NightWatch Selenium Socket Hangup in Jenkins / Ubuntu16 - selenium

Issue :
We are running NightWatch tests in Jenkins, and once in a while, we get this error :
[0;31mConnection refused! Is selenium server started?
[0m[0;90m{ Error: socket hang up
at createHangUpError (_http_client.js:253:15)
at Socket.socketCloseListener (_http_client.js:285:23)
at emitOne (events.js:101:20)
at Socket.emit (events.js:188:7)
at TCP._handle.close [as _onclose] (net.js:501:12) code: 'ECONNRESET' }[0m
The fix recommended in other places is already there :
export DBUS_SESSION_BUS_ADDRESS=/dev/null
This is how we run the tests :
sudo npm install selenium-standalone -g
selenium-standalone install --version=3.0.1 --baseURL=https://selenium-release.storage.googleapis.com --drivers.chrome.version=2.28 --drivers.chrome.baseURL=https://chromedriver.storage.googleapis.com --basePath=bin
sudo npm install nightwatch#0.9.13 -g && nightwatch --tag sanity --retries 1 --suiteRetries 1
Rebuilding the job usually run the tests successfully.
Softwares' version:
Chrome : 57
Chrome driver : 2.28
Selenium: 3.0.1
NightWatch : 0.9.13 (bumped down from 0.9.14 in case the latest release had a bug but no change.)
Jenkins : 1.6x
Ubuntu : 16.04
Using xvfb
Error Scenarios
The tests were running fine until about 2 weeks ago, just when we switched to Ubuntu 16, but that day it started happening on Ubuntu 14 too.
It especially happens when we've just spun a new Jenkins slave (on AWS cloud.) Happens more often at the very start of the test run but sometimes it happens in the middle of the run too.
I'll appreciate any help or pointers!

I have the same problem yesterday.
Maybe you can change the version of Chrome.
such as Chrome:55.
when i changed the version of chrome the selenium

We ran into a similar issue, but our error message was more specific:
Selenium is already running on port ####. Or some other service is.
The solution was to use the Jenkins Port Allocator Plugin to assign an available port number to an environment variable.
From there, we were able to set the port number in our globals.js like this (with a default of 4444 for running locally):
portNumber: process.env.PORT_NUMBER || 4444,
In our nightwatch.conf.js file, we used a lodash template to replace values in our nightwatch.json like this:
fs = require('fs'),
_ = require('lodash'),
template = _.template(fs.readFileSync('./nightwatch.json', 'utf8')),
settingsString,
settings;
settingsString = template({
portNumber: globals.portNumber
});
settings = JSON.parse(settingsString);
module.exports = settings;
Finally, in our nightwatch.conf file, we set the values for all port references to:
"<%= portNumber %>"

Related

Jenkins run causes Selenium tests to fail with org.openqa.selenium.WebDriverException: unknown error: Chrome failed to start: exited abnormally?

I'm running Selenium tests for Chrome through Jenkins pipeline on CentOs machine. When ran, it gives following error -
org.openqa.selenium.WebDriverException: unknown error: Chrome failed to start: exited abnormally
[testng] (unknown error: DevToolsActivePort file doesn't exist)
[testng] (The process started from chrome location /usr/bin/google-chrome is no longer running, so ChromeDriver is assuming that Chrome has crashed.)
However if I run same tests with same command(which is used through Jenkins pipeline) from that machine locally, it succeeds.
For the same, I went through link1, link2, link3 and other answers. From those answers, only following helped -
chrome_options.add_argument('--headless')
But this run tests with chrome in headless mode which is not my intention.
In my case, having matching Chrome browser and chromedriver version did not help, also Chrome has executable permission for all.
Note: I've my CentOs machine display enabled
Could you please help to understand what is causing this failure? Also please help with what code and/or machine level changes I've to make for its success.
EDIT:
In case, this helps to get more info.
ls | grep 'chrome' under /usr/bin gave following result -
lrwxrwxrwx 1 root root 31 Jan 20 12:29 google-chrome -> /etc/alternatives/google-chrome
lrwxrwxrwx 1 root root 32 Jan 20 12:29 google-chrome-stable -> /opt/google/chrome/google-chrome
Do you run chromedriver locally on a linux machine? What version of CentOS are you using? If you haven't already you need to install the correct chromedriver for linux. As far as I know if you want to run via Jenkins you have to run in headless mode. But I may be wrong. Below is another post that can help:
Steps to Install and run headless chrome browser on centos 6.5 using chrome driver
Below are my own settings for chrome options:
options.addArguments("--headless");
options.addArguments("--disable-dev-shm-usage");
options.addArguments("--allow-insecure-localhost");
options.addArguments("--window-size=1920,1080");
options.addArguments("start-maximized");
options.addArguments("enable-automation");
options.addArguments("--disable-infobars");
options.addArguments("--disable-browser-side-navigation");
options.addArguments("--disable-gpu");
options.addArguments("--no-sandbox");
options.addArguments("--force-device-scale-factor=1");

session not created: Chrome version must be between 71 and 75 [duplicate]

After update of chromedriver to version 2.46 my tasts fail to initialize.
I got message like this:
Starting ChromeDriver 2.46.628402 (536cd7adbad73a3783fdc2cab92ab2ba7ec361e1) on port 44269
Only local connections are allowed.
Please protect ports used by ChromeDriver and related test frameworks to prevent access by malicious code.
Failed to invoke configuration method com.personal.CustomTest.initTests not created: Chrome version must be between 71 and 75
(Driver info: chromedriver=2.46.628402,platform=Windows NT 10.0.16299 x86_64) (WARNING: The server did not provide any stacktrace information)
Command duration or timeout: 1.58 seconds
Build info: version: '2.53.1'
It is clearly saying that my browser version is not valid. But I am using Chrome 72.0.3626.119 so it is between 71 and 75.
Selenium version is 2.53.1.
And I am running test through console command with the help of testNG.
Any idea? Every ideas that I found was about changing selenium version but I cant do it.
For me to resolve this problem :
On Windows
cd C:\Users\[myname]\AppData\Roaming\npm\node_modules\protractor
npm i webdriver-manager#latest
webdriver-manager update
webdriver-manager start &
On Cent-OS (I used Cent-OS 7.4.* and it worked fine.)
cd /usr/lib/node_modules/protractor/
sudo npm i webdriver-manager#latest
sudo webdriver-manager update
sudo webdriver-manager start &
I hope this helps you in any way.
This error message...
Starting ChromeDriver 2.46.628402 (536cd7adbad73a3783fdc2cab92ab2ba7ec361e1) on port 44269
Only local connections are allowed.
Please protect ports used by ChromeDriver and related test frameworks to prevent access by malicious code.
Failed to invoke configuration method com.personal.CustomTest.initTests not created: Chrome version must be between 71 and 75
...implies that the ChromeDriver v2.46 is not compatible with the Chrome Browser version which is being accessed by your program/webdriver.
Your main issue is the incompatibility between the version of the binaries you are using as follows:
You are using chromedriver=2.46
Release Notes of chromedriver=2.46 clearly mentions the following :
Supports Chrome v71-73
Though you mentioned you are using Chrome 72.0.3626.119 possibly there are multiple instances of Chrome Browser installed within your system and your program by default is accessing the Chrome Browser whose version is not between v71.x and v75.x
You are using chrome=67.0
Release Notes of ChromeDriver v2.38 clearly mentions the following :
Supports Chrome v65-67
Solution
Keep JDK upgraded to recent levels JDK 8u201.
Uninstall all the instances of Chrome Browser (you can opt to use Revo Uninstaller).
Upgrade ChromeDriver to current ChromeDriver v2.46 level.
Keep Chrome version between Chrome v71-73 levels. (as per ChromeDriver v2.46 release notes)
For me, I had to update my chrome driver in my project to match the version of Chrome on my local machine.
yarn add chromedriver#76.0.0 -D
https://www.npmjs.com/package/chromedriver
In my case I was getting the same error after my chrome was updated to version 76. Which was happening when chimp tries to use chromedriver to execute the selenium test.
DevTools listening on ws://127.0.0.1:49220/devtools/browser/e88586cb-ed67-44fc-a742-43b767e2b8f9
[chimp][helper] setupBrowserAndDDP had error
{ Error: session not created: Chrome version must be between 71 and 75
at Object.wait (C:\ACPMS\ELS_AT\node_modules\fibers\future.js:449:15)
at Object.<anonymous> (C:\ACPMS\ELS_AT\node_modules\wdio-sync\build\index.js:344:27)
at Object.<anonymous> (C:\ACPMS\ELS_AT\node_modules\chimp\dist\lib\session-manager.js:145:21)
at initBrowser (C:\ACPMS\ELS_AT\node_modules\chimp\dist\lib\chimp-helper.js:189:43)
at Object.setupBrowserAndDDP (C:\ACPMS\ELS_AT\node_modules\chimp\dist\lib\chimp-helper.js:264:7)
at Context.<anonymous> (C:\ACPMS\ELS_AT\node_modules\chimp\dist\lib\mocha\mocha-helper.js:13:15)
at C:\ACPMS\ELS_AT\node_modules\chimp\dist\lib\utils\fiberize.js:29:22
(Driver info: chromedriver=2.46.628402 (536cd7adbad73a3783fdc2cab92ab2ba7ec361e1),platform=Windows NT 10.0.17134 x86_64)
at new RuntimeError (C:\ACPMS\ELS_AT\node_modules\webdriverio\build\lib\utils\ErrorHandler.js:143:12)
at Request._callback (C:\ACPMS\ELS_AT\node_modules\webdriverio\build\lib\utils\RequestHandler.js:318:39)
at Request.self.callback (C:\ACPMS\ELS_AT\node_modules\request\request.js:185:22)
at emitTwo (events.js:106:13)
at Request.emit (events.js:191:7)
at Request.<anonymous> (C:\ACPMS\ELS_AT\node_modules\request\request.js:1161:10)
at emitOne (events.js:96:13)
at Request.emit (events.js:188:7)
at IncomingMessage.<anonymous>
For me it didn't work when I updated the chrome driver by npm (both globally and locally) and then I followed these steps to resolve the issue, which might give some idea on where to check in this kind a issue:
Because exception was thrown in node_modules\chimp\dist\lib\utils\fiberize.js i navigated to that file and found one level above this file node_modules\chimp\dist\lib\chromedriver.js which has the code to start chromedriver.
I added a console.log to chromedriverPath which can be seen below and re-executed the tests to get the chromedriver path in console logs.
Chromedriver.prototype.start = function(callback) {
var self = this;
var port = self.options.port;
if (this.child) {
callback();
return;
}
var chromedriverPath = chromedriver.path;
//this was added my me the see that chrome drive path
console.log("[chimp] " + chromedriverPath);
if (fs.existsSync(chromedriverPath)) {
this.child = processHelper.start(
{
bin: chromedriverPath,
prefix: "chromedriver",
args: ["--port=" + port, "--url-base=wd/hub"],
waitForMessage: /Starting ChromeDriver/,
errorMessage: /Error/
},
callback
);
} else {
callback("[chimp][chromedriver] Chromedriver executable not found.");
}
};
When i got the path of the chromedriver which is used to execute the tests, i just navigated to that folder and replaced chromedriver executable with the most recent version.
I hope this helps.
Try this: Download the compatible ChromeDriver.exe file for the updated version of Chrome browser and replace it. It works for me.

Selenium with Firefox-ESR in Alpine Docker image

I'm using Selenium for testing from Java code, it's started automatically with maven, when I run mvn verify it uses gecko-driver to start new firefox instance for selenium:
final FirefoxOptions opts = new FirefoxOptions();
if (HEADLESS_MODE) {
opts.addArguments(ARG_HEADLESS);
}
return new FirefoxDriver(opts);
It's working fine on my machine (and on machines of other developers) in both modes "headless" and normal. Now I'm trying to run Selenium tests with CI builds in Alpine Linux Docker container. When I tried to run it as-is, I've got an error:
mozrunner::runner INFO Running command: "/usr/bin/firefox"
"-marionette" "--headless" "-foreground" "-no-remote"
"-profile" "/tmp/rust_mozprofile.pW1sbkcv98QN"
Error: GDK_BACKEND does not match available displays.
Then I installed all suggested tools (dbus and xvfb) from this post and started it before running tests:
apk add xvfb dbus
dbus-uuidgen > /etc/machine-id
export DISPLAY=:99
Xvfb $DISPLAY -ac &
mvn verify
but now Selenium tests are hadning for about 10 seconds and failing with error
1565433736375 mozrunner::runner INFO Running command: "/usr/bin/firefox" "-marionette" "--headless" "-foreground" "-no-remote" "-profile" "/tmp/rust_mozprofile.1ZMsJ0Hy95V0"
1565433738424 Marionette INFO Listening on port 2828
!!! [Child][MessageChannel] Error: (msgtype=0x3E0003,name=PCompositable::Msg_Destroy) Channel error: cannot send/recv
[Child 2191] ###!!! ABORT: Aborting on channel error.: file /home/buildozer/aports/community/firefox-esr/src/firefox-52.8.1esr/ipc/glue/MessageChannel.cpp, line 2152
[Child 2191] ###!!! ABORT: Aborting on channel error.: file /home/buildozer/aports/community/firefox-esr/src/firefox-52.8.1esr/ipc/glue/MessageChannel.cpp, line 2152
!!! [Child][MessageChannel] Error: (msgtype=0x3E0003,name=PCompositable::Msg_Destroy) Channel error: cannot send/recv
It seems I'm doing something wrong here, what is the correct way to run headless Selenium tests in Alpine Linux Docker container? I'd prefer to configure existing Docker image instead of using Selenium images, because the migration will be expensive.
Update:
Selenium version is 3.141.59
Gecko-driver version is v0.24.0-linux64
Firefox-ESR version is: 52.8.1-r0
Update2:
xvfb, dbus and exported DISPLAY are not actually needed to run Firefox in headless mode - just run firefox --headless and it's all.
As per Mozilla --headless flag is not supported in Firefox 52-esr.Headless flag was introduced in version 55 (for Linux) and 56 (Mac/Windows) . Please try on latest Firefox ESR release.

e2e tests fail on server with selenium

I'm about to create an application full js based on vuejs. Following https://github.com/vuejs-templates/webpack, I have started a project:
$ vue init webpack my-project && cd my-project && yarn install
When I run unittest yarn run unit everything goes well, but when I run yarn run e2e which runs nightwatch.js, I get this error:
> Starting dev server...
Starting to optimize CSS...
> Listening at http://localhost:8889
Starting selenium server... started - PID: 23451
started - PID: 23451
[Test] Test Suite
=====================
Running: default e2e tests
[client :: capabilities] Test Suite
=======================================
Error retrieving a new session from the selenium server
Connection refused! Is selenium server started?
{ Error: socket hang up
at createHangUpError (_http_client.js:344:15)
at Socket.socketOnEnd (_http_client.js:436:23)
at emitNone (events.js:110:20)
at Socket.emit (events.js:207:7)
at endReadableNT (_stream_readable.js:1047:12)
at _combinedTickCallback (internal/process/next_tick.js:102:11)
at process._tickCallback (internal/process/next_tick.js:161:9) code: 'ECONNRESET' }
error Command failed with exit code 1.
info Visit https://yarnpkg.com/en/docs/cli/run for documentation about this command.
The selenium server is started on the fly when the e2e tests start and is shutdown when they finish. I'm sure because when I stop nightwatch.js before it fails I see :
$ ps aux | grep -i selenium
java -Dwebdriver.chrome.driver=node_modules/chromedriver/lib/chromedriver/chromedriver -jar node_modules/selenium-server/lib/runner/selenium-server-standalone-2.53.1.jar -port 4444
So the real error based on the output I get is mostlikely Error retrieving a new session from the selenium server. How do I fix this?
package.json
"chromedriver": "^2.27.2",
"cross-spawn": "^5.0.1",
"nightwatch": "^0.9.12",
"selenium-server": "^2.53.1",
test/e2e/nightwatch.conf.js
require('babel-register')
var config = require('../../config')
// http://nightwatchjs.org/gettingstarted#settings-file
module.exports = {
src_folders: ['test/e2e/specs'],
output_folder: 'test/e2e/reports',
custom_assertions_path: ['test/e2e/custom-assertions'],
selenium: {
start_process: true,
server_path: require('selenium-server').path,
host: '127.0.0.1',
port: 4444,
cli_args: {
'webdriver.chrome.driver': require('chromedriver').path
}
},
I recently started a Vue.js project with Webpack template. Your problem sounds familiar, but surprisingly, I forgot how I solved it...
The only difference I can see between your package.json and mine is the selenium-server version (^3.0.1). You should also check your Node version because Nightwatch is based on it. I remember that I had to update Node and npm for some reason (I do not use Yarn).
This is my environment (and it is working fine):
Node.js (v8.1.3)
npm (5.0.3)
Did you run node nightwatch.config.js, or node {whatever your nightwatch config file is called}. This will install Selenium and ChromeDriver correctly.

seleniumAddress not taking effect in Protractor config

I am having a issue where seleniumAddress is not taking effect in protractor config. I use seleniumAddress: 'http://127.0.0.1:4444/wd/hub' but I see:
[18:52:35] E/launcher - Timed out waiting for the WebDriver server at http://127.0.0.1:47780/hub
The port should be 4444 instead of 47780. Any ideas?
one#development ~/github/drop $ protractor --version
Version 4.0.7