Selected objects not set as values anymore

[crnine]

Hi,

I just updated from 7.3.0 to the 7.3.1 orchestrator version and since this update, selected vcenter objects won't show up as values in their fields, the fileds for vms, folder and resourcepools stay empty.

If it helps I joined the vcenter server to an active directory before and so removed the old name of it from the orchestrator and added it new. I readded the plugin to the vcenter again and workflows in general are working just this value transport not.

I tested with a brand new workflow with just one input value vm = VC:VirtualMachine and a system.log message as output and the same "error" occured, the selected vm doesn't show up as name-value in the workflow field.

Could the isse be related to the domain join of the vcenter server or is it a bug/unknown issue with the orchestrator plugin/server?

Thanks and Regards

cr

[iiliev]

Hi,

Just to clarify - in which UI are these object names not visible? In vRO Java client UI, or in vSphere Web client UI?

If this is related to vSphere Web client UI, and you are not seeing object names when you launch the workflow from a context menu, please ensure that when you added the vCenter Server instance in vRO, you specified the vCenter by its FQDN, and the FQDN has been spelled exactly the same as it is vissible in vSphere Web client UI (uppercase/lowercase matters).

[crnine]

Hi,

in the Web Client UI with vsphere 6.5. The process is, I login into the flash web client, mark a vm and start a workflow. Normally the name of the vm shows up automatically in the appropriate field but currently the fields stays empty.

As I said, I joined the vcenter to an AD domain but it is still pingable with its old FQDN and new one and both are currently still working as web address for the web client and both are reachable from the orchestrator server via ping and nslookup.

Are there any logs or methods to check this FQDN issue?

[iiliev]

In vRO server's Tomcat access log file (/var/log/vco/app-server/localhost_access_log.txt) there should be a request looking something like the following when vSphere Web client tries to fetch virtual machine name:

GET /vco/api/catalog/VC/VirtualMachine/corp-123.domain.com%252Fvm-106/

Note that corp-123.domain.com above represents the vCenter address as vSphere Web client sees it. When you add vCenter instance in vRO, you should use exactly the same address, otherwise vCenter plug-in will not be able to properly match the vCenter, no matter that it is accessible/working using some other names/format.

Casing matters; Corp-123.Domain.com is not the same as corp-123.domain.com. Format also matters; corp-123 is not the same as corp-123.domain.com. vCenter plug-in requires to be configured with a name that matches exactly the one used by vSphere Web client.

There should be also some traces to failed call to vRO in vSphere Web client log file - vsphere_client_virgo.log

[crnine]

Hi,

so, after serveral hours of testing, log analysing, killing domain names and certificates up to a point where nothing was working anymore, I reverted back to the safepoint snapshots and set the old domain name in the orchestrator and it is working again as normal^^

It seems vcenter isn't really ready for domain changes without new installation..

Thx anyways

cr