aboutsummaryrefslogtreecommitdiffstats
path: root/src/remake/catchup.rs
Commit message (Collapse)AuthorAgeFilesLines
* remake: holding a store, a node rejoins a lower epoch a peer servestv17 hours1-29/+17
| | | | | | | | Of two epochs decided for one, a node holding a store follows the lower ID, and a node on the other rejoins it, however many peers are on each (`docs/store.md`, Remakes): one peer on it is enough, so that two sides of unequal size come together. A made-up epoch keeps a node apart at worst, and makes it take nothing it wouldn't (`Holding::follows`).
* remake: holding a store, a node follows an epoch only from a checkpoint it ↵tv17 hours1-0/+16
| | | | | | | | | | | | | builds A node of a swarm holding a store rejoining an epoch (`docs/store.md`, Remakes) follows it only where its snapshot has the swarm's policy and store, and holds the node's document's checkpoint, one it builds from that document, or one its operator pinned; with an empty document, the root of its store too (`Holding::follows`). A made-up epoch holding another, the root after the store moved on, say, would leave it holding events whose history nobody puts again: it stays on its own, as when a rejoin's check fails, and tries again later.
* remake: holding a store, of two epochs decided for one, the lower IDtv17 hours1-3/+85
| | | | | | | | | | In a swarm holding a store, of two epochs decided for one, by a coordinator that wrote two decisions or by remakes decided at once, a node follows the one with the lower ID, and a node on the other rejoins it, however many peers are on each (`docs/store.md`, Remakes): a node switches to the lower of two committed remakes of its epoch, follows the lowest of several newer epochs, and a ready node rejoins a lower one of its own generation that enough peers serve.
* remake: holding a store, a node puts again what the document it left holdstv17 hours1-0/+11
| | | | | | | | | | After a switch or a rejoin, a node of a swarm holding a store puts again what the document it left holds that a store starting at the new document's checkpoint takes, by its own refusals and drops, and the new one lacks (`docs/store.md`, Remakes, `Holding::left`): what it held beyond the base, and, where it moved to a later checkpoint since, what a store starting at the snapshot's takes to reach it. Its events aren't replayed into the new document, so none nobody takes go along.
* remake: a holding swarm's snapshot holds the store read at its basetv18 hours1-0/+2
| | | | | | | | | | | A compaction's snapshot in a swarm holding a store holds the store as the node reads it from the document at the base, by its own refusals and its operator's drops (`docs/store.md`, Remakes): the checkpoint it reaches there, where it isn't the root, at `_continuum/checkpoint`; the root, the events that checkpoint keeps and those after it, each at its ID; and of the entries, the members'. What the checkpoint forgot, events nobody takes, and the entries of nodes that aren't members stay behind (`Holding::at_base`, `events::snapshot`, `Host::holding`).
* remake: earlier epochs' metadata read as the snapshot built ittv9 days1-3/+10
| | | | | | | | | | | | | | A snapshot leaving out earlier epochs' metadata (`snapshot::left_out`) and a rejoining node (`catchup::prepare`) read it from the document as it stands. It is never writable through the store, and under a policy the rules take a change to it only from an epoch's first; but with no policy, screening judges only an epoch's root, so a member's own change could rewrite or remove an entry: a replay window narrowed or widened, a translation or a clock changed, a strict remake's entry gone. Now both read it at the epoch's snapshot change (`snapshot::as_built`), as `verify_document` reads the epoch's own: a snapshot from its base's epoch's, a rejoining node from the target's. Every earlier entry is a copy in that change, so what was written there since doesn't count. The first epoch follows none and has none.
* remake: leave out earlier epochs' metadata past the replay windowtv9 days1-1/+5
| | | | | | | | | | | | | | | | | | | | | Every snapshot carried the metadata of all earlier epochs (`_continuum/epochs`) forward, so it grew by an entry, a clock per actor among it, with every remake. Only a rejoining node reads an earlier entry: a node on epoch P reads P+1's (its clock, when it was proposed, its replay window) and every one up to the swarm's epoch (kind, translation, whether it was strict). It replays only while P+1's replay window covers it; beyond, its writes are quarantined whatever the entries say. So a snapshot now leaves out the generations older than the oldest whose replay window still covers a node on the epoch before it (`snapshot::left_out`), judged by the remake's `proposedAt`: the same on every node, and earlier than any rejoining node's clock, so nothing a node can still replay through goes. Metadata that doesn't read stays, with all after it. A strict remake's metadata stays for good: by it a rejoining node, however far behind, tells a strict remake of a swarm strict already (`catchup::prepare`, which reads the generations it finds). There is at most one, short of such a remake. No new key and no change to the rejoin: a node beyond the window that finds some entries and not others quarantines with "cannot translate" rather than "outside the replay window".
* remake: a strict remake of a strict swarm isn't followedtv9 days1-0/+16
| | | | | | | | | | | | | | | Review 3, R3: 1186 kept honest nodes from proposing or voting for a strict remake of a swarm strict already, but a majority of members, each with keys of its own, could still commit one, and switch and catchup followed it without asking what kind of remake made the epoch: the registry, policy and bans of a strict swarm changing by a member majority, the honest minority following. A node now refuses to switch to such an epoch (its document's policy strict, the remake's kind strict), and catchup refuses to rejoin one whose chain of epochs holds a strict remake after the swarm was strict (as the old document is, or once a remake since made it so). access.md says so.
* rules: a carried remake record taken from the coordinator its snapshot namestv9 days1-0/+1
| | | | | | | | | | | | | | | | | Review finding 7, second part: a record carried into a new epoch whole (proposal and decision in one change) was taken from any member, for any remake ID. A member could write a forged record, decision included, for a fresh ID in a single change. The snapshot's epoch metadata now names the remake's coordinator, as the quorum built it. The carried-record exception holds only in a later epoch, only for the remake that made it, and only from the coordinator its snapshot names, where there is no record yet. Wire version 43: the snapshot, and so the epoch ID, changes. The review suggested carrying the record inside the snapshot instead. The snapshot can't hold the decision, which names the snapshot's own hash. So the snapshot vouches for who may carry the record instead.
* remake: replays judged before they go intv9 days1-0/+2
| | | | | | | | | | | | | | | A switching or rejoining node replays late writes as fresh changes naming each write's user, and nothing judged them before they left: a replay every node would refuse (a revoked capability, a second envelope at a signed path, readers changed) was refused everywhere, and with it everything the node wrote after it. Under a policy, replay_newest now judges each group it commits as every node will (rules::judge_held, in the new epoch); one they would refuse is taken back (forked at the heads before it, same actor) and its paths quarantined with the reason. replay_newest takes the epoch. Review finding 3 (remake replays).
* remake: replays keep their writerstv10 days1-0/+3
| | | | | | | | A remake's replay writes each value as the user who wrote it, splitting where several users wrote under a path, so replays are judged as the original writes. Auth plan M6 (remakes, 2.4).
* document: allow writes during rejointv2026-11-011-178/+210
|
* LWW everywhere, mostlytv2026-11-011-9/+25
|
* remake: copy stale view when writes stoptv2026-11-011-5/+6
| | | | This ensures that stale reads show every write that was answered.
* remake: introduce follow-the-majority ruletv2026-11-011-27/+139
|
* document: reject writes during rejointv2026-11-011-17/+35
|
* use Automerge as Document instead of AutoCommittv2026-10-311-17/+15
| | | | This allows a lot of operations to drop mut.
* tests write through document::testing helperstv2026-10-311-6/+6
|
* make clippy happy againtv2026-10-261-8/+11
|
* wrap unix timestamps in newtypetv2026-10-261-7/+22
|
* add clock moduletv2026-10-251-9/+25
| | | | | The Clock module is the system clock that every unix timestamp new goes through. It comes with an offset that only tests can set.
* only keep list of actor IDs within one epoch (fixup)tv2026-10-251-0/+2
|
* nodes keep list of every actor ID they've usedtv2026-10-251-9/+23
|
* corroberate newer epochstv2026-10-241-11/+53
|
* follow and download from ready members onlytv2026-10-241-1/+3
|
* put remake content into blobstv2026-10-241-1/+1
|
* add chunked download and use it during rejointv2026-10-241-2/+47
|
* rename cluster to swarm everywhere!tv2026-10-211-14/+14
|
* turn compaction into a kind of remaketv2026-10-211-22/+49
|
* rename compaction to remaketv2026-10-211-0/+506