Selenium Grid Timeout - ruby

Question

The following selenium-client (1.2.18) script times out in our test environment. The script works locally, but not in our test environment. I've only included the bare minimum code necessary to reproduce the timeout (which is why I'm doing a http.post directly).
On the test environment, I've looked at the firefox sessionstore.js and seen the browser is getting to the site, but it seems like it never communicates back to the node.
EDIT: curl 'www.google.com' works from the test environment.
require 'selenium/client'
require 'net/http'
#browser = Selenium::Client::Driver.new(:host => 'localhost',
:port => 4445,
:browser => '*firefox',
:url => 'http://www.google.com/',
:timeout_in_second => 60)
#browser.start_new_browser_session
session_id = #browser.session_id
p "SESSION ID: #{session_id}"
http = Net::HTTP.new('localhost', 4445)
begin
response = http.post('/selenium-server/driver/',"cmd=open&1=landing/summer/index.html&2=true&sessionId=#{session_id}")
p "RESPONSE: #{response}"
ensure
p "CLOSING SESSION..."
#browser.close_current_browser_session
end
Grid hub was started with the following:
selenium-server-standalone-2.25.0.jar -port 4445 -timeout 180 -singleWindow -role hub -browserTimeout 300 -newSessionMaxWaitTimeInSeconds 300 -newSessionWaitTimeout 300 -firefoxProfileTemplate /var/tmp/selenium-grid-start-stop/releases/20130802031034/firefox_profile -trustAllSSLCertificates
Node started with the following:
selenium-server-standalone-2.25.0.jar -role node -port 6666 -hub http://localhost:4445/grid/register -maxSession 1 -firefoxProfileTemplate /var/tmp/selenium-grid-start-stop/releases/20130802031034/firefox_profile -browser browserName=*firefox,seleniumProtocol=Selenium
And running the test script...
$ruby test2.rb
"SESSION ID: 18609e43667a4aa094fe81ef30551b84"
"CLOSING SESSION..."
/usr/local/rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/net/protocol.rb:140:in `rescue in rbuf_fill': Timeout::Error (Timeout::Error)
from /usr/local/rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/net/protocol.rb:134:in `rbuf_fill'
from /usr/local/rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/net/protocol.rb:116:in `readuntil'
from /usr/local/rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/net/protocol.rb:126:in `readline'
from /usr/local/rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/net/http.rb:2219:in `read_status_line'
from /usr/local/rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/net/http.rb:2208:in `read_new'
from /usr/local/rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/net/http.rb:1191:in `transport_request'
from /usr/local/rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/net/http.rb:1177:in `request'
from /usr/local/rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/net/http.rb:1170:in `block in request'
from /usr/local/rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/net/http.rb:627:in `start'
from /usr/local/rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/net/http.rb:1168:in `request'
from /usr/local/rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/net/http.rb:978:in `post'
Anyone seen this kind of timeout before? The site (google) is obviously up. Perhaps something is misconfigured in the test environment?

Answer 1

I've found the solution. We accidently upgraded the Firefox in our test environment from 10 -> 17.
Our Grid nodes were providing a Firefox prefs.js that did not work with Firefox 17. This allowed the browser to start up and connect to the site successfully, but unable (for whatever reason) to respond back to the node.
The solution was to provide an updated prefs.js that worked for FF 17. We obtained this by running a test without providing the preferences ourselves. This forced Firefox to create a default preferences file we copied and edited.