Owner

pchrysostomou

1 indexed repository · View on GitHub

ananke
A distributed SQL database in Rust, built from the ground up to be deterministically testable.

Rust · 0 stars

Beginner friendly Everything

Open All

Issues somebody may already be working on are listed last. List everything by date

A split's apply narrows P's descriptor and inserts R's under two locks; on a multi-threaded runtime a placeholder's install of R can pass between them
Open
- Is something wrong?
phase:3

Difficulty 2/5 1-3 hours Newbie friendliness 72/100

pchrysostomou/ananke#183 · Opened Oct 25, 2026

Maintainers usually reply within 1 day
A replica created twice on one node, by a split's apply and an install, is seen by no range-layer check (seed 219 under SnapshotOverlapsReplica); the apply's "R already initialised here" assertion
Open
- Is something wrong?
phase:3

Difficulty 4/5 3-5 days Newbie friendliness 43/100

pchrysostomou/ananke#181 · Opened Oct 25, 2026

Maintainers usually reply within 1 day
A re-seed refuses and marks only configuration's ranges, so a replica outside `hosted` gets no refused mark and no RaftReplicaRefused (Q15) — awaiting the owner
Open
- Is something wrong?
phase:3 question

Difficulty 5/5 Over a week Newbie friendliness 25/100

pchrysostomou/ananke#178 · Opened Oct 25, 2026

Maintainers usually reply within 1 day
After a re-seed, until its next start a node stages, takes and sweeps snapshot directories in the directory it refused (D-041)
Open
- Is something wrong?
phase:3

Difficulty 4/5 3-5 days Newbie friendliness 42/100

pchrysostomou/ananke#176 · 1 comment · Opened Oct 19, 2026

Maintainers usually reply within 1 day
A node re-seeded after it applied a split replays the split and creates the right half unquarantined, though the right half it lost may have voted (Q26)
Open
- Is something wrong?
phase:3 question

Difficulty 5/5 Over a week Newbie friendliness 25/100

pchrysostomou/ananke#175 · 1 comment · Opened Oct 19, 2026

Maintainers usually reply within 1 day
No run re-seeds a node after a split, so the install creation's generation is exercised by no scenario
Open
- Is something wrong?
phase:3

Difficulty 4/5 1-2 days Newbie friendliness 48/100

pchrysostomou/ananke#172 · 1 comment · Opened Oct 18, 2026

Maintainers usually reply within 1 day
The studio cannot tell a descriptor stated at a restart or an install from an apply's change
Open
- Is something wrong?
phase:3

Difficulty 4/5 3-5 days Newbie friendliness 44/100

pchrysostomou/ananke#171 · Opened Oct 18, 2026

Maintainers usually reply within 1 day
The non-voter's split apply rewrites most of its range's tables while every apply waits, and restarts a live install it meets (Stage D)
Open
- Is something wrong?
phase:3

Difficulty 4/5 3-5 days Newbie friendliness 45/100

pchrysostomou/ananke#169 · Opened Oct 18, 2026

Maintainers usually reply within 1 day
routing_variant's first-violation print cuts a verdict by bytes and can panic inside a multibyte character
Open
- Is something wrong?
phase:3

Difficulty 1/5 Under an hour Newbie friendliness 85/100

pchrysostomou/ananke#168 · Opened Oct 18, 2026

Maintainers usually reply within 1 day
The node restates every snapshot as installed: restating each record's taken catches SnapshotWithoutCurrentLast (#126), and needs the inbox to carry the net task's receipt stamp first
Open
- Is something wrong?
phase:3

Difficulty 5/5 Over a week Newbie friendliness 25/100

pchrysostomou/ananke#167 · Opened Oct 18, 2026

Maintainers usually reply within 1 day
SHARD.md:1379-1381 gives n{id}/r{range}/protocol as a stream derived anew at each creation; range_rng is one continuing stream per node and range
Open
- Is something wrong?
phase:3

Difficulty 1/5 Under an hour Newbie friendliness 45/100

pchrysostomou/ananke#165 · Opened Oct 18, 2026

Maintainers usually reply within 1 day
Phase 3 Stages C–E: the plan, where it stands, what is left (tracking)
Open
- Is something wrong?
phase:3

Difficulty 5/5 Over a week Newbie friendliness 15/100

pchrysostomou/ananke#162 · Opened Oct 18, 2026

Maintainers usually reply within 1 day
IncarnationPerRangeStream's read at a split's right half (D-100) is run by no test
Open
- Is something wrong?
phase:3

Difficulty 3/5 1-2 days Newbie friendliness 48/100

pchrysostomou/ananke#160 · 2 comments · Opened Oct 12, 2026

Maintainers usually reply within 1 day
Liveness: a pending post-heal write older than the bound is dropped whenever its key has a completed one
Open
- Is something wrong?
phase:3

Difficulty 4/5 1-2 days Newbie friendliness 45/100

pchrysostomou/ananke#158 · Opened Oct 11, 2026

Maintainers usually reply within 1 day
The liveness per-range reading keys a split's halves by the run's starting map, so a wedged half beside a live sibling passes it
Open
- Is something wrong?
phase:3

Difficulty 5/5 Over a week Newbie friendliness 12/100

pchrysostomou/ananke#148 · Opened Oct 10, 2026

Maintainers usually reply within 1 day
SplitFromRestarted's no-split branch skips its gap and leaves last_heal at the arm before it: say whether that is meant
Open
- Is something wrong?
phase:3

Difficulty 4/5 1-2 days Newbie friendliness 22/100

pchrysostomou/ananke#147 · Opened Oct 10, 2026

Maintainers usually reply within 1 day
After a shrink, the sweep client can cycle between a range's two removed servers and abandon every write it routes there
Open
- Is something wrong?
phase:3

Difficulty 4/5 3-5 days Newbie friendliness 35/100

pchrysostomou/ananke#146 · Opened Oct 10, 2026

Maintainers usually reply within 1 day
The node membership scenario's per-range liveness bound cannot trip: its post-heal window is 2.4 s against RANGE_LIVENESS_TIMEOUTS' 6 s
Open
- Is something wrong?
phase:3

Difficulty 4/5 3-5 days Newbie friendliness 15/100

pchrysostomou/ananke#145 · Opened Oct 10, 2026

Maintainers usually reply within 1 day
The sharded quorum scenario asks its hold clauses of the four user ranges only: a refused root replica that votes passes 10 000 seeds
Open
- Is something wrong?
phase:3

Difficulty 4/5 3-5 days Newbie friendliness 30/100

pchrysostomou/ananke#143 · Opened Oct 10, 2026

Maintainers usually reply within 1 day
Stage C: a node check that observes SnapshotWithoutCurrentLast — the arm fires on 14 of 100 node seeds and nothing catches it
Open
- Is something wrong?
phase:3

Difficulty 5/5 Over a week Newbie friendliness 35/100

pchrysostomou/ananke#126 · 1 comment · Opened Sep 24, 2026

Maintainers usually reply within 1 day
READS_OUTSTANDING counts a client's superseded retries as distinct outstanding reads; revisit once clients get sessions (Phase 4)
Open
- Is something wrong?
help wanted phase:4

Difficulty 5/5 Over a week Newbie friendliness 20/100

pchrysostomou/ananke#125 · Opened Sep 24, 2026

Maintainers usually reply within 1 day
The node's snapshot resend bound is tripped by the correct system on every install: the receiver's deliberate silence is counted as a lost chunk
Open
- Is something wrong?
bug phase:3

Difficulty 4/5 3-5 days Newbie friendliness 45/100

pchrysostomou/ananke#120 · Opened Sep 23, 2026

Maintainers usually reply within 1 day
D-049's rule has no site on the node: a refused replica answers with a store incarnation, never 0, and the snapshot wiring will not change that
Open
- Is something wrong?
Difficulty 5/5 Over a week Newbie friendliness 25/100

pchrysostomou/ananke#116 · Opened Sep 23, 2026

Maintainers usually reply within 1 day
The snapshot task serves a chunk deadline only when its queue is idle
Open
- Is something wrong?
phase:3

Difficulty 4/5 3-5 days Newbie friendliness 35/100

pchrysostomou/ananke#114 · Opened Sep 22, 2026

Maintainers usually reply within 1 day
A node restarted over replicas an install filled can apply from an index its core no longer holds
Open
- Is something wrong?
phase:3

Difficulty 5/5 Over a week Newbie friendliness 35/100

pchrysostomou/ananke#113 · Opened Sep 22, 2026

Maintainers usually reply within 1 day
RealEnv::spawn drops a panicking task's result, so a production assertion that is fatal under simulation is silent in a shipped server
Open
- Is something wrong?
Difficulty 4/5 3-5 days Newbie friendliness 45/100

pchrysostomou/ananke#112 · Opened Sep 22, 2026

Maintainers usually reply within 1 day
Put sim/tests/install.rs under the four tiers, with a weight measured on a quiet machine
Open
- Is something wrong?
phase:3

Difficulty 4/5 3-5 days Newbie friendliness 38/100

pchrysostomou/ananke#108 · Opened Sep 22, 2026

Maintainers usually reply within 1 day
commit_majority takes a replica created outside the configuration as the group's birth configuration
Open
- Is something wrong?
phase:3

Difficulty 3/5 1-2 days Newbie friendliness 68/100

pchrysostomou/ananke#105 · Opened Sep 22, 2026

Maintainers usually reply within 1 day
shard: the snapshot wiring must step SnapshotAcked into the acknowledged range's core, and nothing in the tree can catch it if it does not
Open
- Is something wrong?
phase:3

Difficulty 4/5 3-5 days Newbie friendliness 35/100

pchrysostomou/ananke#103 · Opened Sep 22, 2026

Maintainers usually reply within 1 day
sim: a seed's run depends on whether rustc compiled incrementally, so the gate's green depends on the state of target/
Open
- Is something wrong?
phase:3

Difficulty 5/5 Over a week Newbie friendliness 35/100

pchrysostomou/ananke#102 · 1 comment · Opened Sep 22, 2026

Maintainers usually reply within 1 day
shard: a re-seed writes one synced batch per replica, so a crash can leave some of a node's replicas marked and others not
Open
- Is something wrong?
phase:3

Difficulty 4/5 3-5 days Newbie friendliness 44/100

pchrysostomou/ananke#100 · Opened Sep 22, 2026

Maintainers usually reply within 1 day
shard: a re-seed's new directory is openable while it is still being filled, so OpenNewestEvenIfLost has no reachable case
Open
- Is something wrong?
phase:3

Difficulty 4/5 3-5 days Newbie friendliness 42/100

pchrysostomou/ananke#99 · Opened Sep 22, 2026

Maintainers usually reply within 1 day
Two pinned tests in sim/tests/raft.rs say something other than what they assert
Open
- Is something wrong?
phase:3

Difficulty 3/5 1-2 days Newbie friendliness 57/100

pchrysostomou/ananke#98 · 1 comment · Opened Sep 22, 2026

Maintainers usually reply within 1 day
The node admits an InstallSnapshot chunk to its inbox and then drops it in silence
Open
- Is something wrong?
phase:3

Difficulty 4/5 3-5 days Newbie friendliness 42/100

pchrysostomou/ananke#96 · Opened Sep 22, 2026

Maintainers usually reply within 1 day
Re-audit the RAFT.md citations D-075's insertion moved (+13 below line 207)
Open
- Is something wrong?
phase:3

Difficulty 2/5 1-3 hours Newbie friendliness 78/100

pchrysostomou/ananke#94 · Opened Sep 22, 2026

Maintainers usually reply within 1 day
lin::check's Timeline return value has no consumer, so the order it returns is unchecked
Open
- Is something wrong?
Difficulty 5/5 Over a week Newbie friendliness 35/100

pchrysostomou/ananke#93 · Opened Sep 22, 2026

Maintainers usually reply within 1 day
A linearizability search that exhausts its budget is counted as a catch by every variant counter
Open
- Is something wrong?
Difficulty 4/5 3-5 days Newbie friendliness 52/100

pchrysostomou/ananke#92 · Opened Sep 22, 2026

Maintainers usually reply within 1 day
D-065's compaction record has no node-level test: the fields the store really holds are invisible to every sweep
Open
- Is something wrong?
phase:3

Difficulty 4/5 3-5 days Newbie friendliness 42/100

pchrysostomou/ananke#88 · Opened Sep 22, 2026

Maintainers usually reply within 1 day
docs: SHARD.md cites the wrong ananke-raft lines for the ticker loop and for `execute`
Open
- Is something wrong?
phase:3

Difficulty 1/5 Under an hour Newbie friendliness 85/100

pchrysostomou/ananke#80 · Opened Sep 21, 2026

Maintainers usually reply within 1 day
shard: a directed scenario for the applied-stream gap, once the node has a compaction path
Open
- Is something wrong?
phase:3

Difficulty 4/5 3-5 days Newbie friendliness 48/100

pchrysostomou/ananke#79 · Opened Sep 21, 2026

Maintainers usually reply within 1 day
A read served from a pinned engine version can straddle a live span install's switch (D-069's caveat)
Open
- Is something wrong?
phase:3

Difficulty 5/5 Over a week Newbie friendliness 35/100

pchrysostomou/ananke#72 · Opened Sep 20, 2026

Maintainers usually reply within 1 day
The nightly's shard table: find sweeps by cargo target, not by glob, and exclude them from rest by target
Open
- Is something wrong?
phase:3

Difficulty 4/5 3-5 days Newbie friendliness 48/100

pchrysostomou/ananke#66 · 1 comment · Opened Sep 19, 2026

Maintainers usually reply within 1 day
D-063's exemption is fenced by nothing: a widened arm passes every test
Open
- Is something wrong?
phase:2

Difficulty 4/5 1-2 days Newbie friendliness 48/100

pchrysostomou/ananke#65 · Opened Sep 16, 2026

Maintainers usually reply within 1 day
Nothing bounds an adoption now that D-063 stops measuring it: a bound of its own on completion to restatement
Open
- Is something wrong?
phase:2

Difficulty 4/5 3-5 days Newbie friendliness 38/100

pchrysostomou/ananke#64 · Opened Sep 16, 2026

Maintainers usually reply within 1 day
The crash model traces nothing when it drops a pending truncation, and Wal::open's firsts map over-estimates
Open
- Is something wrong?
phase:1

Difficulty 3/5 1-2 days Newbie friendliness 64/100

pchrysostomou/ananke#63 · Opened Sep 16, 2026

Maintainers usually reply within 1 day
A resurrected WAL segment whose numbers abut the live ones needs a per-segment generation stamp (the limit of D-062)
Open
- Is something wrong?
phase:1

Difficulty 5/5 Over a week Newbie friendliness 25/100

pchrysostomou/ananke#62 · Opened Sep 16, 2026

Maintainers usually reply within 1 day
The correct engine fails seed 30490 on the seek schedule: a record gone with no WAL stop, pointing at the manifest fallback
Open
- Is something wrong?
phase:1

Difficulty 5/5 Over a week Newbie friendliness 35/100

pchrysostomou/ananke#61 · Opened Sep 16, 2026

Maintainers usually reply within 1 day
A pinned seed that asserts only green: seed 420, and a sweep of sim/tests for the others
Open
- Is something wrong?
phase:2

Difficulty 4/5 1-2 days Newbie friendliness 55/100

pchrysostomou/ananke#59 · Opened Sep 16, 2026

Maintainers usually reply within 1 day
adopt_checked reads the staged format record after the manifest, so a newer-format install behind damaged staging is re-seeded instead of refused
Open
- Is something wrong?
phase:3

Difficulty 3/5 1-2 days Newbie friendliness 55/100

pchrysostomou/ananke#58 · Opened Sep 16, 2026

Maintainers usually reply within 1 day
Membership: the truncation revert floor and the kept-tail key repair under snapshots (split from #46)
Open
- Is something wrong?
phase:3

Difficulty 4/5 3-5 days Newbie friendliness 45/100

pchrysostomou/ananke#56 · 2 comments · Opened Sep 15, 2026

Maintainers usually reply within 1 day

pchrysostomou

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Is something wrong?

Get new issues in your inbox