mirror of
https://gitlab.com/wgp/dougal/software.git
synced 2025-12-06 11:17:08 +00:00
Refactor Selenium to make it more robust.
It should stop runaway Firefox processes.
This commit is contained in:
@@ -1,52 +1,110 @@
|
||||
// TODO Append location to PATH
|
||||
|
||||
const path = require('path');
|
||||
const fs = require('fs');
|
||||
const {Builder, By, Key, until} = require('selenium-webdriver');
|
||||
const firefox = require('selenium-webdriver/firefox');
|
||||
const { Builder, By, Key, until } = require('selenium-webdriver');
|
||||
const firefox = require('selenium-webdriver/firefox');
|
||||
const { execSync } = require('child_process');
|
||||
|
||||
const geckodriverPath = path.resolve(__dirname, "geckodriver");
|
||||
|
||||
// We launch a browser instance and then start an activity timer.
|
||||
// We shut down the browser after a period of inactivity, to
|
||||
// save memory.
|
||||
// State to prevent race conditions
|
||||
let driver = null;
|
||||
let timer = null;
|
||||
let isShuttingDown = false;
|
||||
|
||||
function resetTimer () {
|
||||
clearTimeout(timer);
|
||||
timer = setTimeout(shutdown, 120000); // Yup, hardcoded to two minutes. For now anyway
|
||||
// Verify GeckoDriver exists
|
||||
if (!fs.existsSync(geckodriverPath)) {
|
||||
throw new Error(`GeckoDriver not found at ${geckodriverPath}`);
|
||||
}
|
||||
|
||||
async function launch () {
|
||||
function resetTimer() {
|
||||
clearTimeout(timer);
|
||||
timer = setTimeout(shutdown, 120000); // 2 minutes inactivity timeout
|
||||
}
|
||||
|
||||
async function launch() {
|
||||
if (isShuttingDown) {
|
||||
console.log("Shutdown in progress, waiting...");
|
||||
await new Promise(resolve => setTimeout(resolve, 1000));
|
||||
return launch(); // Retry after delay
|
||||
}
|
||||
resetTimer();
|
||||
if (!driver) {
|
||||
console.log("Launching Firefox");
|
||||
const options = new firefox.Options();
|
||||
// Explicitly set headless mode and optimize for server
|
||||
options.addArguments('--headless', '--no-sandbox', '--disable-gpu');
|
||||
// Limit content processes to reduce resource usage
|
||||
options.setPreference('dom.ipc.processCount', 1);
|
||||
|
||||
const service = new firefox.ServiceBuilder(geckodriverPath);
|
||||
driver = await new Builder()
|
||||
.forBrowser('firefox')
|
||||
.setFirefoxService(new firefox.ServiceBuilder(geckodriverPath))
|
||||
.setFirefoxOptions(options.headless())
|
||||
.setFirefoxService(service)
|
||||
.setFirefoxOptions(options)
|
||||
.build();
|
||||
}
|
||||
}
|
||||
|
||||
async function shutdown () {
|
||||
if (driver) {
|
||||
async function shutdown() {
|
||||
if (driver && !isShuttingDown) {
|
||||
isShuttingDown = true;
|
||||
console.log("Shutting down Firefox");
|
||||
// This is an attempt at avoiding a race condition if someone
|
||||
// makes a call and resets the timer while the shutdown is in
|
||||
// progress.
|
||||
const d = driver;
|
||||
driver = null;
|
||||
await d.quit();
|
||||
try {
|
||||
const d = driver;
|
||||
driver = null;
|
||||
await d.quit();
|
||||
// Explicitly stop the service
|
||||
const service = d.service;
|
||||
if (service) {
|
||||
service.stop();
|
||||
}
|
||||
console.log("Firefox shutdown complete");
|
||||
} catch (error) {
|
||||
console.error("Error during shutdown:", error);
|
||||
// Forcefully kill lingering processes (Linux/Unix)
|
||||
try {
|
||||
execSync('pkill -u $USER firefox || true');
|
||||
execSync('pkill -u $USER geckodriver || true');
|
||||
console.log("Terminated lingering Firefox/GeckoDriver processes");
|
||||
} catch (killError) {
|
||||
console.error("Error killing processes:", killError);
|
||||
}
|
||||
} finally {
|
||||
isShuttingDown = false;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
async function url2pdf (url) {
|
||||
async function url2pdf(url) {
|
||||
await launch();
|
||||
await driver.get(url);
|
||||
return await driver.printPage({width: 21.0, height: 29.7});
|
||||
try {
|
||||
console.log(`Navigating to ${url}`);
|
||||
await driver.get(url);
|
||||
// Add delay to stabilize Marionette communication
|
||||
await driver.sleep(3000);
|
||||
const pdf = await driver.printPage({ width: 21.0, height: 29.7 });
|
||||
resetTimer(); // Reset timer after successful operation
|
||||
return pdf;
|
||||
} catch (error) {
|
||||
console.error("Error in url2pdf:", error);
|
||||
await shutdown(); // Force shutdown on error
|
||||
throw error;
|
||||
}
|
||||
}
|
||||
|
||||
// Periodically clean up orphaned processes (every 5 minutes)
|
||||
setInterval(() => {
|
||||
try {
|
||||
const firefoxCount = execSync('pgrep -c firefox || echo 0').toString().trim();
|
||||
if (parseInt(firefoxCount) > 0 && !driver) {
|
||||
console.log(`Found ${firefoxCount} orphaned Firefox processes, cleaning up...`);
|
||||
execSync('pkill -u $USER firefox || true');
|
||||
execSync('pkill -u $USER geckodriver || true');
|
||||
console.log("Cleaned up orphaned processes");
|
||||
}
|
||||
} catch (error) {
|
||||
console.error("Error checking orphaned processes:", error);
|
||||
}
|
||||
}, 300000);
|
||||
|
||||
module.exports = { url2pdf };
|
||||
|
||||
Reference in New Issue
Block a user