OXYGENHUB
# f7510b63cceb oxygen / crates / oxy-core / src / repo / gc.rs
5.9 KiB · Rust · last changed Checkpoint payload version 2: large file graphs as shared blobs named from an index, read in parallel (loose and packed) and decoded in parallel; version 1 still read; gc, fsck and shallow-clone bundles carry the graph blobs (spec 11, Draft 0.8j) 1 month ago source blame history raw
1 
//! Garbage collection (spec §21): pack loose objects, then delete objects
2 
//! nothing reachable refers to.
3 
//!
4 
//! Roots: the current view, the views of every operation newer than
5 
//! `core.op-retention`, reference attestations, and the policy. Operation
6 
//! objects are always kept so the log stays walkable, but an operation older
7 
//! than the retention window no longer keeps its states alive, so undoing
8 
//! that far back may fail once its objects are gone.
9
10 
use super::*;
11 
use crate::store::{PackReport, PruneReport};
12 
use std::collections::HashSet;
13
14 
#[derive(Debug, Clone, Default, PartialEq, Eq)]
15 
pub struct GcReport {
16 
    pub packed: PackReport,
17 
    pub pruned: PruneReport,
18 
    pub reachable: usize,
19 
    /// Packs folded into one because there were more than `MAX_PACKS`.
20 
    pub merged_packs: usize,
21 
}
22
23 
impl Repo {
24 
    /// Pack loose objects and prune unreachable ones older than the grace
25 
    /// period (`now` ignores the grace period).
26 
    pub fn gc(&self, now: bool) -> Result<GcReport> {
27 
        let packed = self.store.pack_loose()?;
28 
        let keep = self.reachable()?;
29 
        let grace = if now {
30 
            std::time::Duration::ZERO
31 
        } else {
32 
            std::time::Duration::from_secs(self.config.gc_grace_days * 86_400)
33 
        };
34 
        let pruned = self.store.prune(&keep, grace)?;
35 
        let merged_packs = self.store.merge_packs()?;
36 
        Ok(GcReport {
37 
            packed,
38 
            pruned,
39 
            reachable: keep.len(),
40 
            merged_packs,
41 
        })
42 
    }
43
44 
    /// Every object that must survive garbage collection.
45 
    pub fn reachable(&self) -> Result<HashSet<ObjectRef>> {
46 
        let mut keep: HashSet<ObjectRef> = HashSet::new();
47 
        let mut states: Vec<ObjectRef> = Vec::new();
48 
        let mut changes: Vec<ChangeHash> = Vec::new();
49 
        let mut attests: Vec<ObjectRef> = Vec::new();
50 
        let mut policies: Vec<ObjectRef> = Vec::new();
51
52 
        // Operations: always kept; views within retention are roots.
53 
        let cutoff = Hlc::now()
54 
            .wall_ms
55 
            .saturating_sub(self.config.op_retention_days * 86_400_000);
56 
        for (i, (r, op)) in self.op_log()?.into_iter().enumerate() {
57 
            keep.insert(r);
58 
            if i == 0 || op.hlc.wall_ms >= cutoff {
59 
                states.extend(op.view.bookmarks.values().copied());
60 
                states.extend(op.view.remotes.values().flat_map(|b| b.values().copied()));
61 
                changes.extend(op.view.workspaces.values().copied());
62 
                policies.extend(op.view.policy);
63 
            }
64 
        }
65 
        // Checkpoints: the object, its payload blob, and the blob's chunks.
66 
        for r in self.checkpoint_refs()? {
67 
            if let Ok(bytes) = self.store.get_typed(&r, ObjectType::Checkpoint) {
68 
                keep.insert(r);
69 
                let cp = crate::objects::Checkpoint::decode(&bytes)?;
70 
                keep.insert(cp.payload);
71 
                if let Ok(b) = self.store.get_typed(&cp.payload, ObjectType::Blob) {
72 
                    keep.extend(Blob::decode(&b)?.chunks);
73 
                }
74 
                // A version 2 payload's graph blobs and their chunks.
75 
                for g in self.checkpoint_graph_blobs(&cp).unwrap_or_default() {
76 
                    keep.insert(g);
77 
                    if let Ok(b) = self.store.get_typed(&g, ObjectType::Blob) {
78 
                        keep.extend(Blob::decode(&b)?.chunks);
79 
                    }
80 
                }
81 
            }
82 
        }
83 
        // Tags: their states and attestations.
84 
        let view_now = self.view()?;
85 
        states.extend(view_now.tags.values().copied());
86 
        for name in view_now.tags.keys() {
87 
            if let Some((r, _)) = self.tag_attest(name)? {
88 
                attests.push(r);
89 
            }
90 
        }
91 
        // Reference attestations.
92 
        for name in self.view()?.bookmarks.keys() {
93 
            if let Some((r, _)) = self.latest_ref_attest(name)? {
94 
                attests.push(r);
95 
            }
96 
        }
97 
        for st in states {
98 
            if keep.insert(st) {
99 
                changes.extend(self.store.get_state(&st)?.heads);
100 
            }
101 
        }
102 
        for p in policies {
103 
            if keep.insert(p) {
104 
                let policy = crate::policy::Policy::decode(&self.store.get_typed(&p, ObjectType::Policy)?)?;
105 
                for id in &policy.identities {
106 
                    keep.insert(id.hash());
107 
                }
108 
                for sig in self.store.signatures(&p)? {
109 
                    keep.insert(sig.key);
110 
                }
111 
            }
112 
        }
113 
        // Changes: their closure, identities, blobs, chunks, seals.
114 
        let mut stack = changes;
115 
        while let Some(h) = stack.pop() {
116 
            if !keep.insert(h) {
117 
                continue;
118 
            }
119 
            // Beyond the shallow boundary there is nothing to keep.
120 
            let c = match self.store.get_change(&h) {
121 
                Ok(c) => c,
122 
                Err(StoreError::NotFound(_)) => continue,
123 
                Err(e) => return Err(e.into()),
124 
            };
125 
            stack.extend(c.deps.iter().copied());
126 
            if !c.author.is_zero() {
127 
                keep.insert(c.author);
128 
            }
129 
            for sig in self.store.signatures(&h)? {
130 
                keep.insert(sig.key);
131 
            }
132 
            for e in &c.edits {
133 
                if let Edit::Replace { old, new, .. } = e {
134 
                    for b in [old, new].into_iter().flatten() {
135 
                        if keep.insert(*b) {
136 
                            if let Ok(bytes) = self.store.get_typed(b, ObjectType::Blob) {
137 
                                keep.extend(Blob::decode(&bytes)?.chunks);
138 
                            }
139 
                        }
140 
                    }
141 
                }
142 
            }
143 
            attests.extend(self.store.seals_for(&h)?);
144 
        }
145 
        for a in attests {
146 
            if keep.insert(a) {
147 
                for sig in self.store.signatures(&a)? {
148 
                    keep.insert(sig.key);
149 
                }
150 
            }
151 
        }
152 
        Ok(keep)
153 
    }
154 
}