Add documentation about slave failure due to slow performance

Signed-off-by: Josh Barker <[email protected]>

kitchen.yml

@@ -22,8 +22,9 @@ provisioner:
         mirror: https://updates.jenkins.io
         runit:
           sv_timeout: 30
-        # Set httpKeepAliveTimeout to 5 minutes so that connections are not closed too early.
-        # This can happen when there is slow performance
+        # This is to help slaves to connect to the master
+        # Sometimes with slow performance the connection is dropped/cut too early
+        # 30 seconds was too low, so I bumped it to 5 minutes
         # https://issues.jenkins-ci.org/browse/JENKINS-38623
         jenkins_args: '--httpKeepAliveTimeout=300000'
 

test/fixtures/cookbooks/jenkins_slave/attributes/default.rb

@@ -4,3 +4,8 @@
 when 'rhel', 'amazon'
   default['sshd_service'] = 'sshd'
 end
+
+# launch timeout is 2 minutes so that there's plenty of time for
+# the slave to connect to the master
+
+default['jenkins_slave']['launch_timeout'] = 120

test/fixtures/cookbooks/jenkins_slave/recipes/connect.rb

@@ -6,7 +6,7 @@
   # SSH specific attributes
   host        'localhost'
   credentials 'jenkins-ssh-password'
-  launch_timeout   120
+  launch_timeout   node['jenkins_slave']['launch_timeout']
   ssh_retries      5
   ssh_wait_retries 60
 end

test/fixtures/cookbooks/jenkins_slave/recipes/create_ssh.rb

@@ -76,7 +76,7 @@
   # SSH specific attributes
   host        'localhost'
   credentials credentials
-  launch_timeout   120
+  launch_timeout   node['jenkins_slave']['launch_timeout']
   ssh_retries      5
   ssh_wait_retries 60
 end
@@ -90,7 +90,7 @@
   # SSH specific attributes
   host        'localhost'
   credentials '38537014-ec66-49b5-aff2-aed1c19e2989'
-  launch_timeout   120
+  launch_timeout   node['jenkins_slave']['launch_timeout']
   ssh_retries      5
   ssh_wait_retries 60
 end
@@ -104,7 +104,7 @@
   # SSH specific attributes
   host        'localhost'
   credentials 'jenkins-ssh-password'
-  launch_timeout   120
+  launch_timeout   node['jenkins_slave']['launch_timeout']
   ssh_retries      5
   ssh_wait_retries 60
 end

test/fixtures/cookbooks/jenkins_slave/recipes/disconnect.rb

@@ -1,10 +1,11 @@
+
 jenkins_ssh_slave 'ssh-to-disconnect' do
   remote_fs   '/tmp/ssh-to-disconnect'
   user        'jenkins-ssh-password'
   # SSH specific attributes
   host        'localhost'
   credentials 'jenkins-ssh-password'
-  launch_timeout   120
+  launch_timeout   node['jenkins_slave']['launch_timeout']
   ssh_retries      5
   ssh_wait_retries 60
 end

test/fixtures/cookbooks/jenkins_slave/recipes/offline.rb

@@ -1,10 +1,11 @@
+
 jenkins_ssh_slave 'ssh-to-offline' do
   remote_fs   '/tmp/ssh-to-offline'
   user        'jenkins-ssh-password'
   # SSH specific attributes
   host        'localhost'
   credentials 'jenkins-ssh-password'
-  launch_timeout   120
+  launch_timeout   node['jenkins_slave']['launch_timeout']
   ssh_retries      5
   ssh_wait_retries 60
 end

test/fixtures/cookbooks/jenkins_slave/recipes/online.rb

@@ -1,10 +1,11 @@
+
 jenkins_ssh_slave 'ssh-to-online' do
   remote_fs   '/tmp/ssh-to-online'
   user        'jenkins-ssh-password'
   # SSH specific attributes
   host        'localhost'
   credentials 'jenkins-ssh-password'
-  launch_timeout   120
+  launch_timeout   node['jenkins_slave']['launch_timeout']
   ssh_retries      5
   ssh_wait_retries 60
 end

test/fixtures/cookbooks/jenkins_smoke/recipes/default.rb

@@ -82,8 +82,9 @@
 
       raise 'Jenkins is starting up' if response.include?('Starting Jenkins')
 
-      # after Jenkins has started the slaves will connect, but this can take some time
-      # So we check all the slaves are online before running inspec
+      # After Jenkins has started the slaves will connect, but this can take some time
+      # So we check that the expected slaves are online before running Inspec
+
       [
         'jnlp-builder',
         'jnlp-executor',
@@ -99,6 +100,10 @@
         raise "cannot find slave: #{slave}" unless json_body
         raise "slave: #{slave} isn't online" if json_body[:offline]
       end
+
+      # Note:
+      # Testing that slaves are connected and/or online is prone to failure
+      # due to slow performance, different virtualization, etc
     rescue StandardError
       # re-raise exceptions so that the ruby_block triggers a retry
       raise

test/integration/jenkins_smoke/controls/jenkins_slave.rb

@@ -2,6 +2,11 @@
 # copyright: 2018, The Authors
 
 title 'Jenkins Slaves'
+
+# Note:
+# Testing that slaves are connected and/or online is prone to failure
+# due to slow performance, different virtualization, etc
+
 control 'jenkins_slave-1.0' do
   impact 0.7
   title 'Jenkins JNLP Slaves are created and configured'