GitHub user RyanP-Testing added a comment to the discussion: Cloudstack 4.22 - 
Configuration Wizard failing - Fix Error

Hi Mike,

I went through the details and the follow-up information in the thread. Based 
on what you've described, I would not focus on changing the Guest CIDR again at 
this point.

The important part of the failure appears to be the `isServiceProviderEnabled` 
check while CloudStack is creating the Guest Network. That check is tied to the 
Virtual Router provider and its DHCP capability on the physical network 
associated with the Guest traffic.

Since you've already confirmed that the Virtual Router is enabled on both 
physical networks, and the wizard has been run multiple times, I would first 
verify that CloudStack is actually using the expected physical network and that 
there aren't stale physical-network records left over from previous wizard 
attempts.

In particular, I would check:

1. Which physical network the Guest traffic type is actually attached to.
2. That the VirtualRouter provider on that exact physical network is `Enabled`.
3. That `dhcp_service_provided` is `1` for that provider.
4. That the physical network itself is in an enabled/active state.
5. Whether there are multiple old physical networks from the previous 
configuration attempts.

The fact that your first physical network has a relatively high ID is also 
worth checking if this was intended to be a fresh installation. Repeated wizard 
attempts can leave configuration objects behind, so the UI may appear correct 
while the guest-network creation is resolving against a different 
physical-network record than expected.

I would also avoid manually changing CloudStack's database records as the first 
fix. If this is a test installation and you don't have anything important 
deployed yet, a clean database/configuration reset is safer than trying to 
repair partially-created zone, pod, physical-network and service-provider 
records manually.

There is another important point: the fact that you can subsequently create a 
VM does not necessarily mean the original configuration completed correctly. 
The VM console HTTP errors and inability to upload/use ISOs should be 
investigated separately after the underlying network configuration is 
confirmed. Those functions involve other CloudStack components, such as the 
Console Proxy and Secondary Storage/SSVM path, so I wouldn't assume they are 
simply caused by the wizard timeout.

For the next troubleshooting step, I would compare the following three things:

`Guest traffic → Physical Network → VirtualRouter provider`

and then verify the corresponding state directly in the database and 
management-server log.

If you can post the actual output of the SQL query and the relevant 
`management-server.log` section around the `isServiceProviderEnabled` failure 
(with credentials/IPs redacted), that should be enough to determine whether 
this is a stale physical-network configuration, a provider-state problem, or 
something else in the wizard workflow.

At this point, I would stop changing Guest subnets and concentrate on the 
physical-network/provider mapping. The `172.16.0.0/16` network you described is 
also your existing infrastructure network, so I would keep the CloudStack Guest 
network completely separate from that CIDR regardless. Overlapping the 
physical/infrastructure and Guest networks can create routing and Virtual 
Router problems later even if it gets past the wizard.

Hope this helps narrow it down.

Ryan P.
Cloud & DevOps Engineer 
AccuWeb.Cloud

GitHub link: 
https://github.com/apache/cloudstack/discussions/13956#discussioncomment-18304820

----
This is an automatically sent email for [email protected].
To unsubscribe, please send an email to: [email protected]

Reply via email to