Worker does not exit on EHOSTUNREACH error?
#62
Labels
Comments
|
After restarting the backend, tasks down to 34 and the journal is flooded with disconnects: |
|
A quick fix would be a const net = require("net");
module.exports = ({ host, port }) => {
const logger = require("../system/logger.js");
let socket = new net.Socket();
// keep socket alive
socket.setKeepAlive(true, 5000);
//socket.unref(); // test for #62, did nothing!
socket.on("error", (err) => {
logger.error(`[error] tcp://${host}:${port}`, err);
// why not cal process.exit() here?
// this file/function should only be called in a worker thread
// so it dosnt matter if the process exits or not...
// if the socket.unref() approach fails, try the exit method
process.exit(1);
});
socket.on("close", () => {
logger.debug(`[closed] tcp://${host}:${port}`);
});
socket.on("connect", () => {
logger.info(`[connected] tcp://${host}:${port}`);
});
socket.connect(port, host);
return socket;
}; |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Trying to connect to a invalid/non exisiting/offline tcp socket/server, causes to spwan worker, but do not exit.
This creates a memory leak.
Related/debugging information: #61
A watch on
watch -n 10 "systemctl status connector"and trying to connect to a offline raspberry increases the Tasks number continusly.start to watch:
16:23 = 49
16:25 = 51
16:28 = 54
16:33 = 63
16:37 = 69
The text was updated successfully, but these errors were encountered: