The Zcash Foundation opened Zebra 7.2.0-beta.1 for testing on November 19, 2026, combining mining and synchronization work with optional pruning of historical block bodies. Its recommendation is specific: keep production nodes on Zebra 7.1.0 and run the beta alongside them to gather results.
The headline storage example is promising but limited. The Foundation reports that pruning reduced one mainnet database from roughly 282 GB to 15 GB. That is a result from its test, not a storage guarantee for every deployment or an independently reproduced benchmark.
Less stored history, with a deliberate recovery boundary
Pruning removes older block bodies from a node's local storage. It does not mean deleting funds or rewriting the network's transaction history. Zebra's release notes retain archive storage as the default and make pruning an explicit choice.
For an existing archive, the notes require a stopped-node move into a separate pruning directory and a separate archive backup first. Once historical bodies have been deleted, turning the option off cannot recreate them. Restoring an archive snapshot or synchronizing again into a separate cache is the documented recovery path.
The retention setting has another practical limit: increasing it changes future deletion behavior, rather than bringing back data already removed. A node operator should decide what history its dependent services require before treating a smaller database as a straightforward saving.
Mining improvements need workload evidence
The Foundation describes faster preparation of mining work, faster block propagation and changes to verification and initial synchronization. It asks miners and pools to compare when new work arrives and whether solved blocks are accepted and relayed. Those observations are more useful than assuming that a software announcement establishes a fixed increase in mining income.
The testing instruction also keeps the scope clear. A beta can be exercised with a limited portion of a workload while the existing production setup remains available. Results from one pool, machine or network connection do not automatically describe another. The Zcash profile explains the network context, while the consensus guide separates mining from staking.
A smaller node serves a different set of requests
The detailed release notes explain that pruned nodes advertise limited-history capabilities and cannot supply block bodies they no longer hold. This matters to services and other peers that expect an archive. Storage requirements and the ability to serve historical data are linked decisions, not independent checkboxes.
The same beta offers optional fast tree synchronization, but that is a separate feature with its own compatibility limits. The notes warn that a database synchronized that way cannot simply be opened by the older version, and that some wallet-facing tree-state requests are unavailable at affected heights. Enabling every new option at once would make it harder to isolate which change caused a service problem.
Release availability is only the first checkpoint
For a controlled evaluation, operators can preserve the original archive, document the enabled options, compare the same workload and review the release's compatibility notes before connecting dependent services. Wallet users should follow their wallet provider's supported software path rather than infer that they need to run a node beta.
The constructive opportunity is lower resource use without obscuring what a node can still provide. The remaining work is to establish that trade-off across real deployments. Until the Foundation changes its production recommendation, the beta designation and the instruction to keep production on 7.1.0 belong beside every performance claim.