failed to install on GCE w Centos7 - RuntimeError: Supervisor service service-analytics_poller was unable to be started

[Bob Beliveau]

Using a 2 VCPU/7,5GB machine on GCE, with Centos 7, I fail to install scalr-server.

I used instructions at: https://scalr-wiki.atlassian.net/wiki/display/docs/Installation+and+Upgrade+Instructions

I installed the repo, and then installed scalr, and generated the default config file.

When I try to "Reload/restart all scalr components with the new configuration" using 

sudo /opt/scalr-server/bin/scalr-server-ctl reconfigure

After many items seem to work, I get the error:

------------ START OF ERROR OUTPUT ------------

* supervisor_service[service-analytics_poller] action start
    
    ================================================================================
    Error executing action `start` on resource 'supervisor_service[service-analytics_poller]'
    ================================================================================
    
    RuntimeError
    ------------
    Supervisor service service-analytics_poller was unable to be started
    
    Cookbook Trace:
    ---------------
    /opt/scalr-server/embedded/cookbooks/supervisor/providers/service.rb:49:in `block (2 levels) in class_from_file'
    /opt/scalr-server/embedded/cookbooks/supervisor/providers/service.rb:47:in `block in class_from_file'
    
    Resource Declaration:
    ---------------------
    # In /opt/scalr-server/embedded/cookbooks/scalr-server/recipes/group_service_always_post.rb
    
      5:   supervisor_service name do
      6:     command         "#{bin_dir_for node, 'service'}/scalrpy_proxy" \
      7:                     " #{run_dir_for node, 'service'}/#{svc[:name]}.pid" \
      8:                     " #{node[:scalr_server][:install_root]}/embedded/bin/python" \
      9:                     " #{scalr_bundle_path node}/app/python/scalrpy/#{svc[:service_module]}.py" \
     10:                     " --pid-file=#{run_dir_for node, 'service'}/#{svc[:name]}.pid" \
     11:                     " --log-file=#{log_dir_for node, 'service'}/python-#{svc[:name]}.log" \
     12:                     " --user=#{node[:scalr_server][:app][:user]}" \
     13:                     " --group=#{node[:scalr_server][:app][:user]}" \
     14:                     " --config=#{scalr_bundle_path node}/app/etc/config.yml" \
     15:                     ' --verbosity=INFO' \
     16:                     " #{svc[:service_extra_args]}" \
     17:                     # Note: 'start' is added by the proxy.
     18:     stdout_logfile  "#{log_dir_for node, 'supervisor'}/#{name}.log"
     19:     stderr_logfile  "#{log_dir_for node, 'supervisor'}/#{name}.err"
     20:     autostart       true
     21:     action          [:enable, :start]
     22:     subscribes      :restart, 'file[scalr_config]' if should_restart
     23:     subscribes      :restart, 'file[scalr_code]' if should_restart
     24:     subscribes      :restart, 'file[scalr_cryptokey]' if should_restart
     25:     subscribes      :restart, 'file[scalr_id]' if should_restart
     26:     subscribes      :restart, 'user[scalr_user]' if should_restart
     27:   end
     28: end
    
    Compiled Resource:
    ------------------
    # Declared in /opt/scalr-server/embedded/cookbooks/scalr-server/recipes/group_service_always_post.rb:5:in `block in from_file'
    
    supervisor_service("service-analytics_poller") do
      action [:enable, :start]
      updated true
      retries 0
      retry_delay 2
      default_guard_interpreter :default
      declared_type :supervisor_service
      cookbook_name :"scalr-server"
      recipe_name "group_service_always_post"
      command "/opt/scalr-server/libexec/service/scalrpy_proxy /opt/scalr-server/var/run/service/analytics_poller.pid /opt/scalr-server/embedded/bin/python /opt/scalr-server/embedded/scalr/app/python/scalrpy/analytics_poller.py --pid-file=/opt/scalr-server/var/run/service/analytics_poller.pid --log-file=/opt/scalr-server/var/log/service/python-analytics_poller.log --user=scalr-app --group=scalr-app --config=/opt/scalr-server/embedded/scalr/app/etc/config.yml --verbosity=INFO "
      stdout_logfile "/opt/scalr-server/var/log/supervisor/service-analytics_poller.log"
      stderr_logfile "/opt/scalr-server/var/log/supervisor/service-analytics_poller.err"
      autostart true
      startsecs 1
      service_name "service-analytics_poller"
      process_name "%(program_name)s"
      numprocs 1
      numprocs_start 0
      priority 999
      autorestart :unexpected
      startretries 3
      exitcodes [0, 2]
      stopsignal :TERM
      stopwaitsecs 10
      stdout_logfile_maxbytes "50MB"
      stdout_logfile_backups 10
      stdout_capture_maxbytes "0"
      stderr_logfile_maxbytes "50MB"
      stderr_logfile_backups 10
      stderr_capture_maxbytes "0"
      serverurl "AUTO"
    end
    

Running handlers:
[2016-10-20T23:40:36+00:00] ERROR: Running exception handlers
  - Raven::Chef::SentryHandler
Running handlers complete
[2016-10-20T23:40:36+00:00] ERROR: Exception handlers complete
[2016-10-20T23:40:36+00:00] FATAL: Stacktrace dumped to /opt/scalr-server/embedded/cookbooks/cache/chef-stacktrace.out
Chef Client failed. 170 resources updated in 47.955644198 seconds
[2016-10-20T23:40:36+00:00] ERROR: supervisor_service[service-analytics_poller] (scalr-server::group_service_always_post line 5) had an error: RuntimeError: Supervisor service service-analytics_poller was unable to be started
[2016-10-20T23:40:36+00:00] FATAL: Chef::Exceptions::ChildConvergeError: Chef run process exited unsuccessfully (exit code 1)

------------ END OF ERROR OUTPUT ------------

[Marc O'Brien]

Hi Bob,

I am not finding any occurrences of this in any of our issue archives.  Is there any change to behavior on a second run if you attempt to reconfigure again with another "sudo scalr-server-ctl reconfigure"?  Let us know if you have the same experience once more.

Many thanks,
Wm. Marc O'Brien
Scalr Technical Support

[Marc O'Brien]

Hi Bob,

I was able to replicate this problem on CentOS7 on both GCE and Azure.  However, this issue is not presenting on CentOS6 on GCE.  Can you test once more with CentOS6 or Ubuntu?  Let us know if issues persist on these OS versions.

Many thanks,
Wm. Marc O'Brien
Scalr Technical Support

Yes, issue is same on second run.
I also tried another brand new vm as well.

[Marc O'Brien]

Hi Bob,

I noticed your whitelisting request just came through.  Glad to see you are up and running now.  Just let us know if you have any issues with the Shared Roles Sync once whitelisting has been processed.

Cheers,

Wm. Marc O'Brien
Scalr Technical Support

[Ninad Date]

I tried to install Scalr server OSS on Google Cloud on CentOS6, CentOS7 and Ubuntu 14.04 - all result into same error as this thread has been started for. When checked one of the error file noticed that the script is trying to use AWS boto ec2 while this instance is on Google cloud and hence possibly failing. Hence to re-confirm, I did this test by launching an instance on AWS EC2 with Ubuntu 14.04 and trying the same installation steps and it went through fine successfully.

Thus the problem seems to be in the installation scripts somewhere where it is not installing properly on Google cloud instance

Pasting the error below

# cat /opt/scalr-server/var/log/supervisor/service-analytics_poller.err

Traceback (most recent call last):

  File "/opt/scalr-server/embedded/scalr/app/python/scalrpy/analytics_poller.py", line 39, in <module>

    import boto.ec2

  File "/opt/scalr-server/embedded/lib/python2.7/site-packages/boto/__init__.py", line 1216, in <module>

    boto.plugin.load_plugins(config)

  File "/opt/scalr-server/embedded/lib/python2.7/site-packages/boto/plugin.py", line 93, in load_plugins

    _import_module(file)

  File "/opt/scalr-server/embedded/lib/python2.7/site-packages/boto/plugin.py", line 75, in _import_module

    return imp.load_module(name, file, filename, data)

  File "/usr/lib/python2.6/site-packages/google_compute_engine/boto/compute_auth.py", line 19, in <module>

    from google_compute_engine import logger

ImportError: No module named google_compute_engine

Regards,

Ninad