crates/oxy-core/src/repo/gc.rs at v0.4.5 · andymccall/oxygen

◆ v0.4.5 oxygen / crates / oxy-core / src / repo / gc.rs

9.1 KiB · Rust · last changed The issuer countersigns when it accepted a certified update (keyless step 4c) 18 days ago source blame history raw

//! 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};
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,
}

impl Repo {
    /// Prune unreachable objects older than the grace period (`now`
    /// ignores it), then repack everything into one pack (spec §13.4).
    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> {
        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,
        })
    }

    /// 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};
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	}
24
25	impl Repo {
26	/// Prune unreachable objects older than the grace period (`now`
27	/// ignores it), then repack everything into one pack (spec §13.4).
28	pub fn gc(&self, now: bool) -> Result<GcReport> {
29	self.gc_with(now, &mut \|_\| {})
30	}
31
32	/// [`Repo::gc`], reporting each phase and the packing's progress
33	/// (oxygen#66), for a spinner or a log.
34	pub fn gc_with(&self, now: bool, progress: crate::store::Progress) -> Result<GcReport> {
35	self.gc_keeping(now, &[], progress)
36	}
37
38	/// [`Repo::gc_with`], keeping `extra` blobs and their chunks as well:
39	/// objects a host stored for itself and names from outside the
40	/// repository's history, such as a forge's job artifacts
41	/// (oxygenhub#126).
42	pub fn gc_keeping(
43	&self,
44	now: bool,
45	extra: &[ObjectRef],
46	progress: crate::store::Progress,
47	) -> Result<GcReport> {
48	progress("finding what is reachable");
49	let mut keep = self.reachable()?;
50	for r in extra {
51	if !self.store.has(r) \|\| !keep.insert(*r) {
52	continue;
53	}
54	if let Ok(body) = self.store.get_typed(r, ObjectType::Blob) {
55	if let Ok(blob) = Blob::decode(&body) {
56	keep.extend(blob.chunks);
57	}
58	}
59	}
60	let grace = if now {
61	std::time::Duration::ZERO
62	} else {
63	std::time::Duration::from_secs(self.config.gc_grace_days * 86_400)
64	};
65	let pruned = self.store.prune_with(&keep, grace, progress)?;
66	// Then everything left in one pack, compressed harder than a
67	// commit's (spec §13.4), as `git gc` repacks.
68	let (packed, merged_packs) = self
69	.store
70	.repack_all_with(crate::store::REPACK_ZSTD_LEVEL, progress)?;
71	Ok(GcReport {
72	packed,
73	pruned,
74	reachable: keep.len(),
75	merged_packs,
76	})
77	}
78
79	/// Every object that must survive garbage collection.
80	pub fn reachable(&self) -> Result<HashSet<ObjectRef>> {
81	let mut keep: HashSet<ObjectRef> = HashSet::new();
82	let mut states: Vec<ObjectRef> = Vec::new();
83	let mut changes: Vec<ChangeHash> = Vec::new();
84	let mut attests: Vec<ObjectRef> = Vec::new();
85	let mut policies: Vec<ObjectRef> = Vec::new();
86
87	// Operations: always kept; views within retention are roots. Past
88	// working copies only within the shorter `core.snapshot-retention`
89	// (oxygen#98): a file only ever snapshotted, such as a build output
90	// a status took in, goes once they expire.
91	let now_ms = Hlc::now().wall_ms;
92	let cutoff = now_ms.saturating_sub(self.config.op_retention_days * 86_400_000);
93	let wc_cutoff = now_ms
94	.saturating_sub(self.config.snapshot_retention_days * 86_400_000)
95	.max(cutoff);
96	for (i, (r, op)) in self.op_log()?.into_iter().enumerate() {
97	keep.insert(r);
98	if i == 0 \|\| op.hlc.wall_ms >= cutoff {
99	states.extend(op.view.bookmarks.values().copied());
100	states.extend(op.view.remotes.values().flat_map(\|b\| b.values().copied()));
101	policies.extend(op.view.policy);
102	}
103	if i == 0 \|\| (self.config.snapshot_retention_days > 0 && op.hlc.wall_ms >= wc_cutoff) {
104	changes.extend(op.view.workspaces.values().copied());
105	}
106	}
107	// Checkpoints: the object, its payload blob, and the blob's chunks.
108	for r in self.checkpoint_refs()? {
109	if let Ok(bytes) = self.store.get_typed(&r, ObjectType::Checkpoint) {
110	keep.insert(r);
111	let cp = crate::objects::Checkpoint::decode(&bytes)?;
112	keep.insert(cp.payload);
113	if let Ok(b) = self.store.get_typed(&cp.payload, ObjectType::Blob) {
114	keep.extend(Blob::decode(&b)?.chunks);
115	}
116	// A version 2 payload's graph blobs and their chunks.
117	for g in self.checkpoint_graph_blobs(&cp).unwrap_or_default() {
118	keep.insert(g);
119	if let Ok(b) = self.store.get_typed(&g, ObjectType::Blob) {
120	keep.extend(Blob::decode(&b)?.chunks);
121	}
122	}
123	}
124	}
125	// Tags: their states and attestations.
126	let view_now = self.view()?;
127	states.extend(view_now.tags.values().copied());
128	for name in view_now.tags.keys() {
129	if let Some((r, _)) = self.tag_attest(name)? {
130	attests.push(r);
131	}
132	}
133	// Reference attestations.
134	for name in self.view()?.bookmarks.keys() {
135	if let Some((r, _)) = self.latest_ref_attest(name)? {
136	// With a host's acceptance of it (spec §14.7).
137	if let Some(acc) = self.acceptance_of(&r)? {
138	attests.push(acc.attest.hash());
139	}
140	attests.push(r);
141	}
142	}
143	// A historical view may name objects pruned under an earlier, shorter
144	// retention; there is nothing to keep for those.
145	for st in states {
146	if keep.insert(st) {
147	match self.store.get_state(&st) {
148	Ok(s) => changes.extend(s.heads),
149	Err(StoreError::NotFound(_)) => continue,
150	Err(e) => return Err(e.into()),
151	}
152	}
153	}
154	for p in policies {
155	if keep.insert(p) {
156	let bytes = match self.store.get_typed(&p, ObjectType::Policy) {
157	Ok(b) => b,
158	Err(StoreError::NotFound(_)) => continue,
159	Err(e) => return Err(e.into()),
160	};
161	let policy = crate::policy::Policy::decode(&bytes)?;
162	for id in &policy.identities {
163	keep.insert(id.hash());
164	}
165	for sig in self.store.signatures(&p)? {
166	keep.insert(sig.key);
167	}
168	}
169	}
170	// Changes: their closure, identities, blobs, chunks, seals.
171	let mut stack = changes;
172	while let Some(h) = stack.pop() {
173	if !keep.insert(h) {
174	continue;
175	}
176	// Beyond the shallow boundary, or pruned under an earlier
177	// retention, there is nothing to keep.
178	let c = match self.store.get_change_parts(&h) {
179	Ok((c, texts)) => {
180	keep.extend(texts);
181	c
182	}
183	Err(StoreError::NotFound(_)) => continue,
184	Err(e) => return Err(e.into()),
185	};
186	stack.extend(c.deps.iter().copied());
187	if !c.author.is_zero() {
188	keep.insert(c.author);
189	}
190	for sig in self.store.signatures(&h)? {
191	self.keep_signer(&mut keep, sig.key)?;
192	}
193	for e in &c.edits {
194	if let Edit::Replace { old, new, .. } = e {
195	for b in [old, new].into_iter().flatten() {
196	if keep.insert(*b) {
197	if let Ok(bytes) = self.store.get_typed(b, ObjectType::Blob) {
198	keep.extend(Blob::decode(&bytes)?.chunks);
199	}
200	}
201	}
202	}
203	}
204	attests.extend(self.store.seals_for(&h)?);
205	}
206	self.reachable_receipts(&mut keep)?;
207	self.reachable_approvals(&mut keep)?;
208	for a in attests {
209	if keep.insert(a) {
210	for sig in self.store.signatures(&a)? {
211	self.keep_signer(&mut keep, sig.key)?;
212	}
213	}
214	}
215	Ok(keep)
216	}
217
218	/// Keep a signer's identity and, for a short-lived key, its
219	/// certificate and the certificate's issuer (spec §14.7).
220	pub(crate) fn keep_signer(&self, keep: &mut HashSet<ObjectRef>, key: ObjectRef) -> Result<()> {
221	keep.insert(key);
222	if let Some(c) = self.store.certificate_of(&key)? {
223	keep.insert(c);
224	if let Ok(cert) = self.store.get_certificate(&c) {
225	keep.insert(cert.cert.issuer);
226	}
227	}
228	Ok(())
229	}
230	}

Keyboard