Foreman » Packaging

Underlying errors:

2016-01-29T10:29:08 [app] [W] Unhandled Libvirt error
 | Fog::Errors::Error: Error saving the server: undefined method `Error' for Fog::Errors:Module
| /usr/share/foreman/vendor/ruby/1.9.1/gems/fog-libvirt-0.0.3/lib/fog/libvirt/models/compute/server.rb:67:in `rescue in save'
 | /usr/share/foreman/vendor/ruby/1.9.1/gems/fog-libvirt-0.0.3/lib/fog/libvirt/models/compute/server.rb:60:in `save'
| /usr/share/foreman/app/models/compute_resources/foreman/model/libvirt.rb:125:in `create_vm'
 | /usr/share/foreman/app/models/concerns/orchestration/compute.rb:76:in `setCompute'
| /usr/share/foreman/app/models/concerns/orchestration.rb:168:in `execute'
 | /usr/share/foreman/app/models/concerns/orchestration.rb:113:in `block in process'
| /usr/share/foreman/app/models/concerns/orchestration.rb:105:in `each'
 | /usr/share/foreman/app/models/concerns/orchestration.rb:105:in `process'
| /usr/share/foreman/app/models/concerns/orchestration.rb:35:in `on_save'

Possibly the fog-core errors file isn't loaded or something for that not to be defined.

As a workaround you might want to try checking the Libvirt server's logs for the error thrown, as this is a bug in Foreman/Fog's error reporting.

Any suggestions for logs to check?

I've checked /var/log/libvirt/libvirtd.log and /var/log/syslog and there was no output. The only time I get output is if I try to reuse a host name, which causes it to complain that the storage volume already exists.

Thank you for your help.

The broken error handling (undefined method `Error') is a severe problem that'll prevent us investigating this much. I don't know why that's happening from the code, and I can't reproduce it.

A couple of thoughts from things I've seen:

Do you have a "default" pool, and what directory is it at? There's a bug in fog-libvirt (we're tracking under #13778) where it assumes the pool is at /var/lib/libvirt/images/, so when it creates an ISO image in the real 'default' location, it specifies the wrong path when creating the VM.

Is there a cloud-init disk in the pool already for the host you're building? You may need to delete it, then call virsh pool-refresh default.

@Dominic
I think you need a couple of things for this problem to manifest. I´ll go over my setup. I have:
Smart-proxies - TFTP, DNS, Puppet, PuppetCA (and not DHCP as it cannot be managed)
Subnet - IPAM (Internal DB), Boot mode DHCP, Proxies DHCP (None), TFTP, DNS
Domain - DNS Proxy
Libvirt Compute Resource

During provision I have for network:
Managed, Primary, Provision - Checked

Where the "Managed" part says:
"Should this interface be managed via DHCP and DNS smart proxy and should it be configured during provisioning?"

"DHCP and DNS". Could it be that "Managed" really needs them both or is it ok with just DNS? What does the code say about that?

Anyway, yes I had the standard "default" pool. I tried removing it but it didn´t matter.

@Christian Holzheimer:
How similar is your setup compared to mine?

/K

Karli Sjöberg wrote:

@Dominic
I think you need a couple of things for this problem to manifest. I´ll go over my setup. I have:
Smart-proxies - TFTP, DNS, Puppet, PuppetCA (and not DHCP as it cannot be managed)
Subnet - IPAM (Internal DB), Boot mode DHCP, Proxies DHCP (None), TFTP, DNS
Domain - DNS Proxy
Libvirt Compute Resource

During provision I have for network:
Managed, Primary, Provision - Checked

Where the "Managed" part says:
"Should this interface be managed via DHCP and DNS smart proxy and should it be configured during provisioning?"

"DHCP and DNS". Could it be that "Managed" really needs them both or is it ok with just DNS? What does the code say about that?

None of this is likely to be relevant, the error is in creating a VM, not to do with DNS or DHCP. The text means that Managed must be ticked for DNS and DHCP to be managed, not that you require DNS and DHCP to have a Managed interface.

Anyway, yes I had the standard "default" pool. I tried removing it but it didn´t matter.

No, it must exist - and at the moment, in /var/lib/libvirt/images.

Oh, OK, I misunderstood. Well, if that´s the case then I had it before, so that´s not it then, right?

/K

Indeed, I don't have any more ideas at the moment and still can't see why the error during error reporting is happening.

Hi

Having exactly same issue on Ubuntu with same setup , Cloud Image, with user_data, have default pool and even have symlinks to /var/lib/libvirt/images still happening.. log is very useless as it just says (undefined method `Error')

Could you guys check if this is https://github.com/fog/fog-libvirt/pull/31? It's solved by installing genisoimage.

I've got genisoimage installed on the hostmachine, still have the error.

could you install genisoimage also on the Foreman server?

What's the status here?