continuum - P2P replicated document storage service

	Commit message (Collapse)	Author	Age	Files	Lines
*	allow superseding expired remakes	tv	16 hours	4	-26/+279
\|
*	remake coordinators resume after restart	tv	16 hours	3	-5/+49
\|
*	don't commit expired remakes	tv	16 hours	1	-58/+93
\|
*	add votes register	tv	16 hours	8	-124/+507
\|
*	bound stable-hub degree	tv	17 hours	1	-4/+46
\| \| \| \| \| \|	Each stable member can be preferred at most by 4*k members. The budget is spent in key order, which is the same on every node, so the graph stays deterministic, symmetric and connected.
*	coalesce announcements	tv	17 hours	3	-2/+60
\|
*	cache heads	tv	17 hours	7	-24/+61
\| \| \| \|	To reduce the number of write locks, heads are now cached.
*	corroberate newer epochs	tv	18 hours	1	-11/+53
\|
*	follow and download from ready members only	tv	18 hours	3	-3/+29
\|
*	being caught up needs some heads	tv	18 hours	1	-5/+9
\| \| \| \| \| \| \| \| \| \| \|	A sync message from a ready peer only completes the initial sync if that peer has at least one head. A peer with an empty document can no longer make a joiner ready on an empty store: a misconfigured second swarm, or a genesis node in its first seconds. Effect on a brand-new swarm: a joiner now waits until the genesis node has registered itself in the member list, which happens within about 5 s of it becoming ready.
*	only protocol-writers may modify _continuum	tv	18 hours	6	-14/+63
\|
*	put remake content into blobs	tv	20 hours	15	-79/+319
\|
*	add chunked download and use it during rejoin	tv	22 hours	15	-3/+844
\|
*	allow unsetting config options	tv	22 hours	2	-46/+264
\|
*	add API and CLI to print effective config	tv	22 hours	4	-2/+106
\|
*	add configuration files	tv	23 hours	4	-107/+1261
\|
*	wire up --mycelium-tcp-listen-port	tv	23 hours	1	-2/+5
\| \| \| \| \|	It wasn't used before, causing the supervised daemon to always pick a random one.
*	introduce majority voting on remakes	tv	25 hours	7	-95/+406
\|
*	add electorate	tv	25 hours	3	-8/+109
\| \| \| \| \|	For an upcoming change in remake coordination, nodes now maintain their self-assessed liveness in the document.
*	aliases are not in logs (fixup)	tv	25 hours	2	-2/+2
\|
*	measure document saved-size only after changes	tv	25 hours	2	-18/+43
\|
*	allow import remake using the raw document	tv	31 hours	1	-6/+85
\|
*	add the import kind of remake	tv	31 hours	3	-49/+200
\|
*	add endpoint to export the document	tv	31 hours	3	-0/+88
\| \| \| \|	GET /admin/document to retrieve the automerge document.
*	add node aliases	tv	31 hours	6	-11/+147
\|
*	expose stats for prometheus to scrape	tv	31 hours	2	-7/+514
\|
*	add throughput stats	tv	31 hours	8	-2/+231
\|
*	add propagation stats	tv	31 hours	4	-1/+126
\|
*	add per peer incident stats	tv	31 hours	4	-4/+52
\|
*	expose stats via the API	tv	32 hours	2	-0/+136
\| \| \| \|	GET /admin/stats to retrieve all stats.
*	add swarm stats	tv	32 hours	9	-1/+238
\|
*	add traffic counter	tv	32 hours	9	-12/+270
\| \| \| \| \| \| \| \| \| \|	This introduces a new wrapper that counts messages and payload bytes per peer, split by kind (hello, helloReply, sync, ping, pong), each sent, received and refused. The traffic counter goes around the firewall, which means messages dropped outbound are still counted as sent, and messages dropped inbound don't count as received.
*	nodes prefer stable neighbours	tv	32 hours	1	-10/+89
\|
*	assess swarm stability	tv	32 hours	8	-11/+330
\|
*	record RTT sampling instant	tv	33 hours	4	-30/+104
\|
*	prefer closer sources to catch up	tv	33 hours	3	-3/+49
\|
*	record RTT from keepalive messages	tv	33 hours	7	-20/+148
\|
*	pull missing changes	tv	33 hours	7	-5/+162
\| \| \| \| \| \|	Nodes push changes to their neighbours but don't relay them directly. In a partial mesh, nodes can learn about missing changes via keepalive messages and can pull them from their neighbours.
*	find reachable neighbours	tv	34 hours	3	-55/+162
\|
*	sync with neighbours	tv	34 hours	6	-4/+171
\|
*	add neighbour graph	tv	34 hours	2	-0/+180
\|
*	add member registry	tv	34 hours	4	-1/+73
\|
*	use nonce in handshake	tv	35 hours	5	-2/+39
\| \| \| \| \|	To prevent stale hello replies to cause joins and thus stale sessions, each run gets a random ID, which used as nonce in the handshake.
*	drain queued messages on startup	tv	35 hours	1	-2/+48
\|
*	consolidate backoff code	tv	35 hours	4	-49/+99
\|
*	retry using fresh messages	tv	35 hours	3	-170/+180
\| \| \| \| \|	Don't store encoded messages but retry using a newly generated one. That way any accumulated changes can be delivered at once.
*	heal from noticed divergence	tv	35 hours	1	-6/+27
\| \| \| \| \|	Wait 2*keepalive but at least 20 s to trigger resync so that lost messages can be resent first.
*	heal after reply timeout	tv	37 hours	1	-0/+3
\|
*	drop stall breaker; resend needed changes	tv	38 hours	2	-82/+21
\| \| \| \| \| \|	The stall breaker was superseded by regular re-handshaking Re-handshaking also gets triggered, when a node needs some of the heads its peer believes it has already sent. In that case the peer now resends what's needed.
*	no heal during catch-up	tv	39 hours	1	-1/+6
\|