ECONNRESET on long idle time

[jumpingaccess]

HI,everyone,
I got this on nodecg when no use of the dashboard is done for long time ( example between evening and next morning).

UNCAUGHT EXCEPTION! NodeCG will now exit.
Error: read ECONNRESET
at __node_internal_captureLargerStackTrace (node:internal/errors:496:5)
at __node_internal_errnoException (node:internal/errors:623:12)
at TCP.onStreamRead (node:internal/stream_base_commons:217:20) {
errno: -104,
code: 'ECONNRESET',
syscall: 'read',
fatal: true
}
UNCAUGHT EXCEPTION! NodeCG will now exit.
Error: Connection lost: The server closed the connection.
at Socket. (/home/laurent/nodecg/node_modules/mysql2/lib/connection.js:117:31)
at Socket.emit (node:events:514:28)
at Socket.emit (node:domain:489:12)
at TCP. (node:net:323:12) {
fatal: true,
code: 'PROTOCOL_CONNECTION_LOST'
}

Is it a way to have a keepalive or something like this ? the bundle I've made use a websocket connection and i have to re setup everything when I've to restart it

thanks in advance

[from-the-river-to-the-sea]

My guess is that somewhere in your bundle's extension is a long-living object (such as a MySQL connection using that mysql2 library I see in the stack trace) and that this object needs to have an error event handler added to it so that any errors it throws don't cause an uncaught exception. This is a somewhat common pattern in Node.js -- if there's an error handler, it gets the errors. Else, errors are uncaught and crash the program.

[from-the-river-to-the-sea]

To clarify, I mean you might need only do this:

yourSocketOrWhatever.on('error', (error) => {
    nodecg.log.error(error)
})

[jumpingaccess]

this is not the websocket client that close , but nodecg itself , websocket server is created on another software

[jumpingaccess]

Websocket stream get information every 100 ms with a timetick

[jumpingaccess]

mysql2 is an error I need to clean , not used anymore

[jumpingaccess]

More on this , even my bundle is empty:

UNCAUGHT EXCEPTION! NodeCG will now exit.
Error: read ECONNRESET
at __node_internal_captureLargerStackTrace (node:internal/errors:496:5)
at __node_internal_errnoException (node:internal/errors:623:12)
at TCP.onStreamRead (node:internal/stream_base_commons:217:20) {
errno: -104,
code: 'ECONNRESET',
syscall: 'read',
fatal: true
}

[jumpingaccess]

HI ,
Solution found :
var KeepAliveAgent = require('keep-alive-agent'),
agent = new KeepAliveAgent();
in extension/index.js
and no more crashes after longperiod of inactivities

:)