How to Recover When the Registry Cell Moves

The registry cell outpoint changes every time a governance update is confirmed. If your application caches the outpoint, it will eventually become stale. This page shows how to detect and recover from that situation.

Detection

fetchRegistryPayload throws when the cell is no longer live:

try {
  const registry = await fetchRegistryPayload(rpcUrl, cachedTxHash, cachedIndex);
  // use registry
} catch (err) {
  if (err instanceof Error && err.message.includes("is not live")) {
    // Registry was updated — recover
  }
}

Recovery

Use findRegistryCell to discover the current outpoint, then re-fetch:

import { findRegistryCell, fetchRegistryPayload } from "@ckb-firewall/sdk";

const REGISTRY_SPEC = {
  codeHash:    "0x493f1700508125b0e281b8fb1d168b03bd5ef71480399dd59221224901a9cd09",
  hashType:    "type" as const,
  typeIdValue: "0x9be0ad6e4e5039a64d9725ff037057c16ef59f126e3bdd9841b802f0e0a112fe",
  required:    true,
};

let cachedTxHash = "0x...";
let cachedIndex  = 0;

async function getRegistry() {
  try {
    return await fetchRegistryPayload(rpcUrl, cachedTxHash, cachedIndex);
  } catch (err) {
    if (err instanceof Error && err.message.includes("is not live")) {
      const { txHash, index } = await findRegistryCell(rpcUrl, REGISTRY_SPEC);
      cachedTxHash = txHash;
      cachedIndex  = index;
      await persistOutpoint(txHash, index); // update your cache
      return await fetchRegistryPayload(rpcUrl, txHash, index);
    }
    throw err;
  }
}

findRegistryCell queries get_cells on the CKB indexer and filters by typeIdValue — it finds the correct cell regardless of how many governance updates have occurred since your last fetch.

Prefer findRegistryCell always

For applications that need robustness, call findRegistryCell before every transaction build rather than caching:

const { txHash, index } = await findRegistryCell(rpcUrl, registrySpec);
const registry = await fetchRegistryPayload(rpcUrl, txHash, index);
// build and check transaction

The extra indexer call adds a small latency (one round-trip) but eliminates the stale-outpoint failure mode entirely.

For in-flight transaction failures caused by a governance update, see Fix: in-flight transaction fails after a governance update.