| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
| |
The current version may name a checkpoint, whose heads are behind: a version naming them
again is never written. Writing a version leaves it out unless the version moves the
store on.
|
| |
|
|
|
|
| |
The next version begun from the current one, and recipes as jq filters for letting a
party in, taking one out, giving and taking back a place, zones and blocks, and keys
named anew; what a party changes by itself; and a version that only moves the store on.
|
| |
|
|
|
| |
Who a party and the adversary are, the three promises in plain words, what isn't
promised, and what that asks of a party, an operator and an application.
|
| |
|
|
|
|
|
| |
When a group forks, choosing the heads, the fork's version, its root as a node builds
it, starting the fork's nodes and beginning a record there; what a node holding the
history finds of a fork's root; and following a fork, coming back, and groups that
parted coming together.
|
| |
|
|
|
|
|
| |
A version a node refused by what it took before, finding from the parties' records who
voted twice, having a node drop one side's version and take the other's, answering by a
fork, and what keeps a node from holding its store, syncing, taking an event, or
starting.
|
| |
|
|
|
|
|
| |
What a move to a checkpoint keeps and loses, a version only moving the store on and the
heads to name, a node proposing checkpoints by its operator's rules, backing only heads
that leave out nothing held for a while (which a node voting by itself checks and a
party's client doesn't), signing again after a move, and nodes that weren't there.
|
| |
|
|
|
|
| |
The keys a version names, kept apart with the party's record, and those of its key set,
signing its writes; changing a set and signing again what should stay; changing the keys
a version names; a stolen key of either kind; and the node key, which is no party's.
|
| |
|
|
|
|
|
| |
Writing a version from the current one, its `since`s and what makes one none, proposing
it, voting through the party's client by its record and what the client refuses to
cast, voting through a node that votes for its party, a lost record, and when a proposal
passes or dies, with the counts for some numbers of parties.
|
| |
|
|
|
|
| |
Places and the rules naming their writers, a write through the party's client and what
it may hold, what counts and what stops counting, the refusals a node names, signing
again what a move refused, and what `GET /events/counts` shows.
|
| |
|
|
|
|
|
| |
What a node of a federated swarm does otherwise than any: its store options and file,
members syncing with members only, its API letting in whoever its operator lets in to do
everything, the store API refusing writes, joining by ID and from a pinned checkpoint,
and what it writes by itself.
|
| |
|
|
|
|
|
| |
`docs/guides/federation/` holds the guides to a federated swarm, the design saying what
the rules are and they how to live by them. The first: two parties start a federation, a
node each, write, read, and let a third party in by their votes. The README points to
them.
|
| |
|
|
|
|
| |
The federation spells a key without a comment, so a key set holding a line's comment is
none. `continuum-party keys` takes each key as an `authorized_keys` line, one argument
each, and writes it as the federation spells it.
|
| |
|
|
|
|
|
|
|
|
| |
A record casts nothing on the version current when it began, as a `true` sent before
may be held back somewhere. A federation's first version is current from its start, so
a record begun at a node casts nothing on it, and with every party's begun so, no
proposal on it can pass. `continuum-party init PARTY --root FILE` begins a record from
the file the store's swarm starts from, for a party that never voted in it
(`Record::before`): nothing was sent before it, and it casts on every version, the
first too. A node voting by itself takes such a record as its `--store-vote-record`.
|
| |
|
|
|
|
| |
What a federation is used for is read and written through a node's API: kartei's
entries are signed in by their parties and read from what counts there, with no stage
of their own.
|
| |
|
|
|
|
|
|
|
| |
The registry holds each kartei namespace in its party's, a place per file. What kartei's
consumers read as one for all (host and user names, the names hosts go by, addresses and
blocks in every net but `internet`) is shared, at places of its own, given by krebs's
rules: a party's by its name, and what kartei gives it beyond them by a rule each in the
first version. What kartei's consumers read is written out from what counts, but for
what names an identifier the current version doesn't give its party.
|
| |
|
|
|
|
|
|
|
|
|
| |
A rejoin makes the node's document writable again only as it ends, in the
lock of the swap, once the node checked the epoch (`Holding::follows`) and
follows it or gives it up. The store's worker reads nothing, and writes
nothing, while the document isn't writable, checked in the lock it reads
under. Whether a rejoin runs isn't enough: its state is taken up for its
last checks, while the epoch's document is still the node's, and a store
started from that document would keep the checkpoint it read there, and put
again what it took into the document the node goes back to.
|
| |
|
|
|
|
|
|
|
|
| |
A node with an empty document can't tell an epoch made up from another it
can't check: either way it takes the snapshot's checkpoint only where its
operator named or pinned it. The others move to a checkpoint and compact;
f, joining and pinning nothing, doesn't take their epoch, and g, pinning
the checkpoint, takes it and reads what they read. A made-up epoch served
by d, an honest node besides, is row 31's alone: d follows the others'
epoch wherever it reaches them.
|
| |
|
|
|
|
| |
Reaching the others, d rejoins their epoch, the lower, as any node holding
a store does, and e, reaching only d, finds their epoch there, not d's. Cut
off from them, d stays on the one it made up.
|
| |
|
|
|
|
|
| |
Of a lower one, the others try to rejoin it every so often, as in row 31,
and f, rejoining their epoch, finds them away mid-sync and gives up for
that, not for its checkpoint. Of a higher ID, they leave it alone; e,
reaching only d, still takes it, as one joining takes the newer generation.
|
| |
|
|
|
| |
Its target, attempts, whether the node rejoins now, and why the last try
failed, before the lines of the remake records.
|
| |
|
|
|
|
| |
d, serving the epoch it made up, puts again there what its store holds, as
every node holding a store does: e, judging every event, reads what d reads,
the checkpoint the others moved to among it, not the root alone.
|
| |
|
|
|
| |
It builds on tv's write, the checkpoint's head: one built on the root isn't
above the heads of the checkpoint the nodes moved to, and they refuse it.
|
| |
|
|
|
|
|
|
|
| |
While a node rejoins, its document is that of the epoch it tries, which it
follows only once it checks it (`Holding::follows`). The store's worker
reads nothing of it, and writes nothing, while the node rejoins: else the
store, starting there from the snapshot's checkpoint, would take one the
node neither builds nor has pinned, and keep it should the node give up
the rejoin.
|
| | |
|
| |
|
|
|
|
|
|
|
|
| |
d serves an epoch of the next generation it made up, of a lower ID than the
others', its snapshot holding the root as the document stood before the
checkpoint they moved to. A node joining with nothing that reaches only d
takes it, the root being what its operator named, and reads only what it
holds (15); one reaching only the others takes not their epoch, whose
checkpoint it neither builds nor has pinned (15); the others try d's, find
its root before their document's checkpoint, and stay on theirs (31).
|
| |
|
|
|
| |
`testing::serve` switches a node to an epoch whose document is a snapshot
it built itself, as a node serving a made-up epoch would.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
Simulations of the store's attack rows on remakes (`docs/store.md`,
Attacks): a snapshot other than its base's isn't followed (11); one
coordinator's two decisions for an epoch, shown to different nodes, end on
the lower epoch (13); a node away while the others switch rejoins and puts
again what it wrote meanwhile (16); after a compaction whose snapshot holds
a checkpoint, a node away all along follows only once its operator pins it
(24); a compaction whose base comes before a checkpoint, switched after
nodes moved to it, leaves them there, and a node starting with nothing
reads the new document to it (28); a member deciding many compactions at
once moves every node to the lowest epoch, losing no event (30).
|
| |
|
|
|
|
|
|
| |
For simulations of a coordinator that doesn't keep to the protocol:
`testing::propose_again` proposes a remake's proposal again under a new ID,
whatever is open, and `testing::commit_as` commits a remake at the node's
heads now, whatever the votes and another decision, to a given epoch or to
the epoch of the snapshot the node builds there.
|
| |
|
|
|
|
|
|
| |
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`).
|
| |
|
|
|
|
| |
A swarm holding a store compacts: events written beyond the base, on nodes
that switch later, are put again in the new epoch, and every node holds
every event, reads the same, and refused no change.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
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.
|
| |
|
|
|
|
|
|
|
|
| |
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.
|
| |
|
|
|
|
|
|
|
|
| |
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.
|
| | |
|
| |
|
|
|
|
|
|
| |
A writer's changes are its first change's author's, so a node writing in a
test after another under the same writer wrote as that one: a decision by
another than the coordinator was taken as the coordinator's. Another node
now writes under a fresh writer. The proposal written again differs, as one
written the same makes no change.
|
| |
|
|
|
|
|
| |
The snapshot of a compaction after the store moved to a checkpoint, built
from a document every run builds alike: its change's hash and bytes, which a
run of this code prints, as no script apart encodes Automerge. Left empty
until a run prints them.
|
| |
|
|
|
| |
The same events put by the same writer make the same change, so that a
document built of them in a test is the same bytes on every run.
|
| |
|
|
|
|
|
| |
The entry of a node that isn't a member counts for nothing where continuum
reads entries (`docs/store.md`, Members): a remake proposed in a swarm
holding a store counts, of the members alive recently enough, those the
fold names with the party their entries name.
|
| |
|
|
|
|
|
| |
Every node builds a compaction's snapshot itself from the base; another
kind's is what someone gives (`docs/store.md`, Roads not taken). A node of
a swarm holding a store refuses to propose one at its API, as every node
refuses its proposal.
|
| |
|
|
|
|
|
|
|
|
|
| |
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`).
|
| |
|
|
|
|
|
|
| |
A remake's snapshot holds the store's checkpoint at `_continuum/checkpoint`
where it isn't the root (`docs/store.md`, What the document holds), and
none of what came before it. A node reading its store anew from such a
document starts there, with what the checkpoint keeps; one pinning a
checkpoint finds it there too, where it can't build it from the events.
|
| |
|
|
|
|
|
|
|
|
|
| |
A change of a swarm holding a store may write its node's own protocol data
a remake needs (`docs/store.md`, What a change writes): its acks, a
remake's record and proposal, where there is none, of a compaction it
coordinates, and, as the coordinator, its decision, or the record of the
remake that made the epoch, carried into it whole; and the maps holding
them, where there are none. An abort asked for by another than the
coordinator stays closed, as does any other kind of remake. Rows 14 and 21
are judged here.
|
| |
|
|
|
|
|
|
|
| |
while
With no time to hold an event for a while, lass's write, taken just before
the proposal, counted as held for a while before it was proposed, and the
node backed no heads leaving it out. The test's node now holds an event for
an hour before it holds it for a while.
|
| |
|
|
|
|
|
|
|
|
| |
A node voting by itself, and a party's client, signed a vote on the heads
the node held: where those held what is beside a checkpoint's heads (a
party's writes made while the swarm was parted, say), the vote wasn't above
them, and a version event written on it was refused, so the checkpoint
never moved, though it passed. Both now vote on the events proposing what
they vote on, and a node proposes a checkpoint on the heads it names: so
their votes are above those heads wherever the proposal is.
|
| |
|
|
|
|
| |
For a node holding a store, the dump shows each event it refused and why,
the proposals on the current version with the parties proposing them and
their votes, whether each passes there, and the version events it took.
|
| |
|
|
|
|
|
| |
Automerge encodes the hashes of a sync message sorted, and asserts so: the
hashes asked for that changes waiting in screening lack were added after
those Automerge asks for, unsorted, and the heads a peer last said that we
hold were taken as the peer's `have` as they came. Both are sorted now.
|
| |
|
|
|
|
|
|
|
|
|
| |
Messages taken as delivered and lost, more of them than still count as
sent, leave a peer lacking changes we take it to have. Asking for the
first of them, it got that one alone, and asked for the next once that
arrived, a change per round trip; or, asking for our heads alone, the last
of them again and again. When a peer needs a change in the history of the
heads we share that no longer counts as sent, the session answers as if
it had said it has nothing since the latest heads it said that we hold
(`lost_beyond_sent`): everything since goes in one message.
|
| |
|
|
|
|
|
|
|
|
| |
Screening keeps a change whose dependencies haven't come from Automerge,
waiting, so Automerge never learns what it lacks: our next message asks for
the peer's heads alone, and the peer, forgetting having sent those, sends
them again, and what they build on only as far as it still counts it as
sent. So the session asks for what changes waiting build on and the
document lacks too (`Waiting::lacking`, `SyncSession::lacks`), as Automerge
asks for what the changes it holds back lack.
|
| |
|
|
|
|
|
|
| |
A failed simulation prints, for each node, the writer it writes as, the
changes its screening keeps waiting (the latest few: whose writer and seq,
how many of their dependencies it holds, whether a signature vouches for
them) and the latest changes it refused, and, holding a store, the store's
checkpoint, current version and what it took and refused.
|