Ok(b) => b,
171	0×	Err(StoreError::NotFound(_)) => continue,
172	0×	Err(e) => return Err(e.into()),
173		};
174	3×	let policy = crate::policy::Policy::decode(&bytes)?;
175	4×	for id in &policy.identities {
176	4×	keep.insert(id.hash());
177	4×	}
178	3×	for sig in self.store.signatures(&p)? {
179	3×	keep.insert(sig.key);
180	3×	}
181	6×	}
182		}
183		// Changes: their closure, identities, blobs, chunks, seals.
184	23×	let mut stack = changes;
185	386×	while let Some(h) = stack.pop() {
186	363×	if !keep.insert(h) {
187	203×	continue;
188	160×	}
189		// Beyond the shallow boundary, or pruned under an earlier
190		// retention, there is nothing to keep.
191	160×	let c = match self.store.get_change_parts(&h) {
192	150×	Ok((c, texts)) => {
193	150×	keep.extend(texts);
194	150×	c
195		}
196	10×	Err(StoreError::NotFound(_)) => continue,
197	0×	Err(e) => return Err(e.into()),
198		};
199	150×	stack.extend(c.deps.iter().copied());
200	150×	if !c.author.is_zero() {
201	32×	keep.insert(c.author);
202	118×	}
203	150×	for sig in self.store.signatures(&h)? {
204	13×	self.keep_signer(&mut keep, sig.key)?;
205		}
206	150×	for e in &c.edits {
207	133×	if let Edit::Replace { old, new, .. } = e {
208	18×	for b in [old, new].into_iter().flatten() {
209	18×	if keep.insert(*b) {
210	18×	if let Ok(bytes) = self.store.get_typed(b, ObjectType::Blob) {
211	18×	keep.extend(Blob::decode(&bytes)?.chunks);
212	0×	}
213	0×	}
214		}
215	115×	}
216		}
217	150×	attests.extend(self.store.seals_for(&h)?);
218		}
219	23×	self.reachable_receipts(&mut keep)?;
220	23×	self.reachable_approvals(&mut keep)?;
221	23×	for a in attests {
222	9×	if keep.insert(a) {
223	8×	for sig in self.store.signatures(&a)? {
224	8×	self.keep_signer(&mut keep, sig.key)?;
225		}
226	1×	}
227		}
228	23×	Ok(keep)
229	23×	}
230
231		/// Keep a signer's identity and, for a short-lived key, its
232		/// certificate and the certificate's issuer (spec §14.7).
233	23×	pub(crate) fn keep_signer(&self, keep: &mut HashSet<ObjectRef>, key: ObjectRef) -> Result<()> {
234	23×	keep.insert(key);
235	23×	if let Some(c) = self.store.certificate_of(&key)? {
236	3×	keep.insert(c);
237	3×	if let Ok(cert) = self.store.get_certificate(&c) {
238	3×	keep.insert(cert.cert.issuer);
239	3×	}
240	20×	}
241	23×	Ok(())
242	23×	}
243		}

Coverage / crates/oxy-core/src/repo/gc.rs 93.0%

Keyboard