Custom Lua scripts are not loaded in cluster mode

[stavalfi]

I created a local cluster on my macbook pro with 3-8 nodes using redis-cli and redis-server (From homebrew).

When I shutdown all nodes and re-create them (manually) - 9/10 runs this bug occurs:

const client = new IORedis.Cluster([
  {
    host: 'localhost',
    port: 7000,
  },
])

const scriptName = 'script1'
const script1 = 'return {KEYS[1]}'
client.defineCommand(scriptName, { numberOfKeys: 1, lua: script1 })
console.log(await client[scriptName]('aaaa', 'hi', console.log))

Error

(node:54658) UnhandledPromiseRejectionWarning: ReplyError: NOSCRIPT No matching script. Please use EVAL.
    at parseError (/Users/stavalfi/projects/singer/.yarn/cache/redis-parser-npm-3.0.0-7ebe40abcb-45dbcb05be.zip/node_modules/redis-parser/lib/parser.js:179:12)
    at parseType (/Users/stavalfi/projects/singer/.yarn/cache/redis-parser-npm-3.0.0-7ebe40abcb-45dbcb05be.zip/node_modules/redis-parser/lib/parser.js:302:14)
(Use `node --trace-warnings ...` to show where the warning was created)
(node:54658) UnhandledPromiseRejectionWarning: Unhandled promise rejection. This error originated either by throwing inside of an async function without a catch block, or by rejecting a promise which was not handled with .catch(). To terminate the node process on unhandled promise rejection, use the CLI flag `--unhandled-rejections=strict` (see https://nodejs.org/api/cli.html#cli_unhandled_rejections_mode). (rejection id: 2)
(node:54658) [DEP0018] DeprecationWarning: Unhandled promise rejections are deprecated. In the future, promise rejections that are not handled will terminate the Node.js process with a non-zero exit code.

(Very?) Bad Workaround

It took me a while to figure out how to fix it: I need to manually update all nodes about the script

const masters = client.nodes('master')
await Promise.all(
  masters.map(node => Promise.all([node.script('load', script1), node.script('load', persistMollyLiveBmss)])),
)

I'm not sure I can use this fix in production.

---

More info

Note: In non-cluster runs, all works great: using local,remote,docker redis-servers.

System:
    OS: macOS 10.15.1
    CPU: (12) x64 Intel(R) Core(TM) i7-9750H CPU @ 2.60GHz
    Memory: 675.51 MB / 16.00 GB
    Shell: 3.1.2 - /usr/local/bin/fish
  Binaries:
    Node: 14.15.4 - /var/folders/fh/lk3kyv9n29v3w36bqyv_y7d00000gn/T/fnm-shell-5135596/bin/node
    Yarn: 2.4.2 - /var/folders/fh/lk3kyv9n29v3w36bqyv_y7d00000gn/T/fnm-shell-5135596/bin/yarn
    npm: 6.14.10 - /var/folders/fh/lk3kyv9n29v3w36bqyv_y7d00000gn/T/fnm-shell-5135596/bin/npm

ioredis - 4.27.6 and 4.27.7

redis-cli 6.2.4
Redis server v=6.2.4 sha=00000000:0 malloc=libc bits=64 build=f6e7aafbcc3dc4bf

redis.conf for all nodes:

port <some port>
cluster-enabled yes
cluster-config-file nodes.conf
cluster-node-timeout 5000
appendonly no
save ""

[marcbachmann]

This could be caused by what I've described in the comment here: #1438 (comment)

Once a server restarts, it doesn't have the script anymore and ioredis only tries to reload them after 60s.

[ioredis-robot]

🎉 This issue has been resolved in version 4.28.5 🎉

The release is available on:

• npm package (@latest dist-tag)
• GitHub release

Your semantic-release bot 📦🚀

[marcbachmann]

This accidentally got closed. #1499 would fix the behavior