LatticeLattice

Not generally available. Product and docs are public. Sales are closed.

Docs / Operations

Runbooks

Add/drain dhcpd, Postgres failover, clock, pools, outbox. Shared DB, no pairwise failover.

There is no ISC failover protocol. Shared Postgres is HA for dhcpd.

Add a dhcpd node

  1. Host with chrony. Offset < 500ms vs the writer.
  2. /etc/lattice/lattice.env with a new LATTICE_NODE_NAME.
  3. Enable lattice-dhcpd. curl -sf http://<node>:8080/readyz.
  4. Add to the UDP VIP backend. Relays stay on the VIP.
  5. Confirm a node row with recent last_seen_at.

Drain

  1. Remove from VIP / withdraw anycast.
  2. Wait until heartbeat is stale and offer/ack logs stop.
  3. Stop the unit. Do not delete leases. Do not DELETE FROM node unless retiring the name.

Postgres failover

  1. Fence the old writer (Patroni) or wait for the managed endpoint to flip.
  2. Processes reconnect. /readyz is 503 until ping works.
  3. Check lattice_outbox_lag. Pending rows retry.

More: Clock · Pool exhaustion · Outbox · Backup.