Skip to content

[question] why ECONRESET / ECONREFUSED - how to consume response body with fetch #4063

@blephy

Description

@blephy

Hi,

I have a nodejs server running with a DB. They are communicating throught HTTP with undici client

Here is my undici global agent :

const dnsCache = new CacheableLookup();
const agent = new Agent({
      // @ts-expect-error -- this is safe https://github.com/nodejs/undici/tree/main/docs/examples#using-cacheablelookup-to-cache-dns-lookups-in-undici
      connect: {
        lookup: dnsCache.lookup.bind(dnsCache),
        keepAlive: true,
        timeout: 10_000,
        allowH2: true,
      },
      bodyTimeout: 60_000,
      headersTimeout: 60_000,
      allowH2: true,
      keepAliveTimeout: 10_000,
    });

Sometime, i run into this error :

"err":{"type":"SocketError","message":"other side closed","stack":"SocketError: other side closed\n    at Socket.onHttpSocketEnd (/home/node/app/node_modules/.pnpm/undici@7.3.0/node_modules/undici/lib/dispatcher/client-h1.js:902:22)\n    at Socket.emit (node:events:530:35)\n    at endReadableNT (node:internal/streams/readable:1698:12)\n    at process.processTicksAndRejections (node:internal/process/task_queues:90:21)","name":"SocketError","code":"UND_ERR_SOCKET","socket":{"localAddress":"127.0.0.1","localPort":37560,"remoteAddress":"127.0.0.1","remotePort":8080,"remoteFamily":"IPv4","bytesWritten":62192889,"bytesRead":3705674}

and subsequence call produce then this error 3 or 4 times :

"err":{"type":"Error","message":"read ECONNRESET","stack":"Error: read ECONNRESET\n    at TCP.onStreamRead (node:internal/stream_base_commons:216:20)\n    at TCP.callbackTrampoline (node:internal/async_hooks:130:17)","errno":-104,"code":"ECONNRESET","syscall":"read"}

and after again, always this error in loop :

"err":{"type":"Error","message":"connect ECONNREFUSED 127.0.0.1:8080","stack":"Error: connect ECONNREFUSED 127.0.0.1:8080\n    at TCPConnectWrap.afterConnect [as oncomplete] (node:net:1634:16)\n    at TCPConnectWrap.callbackTrampoline (node:internal/async_hooks:130:17)","errno":-111,"code":"ECONNREFUSED","syscall":"connect","address":"127.0.0.1","port":8080}

To fix it, i need to resart the pod containing the app and the DB.

Can somebody explain to me if my configuration is not OP ?

Is this an undici bug which not re-open a socket after the previous one was closed by the DB ?

This can be a DB error too, but it would be great if someone here can explain to me what happens, i'm not even sure what to think about those errors.

Not related question : is CacheableLookup does the same as the dns interceptor ? Should i use dns interceptor then ? If so, we probably update your documentation

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions