ProxCenter v1.4.8

vDC storage policies and tenant VLAN networks, a Task Center that shows every job, and warm migration nodes the product prepares itself.

Multi-tenant and vDC

• Storage policies with QoS caps and per-tier quotas, assigned to vDCs and stamped on the disks a tenant creates.
• Tenant VLAN networks with provider-managed pools, self-service inside the ranges the provider hands out.
• Optional VMID ranges extended to vDC tenants.

Task Center

• Migrations are visible and cancellable, as a fifth job source scoped by target connection.
• Logs and actions are handled per job type, and recovery plan executions are reported as Site Recovery.
• The ProxCenter tab of the tasks bar carries the same jobs over the whole history.

Migration

• The product prepares a warm-migration node itself, now that Broadcom has closed the public VDDK download.
• The operator chooses the cutover moment, with a settable downtime budget.
• A refused guest shutdown is recoverable, with a 30 minute budget and an optional hard power off.
• virt-v2v never prompts for the root filesystem, and a root filesystem can be pinned up front.
• A failed create no longer destroys a pre-existing VM that held the requested VMID.
• A slow storage listing no longer empties the Target Storage field.
• VMID availability is scoped to the target connection, so a template can be cloned with two clusters connected.

Site Recovery

• A powered-off VM can be replicated, and a tag-based job no longer drops one in silence.
• A job whose tags stop matching lands in an explicit no match status instead of failing and retrying.
• The boot screenshot delay is a test failover option, 5 to 600 seconds.
• A DR test can leave the NICs connected.
• Replication RPO and Ceph OSD latency alerts, the OSD check disabled by default.
• A cleared alert sends a recovery notification, and resolution waits for hysteresis.

Access control and security

• Community installs grant the super-admin role again, with a sign-in backfill for the accounts already created without one.
• API tokens: offboarding, scope audit and real deletion.
• A pool or tag scoped user can create a guest.
• Two unguarded routes closed: alert creation and upload progress.
• An unparseable branding colour no longer turns every page into a 500.
• The orchestrator no longer trusts X-Forwarded-For blindly (GO-2026-5777).
• CVE-2026-40345: a deepmerge-ts >= 8.0.2 floor.

Monitoring and inventory

• SMART attributes, ZFS pool state, scrub and vdev tree on the node view.
• Node temperatures, read from the kernel hwmon tree since Proxmox exposes none.
• A dedicated ZFS ARC dashboard widget.
• The real disk format is reported, and the Cloud-Init drive is labelled as such.
• A slow RAM hotplug no longer reports a failed save.

Backups

• Pool-based selection for backup jobs.
• The PBS token guidance is corrected: Audit is needed alongside DatastoreReader.
• A PBS server declared by IP keeps working on Node 26.

Interface

• Widgets stay where they are dropped, plus a theme-logo widget.
• Appearance preferences are stored in the user account, per tenant and per user.
• Text stays readable on a light branding colour.

Dependencies

Eight Dependabot updates batched, including Next 16.3 and MUI X Data Grid 9.12. CI runs on Node 26.

Upgrading

Pull the images. Four schema migrations ship with this release, applied by the frontend entrypoint at first boot under an advisory lock: the user appearance preferences table, the API token creator provenance column, the vDC VLAN support tables and the vDC storage policy tables.

Enterprise: re-run the installer once instead of pulling. This release adds six settings to the compose file, and docker compose pull never touches that file. Five keep their defaults either way, but without GHCR_TOKEN reaching the container, the new Prepare the node action of warm migration is not offered. Re-running the standard install command fixes all of it in one pass: it refreshes docker-compose.yml, writes GHCR_TOKEN into .env, preserves your secrets, your customised URLs, your license key and your data, then pulls and restarts.

curl -fsSL https://proxcenter.io/install/enterprise | sudo bash -s -- --token YOUR_TOKEN

An installation pinned to a specific version should pass it again with --version, since a re-run otherwise resets it to latest. Community installs just pull: their images are public and warm migration is an Enterprise feature. Details in the update guide.

Docker Images

docker pull ghcr.io/adminsyspro/proxcenter-frontend:v1.4.8
docker pull ghcr.io/adminsyspro/proxcenter-orchestrator:v1.4.8
docker pull ghcr.io/adminsyspro/proxcenter-weasyprint:v1.4.8