I can confirm that the same issue happened to me on VMware 4.1 and Foreman 1.4.1.
As Dominic Cleal was saying to me in irc, it is possibly related to my foreman vmware compute user not being able to see the distributed vswitch in Vmware.
I will update this ticket once I get full admin permissions to Vmware for the foreman compute cluster user.
This is the stacktrace that I get, the error in logs and UI is failed to create vm: undefined method `uuid' for nil:NilClass:
https://gist.github.com/stefancocora/9668826

I can confirm that I'm seeing the same issue with Foreman 1.4.1 and Vsphere 5.5. For me it seems to be same scsi_controller_type issue that Domonic is seeing.

RuntimeError
missing required parameter uuid
app/models/concerns/fog_extensions/vsphere/server.rb:36:in `scsi_controller_type='
app/models/compute_resources/foreman/model/vmware.rb:105:in `new_vm'
app/models/concerns/orchestration/compute.rb:24:in `compute_object'
app/views/hosts/_compute.html.erb:1:in `_app_views_hosts__compute_html_erb___1796274211_70115543473140'
app/views/hosts/_unattended.html.erb:2:in `_app_views_hosts__unattended_html_erb___1230911890_70115544936680'
app/views/hosts/_form.html.erb:91:in `_app_views_hosts__form_html_erb___1642109057_70115543637280'
app/helpers/layout_helper.rb:218:in `form_for'
app/views/hosts/_form.html.erb:7:in `_app_views_hosts__form_html_erb___1642109057_70115543637280'
app/models/taxonomy.rb:56:in `as_taxonomy'
app/models/concerns/foreman/thread_session.rb:143:in `as_location'
app/models/taxonomy.rb:55:in `as_taxonomy'
app/models/concerns/foreman/thread_session.rb:108:in `as_org'
app/models/taxonomy.rb:54:in `as_taxonomy'
app/views/hosts/_form.html.erb:5:in `_app_views_hosts__form_html_erb___1642109057_70115543637280'
app/views/hosts/new.html.erb:7:in `_app_views_hosts_new_html_erb___1934921432_70115542850000'
app/controllers/application_controller.rb:282:in `process_error'
app/controllers/hosts_controller.rb:94:in `create'
app/models/concerns/foreman/thread_session.rb:33:in `clear_thread'
lib/middleware/catch_json_parse_errors.rb:9:in `call' 

I can confirm that I was getting this error failed to create vm: undefined method `uuid' for nil:NilClass: because Foreman's user couldn't access the VMware distributed virtual switch.
Once I've got permission to get to that part this error went away.
Ofcourse another error came out related to InvalidProperty: but that is another story

I confirm this with Foreman 1.4.2 and VMware 5.1.

It also occurs when trying to create or edit a Compute Profile:

RuntimeError
missing required parameter uuid
app/models/concerns/fog_extensions/vsphere/server.rb:36:in `scsi_controller_type='
app/models/compute_resources/foreman/model/vmware.rb:105:in `new_vm'
app/models/compute_attribute.rb:21:in `new_vm'
app/models/compute_attribute.rb:26:in `pretty_vm_attrs'
app/models/compute_attribute.rb:32:in `update_name'
app/controllers/compute_attributes_controller.rb:10:in `create'
app/models/concerns/foreman/thread_session.rb:33:in `clear_thread'
lib/middleware/catch_json_parse_errors.rb:9:in `call'

It doesn't seem to be a permissions issue in my case. I tried with admin access to vSphere but still the same.

Very similar errors, but Stefan's case was in the networking code so is unrelated. This is an issue in Foreman and/or Fog rather than permissions.

Confirming same error for 1.4.2 & vSphere 5.1 update 2.

RuntimeError
missing required parameter uuid
app/models/concerns/fog_extensions/vsphere/server.rb:36:in `scsi_controller_type='
app/models/compute_resources/foreman/model/vmware.rb:105:in `new_vm'
app/models/concerns/orchestration/compute.rb:21:in `compute_object'
app/views/hosts/_compute.html.erb:1:in `_app_views_hosts__compute_html_erb__1891142222_70023025250160'
app/views/hosts/_unattended.html.erb:2:in `_app_views_hosts__unattended_html_erb___145426765_70023026613480'
app/views/hosts/_form.html.erb:91:in `_app_views_hosts__form_html_erb__1729657159_70023025903020'
app/helpers/layout_helper.rb:218:in `form_for'
app/views/hosts/_form.html.erb:7:in `_app_views_hosts__form_html_erb__1729657159_70023025903020'
app/models/taxonomy.rb:56:in `as_taxonomy'
app/models/concerns/foreman/thread_session.rb:143:in `as_location'
app/models/taxonomy.rb:55:in `as_taxonomy'
app/models/concerns/foreman/thread_session.rb:108:in `as_org'
app/models/taxonomy.rb:54:in `as_taxonomy'
app/views/hosts/_form.html.erb:5:in `_app_views_hosts__form_html_erb__1729657159_70023025903020'
app/views/hosts/new.html.erb:7:in `_app_views_hosts_new_html_erb___1050982719_70023025365760'
app/controllers/application_controller.rb:287:in `process_error'
app/controllers/hosts_controller.rb:94:in `create'
app/models/concerns/foreman/thread_session.rb:33:in `clear_thread'
lib/middleware/catch_json_parse_errors.rb:9:in `call' 

Getting the same error with 1.4.2 and vsphere 5.5
is there a work around for this?

RuntimeError
missing required parameter uuid
app/models/concerns/fog_extensions/vsphere/server.rb:36:in `scsi_controller_type='
app/models/compute_resources/foreman/model/vmware.rb:105:in `new_vm'
app/models/concerns/orchestration/compute.rb:21:in `compute_object'
app/views/hosts/_compute.html.erb:1:in `_app_views_hosts__compute_html_erb__1927501353_70289390738340'
app/views/hosts/_unattended.html.erb:2:in `_app_views_hosts__unattended_html_erb___674433531_70289393996580'
app/views/hosts/_form.html.erb:91:in `_app_views_hosts__form_html_erb__1711599656_70289393091600'
app/helpers/layout_helper.rb:218:in `form_for'
app/views/hosts/_form.html.erb:7:in `_app_views_hosts__form_html_erb__1711599656_70289393091600'
app/models/taxonomy.rb:56:in `as_taxonomy'
app/models/concerns/foreman/thread_session.rb:143:in `as_location'
app/models/taxonomy.rb:55:in `as_taxonomy'
app/models/concerns/foreman/thread_session.rb:108:in `as_org'
app/models/taxonomy.rb:54:in `as_taxonomy'
app/views/hosts/_form.html.erb:5:in `_app_views_hosts__form_html_erb__1711599656_70289393091600'
app/views/hosts/new.html.erb:7:in `_app_views_hosts_new_html_erb__308404150_70289391525100'
app/controllers/application_controller.rb:287:in `process_error'
app/controllers/hosts_controller.rb:94:in `create'
app/models/concerns/foreman/thread_session.rb:33:in `clear_thread'
lib/middleware/catch_json_parse_errors.rb:9:in `call'

Dmytro Dondyk wrote:

Getting the same error with 1.4.2 and vsphere 5.5
is there a work around for this?

I was able to downgrade to 1.4.0 & the issue went away.

Any chance to fixed in 1.5.0 ? It's a big blocker for people using VmWare if we want to upgrade > 1.4.0

I think what is happening is:
- scsi_controller_type attribute is passed in during compute resource creation (defined in app/views/compute_resources_vms/form/_vmware.html.erb form)
- during instantiation of Fog::Compute::Vsphere::Server, when attributes are being initialized, scsi_controller_type setter is being called (defined in FogExtensions::Vsphere::Server)
- It in turn calls Fog::Compute::Vsphere::Server#scsi_controller method, which attempts to retrieve controller configuration from the server, as it hasn't been initialized yet
- The call fails, as the image id is nil for the new image.

This might fix the issue (I have no access to vmware to test it) https://github.com/witlessbird/foreman/tree/4590. The fix above relies on a certain order of initialization, however, and it's possible for this data to get overwritten. I think that a better approach would be to remove scsi_controller_type attribute from FogExtensions::Vsphere::Server and merge controller type data into scsi_controller hash in the controller.

I applied the patch on 1.4.1 (found here: https://github.com/witlessbird/foreman/commit/4efcc0d138.patch ), restarted apache and get the same errors as before.

I see this behavior when editing the large compute profile as well as trying to create a new host using any of the compute profiles.

The error I get when editing the compute profile can be found here: https://www.dropbox.com/s/5j9fopqj29j6anr/Screenshot%202014-04-17%2011.03.44.png

And the error when creating a new host:
https://www.dropbox.com/s/f9gmwkj8czzchql/Screenshot%202014-04-17%2010.50.24.png

I'll try any other patches as they become available.

Nate reported on IRC that this now seems to be working. Dmitri, do you want to submit that as a PR and we'll see how it goes?

This is still an intermittent issue for me. It worked great for ~7 builds and then started exhibiting one of three behaviors. They are as follows:

1. It actually trys to provision, but errors out with 'missing requires parameter uuid' - Stacktrace: https://gist.github.com/25190d4a032575b658a5
2. It fails to load the compute profiles while configuring the new host (before clicking submit to provision it). Screenshot: https://www.dropbox.com/s/skfmusqj4v6yik2/Screenshot%202014-04-22%2009.23.36.png
3. The third behavior is that it provisions and works 100%.

It is of note that the error found in the stack trace is identical, regardless of which error it is (both Internal server error and missing uuid show the same stacktrace in the debug log).

I've done precisely the same workflow every time I've tested it.

Prereqs: Have a hostgroup setup with all vmware settings specified.

1. Click New Host
2. Select your provisioning hostgroup
3. Select 'VMWare' under 'Deploy On' (Leave the default compute profile)
4. Enter the hostname
5. Check 'Virtual Machine' Section to confirm computer profile loaded. 50% of the time, it says ''Error loading virtual machine information: Internal Server Error, the other 50% it shows the profile.
6. Click Submit
7. Starts spinning up the machine, but throws 'missing required parameter uuid'.

Let me know if you need any additional information.

Also, I am on Foreman 1.4.1 and Running vCenter + ESXi 5.5.

One last update!

When I enable debugging, provisioning works 80% of the time. When I disable debugging in production.rb, provisioning works 0% of the time. It is a guaranteed failure. I am not sure what the different between debug and non-debug is, but this is behavior we noticed.

I updated to 1.5.0 RC2 and noticed it was lacking the patch provided by Dmitri.

Will 1.5.0 stable have the patch included? In the meantime, I've applied the patch to 1.5.0 RC2 to see if it increases success rate with VM provisioning on VMWare.

Confirmed that this patch does not completely fix the issue. When selecting "VMWare" from the "Deploy On" menu, I get the following in the log, almost instantly:

Started POST "/hosts/compute_resource_selected" for SOME_IP at Mon Apr 28 22:12:16 -0700 2014
Processing by HostsController#compute_resource_selected as */*
  Parameters: {"utf8"=>"✓", "host"=>{"environment_id"=>"1", "architecture_id"=>"1", "build"=>"1", "managed"=>"true", "ip"=>"", "hostgroup_id"=>"24", "type"=>"Host::Managed", "overwrite"=>"false", "progress_report_id"=>"[FILTERED]", "domain_id"=>"1", "mac"=>"", "operatingsystem_id"=>"2", "provision_method"=>"build", "ptable_id"=>"7", "compute_profile_id"=>"1", "model_id"=>"", "comment"=>"", "puppet_ca_proxy_id"=>"1", "puppetclass_ids"=>[""], "realm_id"=>"", "is_owned_by"=>"13-Users", "name"=>"", "compute_resource_id"=>"1", "interfaces_attributes"=>{"new_interfaces"=>{"ip"=>"", "type"=>"Nic::Managed", "_destroy"=>"false", "mac"=>"", "domain_id"=>"", "provider"=>"IPMI", "name"=>""}}, "root_pass"=>"[FILTERED]", "puppet_proxy_id"=>"5", "subnet_id"=>"1", "medium_id"=>"5", "disk"=>"", "enabled"=>"1"}, "authenticity_token"=>"REDACTED"}
  Rendered hosts/_compute.html.erb (295.2ms)
Operation FAILED: missing required parameter uuid
  Rendered common/500.html.erb (4.4ms)
Completed 500 Internal Server Error in 315.3ms (Views: 5.0ms | ActiveRecord: 0.0ms)

https://github.com/domcleal/foreman/commit/518a39c10.patch is patch against Foreman 1.4 (the one in comment 22 is against 1.5). Nate's reported promising results so far.

I can confirm that this patch worked in my environment. vCenter 5.5 ESXi 5.5. Foreman 1.4 puppet 3.5.1

I meant to update this sooner.

The patch provided works great! It completely fixed the problem as far as I can tell and VMWare provisioning is now rock solid when it comes to creating a new host and modifying compute profiles.

Thanks again for the patch!

Spam removed by Redmine administrator