Node doesn't register with hub when autoscaling is enabled.

[gitissuepost]

Hi,

I am unsing below values.yaml to install helm chart (0.39.1) on minikube for learning. However, the node pod (Chrome in this case) stucks at "Sending Registration Event". But when the autoscaling is disabled, I can see the node available to use. What am I missing to establing the connectivity between node and hub with autoscaling enabled.

autoscaling:
  enabled: true
  scaledOptions:
    minReplicaCount: 1
    maxReplicaCount: 4
    pollingInterval: 20

firefoxNode:
  enabled: false

edgeNode:
  enabled: false

crossBrowsers:
  chromeNode:
    - nameOverride: '{{ $.Release.Name }}-node-chrome-120'
      imageTag: '120.0'
      hpa:
        platformName: 'Linux'
        browserVersion: '120.0'

Upon checking the logs in Hub, I see below exception which I feel is not related to node register

Thank you

[VietND96]

Looks like image tag 120.0 was using an old selenium-server.jar to start Node, where a config is not compatible with latest Hub 4.28.0
In this case, via env var SE_OPTS, can you try to add config flag --session-timeout 300 to see any help?

[gitissuepost]

@VietND96 - I am new to helm. Would you mind help me on how to set env vars through values.yaml. Not sure of it is already documented and I overlooked that.

Thank you

[VietND96]

It follows the structure of chromeNode

crossBrowsers:
  chromeNode:
    - nameOverride: '{{ $.Release.Name }}-node-chrome-120'
      imageTag: '120.0'
      extraEnvironmentVariables:
        - name: SE_OPTS
          value: "--session-timeout 300"
      hpa:
        platformName: 'Linux'
        browserVersion: '120.0'

[gitissuepost]

@VietND96 : With this configuration, the node pod gets terminated saying "Can only specify option --session-timeout once." as session-timeout is getting added twice.

I checked removing the explicit env var and the session timeout is being used with default value of 300 for node.

2025-01-29 09:04:57,244 INFO Included extra file "/etc/supervisor/conf.d/selenium.conf" during parsing
2025-01-29 09:04:57,251 INFO RPC interface 'supervisor' initialized
2025-01-29 09:04:57,251 CRIT Server 'unix_http_server' running without any HTTP authentication checking
2025-01-29 09:04:57,253 INFO supervisord started with pid 20
2025-01-29 09:04:58,256 INFO spawned: 'xvfb' with pid 21
2025-01-29 09:04:58,259 INFO spawned: 'vnc' with pid 22
2025-01-29 09:04:58,261 INFO spawned: 'novnc' with pid 23
2025-01-29 09:04:58,263 INFO spawned: 'selenium-node' with pid 25
2025-01-29 09:04:58,266 INFO success: selenium-node entered RUNNING state, process has stayed up for > than 0 seconds (startsecs)
Appending Selenium options: --session-timeout 300
Appending Selenium options: --session-timeout 300
Appending Selenium options: --log-level INFO
Generating Selenium Config
Tracing is enabled
Classpath will be enriched with these external jars :
Selenium Grid Node configuration:
[events]
publish = "tcp://selenium-grid-selenium-hub:4442"
subscribe = "tcp://selenium-grid-selenium-hub:4443"

[server]
host = "10.244.0.61"
port = "5555"
[node]
grid-url = "http://selenium-grid-selenium-hub.default"
session-timeout = "300"
override-max-sessions = false
detect-drivers = false
drain-after-session-count = 1
max-sessions = 1

[[node.driver-configuration]]
display-name = "chrome"
stereotype = '{"browserName": "chrome", "browserVersion": "120.0", "platformName": "Linux", "goog:chromeOptions": {"binary": "/usr/bin/google-chrome"}}'
max-sessions = 1

Starting Selenium Grid Node...
Can only specify option --session-timeout once.
Usage: node [options]

2025-01-29 09:04:59,253 INFO exited: selenium-node (exit status 0; expected)
2025-01-29 09:04:59,254 WARN received SIGINT indicating exit request
2025-01-29 09:04:59,254 INFO waiting for xvfb, vnc, novnc to die
2025-01-29 09:04:59,271 INFO success: xvfb entered RUNNING state, process has stayed up for > than 1 seconds (startsecs)
2025-01-29 09:04:59,271 INFO success: vnc entered RUNNING state, process has stayed up for > than 1 seconds (startsecs)
2025-01-29 09:04:59,271 INFO success: novnc entered RUNNING state, process has stayed up for > than 1 seconds (startsecs)
2025-01-29 09:04:59,280 INFO stopped: novnc (terminated by SIGTERM)
2025-01-29 09:04:59,782 INFO stopped: vnc (terminated by SIGTERM)
2025-01-29 09:05:00,784 INFO stopped: xvfb (terminated by SIGTERM)

[gitissuepost]

We have requirements to setup a cross browser lab and few browsers are in list starting 110

[VietND96]

I can understand the requirement now. Currently, you have the range from 121.0 to latest that are working?

[gitissuepost]

I was checking randomly. Let me check for the list of versions to see which versions are failing

[VietND96]

In the meantime, I am also able to create a workflow to build node Chrome image with different browser versions on top of base 4.28.1-20250123. It will be launched if range versions from v110 to (the latest - 1) are ready.

[gitissuepost]

@VietND96 - Thank you for the update. However, can we not have a generic solution that will work for any prior version of browser (Chrome/Edge/any supported browser) to work seemless with Selenium hub as we progress? May be a flag like legacy.browser=true? I may be thinking out loud without considering the impacts, but it's just a thought as similar chalanges may be availabel with Edge, which we haven't tried yet