Cluster, Quorum & Fabric Network Commands
Authoritative command reference for am cluster, am mon (Raft Quorum), am network (Latency Mesh), and am upgrade.
This reference provides production syntax, options, and operational snippets for managing holistic cluster topology, Raft consensus monitors, network transport verification, and software upgrades using am.
1. am cluster — Topology, Capacity & Doctor
The am cluster suite manages global topology epochs, membership registration, quorum health, and automated diagnostic scorecards.
1.1 am cluster status
Interrogates the coordinator and Raft consensus leader to output active epoch, monitor leadership, online storage nodes, and raw capacity:
# Summary status
am cluster status
# Detailed per-node breakdown with Kdouja sync state
am cluster status --detailed
# Live refresh mode during maintenance
am cluster status --watch 3s
1.2 am cluster health
Machine-verifiable health check with strict exit code guarantees (0 for Healthy, 8 for Degraded/Critical):
am cluster health --strict --min-nodes 4
1.3 am cluster doctor
Automated 5-point fabric diagnostic scanner evaluating Raft consensus, node capacity headroom (>15% free), rack failure domain diversity, and CPU SIMD acceleration (AVX2 / AVX-512):
am cluster doctor
1.4 Node Registration & Removal
# Register newly provisioned node
am cluster register-node storage-05 https://10.0.1.15:57775 --int-id 5 --type NVMe
# Evict decommissioned node from topology
am cluster remove-node storage-old-01 --yes
2. am mon — Raft Consensus & Election Control
The am mon suite inspects and administers the Raft consensus group operating on Port 57777 / 57774:
2.1 Inspect Quorum Status & Peers
# Display active term, leader node, and consensus state
am mon status
# Probe all peer coordinators and measure round-trip keepalives
am mon peers --ping
2.2 Leadership Election & Planned Failover
# Transfer leadership to nominated peer prior to node reboot
am mon elect --target coord-02 --force
3. am network — Fabric Transport, MTU & Latency Mesh
Verifies inter-node high-bandwidth storage fabrics, Jumbo Frame support, and fabric-wide latency matrices:
# Ping target storage node over gRPC
am network ping storage-02
# Generate cluster-wide latency mesh table
am network matrix
# Test MTU 9000 (Jumbo Frames) payload transmission
am network mtu-test --size 9000
4. am upgrade — Zero-Downtime Rolling Upgrades
Safely plans, validates, and coordinates rolling binary upgrades across the storage substrate without downtime:
# 1. Generate safe sequential upgrade plan
am upgrade plan --package /opt/releases/aarkam-v1.0.0.tar.gz
# 2. Apply rolling upgrade node by node with health gating
am upgrade apply --yes
# 3. Verify cluster protocol version
am upgrade verify