Keryx-Labs/keryx-node v1.6.1-PoM

Keryx-node-v1.6.1-PoM

on GitHub

Mandatory before H13.

H13 activates at DAA 93 550 000 (≈ Sunday 2026-09-06 19:30-20:00 UTC). Upgrade AND resync from the new datadir before Sunday 18:00 UTC.

Fixes

• A restart could lose up to 12 hours of service-bond events (rewards, strikes, burns, suspensions) on a pruned node. The cold-start refold restored the newest sample snapshot and only re-derived what lay above it, while events still waiting for finality between the persisted frontier and that sample were kept nowhere. The refold now restores the newest sample at or below the persisted frontier and re-derives everything above it. Validated on testnet: a miss folded before a sample, node restarted after it, burn finalized at the expected daa, identical on a fresh sync.
• A node stuck below a branch the network keeps building on no longer spins at 100 % CPU. Once that branch is 1 000 blue scores past the sink, the node logs chain divergence with the offending block and exits (code 3). The remedy is a resync from a trusted datadir. KERYX_REVALIDATE_DISQUALIFIED disables it for recovery runs.

H13

• Before the gate, every node trusts the chain for the coinbase and service-state checks, whatever its own trust window.
• At the gate, verification is enforced by every node, and the service-state commitment moves to v4: the production-index part is hashed in a chain-numbering-independent form, so nodes holding the same production state agree whatever their reset history.
• P2P handshake floor 1.6.1 at the gate (24 h ban below).

Why the datadir is mandatory

Nodes that restarted on 1.6.0 lost service-bond rows that cannot be re-derived once the blocks are pruned. Such a node disagrees with the network the moment one of those events matters, and at H13 it stops with the message above. The new datadir carries the complete state (archival lineage, verified against a fresh sync):

• https://huggingface.co/datasets/Keryx-Labs/datadir/resolve/main/datadir.zip
• https://drive.google.com/file/d/1KAmehQ2rK3fkISc19lqOnrbGZiOUw8pu/view?usp=sharing
• https://keryx-labs.com/datadir.zip

Upgrade

1. Install 1.6.1 (keryxd + keryx-cli).
2. Stop the node, replace datadir with the new one, start.
3. Do it before Sunday 18:00 UTC. Pools first.

A node that upgrades without the datadir follows the chain until the gate, then stops with chain divergence.