Proxmox 9 uses HA rules instead of HA groups.
Do not treat it as an error if the HA group (used in Proxmox 8) cannot be retrieved.
Signed-off-by: Serge Logvinov <serge.logvinov@sinextra.dev>
Add the group.topology.proxmox.sinextra.dev/ label to improve support for node selector and affinity rules.
Signed-off-by: Serge Logvinov <serge.logvinov@sinextra.dev>
Enhanced instance existence checks to handle inaccessible Proxmox nodes.
Improved test cases for instance existence and metadata retrieval.
Signed-off-by: Serge Logvinov <serge.logvinov@sinextra.dev>
Added error logging in the InstanceMetadata function to capture failures
when retrieving instance information, enhancing debugging capabilities.
Also includes:
- Added error check for metadata retrieval
- Added unit tests for error handling
- Updated to use errors package for error equality
Signed-off-by: rojanDinc <rojand94@gmail.com>
Signed-off-by: Serge Logvinov <serge.logvinov@sinextra.dev>
Add labels:
* topology.proxmox.sinextra.dev/node
* topology.proxmox.sinextra.dev/region
These labels represent the default topology labels.
They make it possible to use different topologies on the Proxmox side.
Signed-off-by: Serge Logvinov <serge.logvinov@sinextra.dev>
Changes errors created by cloud config to be standardized so that any
other packages relying on the cloud config can check if the error is of
the same "type".
Signed-off-by: Daniel J. Holmes (jaitaiwan) <dan@jaitaiwan.dev>
Changes:
- Increase test coverage of config
- Add networking feature config
- Add ability to find node ip addresses via qemu and specify ips that
should be treated as ExternalIPAddresses
Signed-off-by: Daniel J. Holmes (jaitaiwan) <dan@jaitaiwan.dev>
Signed-off-by: Serge Logvinov <serge.logvinov@sinextra.dev>
We will split the cloud configuration into two parts:
the original cloud controller configuration and a separate function for working with multiple Proxmox clusters.
Signed-off-by: Daniel J. Holmes (jaitaiwan) <dan@jaitaiwan.dev>
Signed-off-by: Serge Logvinov <serge.logvinov@sinextra.dev>
Error now clearly indicates the reasoning for the error message.
Previously the error message suggested a kubelet flag was not set even
when it may have been.
Signed-off-by: Daniel J. Holmes (jaitaiwan) <dan@jaitaiwan.dev>
We will find the node by name more precisely.
Check the UUID and VM name to determine the VM ID.
Signed-off-by: Serge Logvinov <serge.logvinov@sinextra.dev>
In some setups, the Proxmox VM name may differ from the Linux hostname.
To reliably identify a VM within a Proxmox cluster, we can use the system's UUID
Signed-off-by: Serge Logvinov <serge.logvinov@sinextra.dev>
Some method in Proxmox required root permissions (account).
So we can pass it through cluster config.
Signed-off-by: Serge Logvinov <serge.logvinov@sinextra.dev>
