aboutsummaryrefslogtreecommitdiffstats
path: root/docs/peers.md
Commit message (Collapse)AuthorAgeFilesLines
* peers: what each peer said it speaks and knows, in GET /peerstv4 hours1-4/+5
| | | | | | | | | | | | | | | | GET /peers shows, of each peer, what it said with its handshakes (SPEAKS): its `knows`, the wire versions it speaks, the oldest and the newest, and the newest revision of the federation it knows; null where it said nothing, a build from before SPEAKS or no handshake from it since the node started. `localKnows` is the node's own. So every swarm, federated or not, shows which builds its nodes run, as far as each heard. `swarm::speaking::shown` shapes it, for GET /events/counts too, whose `knows` is the same; docs/openapi.yaml has it once, as `Knows`. Without the federation feature, a revision reads as one this continuum doesn't know. `Speaking::of` is now used in every build, not only federated ones, which warned of it as dead code.
* federation: which nodes know a revision, in counts and before a `true`tv5 hours1-3/+4
| | | | | | | | | | | | | | | | | | `GET /events/counts` says, of each node a party lists, what it said it speaks and knows (`knows`, by node ID): the wire versions, the oldest and the newest, and the newest revision of the federation it knows, by number and ID and as people read it, its number null where this continuum doesn't know it; null where it said nothing, a build from before SPEAKS or one the node didn't hear from; the node's own as it is. So the nodes still to upgrade before a revision passes show, every party's, where before a party's operators could check only their own. A party's client, backing a proposal that moves to another revision, warns of every node of any party not known to know it, a line a party naming each and what it said: once it passes, those stop at it till upgraded. `next` warns alike, of the revision it moves to, or of the later one it suggests. A warning, not a refusal: what a node heard is no proof, and a node not heard from may well be upgraded.
* transport: SPEAKS, which wire versions a node speaks and what it knowstv5 hours1-0/+8
| | | | | | | | | | | | | | | | | | | | | | | Before each HELLO and each HELLO_REPLY, a node sends SPEAKS: the oldest and the newest wire version it speaks, and the ID of the newest revision of the federation it knows. It stands outside the wire versions, its first byte 0, which no version is, laid out as no build changes: that byte, the oldest, the newest, then 32 bytes of the revision's ID where there is one; a reader takes those and leaves whatever follows, which a later build may add. So a peer on another build reads it too, and nodes on different builds can say what they run, which a rollout needs. A build from before it takes it for a message of another wire version and logs that once of the peer (version=0), reading the rest of the peer's messages as before: SPEAKS rolls out node by node, with no change of VERSION. The node keeps what each peer said last (`swarm::speaking`, at most 4096 peers). It counts as `hello` traffic. `OLDEST`, the oldest version this build reads, is `VERSION` for now: a range for a later build to widen. operations.md also says what is so of the other-version warning now: every transport logs it once per peer at warn, not Mycelium alone.
* docs: references to another document's section linked to ittv29 hours1-11/+12
| | | | | | | | Where a document names a section of another after linking to it ("([`store.md`](store.md), Pinned checkpoints)"), the name links to that section too, by the id both GitHub and cgit's about filter give its heading; paragraphs that grew past the width are wrapped anew, links and code kept on one line each.
* relay: a STATUS's heads wait at least 50 ms before they are pulledtv6 days1-1/+2
| | | | | | | | | | | | | | A neighbour pulls heads a STATUS listed only if it still lacks them after a round trip to the sender, as they are usually on their way along a writer's tree. With no round trip known, or one measured long ago (busy pairs send no keepalives), it pulled at once; with a short one, a STATUS sent right after a change can overtake it by more than that round trip, delays varying per message. Either way the change came twice (`trees_hold_through_continuous_writes`: c pulled once while every write reached it along x's tree). The wait is the round trip, but 50 ms at least (and 1 s at most, as before).
* dashboard: without LINKS; LINKS removedtv12 days1-3/+2
| | | | | | | | | The dashboard reads every member's links from the members themselves; LINKS and LINKS_REPLY, swarm::links and their wire tags go (wire version bumped). Nodes the browser can't reach are drawn from what the reachable ones measure. Auth plan M3.
* substitutes held longertv13 days1-2/+3
|
* indirect probestv13 days1-2/+10
|
* unlisted members forgottontv14 days1-1/+3
|
* fix doc (working links measured)tv14 days1-1/+2
|
* trees over working linkstv14 days1-3/+5
|
* working links recordedtv14 days1-0/+10
|
* suspend detectedtv14 days1-0/+7
|
* views documentedtv14 days1-0/+11
|
* swarm neighbourstv14 days1-2/+7
|
* slow link a hop moretv14 days1-1/+3
|
* heal is lasttv2026-11-051-0/+6
|
* offer answerstv2026-11-051-1/+2
|
* round trips in treestv2026-11-051-3/+6
|
* suspect pathstv2026-11-051-1/+4
|
* deflate blob chunkstv2026-11-051-2/+3
|
* deflate blob chunkstv2026-11-051-0/+5
|
* epoch tag in synctv2026-11-051-1/+4
|
* no swarm ID in synctv2026-11-051-1/+2
|
* announce only to those behindtv2026-11-051-1/+2
|
* predecessors go alongtv2026-11-051-1/+5
|
* covered neighbours have ittv2026-11-051-0/+8
|
* delta from held headstv2026-11-051-2/+3
|
* offer again ungeneratedtv2026-11-051-1/+2
|
* stale delta offerstv2026-11-051-0/+3
|
* delta known to sendertv2026-11-051-1/+2
|
* one delta at a timetv2026-11-041-1/+5
|
* document messages that overtaketv2026-11-041-4/+20
|
* any message taken as deliveredtv2026-11-041-7/+9
|
* slow round back offtv2026-11-041-4/+8
|
* tree messages need no answertv2026-11-041-3/+10
|
* hello again when heard fromtv2026-11-041-1/+4
| | | | | | While we wait for a peer's HELLO_REPLY, anything else that comes from it makes our HELLO go again at once, at most once a second, instead of when the backoff says.
* unanswered pull frees the slottv2026-11-041-3/+4
| | | | | | When the answer to a pull is overdue (the sync reply timeout: 1 s in the simtests, 10 s in production), the pull gives up the slot, and the next waiting neighbour is pulled at once.
* no heal while both move ontv2026-11-041-1/+1
| | | | | | | The heal clock now also restarts when our heads and the peer's have both changed since the last STATUS that showed them differing. While writes go on, two neighbours's heads differ because of changes still on their way, and that alone no longer leads to a heal.
* pull after gracetv2026-11-041-1/+3
| | | | | | | | Before pulling, a node waits one measured round trip to the announcer (1 s at most) and pulls only what is still missing. If the round trip isn't known yet, it pulls at once, as now. On the clean benchmark links the round trip is a few milliseconds, so those scenarios shouldn't change.
* announcements answered when lackingtv2026-11-041-2/+3
| | | | | | | A neighbour now answers an announcement only if it lacks some of the announced heads. Heads that merely differ are no longer enough, since they do while writes go on. The check reads the document, so it runs off the message consumer, as the pull check already does.
* unanswered members not reachedtv2026-11-031-5/+10
|
* writers from authorstv2026-11-031-5/+7
|
* reach set for the sessiontv2026-11-031-2/+4
|
* tree changes had alreadytv2026-11-031-1/+3
|
* announce what trees misstv2026-11-031-1/+6
|
* order members by distancetv2026-11-031-4/+6
|
* replace eager links by treestv2026-11-031-9/+2
|
* add writer treestv2026-11-031-0/+16
|
* measure 2-hop neighbourhoodtv2026-11-031-0/+9
|