Atualmente em testes: o código do GitHub será aberto quando concluído.

Configuração (Variáveis de Ambiente)

A configuração vem apenas de variáveis de ambiente. Nomes ignoram maiúsculas, vazio é não definido e só 1/true/yes/on é true. Intervalo ou combinação inválida aborta o início. A imagem scratch não usa disco local nem PV.

Configurações comuns e limites de recursos

VariableDefaultAllowedPurpose
SERVER_PORT5225valid u16 portsingle client port; cluster advertises CLUSTER_SELF
SOCKET_BIND0.0.0.0bindable addressclient and Raft bind host
DEBUG_MODEbuild settingflagdebug logging
DOCKER_MODEfalseflagofficial image binds internal 5225/6225
MAX_CONNECTIONS10241..8192concurrent client connections
MAX_RAFT_CONNECTIONS641..256separate Raft/control reserve
MAX_WAITERS20481..16384waiters per key; excess acquire gets B
MAX_TOTAL_WAITERS16384MAX_WAITERS..65536total waiters; excess acquire gets B
Hard capValueSaturation behavior
in-flight acquire per connection64B for that acquire
global in-flight acquire4096B for that acquire
global in-flight release512 separate lanebounded wait until connection close
queued replies per connection256close slow client
coordinator channel4096backpressure
cluster active keys65536B for new-key acquire
Raft frame64MiBreject RPC
encoded snapshot64MiB - 64KiBreject build/install
concurrent Raft decode budget72MiBbounded permit wait/timeout

O perfil padrão visa limite de memória de 512MiB. Hard ceilings não podem ser excedidos e nada é unlimited.

Autenticação do cliente e TLS

VariableDefaultPurpose
CLIENT_TOKENSone empty tokencomma-separated client credentials; production should use nonblank secrets
TLS_CERT / TLS_KEYunsetserver certificate chain/private key; set together
TLS_CAsystem rootsCA for peer certificate verification
TLS_SKIP_VERIFYfalsetest only
CLUSTER_ALLOW_PLAINTEXT_PRIVATEfalseexplicit opt-out for isolated private networks

TLS é recomendado. Cluster plaintext exige opt-out explícito de rede privada e não oferece confidencialidade nem identidade de node por certificado.

Identidade estável de voters

VariableConstraintPurpose
CLUSTER_ID1–128 ASCII alnum/./_/-fencing domain
CLUSTER_NODE_IDu64Raft ID; never reuse after retirement
CLUSTER_INCARNATIONnon-zero u128freshly issued per process start
CLUSTER_SELFhost:client-portstable client slot and redirect address
CLUSTER_RAFT_SELFhost:raft-portunique Raft/control address for this process
CLUSTER_PEERSexactly 3 or 5NodeId@incarnation@client-address@raft-address entries

Todas as variáveis de identidade cluster são exigidas juntas; omitir todas escolhe single mode. Cada início recebe incarnation nova emitida externamente. NodeId retirado nunca é reutilizado: substituto entra learner com NodeId novo. Portas client e Raft são explícitas, sem regra +1000.

Startup admission falha fechado. Cluster inicial abre só quando todo peer prova estado empty/unconsumed e passa barreira Prepared. Substituto abre só após Join one-shot. Pending/Prepared rejeita Vote, Append, Snapshot e client mutation. Peer inalcançável nunca é presumido vazio.

Raft e clock autenticado

VariableDefaultConstraint
CLUSTER_HEARTBEAT_MS500>=10
CLUSTER_ELECTION_TIMEOUT_MS2000>=600, >=4× heartbeat
CLUSTER_SNAPSHOT_TIMEOUT_MS120000>=1000
RAFT_PROGRESS_TIMEOUT_MS10000>=1000
CLUSTER_CLOCK_AGENT_ENDPOINTunsetrequired numeric IP:port in cluster
CLUSTER_CLOCK_AGENT_SECRETunsetrequired, at least 32 bytes, separate domain
CLUSTER_CLOCK_PAIRWISE_DELTA_MS202..=60000, identical on all nodes
CLUSTER_CLOCK_MAX_STALENESS_MS5000greater than sample interval + timeout
CLUSTER_CLOCK_SAMPLE_INTERVAL_MS1000positive and below staleness
CLUSTER_CLOCK_TIMEOUT_MS250positive and below staleness
CLUSTER_CLOCK_MAX_RTT_MS50positive and at most timeout
CLUSTER_CLOCK_MAX_UNCERTAINTY_MS2at most pairwise delta / 2
CLUSTER_CLOCK_MAX_SOURCE_AGE_MS2000positive

Endpoint clock retorna reference time autenticado, uncertainty, sync state e source age. Falha de auth/freshness/RTT/pairwise delta/wall-step para mutations, drena clients e encerra Raft.

Autenticação cluster e join de substituição

VariableRequirement
CLUSTER_TOKENSnonblank list required in cluster
CLUSTER_CONTROL_TOKENSnonblank list required in cluster
CLUSTER_BOOTSTRAP_CREDENTIALinitial cluster process only; recorded once in replicated state

Replacement process values generated by prepare-replacement:

  • CLUSTER_JOIN_LEADER
  • CLUSTER_JOIN_LEADER_NODE_ID
  • CLUSTER_JOIN_LEADER_INCARNATION
  • CLUSTER_JOIN_CREDENTIAL
  • CLUSTER_JOIN_SEED_PEERS
  • fresh CLUSTER_NODE_ID, CLUSTER_INCARNATION, CLUSTER_SELF, CLUSTER_RAFT_SELF, CLUSTER_PEERS

Credentials cluster/control/clock/bootstrap/join são domínios separados sem reutilização. Valores Join apenas de prepare-replacement; credential one-shot, padrão 60s, permitido 5–300s. Peer wire exige exatamente CLUSTER_WIRE_VERSION=1; package version não conta.

See zero-downtime fresh-NodeId learner replacement.