VirtConsole 3.3.2 brings full VM backups to managed databases and Kubernetes control planes, stored on Proxmox Backup Server. It also adds provider controls and flexible billing for these backups, along with faster virtual machine networking and a number of improvements to everyday behaviour. 3.3.2 is available now under updates in your panel.
Managed databases: two kinds of backup, side by side
Every managed database now offers two kinds of backup next to each other: Full VM backups (local) on Proxmox Backup Server, and the existing Database backups (offsite).
Customers turn full VM backups on or off per database, and the price is shown before they confirm. From the same page they can list, restore and delete their full VM backups. After a restore, the platform re-applies the database’s credentials and monitoring, and sets up point-in-time recovery again with a fresh full backup.
When a database is deleted, its full VM backups are kept until the customer deletes them, and the delete dialog says so.
Kubernetes: control-plane backups with one switch
Control-plane VMs are backed up together as one set, with an etcd snapshot taken just before. Customers get a single switch on the cluster. Restores are performed by the provider, and for high-availability clusters the provider can recover the whole control plane to one consistent point.
Turning the switch off, or deleting the cluster, removes the control-plane backups and stops the charge.
Provider controls
For each location, the provider chooses the Proxmox Backup Server destination, the schedule and retention, and whether customers can see the retention.
Flexible billing
Each plan can carry a flat monthly fee, a per-GB rate, or both. The per-GB part is charged on the space the backups actually take on Proxmox Backup Server after deduplication, not on disk size. Backups appear as their own line in usage. The stored size is measured automatically and shown to both the customer and the admin.
Faster networking for virtual machines
Network cards now use the accelerated vhost driver with one queue per vCPU (up to eight), so traffic is spread across cores instead of one. Plans with unlimited bandwidth no longer carry a rate cap on the network card. The new settings apply the next time an instance starts.
Also in this release
- IPv6 address and subnet generation returns exactly the number requested, for every prefix length, and says how many fit when the subnet is too small.
- Single IPv4 and IPv6 addresses can be added in any notation, with clear messages when an address is a duplicate, the gateway, or outside the subnet.
- When the platform generates the root password for an instance created or reinstalled from the admin panel, the panel shows it once.
- Adding a secondary private network interface explains exactly why it cannot proceed, such as no free private address on that node, with proper status codes.
- After a managed database restore, its replicas resync from the restored primary automatically, and a restore asked for while another operation is running on the database is declined with a clear message.
- Clearer status codes and reasons when a VPC, image or ISO request cannot be completed.
- Image size detection, and https checks for image and ISO URLs.
- Faster backup and task pages on large installs.
- More reliable switching of point-in-time recovery for managed databases.
- Team members with database permissions can manage database backups.
- Kubernetes task lists for customers now show a concise view.
Thanks for building on VirtConsole. Questions along the way, reach us any time.