
Proxmox
An Open Source virtualisation platform for organisations that want flexible and independent IT infrastructure. We use Proxmox VE to build HA clusters, Private Cloud environments, Ceph-based storage architectures and infrastructure designed for reliable ba
Proxmox is an open infrastructure platform for building virtualization environments, high-availability clusters, centralized backup systems, and distributed data centers. The core ecosystem consists of Proxmox Virtual Environment, Proxmox Backup Server, and Proxmox Datacenter Manager.
At SLogic Group, we design, deploy, and administer complete Proxmox environments. We build single-host and multi-node clusters, configure local and shared storage, deploy Ceph, prepare independent backup infrastructure, and centralize the management of multiple clusters and locations.
In addition to the virtualization ecosystem, we also deploy Proxmox Mail Gateway. It is a separate product designed to protect mail servers against spam, phishing, viruses, and other threats delivered through email.
Proxmox can replace several separate platforms, but a successful deployment requires much more than installing the system on servers. Cluster architecture, networking, storage, quorum, backup, and verified emergency procedures are critical.
The Proxmox ecosystem in business environments
The platform can be used by small companies operating a single server, organizations maintaining a local high-availability cluster, service providers, data centers, and enterprises using multiple independent locations.
Each product is responsible for a different part of the infrastructure, but they can work together as one operational environment. Proxmox VE runs virtual machines and containers, Proxmox Backup Server protects their data, and Proxmox Datacenter Manager centralizes the view of multiple clusters and servers.
Proxmox Virtual Environment
A server virtualization platform based on KVM virtual machines, LXC containers, clustering, high availability, and software-defined storage.
Proxmox Backup Server
A dedicated system for incremental, deduplicated, and encrypted backups of virtual machines, containers, and physical hosts.
Proxmox Datacenter Manager
A central platform for supervising multiple independent Proxmox clusters and nodes from one interface.
KVM and LXC
We combine full virtualization of Windows and Linux systems with lightweight system containers.
Ceph and ZFS
We build local and distributed storage layers tailored to the required performance, availability, and scale.
Proxmox Mail Gateway
As a separate product, it protects company mail servers against spam, phishing, malware, and unwanted messages.
Our Proxmox expertise
We can deploy a complete environment from scratch, migrate from another hypervisor, expand an existing cluster, modernize storage, or take over ongoing administration.
- design of Proxmox VE architecture,
- selection of servers, processors, RAM, and storage controllers,
- design of single hosts and multi-node clusters,
- installation and configuration of Proxmox VE,
- creation of KVM virtual machines and LXC containers,
- configuration of clusters and quorum mechanisms,
- deployment of high availability,
- configuration of live migration and storage migration,
- design of management, migration, and storage networks,
- configuration of bridges, bonds, VLANs, and Linux SDN,
- deployment of ZFS, LVM, LVM-thin, and directory-based storage,
- integration with NFS, iSCSI, Fibre Channel, and block storage,
- design and deployment of Ceph clusters,
- configuration of Ceph RBD and CephFS,
- design of pools, device classes, and CRUSH rules,
- deployment of Proxmox Backup Server,
- configuration of datastores, retention, prune, and garbage collection,
- synchronization of backups between PBS servers,
- backup encryption and key management,
- recovery testing for virtual machines, containers, and files,
- deployment of Proxmox Datacenter Manager,
- centralization of multiple clusters and locations,
- integration with LDAP, Active Directory, and OIDC systems,
- configuration of roles, ACLs, and multi-factor authentication,
- monitoring through Zabbix, Prometheus, Grafana, and other tools,
- automation through APIs, Ansible, Terraform, and scripts,
- migrations from VMware, Hyper-V, and other platforms,
- cluster and component updates,
- performance, security, and high-availability audits,
- technical documentation and emergency procedures,
- deployment and maintenance of Proxmox Mail Gateway.
An open server virtualization platform
Proxmox Virtual Environment combines the KVM hypervisor, LXC containers, storage management, network configuration, clustering, and high-availability mechanisms in one system. Administrators can manage the infrastructure through a web interface, command line, or API.
The platform can operate as a single host, a small business cluster, or an advanced environment consisting of multiple servers, shared storage, and separate infrastructure networks.
KVM Virtual Machines
We run full Windows, Linux, and other operating systems supported by hardware virtualization.
LXC Containers
We use lightweight system containers for Linux services that do not require a full virtual machine.
Cluster
We combine multiple nodes into one logical environment managed through a shared interface.
High Availability
We configure protected virtual machines to restart automatically on another node after a host failure.
Storage
We integrate local disks, arrays, ZFS, Ceph, NFS, iSCSI, and other shared resources.
Networking
We create bridges, bonds, VLANs, SDN zones, and separate networks for management, migration, and storage.
KVM virtual machines
KVM provides full operating system virtualization. Each virtual machine receives virtual CPUs, memory, disks, network interfaces, and other devices assigned by the administrator.
Windows Systems
We run application, terminal, domain, database, and other Microsoft servers.
Linux Systems
We deploy web servers, databases, monitoring platforms, application containers, and infrastructure services.
VirtIO
We use optimized storage and network drivers to reduce virtualization overhead.
UEFI and TPM
We prepare virtual machines requiring modern firmware, Secure Boot, and a virtual TPM module.
Passthrough
In suitable environments, we pass GPU cards, network interfaces, or other PCIe devices directly to virtual machines.
Snapshots
We create controlled state points before changes while remembering that a snapshot does not replace backup.
LXC containers
LXC containers share the host system kernel and therefore use fewer resources than full virtual machines. They work well for many Linux services, test environments, and small application servers.
- fast creation of new environments,
- lower memory and disk usage,
- templates for popular Linux distributions,
- privileged and unprivileged containers,
- CPU, memory, and storage limits,
- integration with VLANs and network bridges,
- backup and recovery through Proxmox VE and PBS,
- replication on ZFS storage,
- migration between nodes,
- control of permissions and access to host resources.
A container is lighter than a virtual machine, but it is not always a replacement for one. The choice should be based on application, isolation, and operating system requirements.
Multiple servers managed as one environment
A Proxmox VE cluster enables shared management of multiple hosts, virtual machine migration, centralized permission management, and the use of high-availability mechanisms.
The cluster design must take Corosync communication, network latency, quorum, vote distribution, and infrastructure behavior during partial failures into account.
Unified Management
Administrators can view nodes, virtual machines, containers, and storage through one cluster interface.
Migration
We move running or powered-off virtual machines between hosts according to available resources.
Quorum
The voting mechanism protects the cluster against simultaneous operation of separated infrastructure partitions.
Central ACLs
Roles and permissions can apply to the entire cluster, selected resource pools, virtual machines, or storage.
Shared Configuration
Cluster resource information is replicated between nodes.
Expansion
We add more hosts and resources while preserving a common management model.
Corosync and quorum
Corosync is responsible for cluster communication and determining which nodes belong to the active part of the environment. The network carrying this traffic should be stable, predictable, and protected against congestion.
- we design a separate cluster communication network,
- we minimize latency and packet loss,
- we configure redundant Corosync links,
- we analyze node count and vote distribution,
- we deploy QDevice in justified architectures,
- we test cluster behavior after node loss,
- we monitor membership and quorum changes,
- we avoid stretching the cluster across unstable WAN links,
- we document procedures for rejoining hosts,
- we prepare scenarios for a complete site failure.
A cluster should not be based on the assumption that the network will always work perfectly. Quorum and Corosync must be designed specifically for failure and communication-loss scenarios.
Automatic service recovery after a host failure
The HA mechanism monitors the status of nodes and protected resources. If a host fails, a virtual machine or container can be started on another available node.
High availability requires more than enabling a feature. Virtual machines must retain access to their data, and the remaining hosts must have sufficient memory, CPU capacity, and network connectivity.
HA Groups
We define preferred nodes and placement rules for critical virtual machines.
Automatic Restart
A protected system can be started on an available server after the source host is lost.
Fencing
Cluster mechanisms prevent the same resource from running on multiple nodes at the same time.
Priorities
We define the order and importance of services when available resources are limited.
Capacity Planning
We verify that the cluster retains enough resources after a failure to run the most important systems.
Failure Testing
We verify the behavior of virtual machines, networking, storage, and the cluster after a controlled host shutdown.
HA does not mean uninterrupted application operation. After a physical node failure, the virtual machine must start on another host, and the operating system and application require time to become available.
Live migration and storage migration
Migration enables virtual machines to be moved between nodes during maintenance, load balancing, or infrastructure modernization.
Live Migration
We move a running virtual machine between hosts with minimal impact on service availability.
Offline Migration
We move powered-off systems when simplicity is preferred or unsupported online parameters must be changed.
Storage Migration
We move virtual machine disks between local, shared, and distributed storage.
Maintenance Work
We evacuate a host before an update, component replacement, or configuration change.
Load Balancing
We distribute virtual machines across nodes based on memory, CPU, and storage utilization.
Storage Platform Change
We migrate disks between ZFS, Ceph, LVM, NFS, and other supported resources.
Bridges, bonds, VLANs, and software-defined networking
The network layer affects virtual machine availability, migration performance, cluster operation, and storage speed. We design it with traffic separation and physical interface redundancy in mind.
- Linux Bridge for virtual machines and containers,
- active-backup and LACP bonding,
- VLAN-aware bridges,
- separate management networks,
- separate Corosync communication,
- migration networks,
- Ceph public networks,
- Ceph replication networks,
- backup networks,
- MTU aligned across the entire path,
- redundant uplinks to independent switches,
- restricted access to the management panel and API,
- SDN zones, VNets, and subnets,
- routing and firewall integration,
- documentation of VLANs and interfaces.
Software-Defined Networking
SDN features simplify the creation of logical networks available across multiple nodes. They allow zones, virtual networks, and subnets for virtual machines and containers to be defined centrally.
Zones
We group networks by technology, location, or role within the environment.
VNet
We create logical networks assigned to virtual machines on different cluster nodes.
Subnets
We define addressing, gateways, and routing parameters for individual segments.
VLAN
We centralize the configuration of networks based on the existing switched infrastructure.
Overlay
In suitable architectures, we create logical networks over routed transport infrastructure.
Central ACLs
We restrict the use of selected networks by users and resource pools.
Local and shared storage for virtual machines
Storage selection affects performance, availability, migration time, and recovery procedures. A different architecture is appropriate for a single host than for an HA cluster or distributed data center.
ZFS
We use software pools, checksums, snapshots, compression, and replication between nodes.
LVM and LVM-thin
We create local block storage for virtual machine and container disks.
Ceph
We build distributed storage available simultaneously from multiple cluster nodes.
NFS
We integrate external file resources used for disks, ISO images, templates, and backups.
iSCSI and Block Storage
We connect arrays and external storage systems according to vendor requirements.
Local Storage
In simpler environments, we use fast host disks combined with replication and backup.
ZFS in Proxmox environments
ZFS works well as local storage for a single host and for clusters using replication. It provides data integrity protection, snapshots, compression, and software-defined disk groups.
- design of mirror, RAIDZ1, RAIDZ2, and RAIDZ3 layouts,
- selection of vdev width,
- separation of system disks from VM storage,
- compression configuration,
- selection of recordsize and volblocksize,
- monitoring of pool health and checksum errors,
- scheduled scrubs,
- replication of datasets and volumes,
- planning replacement of failed disks,
- monitoring ARC memory usage,
- selection of SLOG and special vdev only where justified,
- backup independent of ZFS snapshots.
ZFS provides storage integrity and redundancy, but RAID, mirrors, and snapshots do not replace an independent backup.
Distributed storage for virtualization clusters
Ceph combines disks located in multiple servers into one distributed storage system. Data is replicated or encoded across devices, so the failure of a single disk or node does not necessarily interrupt access to virtual machines.
A Ceph deployment requires the right number of servers, a fast network, appropriate disks, spare capacity, and monitoring of recovery processes.
Ceph RBD
Provides distributed block storage for virtual machine and container disks.
CephFS
Provides a shared file system for selected data, images, and services.
OSD
OSD processes manage disks, store data, and participate in replication and recovery.
MON and MGR
Monitors maintain cluster maps, while managers provide operational data and management functions.
CRUSH
Placement rules determine which disks, hosts, and device classes store copies of data.
Self-Healing
After a failure, Ceph rebuilds missing copies on available devices according to the pool policy.
Ceph cluster design
- we select the number of nodes and failure domains,
- we separate system disks from OSD devices,
- we select HDD, SSD, and NVMe devices according to workload type,
- we design public and replication networks,
- we configure bonding and redundant switches,
- we select replication size and min_size,
- we create device classes and CRUSH rules,
- we separate high-performance and capacity pools,
- we manage PG counts and autoscaling,
- we define capacity utilization thresholds,
- we monitor recovery, backfill, and scrub operations,
- we plan gradual cluster expansion,
- we test disk and node failure scenarios,
- we maintain an independent backup system.
Ceph provides data availability, but it is not a backup. Deleting a virtual machine or damaging its file system is reflected across all replicas.
Local storage replication
In clusters using local ZFS, we can configure scheduled replication of virtual machine disks between nodes. This allows HA to be used without deploying fully shared storage, although potential data loss depends on the replication frequency.
Local Performance
The virtual machine uses host disks directly without sending every write through the storage network.
Scheduled Replication
Changes are transferred to another node according to the defined schedule.
Controlled RPO
Potential data loss corresponds to the time since the latest successful replication.
Lower Complexity
The solution can be simpler than a complete Ceph cluster in smaller environments.
Capacity Constraints
Each host must have sufficient local space to store replicated disks.
Backup Is Still Required
Replication improves availability, but it also replicates errors and does not replace historical backups.
Dedicated protection for virtual machines, containers, and hosts
Proxmox Backup Server is a system designed for creating and storing backups. It integrates closely with Proxmox VE and supports incremental and deduplicated backups of virtual machines and containers.
PBS should be treated as an independent security layer. Where possible, we install it outside the production cluster, on separate hardware, and with separate administrative credentials.
Incremental Backup
After the first job, mainly new or changed data blocks are transferred.
Deduplication
Repeated data chunks are stored only once within a datastore.
Compression
Data is compressed to reduce storage and network utilization.
Encryption
Backups can be encrypted on the client side before being transferred to the repository.
Integrity Verification
The system uses checksums and scheduled verification jobs to validate stored data.
Granular Recovery
We restore complete virtual machines, containers, disks, and selected files from a specific point.
Proxmox Backup Server architecture
- we select an independent backup server,
- we design a ZFS pool or another appropriate storage platform,
- we create datastores for different environments and retention policies,
- we separate the backup network from user traffic,
- we configure users, API tokens, and ACLs,
- we integrate PBS with Proxmox VE clusters,
- we create backup schedules,
- we configure retention and prune jobs,
- we define garbage collection cycles,
- we configure verify jobs,
- we monitor capacity and data growth,
- we deploy synchronization to a second PBS,
- we secure encryption keys,
- we prepare procedures for full cluster recovery.
Datastores and retention management
A datastore stores data chunks and backup indexes. The retention policy should reflect system importance, available space, and the required recovery history.
Prune
Removes references to restore points that are no longer required by the retention policy.
Garbage Collection
Reclaims space occupied by data chunks no longer referenced by any backup.
Verify
Checks the integrity of stored backups and detects damaged chunks.
Namespaces
We organize backups from multiple environments, customers, or locations within one datastore.
ACLs
We restrict user and token access to specific namespaces and operations.
Capacity Monitoring
We analyze actual data growth after deduplication and forecast when capacity will be exhausted.
Synchronization between PBS servers
An additional Proxmox Backup Server can store a copy of backups in a second location. Synchronization reduces the risk of data loss after failure of the main repository or an entire data center.
Secondary Location
Backups are synchronized to a separate data center, office, or secure recovery location.
Deduplicated Transfer
Only chunks that do not yet exist on the destination server are transferred.
Independent Retention
The secondary location can retain a different restore point history.
Restricted Credentials
The synchronization token receives only the permissions required for the specific job.
Bandwidth Control
We align the schedule and bandwidth usage with the capabilities of the inter-site connection.
Disaster Recovery
We prepare a procedure for connecting the secondary PBS to a cluster rebuilt after a major failure.
A backup stored in the same server room and dependent on the same infrastructure does not protect against every site-failure scenario.
PBS backup encryption
Client-side encryption protects data against being read by a repository administrator or anyone who gains access to its disks. However, the recovery key must be stored independently from the backup server.
- we generate individual keys for each environment,
- we store keys outside the PBS server,
- we create secure copies of recovery documents,
- we restrict administrator access to keys,
- we test recovery of encrypted backups,
- we document responsibility for key storage,
- we include key loss in the risk analysis,
- we do not store the only key copy on the protected cluster.
Recovery from Proxmox Backup Server
Complete Virtual Machine
We restore a complete KVM virtual machine to the original or another cluster.
LXC Container
We restore the entire container together with its configuration and file system.
Individual File
We mount the backup and recover selected files without restoring the entire system.
Selected Disk
We restore a virtual machine volume to an existing or new system.
Alternative Storage
We restore data to Ceph, ZFS, LVM, NFS, or another available storage resource.
New Cluster
In a disaster scenario, we rebuild the infrastructure and restore systems in the defined order.
Central management of multiple Proxmox environments
Proxmox Datacenter Manager enables multiple independent clusters and servers to be supervised from one place. It is particularly useful for organizations operating several data centers, branches, or separate production environments.
PDM does not replace the management interface of an individual Proxmox VE cluster. Instead, it provides a higher-level view of resources, updates, and the status of distributed infrastructure.
Multiple Clusters
We connect independent Proxmox VE environments located in different sites.
Unified Resource View
Administrators obtain an overview of nodes, virtual machines, and infrastructure utilization.
Central Update Visibility
We monitor available updates across connected PVE and PBS environments.
Virtual Machine Operations
We perform selected operational actions without switching between multiple management panels.
Distributed Data Centers
We organize the management of environments operating in several server rooms and branch locations.
Operational Scalability
We reduce the number of manual actions required to maintain multiple independent clusters.
Use cases for Proxmox Datacenter Manager
- central view of several Proxmox VE clusters,
- supervision of independent hosts and locations,
- overview of compute resources and virtual machines,
- central visibility of available updates,
- simplification of daily administrator operations,
- management of distributed enterprise infrastructure,
- standardization of environment naming and organization,
- improved resource utilization planning,
- reduced need to switch between management panels,
- development of infrastructure spanning multiple data centers.
One large cluster stretched across distant locations is not always the best solution. It is often safer to maintain independent clusters and manage them centrally through PDM.
Users, roles, ACLs, and authentication
Access to the virtualization platform allows systems to be started, stopped, copied, and deleted. For this reason, we integrate Proxmox with central identity sources and restrict permissions according to the user’s role.
Local Users
We maintain emergency accounts independent of the external identity system.
LDAP and Active Directory
We integrate authentication with the organization’s user and group directory.
OpenID Connect
We deploy authentication through an external identity provider and SSO mechanisms.
Roles and ACLs
We grant access to specific virtual machines, pools, storage resources, nodes, and operations.
MFA
We require an additional authentication factor for administrators and sensitive accounts.
API Tokens
We create restricted tokens for automation, monitoring, and external systems.
Separation of duties
- a cluster administrator does not need access to every virtual machine,
- an operator can start systems without permission to delete them,
- the backup team can manage jobs without full host administration rights,
- users can receive access to their own resource pool,
- automation uses separate tokens instead of personal accounts,
- storage administrators receive only the permissions they require,
- critical operations are logged,
- unused accounts are reviewed and disabled regularly.
Protection of the management panel, hosts, and virtual machines
Proxmox includes integrated firewall mechanisms at the data center, node, virtual machine, and container levels. We combine them with physical network segmentation and external firewalls.
Data Center Firewall
We create shared rules and security groups applied across the environment.
Node Firewall
We restrict access to the management panel, SSH, Corosync, and host infrastructure services.
Virtual Machine Firewall
We filter traffic for individual virtual machines and containers according to their purpose.
IP Sets
We group addresses and subnets used across multiple rules.
Security Groups
We create reusable rule sets for defined system types.
Logging
We log rejected connections and forward events to a central system.
Proxmox security best practices
- we expose the management panel only through a management network or VPN,
- we restrict SSH access to trusted addresses,
- we use SSH keys instead of relying only on passwords,
- we require MFA for administrators,
- we separate administrator accounts from daily-use accounts,
- we do not reuse the same credentials for production and backup,
- we install security updates regularly,
- we monitor administrative operations,
- we disable unused interfaces and services,
- we create independent backups of configuration and virtual machines,
- we restrict host communication to required networks only,
- we test access recovery procedures after an identity-system failure.
Continuous supervision of the Proxmox environment
Cluster monitoring should cover hosts, virtual machines, storage, networking, backup, and high-availability mechanisms. Management panel availability alone does not mean that the environment is operating correctly.
Nodes
We monitor CPU, RAM, load, temperatures, power supplies, disks, and host availability.
Virtual Machines and Containers
We monitor state, resource utilization, uptime, and unexpected shutdowns.
Storage
We analyze capacity, latency, errors, ZFS status, Ceph health, and external storage resources.
Cluster
We monitor quorum, Corosync, HA status, and membership changes.
Network
We monitor interfaces, bonding, transmission errors, throughput, and packet loss.
Backup
We monitor job status, backup age, PBS capacity, and verification results.
Integration with external monitoring
- Zabbix for monitoring nodes, virtual machines, and services,
- Prometheus and Grafana for metric collection and visualization,
- InfluxDB for historical data,
- syslog and SIEM systems for centralized logging,
- email alerts and webhooks,
- SMART and storage-controller monitoring,
- monitoring of ZFS scrubs and pool errors,
- monitoring of Ceph health and OSD utilization,
- monitoring of PBS backups and synchronization,
- reporting of resource utilization and capacity trends.
APIs, Ansible, Terraform, and operational processes
Proxmox provides an API that allows resources to be created, modified, and deleted without manual work in the management panel. In larger environments, we use automation to standardize deployments.
REST API
We integrate the platform with portals, ordering systems, monitoring, and internal processes.
Ansible
We automate the configuration of hosts, networks, users, and systems running inside virtual machines.
Terraform
We define virtual machines and selected resources as code to maintain deployment repeatability.
Cloud-Init
We automatically configure addressing, users, SSH keys, and parameters of new virtual machines.
Templates
We create reference virtual machines and containers that accelerate the deployment of new services.
Business Integrations
We connect Proxmox with billing systems, CMDB, IPAM, helpdesk platforms, and self-service portals.
GPU and device passthrough
In suitable environments, we can pass a physical graphics card, accelerator, USB controller, or network interface directly to a virtual machine. This makes it possible to run systems that require direct access to specific hardware.
- IOMMU configuration,
- device group analysis,
- PCIe passthrough,
- GPU passthrough,
- network interface passthrough,
- storage controller passthrough,
- driver configuration in the guest system,
- stability testing after host restart,
- consideration of migration limitations,
- planning device availability after node failure.
Direct device passthrough increases a virtual machine’s capabilities, but it may restrict live migration and automatic restart on another host.
Moving virtual machines from other platforms
Migration to Proxmox can involve individual servers, an entire virtualization cluster, or a phased transfer of systems as part of infrastructure modernization.
VMware
We import virtual machine disks and configuration while adapting drivers, networking, and boot settings.
Hyper-V
We convert disks, prepare VirtIO drivers, and recreate virtual machine configuration.
Physical Servers
We perform P2V migrations of Windows and Linux systems to KVM virtual machines.
Another Proxmox Environment
We move virtual machines between independent clusters through backup, replication, or remote migration.
Containers and Applications
We migrate services to LXC, new virtual machines, or application container environments running inside VMs.
Storage Migration
We move systems from external arrays to ZFS, Ceph, or another target platform.
Migration process
- we inventory virtual machines, systems, and dependencies,
- we analyze CPU, RAM, storage, and network utilization,
- we identify passthrough devices and unusual configurations,
- we verify operating systems and drivers,
- we prepare the target cluster, networks, and storage,
- we install required drivers before migration,
- we create safety backups of source systems,
- we perform a test migration,
- we measure copy time and the downtime window,
- we create cutover and rollback plans,
- we migrate systems in the agreed order,
- we test applications, networking, and services,
- we monitor the environment after cutover,
- we update the documentation.
A successful migration does not end when the operating system starts. Applications, backup, monitoring, licensing, and all network dependencies must also be verified.
Controlled Proxmox updates
We update clusters in stages, taking version compatibility, quorum status, host load, storage health, and virtual machine migration capabilities into account.
- we verify the current cluster condition,
- we confirm backup availability,
- we analyze information about the new release,
- we verify Ceph and dependent component compatibility,
- we check resource availability on the remaining hosts,
- we migrate virtual machines away from the updated node,
- we install updates and restart the node,
- we verify network, storage, and cluster communication,
- we continue with the next node,
- we test HA and migration after completion,
- we update PBS and other components,
- we monitor the environment after the change.
Major version upgrades
Moving between major releases requires additional analysis of packages, repositories, the base operating system, and hardware compatibility. Before starting, we run validation checks and resolve issues that could interrupt the process.
Compatibility Review
We verify PVE, Ceph, kernel, firmware, and additional package versions.
Pre-Upgrade Checks
We run validation tools and resolve detected warnings.
Configuration Backup
We preserve host, virtual machine, network, and cluster settings before the change.
Staged Upgrade
We modernize individual nodes without shutting down the entire environment at once.
Virtual Machine Verification
We verify system startup, drivers, migration, and storage access.
Rollback Plan
We prepare a procedure for handling problems during the upgrade.
Architecture, security, and performance verification
An audit helps identify single points of failure, missing backups, incorrect quorum, Ceph problems, overloaded hosts, and other risks that may not be visible during daily operation.
Cluster
We review Corosync, quorum, node status, HA, and behavior after communication loss.
Storage
We analyze ZFS, Ceph, external arrays, capacity, performance, and redundancy.
Network
We verify bonding, VLANs, MTU, uplink redundancy, and traffic separation.
Backup
We compare production resources with the protection scope and test recovery.
Security
We review accounts, MFA, ACLs, administrative access, firewall rules, and updates.
Performance
We analyze CPU, memory, storage, network utilization, and virtual machine distribution.
Proxmox Backup Server audit
- we verify the scope of protected virtual machines and containers,
- we check the age of the latest successful backups,
- we analyze the retention policy,
- we review prune and garbage collection schedules,
- we verify verify-job results,
- we analyze datastore capacity,
- we verify synchronization to a secondary location,
- we review encryption key storage,
- we assess account separation and backup network isolation,
- we perform agreed recovery tests.
Failure testing
The environment should be tested not only during deployment. Controlled failure scenarios show whether the cluster, storage, and procedures behave as intended.
- loss of a single network interface,
- loss of one uplink or switch,
- failure of a Proxmox VE host,
- loss of quorum and Corosync communication,
- ZFS disk failure,
- Ceph OSD or node failure,
- unavailability of external storage,
- virtual machine recovery from PBS,
- recovery in a secondary cluster,
- failure of the primary PBS server,
- service startup from a secondary location,
- measurement of actual RTO.
Documentation and operational procedures
Even a well-designed cluster becomes difficult to maintain when knowledge about its networks, storage, and dependencies exists only in an administrator’s memory.
Physical Topology
We document servers, switches, interfaces, ports, and connections between locations.
Logical Topology
We describe VLANs, bridges, cluster networks, storage, backup, and management networks.
Storage
We document ZFS pools, Ceph, PBS datastores, arrays, and data placement policies.
Virtual Machines and Services
We create an inventory of systems, their owners, importance, and application dependencies.
Emergency Procedures
We describe recovery of the cluster, virtual machines, storage, and administrative access.
Change History
We record key modernizations, updates, and architectural decisions.
A separate product for email protection
Proxmox Mail Gateway is not part of the PVE, PBS, and PDM virtualization ecosystem. It is an independent product operating as an email security gateway.
The system works as an SMTP proxy positioned between the Internet and the target mail server. It analyzes inbound and outbound messages, filters spam, detects malicious attachments, and supports advanced mail-processing rules.
Spam Protection
We analyze reputation, message content, headers, and other characteristics to identify unwanted email.
Antivirus
We scan attachments and message content for known threats.
Phishing Protection
We reduce messages impersonating trusted people, companies, and services.
Quarantine
Suspicious messages can be held and made available to the user or administrator for review.
Mail Rules
We create policies based on sender, recipient, domain, content, analysis score, and other parameters.
Message Tracking
We search for messages and analyze how they were processed by the gateway.
Proxmox Mail Gateway use cases
- protection of an on-premises Exchange server,
- protection of Postfix, Exim, and other SMTP systems,
- filtering of mail for multiple domains,
- support for several internal mail servers,
- protection of inbound messages,
- control of outbound email,
- spam and virus quarantine,
- blocking dangerous attachment types,
- whitelist and blacklist rules,
- SPF, DKIM, and DMARC verification within the mail architecture,
- message flow tracking,
- mail statistics reporting,
- deployment of a mail gateway cluster,
- integration with backup and monitoring systems.
Proxmox Mail Gateway architecture
Inbound Gateway
MX records direct mail to PMG, which analyzes messages before forwarding them to the destination server.
Outbound Smarthost
The mail server can route outbound messages through PMG for control and logging.
PMG Cluster
We deploy multiple nodes that provide redundancy and shared configuration management.
Separate Locations
We place gateways in different data centers to reduce the impact of a single-site failure.
Relay Restrictions
We restrict mail relay to authorized domains and systems only.
TLS
We configure SMTP encryption and appropriate certificates.
Filtering and quarantine
The filtering policy must reduce spam and threats without blocking legitimate business correspondence. We define thresholds, rules, and exceptions based on the organization’s actual mail flow.
- configuration of spam scoring thresholds,
- blocking messages with a high risk score,
- quarantine of suspicious messages,
- quarantine reports for users,
- whitelisting of trusted senders,
- blacklisting of unwanted sources,
- blocking of extensions and MIME types,
- rules for selected domains and recipients,
- message size control,
- analysis of false-positive logs,
- periodic policy adjustment,
- procedure for releasing legitimate messages.
A filter that is too aggressive can block important correspondence, while one that is too permissive can allow threats through. The policy should be adjusted using data and observation.
Proxmox Mail Gateway monitoring
- availability of nodes and SMTP services,
- mail queue length,
- number of processed messages,
- number of rejections and delivery errors,
- CPU, RAM, and storage utilization,
- PMG cluster status,
- TLS certificate validity,
- antivirus signature currency,
- reputation of public IP addresses,
- correctness of DNS and MX records,
- unusual increases in outbound mail,
- detection of attempts to use the gateway as an open relay.
Migration of complete infrastructure to the Proxmox ecosystem
A comprehensive migration can include not only virtual machines, but also networking, storage, backup, multi-site management, and email protection.
Compute Layer
We move virtual machines and containers to Proxmox VE and optimize resource allocation.
Storage
We migrate data to ZFS, Ceph, external arrays, or a hybrid solution.
Backup
We deploy PBS, retention policies, off-site synchronization, and recovery testing.
Management
We connect independent environments through Proxmox Datacenter Manager.
Identity
We integrate authentication, roles, MFA, and a central user directory.
Independently, we deploy Proxmox Mail Gateway to protect the existing mail server.
How can we support your Proxmox environment?
Design and Deployment
We select servers and design the cluster, network, storage, backup, and central management.
Audit and Optimization
We analyze the existing platform, identify risks, and implement agreed improvements.
Migration
We move virtual machines, data, and services from VMware, Hyper-V, other clusters, or physical servers.
Ongoing Administration
We monitor the environment, perform updates, respond to errors, and develop the infrastructure.
How do we begin working together?
- We inventory the environment – we analyze servers, virtual machines, storage, networking, backup, and dependencies.
- We define the requirements – we determine performance, availability, RPO, RTO, retention, and growth plans.
- We design the architecture – we select PVE, PBS, PDM, storage, networks, and the high-availability model.
- We prepare the infrastructure – we configure servers, switches, VLANs, disks, and management systems.
- We deploy the platform – we install the cluster, storage, backup, monitoring, and identity integrations.
- We migrate the systems – we move virtual machines and data according to the cutover plan.
- We test failure and recovery – we verify HA, backup, quorum, storage, and disaster procedures.
- We provide documentation – we describe the topology, configuration, and operational processes.
- We provide ongoing support – we monitor, update, and scale the environment as the organization grows.
Proxmox products and technologies
Proxmox Virtual Environment
A KVM- and LXC-based virtualization platform covering clustering, HA, storage, networking, and virtual machine management.
Proxmox Backup Server
A dedicated solution for incremental, deduplicated, and encrypted backup of virtual machines, containers, and hosts.
Proxmox Datacenter Manager
A central platform for managing multiple independent Proxmox clusters and environments.
KVM
Full virtualization of Windows, Linux, and other operating systems.
LXC
Lightweight system containers for efficient operation of Linux services.
Ceph
Distributed storage providing shared block storage and resilience against disk and host failures.
ZFS
A file system and volume manager providing checksums, snapshots, compression, and replication.
Proxmox SDN
Central management of logical networks, zones, VNets, and subnets for virtual machines and containers.
High Availability
Automatic startup of protected resources on available nodes after a host failure.
Live Migration
Movement of running virtual machines between nodes during maintenance and load balancing.
Proxmox Firewall
Integrated traffic filtering at the data center, node, virtual machine, and container levels.
Proxmox Mail Gateway
An independent product protecting mail servers against spam, phishing, viruses, and malicious messages.
Do you need a Proxmox deployment, migration, or infrastructure expansion?
We will assess your infrastructure, design a Proxmox VE cluster, deploy Proxmox Backup Server and Proxmox Datacenter Manager, and prepare an environment ready for secure operation and further scaling.